Lymphotrophic nanoparticle-enhanced magnetic resonance imaging for nodal clinical

target volume delineation in the radiotherapy treatment planning of pelvic malignancies:

Derivation of a class solution nodal clinical target volume

By

Robert Edward Dinniwell

A thesis submitted in conformity with the requirements

for the degree of M. Sc.

Graduate Department of the Institute of Medical Science

University of Toronto

© Copyright by Robert Edward Dinniwell (2010)

ii

Lymphotrophic nanoparticle-enhanced magnetic resonance imaging for nodal clinical

target volume delineation in the radiotherapy treatment planning of pelvic malignancies:

Derivation of a class solution nodal clinical target volume

Master of Science 2010

Robert Edward Dinniwell

Graduate Department of the Institute of Medical Science

University of Toronto

Dextran-coated ultra-small, superparamagnetic, iron oxide particles (USPIO) have been

proposed as magnetic resonance (MR) lymph node contrast agents. This thesis

analyzed the topographic distributions of the pelvic and inguinal lymph nodes and

quantified their spatial relations with the adjacent vascular system. We hypothesized

that USPIO would facilitate identification of normal lymph nodes in a manner superior to

that afforded by computed tomography or unenhanced MR, but using current clinically

available scanners would be unlikely to identify microscopic nodal metastases. We have

constructed a high quality nodal atlas describing probability distributions for lymph node

number, size and position. Using this model, we then defined a generic three-

dimensional nodal clinical target volume and a means of accurate delineation of this

volume in a three-dimensional representation. This is the most quantitative assessment

of the pelvic and inguinal lymphatics to date and will help to improve the successful

targeting of lymph nodes for radiotherapy.

iii

Acknowledgements

There are many people whom I want to thank for their patience, support and mentorship

during my MSc. studies. Dr. Michael Milosevic, my primary supervisor, whose focus and

insight were fundamental in helping to define this project. Dr. Masoom A. Haider, with

his imaging expertise and guidance, was fundamental to the success of our

collaboration. I would additionally like to thank Drs. Kartik Jhaveri, Michael Jewett,

David Jaffray and Anthony Fyles.

Ami Syed and Anna Kirilova whose help with the execution of the trial enabled it to run

smoothly. Dr. Gregory Czarnota who enabled us to obtain the largest series in

publication and the fastest accrual rate anywhere. Dr. Philip Chan whose assistance

and support has been invaluable in this as well as other endeavors.

This work was made possible with the financial assistance from an ACURA Canadian

Association of Radiation Oncology grant; the Canadian Prostate Cancer Research

Initiative; and the CIHR Excellence in Radiation Research for the 21st Century

(EIRR21). I would additionally like to thank Dr. Mukesh G. Harisinghani, Department of

Abdominal Imaging, Massachusetts General Hospital, Dr. Paula Jacobs, and AMAG

Pharmaceuticals Inc., Lexington, MA for their valued support during this study.

Lastly, I wish to thank my friends, family and wife, Julie, for her support and

understanding.

iv

Table of Contents

Abstract …………………………………………………………………………………….…….ii

Acknowledgements ……………………………..……………………………………………..iii

List of Tables .…………………………………..……………………………….………….......v

List of Figures …………………………………...…………………………….…………..…..vii

List of Abbreviations ……….…………………...…………………………...…………...…..xiii

Chapter 1. Background and Literature Review ..……………….…………....……………...1

Chapter 2. …………………….……………...………………………………………………..46

Pelvic Lymph Node Topography for Radiotherapy Treatment Planning from

Ferumoxtran-10 Contrast-Enhanced MR Imaging

Dinniwell R, Chan P, Czarnota G, Haider MA, Jhaveri K, Jewett M, Fyles A, Jaffray D,

Milosevic M. Pelvic lymph node topography for radiotherapy treatment planning from

ferumoxtran-10 contrast-enhanced magnetic resonance imaging. Int J Radiat Oncol Biol

Phys. 2009 Jul 1;74(3):844-851. Published as the cover feature.

Chapter 3. …………………….………………...………………………………….…………72

Inguinal Lymph Node Topography for Radiotherapy Treatment Planning from

Contrast-Enhanced MR Imaging

Chapter 4. General Discussion ..……………...……………………………….…………..99

Chapter 5. Conclusions and Future Directions ……………………………………….…117

References ………………….…..……………...………………………………..…………..125

v

List of Tables

Table 2.1: Pelvic lymph node frequency and distribution ………………………………...58

Table 2.2: Spatial distribution of lymph node tissue in relation to the closest vessel in

each vascular segment ….……………………………………………………………………63

Table 2.3: Spatial distribution of lymph node tissue in relation to the closest vessel in

each vascular segment and musculoskeletal land markers ……………………….……..64

Table 3.1: Vessel combinations and the spatial distribution of lymph node tissue in

relation to the closest vascular segment as determined by the VHM dataset ...............84

Table 3.2: Spatial distribution of lymph node tissue in relation to the closest vascular

segment (Femoral Artery and Vein) derived from the 30 patient series .......................88

Table 3.3: Spatial distribution of lymph node tissue in relation to the closest vascular

segment (Femoral Artery and Vein and Superficial Circumflex Iliac Vein) derived from

the 30 patient series ...................................................................................…………....89

Table 3.4: Anisotropic analysis of the spatial distribution of lymph node tissue in relation

to the closest vascular segment to the closest vascular segment derived from the 30

patient series ……............................…............…………………………………………....90

vi

Table 4.1: Comparison of recommendations for nodal CTV delineation guidelines ...109

vii

List of Figures

Figure 1.1: Diagrammatic representation of a lymph node in cross-section showing the

entry and passage of lymph, its blood supply and the distribution of macrophages within

its interior. ..................................................................................................………..…….5

Figure 1.2: Representation of lymph (blue and yellow) draining from a lymph territory

(blue) into a first echelon lymph node shown in cross-section. ......................................6

Figure 1.3: Schematic depiction of the interconnections between lymphatic capillaries,

collectors, vessels and nodes for adjacent lymphatic territories. ..........…………....…....7

Figure 1.4: Schematic of the routes of lymph drainage from the lymphatic territories and

the proportion of the territory specific lymph in the lymph nodes downstream. ......…….8

Figure 1.5: (Left) Engraved front plate of Vasorum lymphaticorum corporis humani

historia et ichnographia and (Right) apparatus used to infuse mercury into the cadaveric

specimens and depictions of the efferent and afferent lymphatics with associated lymph

nodes. ....………………………………………………………..………………….…….….…11

Figure 1.6: (Left) Representative engraved plates from Vasorum lymphaticorum corporis

humani historia et ichnographia depicting the inguinal, iliac and aorto-caval lymphatic

vessels and nodes within the pelvis and retroperitoneum and (Right) the left inguinal

region showing the exquisite detail inherent in this work. ...........................…………….12

viii

Figure 1.7: (a) High resolution pelvic axial section from the Visible Human Anatomic

Series illustrating (b) small lymph nodes (arrows) close to the left external iliac vessels.

(c) Three-dimensional computer rendering derived from this dataset showing the

relationship between 177 pelvic and para-aortic lymph nodes (green) and the adjacent

vessels (red and blue). ....……………………………….……………………………………14

Figure 1.8: Frontal projection of the distal lumbar spine and bony pelvis depicting the

major vascular segments of interest. …………………………..………..………………….17

Figure 1.9: Frontal oblique projection of the distal lumbar spine and bony pelvis

depicting the major vascular segments of interest. ..…………………………..……….....18

Figure 1.10: Frontal projection of the distal lumbar spine and bony pelvis depicting the

major vascular segments of interest within the femoral triangle. ...…………..……….....22

Figure 1.11: Diagrammatic representation of the distribution of the deep and superficial

inguinal lymph nodes within the femoral triangle. ...……………………………………….23

Figure 1.12: Diagrammatic representation of the distribution and routes of lymphatic

drainage for the inguinal and pelvic lymph nodes. .........................................................25

Figure 1.13: (A) diagrammatic representation of a normal lymph node in cross-section

following USPIO administration and the corresponding MR appearance of the lymph

node. (B) lymph node with a moderate cancer burden and its post-contrast MR

ix

appearance. (C) lymph node with a minimal cancer burden and its post-contrast MR

appearance. ...................................................................................................................34

Figure 1.14: (a) T2-w pre-contrast axial MR image of the pelvis at the level of the

acetabulum. Inset figure depicts a normal-size proximal inguinal lymph node with similar

signal characteristics to the surrounding tissues. (b) T2-w post-contrast MR image

showing uniform distribution of USPIO contrast agent as evidenced by the pattern of

signal loss evident in the inset image. ...........................................................................35

Figure 1.15 Depiction of radiotherapy target volumes: GTV - GrossTarget Volume, CTV

- Clinical Target Volume, PTV - Planning Target Volume, IM - Internal Margin, and SM -

Set-up Margin. ...............................................................................................................38

Figure 1.16 Conventional radiotherapy plan for a carcinoma of the anal canal (top) and

an intensity modulated plan (bottom) illustrating a steep dose gradient at the edge of the

high dose volume (green) and reduced dose to bladder (B) and genitalia (G).

........................................................................................................................................42

Figure 2.1 A to D: Cranial to caudal axial T2w FSE, T2* GRE and T1w images superior

(A) to inferior (C and D) the bifurcation of the abdominal aorta. (Red arrow denotes the

division of the aorta in the right and left common iliac arteries). …………………………52

x

Figure 2.2: A three-dimensional surface rendering of the contoured pelvic lymph nodes

identified using ferumoxtran-10 in relation to the pelvic vasculature for a typical patient.

Arteries are indicated in red, veins in blue and lymph nodes in green. …………………55

Figure 2.3: A three-dimensional surface rendering of the contoured pelvic lymph nodes

(a) and a schematic representation (b) of the method used to determine the spatial

distribution of lymphatic tissue. The distance from the center of each lymph node

volume unit (green) to the closest artery (red) and vein (blue) was calculated in three

dimensions. ...………………………………………………………………………………….56

Figure 2.4: Histograms for a typical patient illustrating the spatial distribution of nodal

tissue in relation to the closest edge of the nearest artery or vein in each anatomic nodal

region: a) distal para-aortic; b) common iliac; c) external iliac; d) internal iliac; e) pelvic

side-wall and f) the sub-aortic/presacral. Note the different Y-axis scales, reflecting

differences in total lymphatic nodal volume from one region to another. .………………61

Figure 2.5: Pelvic nodal CTV at six cranio-caudal levels in the pelvis developed using

our population-based model of lymph node topography. The vascular, sacral and pelvic

sidewall expansions were trimmed to exclude bone, muscle, fascia, bladder and bowel.

(a) Distal para-aortic, (b) Bifurcation of the aorta and inferior vena cava, (c) Inferior to

the bifurcation of the common iliac vessels, (d) Mid-pelvis, (e) Cranial to the

acetabulum, and (f) Mid-acetabulum. …………………………………….............……67-68

xi

Figure 3.1: (A) A three-dimensional image rendering of the contoured lymph nodes,

vasculature and bony anatomy derived from the VHM with a representative image

adjacent. (B) three-dimensional rendering of pelvic and inguinal vasculature (1 inguinal

ligament, 2 sartorius muscle, 3 adductor longus muscle, 4 femoral artery and vein, 5

superficial circumflex iliac vein, 6 great saphenous vein, 7 inferior epigastric vein and

the 8 external pudendal vein) and (C) the corresponding lymph nodes. .......................80

Figure 3.2: (A) A three-dimensional surface rendering of the contoured left inguinal

lymph nodes with a schematic representation of the method used to determine the

spatial distribution of lymphatic tissue inset. (B) The distance from the center of each

lymph node volume unit (green) to the closest artery (red) and vein (blue) was

calculated in three dimensions. .……………………………………………………………..82

Figure 3.3: Histograms derived from the Visible Human Male illustrating the spatial

proximity of nodal tissue in relation to the closest edge of the nearest artery or vein for

the: a) femoral artery and vein; b) femoral artery and vein and great saphenous vein; c)

femoral artery and vein and superficial circumflex iliac vein and d) femoral artery and

vein, great saphenous vein and superficial circumflex iliac vein. ..................................87

Figure 3.4: (A). A frontal view of a three-dimensional surface rendering for a typical

patient with the left femoral artery and vein and inguinal lymph nodes delineated and (B)

the corresponding scatter plot distribution of the lymph node volume elements in relation

to the point collapse surface of the femoral artery and vein. (C) A frontal view with the

left femoral artery and vein and superficial circumflex iliac vein delineated and (D) the

xii

corresponding scatter plot distribution of the lymph node volume elements in relation to

the point collapse surface of these vascular segments. Note the antero-radial

distribution of the lymph node volume elements in relationship to the femoral vessels

and the effect of the inclusion of the superficial circumflex iliac vein on the distribution.

.......................................................................................................................................92

Figure 3.5: Inguinal nodal clinical target volume at twelve cranio-caudal levels in the

groin developed using our population-based model of lymph node topography. The

vascular expansions were trimmed to exclude bone, muscle, and skin. Legend: Femoral

arteries – red; Femoral vein – navy blue; Superficial circumflex iliac vein – light blue;

Sartorius muscle – brown; Adductor longus muscle – khaki and lymph nodes – green.

..............................................................................................................………………...94

Figure 4.1: Axial MR images for: (A) the common iliac vessels and common iliac lymph

node (red arrow); (C) the external iliac vessels and external iliac lymph node (white

arrow) and (E) the internal iliac vessels and an internal iliac lymph node (blue arrow)

and the corresponding scatter plot distributions of the lymph node volume elements in

relation to the surface of these vascular segments. Note: the red dot (0,0) denotes the

point defining the vessel surface. ...........................................................………………115

Figure 5.1: (Left) T2-w Fast Spin Echo MR image of the pelvis at the level of the

acetabulum with ( ) denoting a left lateral superficial inguinal lymph node. (Right) DWI

MR image at the level of the acetabulum with ( ) denoting a left lateral superficial

inguinal lymph node. ...………………………………………………………………………120

xiii

List of Abbreviations

AP Anterior-Posterior

CT Computed Tomography

CTV Clinical Target Volume

DICOM Digital Imaging and Communications in Medicine

ePLND extended pelvic lymphadenectomy

FOV Field of view

GTV Gross Target Volume

GU Genitourinary

GYN Gynecologic

IM Internal Margin

IMRT Intensity modulated radiotherapy

nCTV nodal Clinical Target Volume

nodal CTV nodal Clinical TargetVolume

nPTV nodal Planning Target Volume

MR Magnetic Resonance

MRI Magnetic Resonance Imaging

PA Posterior-Anterior

PET Positron Emission Tomography

POP Parallel Opposed Pair

PTV Planning Target Volume

RTOG Radiation Therapy Oncology Group

S3 Third Sacral Vertebra

SAD Short-axis diameter

xiv

SM Set-up Margin

T Tesla

T2-w T2-weighted

232Th Thorium

US Ultrasound

USPIO Ultra-small Particles of Iron Oxide

VHM Visible Human Male

1

Chapter 1. General Background and Literature Review

Science is the belief in the ignorance of experts.

- Richard Feynman

1.1 Lymphatic Anatomy and Function

The lymphatic system functions within the immunologic system, playing a key role in

host defense, in addition to the absorption and transportation of nutrients from the

gastrointestinal system and the regulation of interstitial fluid volume. The lymphatic

network arises embryologically from the developing blood vessels and functions in

parallel with the cardiovascular system1. The arterial system acts as a conduit to

distribute oxygen and other nutrients throughout the body to support tissue growth,

development and maintenance. The venous system returns the waste-laden blood to

the kidneys, liver and lungs where the products of metabolism are processed and/or

eliminated. The fluid that remains within the interstitial space enters the lymphatic

system and passes through a network of lymphatic vessels and nodes to ultimately

empty into the venous system before reaching the heart.

The lymphatic system architecture is similar to the venous system, with a fine capillary

network permeating most tissues and transitioning to successively larger lymphatic

vessels1. Lymphatic vessels possess three layered walls comprised of intima, media

and adventitia, as do the veins. Unlike the veins however, the lymphatic vessels tend to

2

be of much smaller diameter and are difficult to appreciate without the use of

specialized imaging techniques2, 3.

Lymph nodes of varying sizes are distributed along the routes of lymphatic return. They

are glandular structures, which vary markedly in both size and shape and are comprised

of a peripheral cortex and a medullary core (Figure 1.1). The parenchyma is surrounded

by a collagenous fiber capsule immediately beneath which is the subcapsular sinus.

Trabeculae extend from the subcapsular sinus to the interior and a coarse network of

reticular fibers that create an interdigitating sinus system. As with the lymphatic

vessels, standard clinical radiographic imaging techniques lack sufficient resolution to

detect normal, non-diseased lymph nodes and are unable to discern their inner

structure.

Communicating with the intercellular space, in which interstitial fluid resides, the

meshed capillary networks afford entry into the lymphatic system2, 4. Lymph arises from

interstitial fluid and enters into the blind-ended sacs at the terminus of a given capillary

network. Once in the lymphatic network, the lymphatic fluid progresses through

lymphatic collectors to reach the lymphatic vessels. The initial lymphatic capillaries,

collectors, vessels and nodes that communicate and direct lymph within a particular

anatomic region together constitute a lymphatic drainage basin. Adjacent lymphatic

basins are often interconnected in complex ways.

Fluid enters the lymphatic network through the route that presents the least amount of

resistance. Under normal physiologic conditions, the fluid within a territory flows to the

3

lymphatics in closest proximity. The system however is complex and lymphatic flow in

one region often influences flow in adjacent regions (Figures 1.2, 1.3 and 1.4)5, 6. For

example, efferent lymphatic vessels, which act as conduits for lymph exiting a lymph

node within a given lymphatic basin, may function as afferent vessels to a node within a

different lymphatic basin These interconnections produce a structural physiologic

reserve that is important during times of increased interstitial fluid load.

Lymph nodes are active in the immunologic filtration of lymph and the generation and

storage of lymphocytes. During embryogenesis, the volume of nodal tissue involved in

the lymphatic drainage for a given region is formed. With growth and development, this

volume may be divided into several small volume lymph nodes or a small number of

larger volume lymph nodes all of which occupy the fatty tissue around the blood

vessels7. As individuals age, the medullary cortex within a lymph node may involute

resulting in senile atrophy. While the nodal capsule remains, the node itself may

become difficult to detect clinically5. When an immune response is initiated, these nodes

may return to function with their peers in a lymphatic basin. Irrespective of the presence

or degree of atrophy, lymph will still continue to pass through all of the lymph nodes

within a given lymphatic basin.

The histologic architecture of the lymph node interior facilitates its ability to filter lymph

fluid. In a given lymphatic territory, lymph will enter into the node through an afferent

vessel and empty into the subcapsular sinus. Once in the lymphatic interior, the lymph

travels through only a part of the nodal interior with the trabeculae, cortical and

medullary sinuses effecting lymph flow direction (Figures 1.1 and 1.4)6. Metastases

4

entering the subcapsular sinus through an afferent vessel will not be dispersed

throughout the entire marginal sinus but instead enter into the porous intermedullary

sinuses in proximity to their point of entry into the lymph node. Nodal filtration during the

period following metastatic seeding is not markedly impaired as only one region is

initially affected. With on-going growth and proliferation, cancer cells may extend

throughout the nodal interior and restrict fluid entry by destroying or effacing the lymph

sinuses.

5

Figure 1.1: Diagrammatic representation of a lymph node in cross-section showing the

entry and passage of lymph, its blood supply and the distribution of macrophages within

its interior.

6

Figure 1.2: Representation of lymph (blue and yellow) draining from a lymph territory

(blue) into a first echelon lymph node shown in cross-section.

7

Figure 1.3: Schematic depiction of the interconnections between lymphatic capillaries,

collectors, vessels and nodes for adjacent lymphatic territories. (Modified from 5, 6)

8

Figure 1.4: Schematic of the routes of lymph drainage from the lymphatic territories and

the proportion of the territory specific lymph in the lymph nodes downstream. (Modified

from 6).

9

Lymph entering peripherally into the lymphatic capillaries will eventually reach a central

location of venous return through either the thoracic duct, which empties into the left

brachiocephalic vein, or the right lymphatic duct, which empties into the right subclavian

vein. Additional lymphaticovenous or lymphovenous anastomoses have been

described8-18 but are infrequent and tend not to function under normal physiologic

conditions. The purpose, frequency and distribution of these anastomoses remains

unclear.

The form of the lymphatic system is such that anatomic and/or physiologic disturbances

may transpire without markedly compromising its function. Were an obstruction to occur

distally, the lymph within the intercellular space would continue to drain to that site

which presented the least amount of resistance. In the region of the obstruction, the

fluid instead of entering into the obstructed capillary network adjacent to it travels in the

direction of lowest pressure and enters into the lymphatic system at some distance from

its original location. Similarly, obstruction of lymphatic collectors and afferent and

efferent lymphatic vessels tend to alter the route and passage of lymph to what would

typically be secondary or tertiary routes of drainage under normal conditions.

Intralymphatic growth of tumor metastases will compromise the sinus system within the

node and retard lymph entry into it. With an increase in intranodal pressure, lymph will

be diverted elsewhere within the system to be filtered by other nodes. An increase in

pressure within the lymph vasculature may open previously dormant lymphaticovenous

anastomoses. All of these responses are adaptive and facilitate continued functioning of

the lymphatic system in spite of local or regional insults or stressors4.

10

An appreciation of the form and function of the lymphatic system in its normal and

aberrant states is instructive in the pathologic and radiographic assessment of lymph

node metastases. Knowledge of the lymphatic drainage of a given organ or region and

an understanding of lymphatic physiology affords an insight as to the probable sites of

initial nodal spread from cancers arising from within it. Therefore, cancer staging and

therapy is predicated upon a sound study and understanding of the lymphatic system.

Despite the importance of the lymphatic system in maintaining normal physiologic

function and its role in host defense, research into the lymphatic system lags behind its

circulatory counterpart. The single most comprehensive study of the lymphatic system,

Vasorum lymphaticorum corporis humani historia et ichnographia, was undertaken by

Paolo Mascagni19. This exhaustive cadaveric study used cannulation and mercury

injection to identify the lymphatic vessels (Figures 1.5 and 1.6). Despite the passage of

223 years since its publication, there exists no modern non-invasive or invasive imaging

equivalent. While further studies such as those of Rouviere, have helped to describe

and refine the patterns and routes of lymphatic transport through the parietal and

visceral lymph nodes, Mascagni is credited with identifying and naming approximately

50% of the known lymphatic vessels7.

11

Figure 1.5: (Left) Engraved front plate of Vasorum lymphaticorum corporis humani

historia et ichnographia by Mascagni, and (Right) apparatus used to infuse mercury into

the cadaveric specimens and depictions of the efferent and afferent lymphatics with

associated lymph nodes.

12

Figure 1.6: Representative engraved plates from Vasorum lymphaticorum corporis

humani historia et ichnographia by Mascagni depicting (Left) the inguinal, iliac and

aorto-caval lymphatic vessels and nodes within the pelvis and retroperitoneum and

(Right) the left inguinal region showing the exquisite detail inherent in this work.

13

While no comparable study has been done in the modern era, the National Institute of

Health has recorded post-mortem anatomy using digitized photomacrographs of male

and female cadavers subjected to serial axial macrotomal sectioning20. The Visible

Human Male data set that resulted from this process is comprised of axial images with

pixel dimensions of 0.14 mm x 0.14 mm and 1 mm axial slice spacing, and the Female

data set with pixels measuring 0.33 mm x 0.33 mm and 0.33 mm axial slice spacing.

The inherent natural contrast between adjacent anatomic structures facilitates both

manual and automatic segmentation21. Delineation of the structures of interest and post-

processing can provide 3-dimensional rendering of the lymph nodes, blood vessels,

organs and clinical volumes of interest (Figure 1.7). While these are high fidelity virtual

representations of lymph nodes, organs and other anatomic structures of interest, the

lymphatic system vasculature is too fine and lacks sufficient inherent contrast to be

readily discernable even at this resolution.

In contemporary clinical practice, there are no methods for the in vivo detection of the

lymphatics (lymphatic capillaries through to lymph nodes and efferent lymphatic

vessels) of the inguinal and pelvic region in their entirety. This has profound implications

for the investigation and management of individuals with disorders and diseases

involving the lymphatic system. In the absence of a comprehensive image based means

for patient specific assessment, incorporation of information that can be obtained from

nuclear medicine and diagnostic radiology studies with knowledge of lymphatic structure

and function is the norm.

14

Figure 1.7: (a) High resolution pelvic axial section from the Visible Human Anatomic

Series illustrating (b) small lymph nodes (arrows) close to the external iliac vessels. (c)

Three-dimensional computer rendering derived from this dataset showing the

relationship between 177 pelvic and para-aortic lymph nodes (green) and the adjacent

vessels (red and blue).

15

1.2 Pelvic Lymphatic Network

Pelvic tumors are among the most common malignancies in women and men and are

associated with a significant burden of suffering worldwide22-24. Many pelvic tumors

including prostate, urinary bladder, uterine, cervical, vulva and anal canal cancers

frequently spread via the lymphatic channels to pelvic and inguinal lymph nodes. Lymph

node involvement at diagnosis generally is associated with a high risk of metastatic

disease at other sites outside of the pelvis and lower overall survival rates following

treatment. In patients undergoing surgical treatment, pelvic lymph node dissection is

frequently performed at the same time as removal of the primary tumor. Lymph node

dissection may be therapeutic as well as providing valuable information about prognosis

and the need for additional treatment like chemotherapy25-48. In patients receiving

radical or adjuvant radiotherapy, pelvic and inguinal lymph nodes are often included in

the clinical target volume (CTV) with the aim of reducing recurrence and improving

survival. The success of radiotherapy in this context is dependent on a detailed

understanding of pelvic and inguinal lymph node anatomy and patterns of cancer

spread.

The pelvic lymphatics are divisible into two primary networks of lymph nodes and

connecting lymph vessels: parietal and visceral25, 26, 28-30, 49-51. The parietal network of

lymphatics receives lymph from the walls of the pelvis that drain the perineum and

muscles of the pelvic girdle. The visceral network drains the pelvic organs via a series

of vessels and nodes that are situated in close relation to the pelvic arteries and veins.

The visceral lymphatics can be divided into major groups based on their anatomic

16

relation to the pelvic vasculature: common iliac, internal iliac, external iliac, superficial

inguinal and deep inguinal (Figures 1.8, 1.9, 1.10 and 1.11). The following subsections

describe in more detail the lymphatic architecture of the regions of relevance to this

body of work.

1.2.1 Common Iliac Lymphatics

The common iliac lymph nodes (4 to 7 nodes in most people) are in close proximity to

the common iliac vessels and comprise three distinct groups: the medial, intermediate

and lateral common iliac lymph node chains49, 51. The medial chain occupies the region

on the inner aspect of the common iliac artery and forms a triangular arrangement with

the corresponding nodes of the opposite side. The intermediate chain (3 to 4 nodes) is

located on the posterior aspect of the common iliac artery and vein7. The lateral chain (2

nodes) is positioned along the lateral aspect of the common iliac artery and the medial

border of the psoas muscle4, 7, 52. The lateral and intermediate common iliac chains

receive drainage from the lateral and intermediate external iliac lymph node chains and

empty into the lateral aortic lymph node chains52. The medial common iliac lymph node

chain may receive lymphatic drainage directly from the bladder neck and cervix uteri7.

17

Figure 1.8: Frontal projection of the distal lumbar spine and bony pelvis depicting the

major vascular segments of interest.

18

Figure 1.9: Frontal oblique projection of the distal lumbar spine and bony pelvis

depicting the major vascular segments of interest.

19

1.2.2 Internal Iliac Lymphatics

The internal iliac nodes are associated with the internal iliac vessels and their

branches51. They are situated anterior to the sacro-iliac joint descending to the level of

the greater sciatic foramen. The most superior node, the superior gluteal, is located in

proximity to the origin of the superior gluteal artery4, 7, 53. The inferior most node of this

chain is situated between the umbilical and obturator arteries4, 7, 53. Intermediate nodes

may be located along the uterine, internal pudendal, inferior gluteal and middle rectal

arteries4, 7, 53. The internal iliac lymph nodes drain the pelvic viscera including the

posterior aspect of the prostate, inferior and lateral aspects of the bladder, seminal

vesicles, middle and inferior aspects of the vagina and the body of the uterus4, 7, 53.

Lymphatic drainage from the perineum, inferior prostate and lower vagina travels to the

internal iliac chain54. The efferent vessels from the internal iliac nodes ascend superiorly

through the hypogastric lamina and pass below the common iliac vein to the

intermediate group of common iliac nodes.

1.2.3 External Iliac Lymphatics

The external iliac nodes are continuous with the external iliac vessels forming three

separate chains: lateral, intermediate and medial4, 7, 51. The lateral external iliac chain is

situated within the cleft formed by the medial border of the psoas muscle and the lateral

aspect of the external iliac artery. The inferior most node of this chain (lateral lacunar

node) is situated deep to the inguinal ligament. The intermediate external iliac chain is

situated on the medial side of the external iliac artery and anterior to the external iliac

20

vein. The medial external iliac chain is situated on the medial and dorsal aspect of the

external iliac vein abutting the lateral pelvic sidewall. The external iliac lymphatics

primarily drain the lower limbs receiving lymphatic fluid from the superficial and deep

inguinal nodes4, 7, 53. Of note, the medial and intermediate chains also receive drainage

from the obturator lymphatics, which in turn receive the visceral lymphatic vessels

arising from the lateral lobes of the prostate gland, bladder fundus, cervix uteri and

upper part of the vagina53. The juxtavisceral nodes of the vagina and cervix primarily

form associations with the medial group of external iliac nodes54.

1.2.4 Superficial Inguinal Lymphatics

The superficial inguinal nodes, anterior to Scarpa’s femoral triangle, typically consist of

10 to 12 large nodes7, 49, 54. They are divided from the femoral nerve, femoral vessels

and the deep inguinal nodes by the cribiform fascia. The medial superficial inguinal

nodes receive afferent vessels from the vulva and inferior vagina54. Additional drainage

from the uterine horns, which travels through the inguinal canal to reach these nodes,

may also occur54.

1.2.5 Deep Inguinal Lymphatics

The deep inguinal lymph nodes are located medial to the femoral vein and posterior to

the fascia lata in the femoral canal49, 51. The superior most deep inguinal node may

reside in the low pelvis and has been referred to as Cloquet’s or Rosenmuller’s node7.

21

The drainage from the deep inguinal nodes is through the femoral canal to the lacunar

nodes of the external iliac chain.

1.3 Lymphatic Drainage of the Pelvic Organs

The lymphatic drainage of the prostate gland, urinary bladder, corpus uterus, cervix,

vulva and anal canal has been studied using a variety of approaches, including detailed

anatomic dissection and surgical staging of patients with cancer to define the

distribution of nodal metastases7, 49, 53-56. Typical patterns of pelvic lymphatic flow are

illustrated in Fig. 1.12. However, lymphatic drainage can be highly variable from patient

to patient, influenced by differences in the distribution of normal lymphatic vessels and

the emergence of collaterals following destruction or obstruction of the usual routes of

flow. This can occur, for example, following surgery that disrupts lymphatic vessels or

as a result of bulky cancer lymph node metastases57, 58. Nevertheless, commonalities

emerged from the anatomic and surgical studies that influenced conventional radiation

treatment planning for many years prior to the widespread implementation of highly

conformal techniques like intensity-modulated radiotherapy (IMRT).

The volumes irradiated with conventional radiotherapy field borders are typically much

greater than those defined for highly conformal treatments. While this limits the dose

that may be delivered and sparing of adjacent normal tissues, it does afford greater

assurance of target coverage. In transitioning to increasingly conformal therapies, an

22

Figure 1.10: Frontal projection of the distal lumbar spine and bony pelvis depicting the

major vascular segments of interest within the femoral triangle.

23

Figure 1.11: Diagrammatic representation of the distribution of the deep and superficial

inguinal lymph nodes within the femoral triangle.

24

understanding of lymphatic topography and routes of lymph flow is of critical

importance. Disease site-specific and patient-specific knowledge of lymphatic drainage

is becoming increasingly important to assure adequate radiotherapy targeting of critical

lymph node regions.

1.3.1 Lymphatic Drainage of the Prostate Gland

Lymphatic drainage originates within the glandular acini of the prostate, traverses the

prostatic capsule and forms lymphatic networks on the posterior and superior aspects of

the gland7, 49, 53, 55, 59-64. Lymphatic flow from the anterior lobe of the prostate occurs

primarily via the medial and intermediate external iliac chains through direct lymphatic

vessels or after passing through the prevesical nodes7, 49, 53, 55. Drainage posterior along

the course of the internal pudendal artery to the inferior gluteal nodes may also occur7,

49, 53, 55. The posterior lobe of the prostate is drained by three lymphatic trunks53. The

first arises from the base of the gland, follows the seminal vesicles, passes superior to

the terminal segment of the ureter and drains into the middle and superior nodes of the

intermediate external iliac chain. The second lymphatic pathway originates from the

posterior and inferior regions of the prostate gland and follows the prostatic artery to

empty into the internal iliac lymph nodes. The third travels posterior to the superior

gluteal, lateral sacral and promontory lymph nodes along the course of the sacro-

prostatic ligament60. The lymphatic drainage of the prostate is known to have many

anastomoses, which may result in flow through the lymphatics of the rectum, bladder,

ampulla of the ductus deferens and seminal vesicles53. Lymph from the seminal vesicles

combines with the efferent

25

Figure 1.12: Diagrammatic representation of the distribution and routes of lymphatic

drainage for the inguinal and pelvic lymph nodes. (Modified from 4).

26

lymphatic vessels of the urinary bladder and prostate to drain into the intermediate and

medial external iliac nodes as well as the internal iliac nodes.

1.3.2 Lymphatic Drainage of the Urinary Bladder

An elaborate network of lymphatic drainage is present within the mucosa and muscular

layers of the bladder. The mucosal lymphatics unite to form collecting lymph vessels

that drain initially to the prevesical space and paravesical lymph nodes before reaching

the middle and superior nodes of the medial and intermediate external iliac groups7, 53,

63, 65-67. Clearance of lymph from the anterior aspect of the bladder wall begins with

collecting lymph vessels situated in proximity to the middle third of the lateral vesicle

border. Once the lymph fluid has reached this location, it flows through vessels towards

the origin of the umbilical artery in addition to the superior vesical artery. The lymph

vessels from the posterior bladder wall unite with vessels from the lateral vesicle border

to reach the external iliac nodes. Additional lymph vessels arise from the posterior

bladder wall and course either to the posterolateral angle of the bladder and then to the

medial and intermediate external nodes or extend to the lymph vessels of the bladder

trigone. Less commonly, lymph may flow to the medial lacunar node or internal iliac

nodes. The bladder trigone predominately drains through vessels located along the

course of the uterine artery to the medial and intermediate groups of external iliac

nodes53, 68.

27

1.3.3 Lymphatic Drainage of the Corpus Uterus

Lymphatic networks exist within the mucosa of the corpus uterus in addition to the

muscularis and subserosa53, 54. Lymph vessels emerge from the lateral aspects of the

uterus to travel along the suspensory ligaments in association with the ovarian branches

of the uterine artery. Ultimately, they unite with the tubo-ovarian lymph vessels and

terminate in the lumbar nodes close to the renal and superior mesenteric arteries at the

level of the lower poles of the kidneys4, 7, 69-72. Lymphatic drainage may also occur

laterally to the external iliac lymph nodes, with associated drainage to parauterine

nodes and via lymphatic trunks in the broad ligament to the superior interiliac lymph

nodes. Drainage channels arising in proximity to the cornua, mesosalpinx and tubular

bed direct lymph to the aortic and superior gluteal nodes. Lymph vessels at the site of

insertion of the round ligament may drain to the superomedial group of superficial

inguinal nodes. Anastomoses among lymph vessels arising in the uterus may also result

in drainage into the external iliac and superior gluteal lymph nodes4, 54.

1.3.4 Lymphatic Drainage of the Cervix

A rich lymphatic network exists within the uterine cervix and the adjacent parametrial

tissues, which terminates in three principal collecting lymphatic vessels54, 63, 69, 71, 73-76.

They include: the preureteral collecting vessels found anteriorly; the retroureteral

vessels; and the uterosacral vessels posteriorly7, 54, 73. The anterior preureteral lymph

vessels course in close association to the uterine artery and travel anterior to the

ureters. They comprise the primary lymphatic chain of the cervix and terminate in the

28

superior and middle nodes of the medial and intermediate external iliac chains. Less

commonly, they may also drain into the obturator group of external iliac lymph nodes.

Traveling with the uterine vein, the retroureteral lymph vessels pass posterior to the

ureter and end in the internal iliac lymph nodal chain. Drainage to the promontory nodes

may also occur. The uterosacral collecting lymph vessels travel posterolateral to the

rectum and then superiorly to reach the lateral sacral, promontory and subaortic lymph

nodes. Uninterrupted drainage from the cervix to the aortic nodes has been observed

via vessels traveling in association with the ureter to the common iliac, subaortic and

aortic nodes at the level of the pelvic brim. The lymphatic drainage of the uterine cervix

may be direct or may be interrupted by passage through the parauterine and

paravaginal lymph nodes73. The ureterouterine node, located at the junction of the

ureter and the uterine artery, is the largest of the parauterine and paravaginal lymph

nodes. Intramural and parauterine anastomoses of the cervical lymphatics occur with

the lymphatic vessels of the corpus uterus and vagina.

1.3.5 Lymphatic Drainage of the Vulva and Anal Canal

Lymphatic drainage from the vulva typically proceeds to the medial superficial inguinal

lymphatics before reaching the deep inguinal lymph nodes and then the external iliac

lymph nodes7, 53-56. Vulvar lesions in the mid-line, in proximity to the clitoris, may drain

to either or both sides and can drain directly into the external iliac lymph nodes

effectively bypassing the inguinal ones. Similarly, posterior lesions may drain directly

into the sacral lymph nodes. Lymph from the anal canal may drain into either the

29

superficial inguinal lymphatics or through the lower rectal lymphatics and the internal

iliac and inferior mesenteric lymph nodes7, 53, 55.

1.4 Imaging of Lymphatics

The variable nature of the lymphatic drainage from the prostate gland, urinary bladder,

corpus uterus, cervix, vulva and anal canal implies the need for patient-specific

lymphatic imaging to define individualized target volumes for highly conformal radiation

treatment. Unfortunately, the fine reticular structure of the small lymph vessels and

nodes, which comprise a significant portion of the lymph vascular system, limit the

usefulness of currently available conventional imaging approaches.

Early radiographic visualization of the lymphatic system followed 35 years after

Roentgen’s discovery of x-rays. In 1930 and 1931, Funaoka and Carvalho each injected

thorium (232Th) dioxide into the subcutaneous tissues and lymph nodes to demonstrate

them radiographically77, 78. While images were obtained from this technique, the

unstable 232Th nuclide decays by alpha particle emission making the imaging agent

highly carcinogenic. It was not until 1952, when Kinmonth used a water-soluble contrast

media that was injected into cannulated lymphatic vessels to study lymphedema, that a

practicable means of imaging became available52. Injection of contrast material was

preceded by subcutaneous injection of a vital dye to opacify the small lymphatic

channels at the local site and facilitate their cannulation52. While the lymphatic vessels

could be demonstrated, the water-soluble contrast agents were limited in their ability to

visualize the lymph nodes because of rapid dilution and clearance. Malek in 1959

30

revised the technique developed by Kinmonth and employed an oil-soluble iodized

contrast material that remained stable in tissue for longer periods79. Prolonged retention

within the lymphatic system enabled oil-based media to delineate a far greater extent of

the lymphovascular system along the drainage path from the site of injection.

Cannulation of the lymphatics in the distal lower extremity allowed depiction of

lymphatic channels and nodes in the pelvis and abdomen. In addition,

lymphangiography allowed visualization of the internal structure of lymph nodes, and

the detection of small metastases that appeared as filling defects on an otherwise

uniform background of contrast uptake. Lymphangiography was a major step forward

compared to earlier imaging techniques, and was used clinically for many years to

delineate lymph nodes in patients with cancer and other ailments.

While lymphangiography had many advantages, it also had limitations that largely were

imposed by the anatomic structure and architecture of the lymphovascular system58.

Cannulation of the lymphatic vessels was invasive, the injection of the agent was

uncomfortable and needed to be done over an extended period and the distribution of

the agent in tissue was determined by patterns of lymphatic drainage from the site of

injection. Clinically relevant lymph nodes anatomically and functionally remote from the

primary drainage pathway often were not visualized. For example, the internal iliac

lymph nodes, to which central pelvic tumors frequently metastasize, were difficult to

discern following injection of lymphatic channels in the leg that drain mainly via the

inguinal and external and common iliac pathways2, 49, 52.

31

With the emergence of cross-sectional imaging, the routine use of direct

lymphangiography waned. Technologic improvements in ultrasound (US), computed

tomography (CT) and magnetic resonance imaging (MRI) supplanted the routine use of

lymphangiography in the provision of clinical care and in research. Ultrasound can

detect imaging features indicative of macrometastases in superficial lymph nodes,

including irregularities in external contour, enlargement, round shape and loss of the

central hilum80. Doppler flow can also provide useful ancillary information. Ultrasound

imaging does however lack a means of providing easily interpretable serial image data

sets. CT and MRI, while non-invasive cross sectional imaging modalities, do not have

sufficient spatial and contrast resolution to detect all of the lymphatic channels and

normal lymph nodes. Neither technique is able to identify nodal metastases less than a

few mm in size81-86. Evolving dynamic CT and MR imaging approaches may improve the

detection of small volume lymph node metastases87-89. While these modalities have

unique strengths in lymphatic imaging, they fail to capture completely the information

that what was previously available from lymphangiography.

Molecular imaging approaches provide the promise of a new paradigm. The non-

invasive measurement of patient specific tumor physiology has the potential to improve

staging and response assessment following radiotherapy and other anti-cancer

treatments and facilitate the implementation of new molecular targeted therapeutic

strategies. The use of the glucose analogue 18F-fluoro-2-deoxy-d glucose (FDG) with

Positron Emission Tomography (PET) has been shown to improve the detection of

lymph node metastases in many malignancies90-109, reflecting the higher glucose uptake

and glycolytic rates in tumors compared to most normal tissues110. A disadvantage of

32

PET is low spatial resolution that can partially be overcome by co-registration with CT

images to enhance the localization of lymph node metastases109. However, PET is

limited by an inability to consistently detect disease in sub-centimeter lymph nodes and

micrometastatic lymphatic spread111.

1.4.1 USPIO-Enhanced MR Lymphatic Imaging

A novel class of MR imaging agents that capitalizes on the properties of iron within a

magnetic field has been explored clinically112-122. Dextran-coated ultra-small,

superparamagnetic, iron oxide particles (USPIO) have been evaluated as MR lymph

node contrast agents for discriminating between malignant and benign lymph nodes.

The particles are injected intravenously and taken up by macrophages in the

reticuloendothelial system, most of which reside in lymph nodes. The iron particles

cause a local change in tissue susceptibility, which results in a signal void or dark region

on post-contrast T1-weighted and T2-weighted MR images. Uniform signal loss is

apparent in normal lymph nodes as well as in those that are enlarged due to

inflammation. However, metastatic tumor in lymph nodes appears as a persistent bright

focus on a darker background in post-contrast scans, reflecting reduced regional

macrophage infiltration and USPIO uptake in cancer (Figures 1.13 and 1.14).

The potential of USPIO as a negative contrast agent for detecting lymph node

metastases has been explored with some success. USPIO-enhanced MRI has been

shown to improve the identification of lymph node metastases in patients with pelvic

malignancies relative to standard CT and/or MR techniques. However, it has limited

33

sensitivity for the detection of metastases in lymph nodes <5 mm in size, and is unable

to identify microscopic nodal disease with the accuracy of surgical staging123. An

important advantage though, in the context of radiotherapy treatment planning, is the

capability of USPIO-enhanced MR to facilitate the identification of apparently normal

lymph nodes relative to CT or unenhanced MR117, 124. The improved identification of

these nodes could aid in the definition of lymph node target volumes for high-precision

adjuvant treatment of patients at high risk of harboring occult nodal metastases.

34

Figure 1.13: (A) diagrammatic representation of a normal lymph node in cross-section

following USPIO administration and the corresponding MR appearance of lymph node.

(B) lymph node with a moderate cancer burden and its post-contrast MR appearance.

(C) lymph node with a minimal cancer burden and its post-contrast MR appearance.

35

Figure 1.14: (a) T2-w pre-contrast axial MR image of the pelvis at the level of the

acetabulum. Inset figure depicts a normal-sized proximal inguinal lymph node with

similar signal characteristics to the surrounding tissues. (b) T2-w post-contrast MR

image showing uniform distribution of USPIO contrast agent as evidenced by the

pattern of signal loss evident in the inset image.

36

1.5 Radiotherapy of Pelvic Lymph Nodes

Radiotherapy has a proven role in the curative management of gynecological and

genitourinary malignancies125-134. It may be used as primary treatment with the intent of

eradicating gross disease or in an adjuvant fashion after primary surgery to reduce the

likelihood of recurrence in high-risk patients. Tumors of the prostate gland, urinary

bladder, corpus uterus, cervix, vulva and anal canal commonly metastasize to iliac or

inguinal lymph nodes, and irradiation of these nodes has been shown to enhance pelvic

control and/or improve patient survival39, 135-150. Historically, pelvic radiation fields that

encompassed lymph nodes were planned using general knowledge of patterns of lymph

node spread, derived from anatomic and surgical staging studies as discussed

previously. High-resolution imaging of lymph node anatomy that could be applied to

individual patients was not available, nor was the technology to allow precise radiation

treatment targeting. Therefore, treatment plans tended to be very simple with four-field,

three-field or a two-field (AP/PA POP) beam arrangements positioned using bony

anatomy. The treatment volumes generally are quite large to minimize the chance of

missing disease. However, large volumes of normal tissue were also irradiated, which

contributed to side effects and limited the dose of radiation that could be used safely151-

153.

Radiotherapy treatment planning is now largely based on CT imaging rather than planar

x-rays. CT affords an ability to obtain high contrast as well as high resolution images

with a marked improvement in the level of anatomic detail that is appreciable in

comparison to projection radiography. With this improvement in imaging, standardized

37

definitions have been developed to facilitate treatment uniformity and standardization

(Figure 1.15). The Gross Target Volume (GTV) constitutes the clinically appreciable

tumor extent with the Clinical Target Volume (CTV) being the volume that includes

microscopic spread into adjacent tissues. The CTV includes direct spread from the GTV

in addition to spread to local and regional lymph nodes (nodal Clinical Target Volume -

nCTV). A further geometric margin of expansion (Planning Target Volume - PTV) is then

applied to account for physiologic movement (Internal Margin - IM) and interfraction

differences in treatment reproducibility (Set-up Margin - SM). Derivation of a treatment

plan using these definitions will produce a 3D radiotherapy treatment volume that not

only accurately and reproducibly encompasses the key target regions but also defines

the regions that can be avoided so as to spare critical anatomical structures.

38

Figure 1.15 Depiction of radiotherapy target volumes: GTV - Gross Target Volume, CTV

- Clinical Target Volume, PTV - Planning Target Volume, IM - Internal Margin, and SM -

Set-up Margin.

39

The use of planning CT has lead to improved targeting of the primary tumor but

has not substantially influenced the delineation of nodal volumes because

conventional CT and MRI scans are imperfect in their ability to detect pelvic

lymph nodes that are not enlarged29, 154, 155. However, the majority of the pelvic

lymph nodes lie in close proximity to adjacent blood vessels. Unlike the lymph

nodes, the blood vessels are more readily discernable with either CT or MR

imaging. Therefore, several investigators have proposed using blood vessels as

surrogates for nodal position with the application of an appropriate margin of

expansion to derive nodal clinical target volumes (nCTV) for radiotherapy156-159.

These margins could either be isotropic (uniform in all radial directions around a

vessel cross-section) or anisotropic (non-uniform). The ideal nCTV margin would

include all clinically relevant nodal tissue, while excluding as much surrounding

normal tissue as possible.

Several studies have now demonstrated that standard pelvic radiotherapy fields based

on bony anatomy often fail to encompass all clinically relevant lymph nodes, despite

large treatment volumes160-163. Bonin et al. used lymphangiography to study the

adequacy of pelvic lymph node coverage, as defined by the Gynecologic Oncology

Group, in patients with cervical carcinoma161. Application of standard field borders

resulted in inadequate coverage of the lower, lateral external iliac lymph node in 45% of

patients. In a similar study, Finlay et al. delineated pelvic blood vessels as a surrogate

for the lymphatics in 43 patients with cervix cancer160. With standard pelvic fields, there

was inadequate lymphatic coverage at the level of the bifurcation of the common iliac

arteries in 79.1% of patients, inadequate lateral coverage of the external iliac arteries in

40

20.9% of patients, and inadequate anterior coverage of the external iliac arteries in

69.7% of patients160. In addition, 44.2% of patients were judged to have excessive

lateral coverage (unnecessarily wide anterior-posterior fields >2 cm lateral to the

external iliac artery), 14% had excessive anterior coverage and 9.3% had excessive

superior coverage. Similar studies using lymphangiograms and surgical clips have

corroborated these findings162, 163. Standard pelvic fields do not adequately treat

portions of the common and external iliac lymph nodes and may result in irradiation of

larger than necessary volumes of normal tissue. These studies did not directly address

the implications of these results on subsequent disease recurrence in the pelvis,

although the highly variable nature of lymph node metastatic spread in individual

patients implies that all potential sites of disease need to be encompassed within the

high dose radiation volume to achieve maximal benefit.

There have been major advances in radiation treatment delivery over the past several

years that have revolutionized the ability to deliver dose in a highly conformal manner to

specific targets, while at the same time excluding surrounding normal tissues. Figure

1.16 is an example of an intensity modulated radiotherapy (IMRT) plan for the treatment

of cancer of the anal canal and the inguinal and pelvic lymph nodes, illustrating steep

dose gradients at the edge of the high dose volume leading to improved sparing of the

sigmoid and high rectum, small bowel and bladder. It is anticipated, as the long term

results following IMRT delivery become available, that reductions in late toxicity will

become apparent because of improved normal tissue avoidance. Similarly, the potential

for dose escalation afforded by IMRT should facilitate improvements in local control and

cure. High precision treatments like this highlight the need for accurate primary and

41

nodal target volume delineation because of the increased potential to underdose tumor.

A comprehensive and reproducible pelvic and inguinal nCTV definition would allow the

potential of IMRT to be fully exploited to improve tumor control and reduce toxicity.

42

Figure 1.16 Conventional radiotherapy plan for a carcinoma of the anal canal

radiotherapy (top) and an intensity modulated plan (bottom) illustrating a steep dose

gradient at the edge of the high dose volume (green) and reduced dose to bladder (B)

and genitalia (G).

43

1.6 Hypothesis, Rationale and Objectives

Radiation therapy has entered a new era of image guidance and precision treatment

delivery, which demands accurate and reproducible primary tumor and nodal target

volume delineation for optimal results. The determination of nodal target volumes has

been particular challenging because of the anatomic and structural nature of the

lymphatics and the difficulties of imaging with conventional techniques. However,

USPIO-enhanced MR imaging now offers the potential of improved lymph node

delineation, thereby enabling more comprehensive anatomic studies of the lymphatic

system. The inherent limitations of USPIO contrast agents and existing MR technology

preclude the identification of microscopic lymph node metastases, which are important

therapeutic targets in patients undergoing radiotherapy for many pelvic malignancies.

Currently, neither this approach nor any other clinically applicable imaging technique

offers sufficient sensitivity to be used as a means of targeting only tumor-bearing lymph

nodes. Nevertheless, the capability of USPIO-enhanced MR to facilitate detection of

apparently normal lymph nodes (sub-centimeter size with uniformly low signal on post-

contrast imaging – Fig. 1.13) that may or may not contain microscopic tumor deposits

could be utilized to develop more comprehensive radiotherapy target volume definitions.

Ideally, every patient requiring irradiation of pelvic lymph nodes would first have a

USPIO-enhanced MR imaging study to define the lymph node target volume. This

currently is not practical, however, because of limited MR availability in many centers

and lack of a clinically approved USPIO contrast agent. An alternative is to use USPIO-

enhanced MR in a large cohort of representative patients to study pelvic lymph node

44

distribution in relation to more easily imaged anatomic surrogates like the tumor, pelvic

and inguinal vasculature, and derive robust three-dimensional population-based nodal

target volume definitions that could be adapted to individual patients. The application of

these anatomic models in clinical practice would ensure a high probability of accurately

targeting volumes at high risk of harboring disease, thus avoiding a geographic

treatment miss. In parallel, it would facilitate the avoidance of adjacent normal

structures and limit treatment toxicity in the context of a highly conformal radiotherapy

plan.

The fundamental hypothesis of this thesis is that USPIO-enhanced MR imaging will

facilitate the derivation of a clinically relevant, population-based pelvic and inguinal

lymph node target volume definition for radiotherapy treatment planning based on

anatomic surrogates for lymph node location. The thesis will focus on nCTV definitions

relevant to the treatment of prostate, bladder, uterine, cervical, vulvar and anal cancers.

These tumors drain mainly to the inguinal and/or iliac lymph nodes. Pelvic tumors that

have a significant component of drainage to mesenteric lymph nodes, such as rectal

cancer, will be excluded because of the greater potential mobility of these nodes and

concerns about the applicability of this methodology.

45

The specific aims of this study are to:

1. Identify iliac and inguinal lymph nodes using USPIO-enhanced MRI

2. Define the topographic relationships and spatial probability distribution functions

for the location of lymph nodes relative to the adjacent vasculature and other

anatomic surrogates.

3. Derive population-based nCTV’s for the treatment of pelvic and inguinal lymph

nodes that can be adapted to individual patients based on the location of

anatomic lymph node surrogates.

46

Chapter 2. Pelvic Lymph Node Topography for Radiotherapy

Treatment Planning from Ferumoxtran-10 Contrast-Enhanced MR

Imaging (Published as cover feature in: Int J Radiat Oncol Biol Phys.

2009 Jul 1;74(3):844-851)

2.1 Introduction

Lymph node metastases are a common site of spread in pelvic malignancies. Patients

with a pelvic primary and lymph node metastases at diagnosis have a significantly

worse prognosis as compared to those with no evidence of lymphatic involvement.

Metastases to local and distant lymph nodes play a major role in tumor staging, therapy

and prognosis. Improvements in image-guided, intensity-modulated radiotherapy

(IMRT) treatment planning and delivery may help to overcome the adverse prognostic

significance of loco-regional lymph node metastases, and contribute to improved patient

outcome.

A barrier to the widespread implementation of IMRT for pelvic malignancies is the

absence of an objective description of lymph node locations in three-dimensional space

that is easily used for radiation treatment planning. Conventional CT and MR do not

reliably detect small pelvic lymph nodes, whether normal or abnormal, and have a low

sensitivity for identifying metastases, particularly in nodes <10 mm in short axis

dimension164-167. Anatomic and embryologic studies have demonstrated the close

developmental and spatial relationships between the lymphatics and vasculature3. The

47

pelvic vessels are readily detected with contrast-enhanced CT or MR, and have been

advocated as surrogates for lymph node location during treatment planning168-170.

However, detailed information about the spatial distribution of nodal tissue in relation to

the vessels at different anatomic levels in the pelvis is lacking.

Dextran-coated ultra-small, superparamagnetic, iron oxide particles (USPIO) have been

evaluated as MR lymph node contrast agents for discriminating between malignant and

benign lymph nodes171-173. While USPIO has the potential to improve the identification

of lymph node metastases in patients with pelvic malignancies relative to standard CT

and/or MR techniques, it is unlikely that USPIO will be able to identify microscopic nodal

metastases with the accuracy of surgical staging124. USPIO has however been shown to

facilitate the identification of normal pelvic lymph nodes relative to CT or unenhanced

MR117. The improved identification of these nodes using USPIO and MR offers the

potential of defining patient-specific nodal CTV’s in a minimally invasive manner that

accounts for the risk of microscopic disease, which cannot be imaged.

The aims of this study were to: 1) construct a high quality nodal atlas describing

probability distributions for normal and diseased pelvic lymph node number, size and

position based upon a 55 patient sample, and 2) define the potential 3D nodal CTV’s to

be irradiated and a means of accurate delineation of these volumes in a three-

dimensional representation.

48

2.2 Materials and Methods

This was a single center pilot study eligible to patients with histologically confirmed

endometrial, cervix, prostate or bladder cancer with no distant metastases. To ensure

uniformity of the sample and to control for artifacts and lymphadenopathy, all subjects

were screened to exclude: men with histologically confirmed carcinoma of the prostate

and a PSA ≤4.0 or >100; evidence of distant metastasis (M1); prior radical surgery or

cryosurgery for carcinoma of the prostate; previous radiation therapy; previous

hormonal manipulation or chemotherapy; patients with biopsy-proven lymph node

involvement; women with histologically confirmed carcinoma of the endometrium or

cervix who have undergone recent pelvic surgery with the exclusion of colposcopy,

dilatation and curettage or biopsy; men or women with histologically confirmed

carcinoma of the bladder who have undergone recent pelvic surgery with the exclusion

of TURBT; previous or concurrent cancers other than superficial basal or squamous cell

skin carcinoma unless disease free for at least 5 years; or a hip prosthesis.

2.2.1 Patient characteristics

The 55 patients included in this prospective study had a biopsy proven pelvic

malignancy. The median patient age was 62 (range 48-80) years. There were 12

women and 43 men (36 prostate cancers, 9 bladder cancers, 5 endometrial cancers,

and 5 cervical cancers). None had previously undergone pelvic lymph node dissection

or received radiotherapy or chemotherapy. The study was approved by the Research

49

Ethics Board of the University Health Network and written informed consent obtained.

The results of the study were not used in therapeutic decision making for these patients.

2.2.2 MR imaging

MR was performed using a 1.5 T magnet (GE EXCITE 1.5 T, Waukesha, WI, USA) with

a four-channel pelvic phased-array coil. Patients were instructed to take nothing by

mouth four hours prior to the study. No other preparation was undertaken. Axial imaging

sequences were obtained before and 24-36 hours after contrast administration at 3 mm

intervals through the pelvis. The pulse sequences consisted of T2w FSE (TR/TE

6500/80; field of view 24-28 cm), and T2*w dual echo GRE (TR/TE1/TE2 1400/14/21

ms; field of view 26-28 cm). Pixel and voxel sizes ranged from 0.9 to 1.2 mm2 and 2.7 to

3.6 mm3 respectively. Motion artifacts from peristalsis were minimized with the

administration of hyoscine-N-butylbromide (Buscopan) 30 mg or glucagon 1 mg i.v. or

i.m. prior to each of the MR imaging sessions. MR imaging remained uniform with no

changes in equipment or technique over the duration of the study period.

2.2.3 Ferumoxtran-10

Ferumoxtran-10 was reconstituted from a lyophilized powder. A dose of 2.6 mg Fe/kg

was dissolved in 100 ml of normal saline (0.9% NaCl) and infused intravenously.

50

2.2.4 Image analysis

Analysis of the DICOM images before and after contrast infusion was performed using

3D-DOCTOR® (Able Software Corp., Lexington, MA). The lymph nodes, pelvic

vasculature, as well as musculoskeletal landmarks for the sub-aortic/presacral space

and the obturator fossa/lateral pelvic sidewall were manually delineated by a single

observer (R.D.) and reviewed by a radiologist. Lymph node size evaluation was

undertaken using the T2w FSE image series as opposed to the T2*w dual echo GRE

which is subject to USPIO induced susceptibility artifacts. To verify the identification and

anatomic accuracy of the manual operator (R.D.) defined regions of interest (vascular

segments and lymph nodes) an expert abdominal-pelvic radiologist (M.H.) reviewed 10

(18%) of the image datasets. No noted variation between the observers was apparent

with regards to either nodal identification and delineation or vessel segment

identification. Voxels defining the edge of each distinct vessel and node were identified

in the x- and y- planes for each axial slice (z-plane). The x-y-z coordinates defining the

segmented structures were then used to generate a 3D stacked image model.

2.2.5 Vascular segmentation

All image sets were reviewed with partitioning of the image data to define the vascular

and lymphatic structures of interest. Manual tracing of the vessel walls was undertaken

using continuous sampling with vertices recorded along the path length. A closed spline

curve was then generated by connection of the vertices. Object specific boundaries on

each axial slice were integrated through the series to define a three-dimensional

51

volume. All points (x, y, and z) at the boundary edge and those pixels enclosed by these

points were considered to be part of the object. Those points and pixels not part of a

target structure were considered to be background. Window width and window level

were standardized to ensure optimization of contrast and brightness.

Segmentation of the pelvic arterial and venous system was performed in a caudal

direction beginning from the abdominal aorta at the origin of the inferior mesenteric

artery. At bifurcations, the first axial section that did not contain any part of the "mother

vessel" was described as the beginning of the "daughter vessel" (Figures 2.1). The T2w

FSE, T2* GRE and T1w image series were reviewed for each patient to identify the

points of bifurcation. The paired external iliac artery and vein were delineated inferiorly

to the level where the inguinal ligament traversed the medial aspect of the artery. To

ensure consistency among patients, segmentation of these structures was continued to

the superior aspect of the femoral heads in all cases.

Vessel topology, course, segment length, and bifurcation location varies within and

between individuals. Bifurcations of the iliac vessels on the right and left sides do

demonstrate varying degrees of spatial asymmetry. Similarly, the vessel bifurcation

angle may occur through a broad range. Vessel courses, lengths, curvatures, and

branching angles will all impact upon object appreciation and segmentation accuracy.

Given that this limitation exists within current clinical radiotherapy treatment planning, its

presence within this study will not limit the clinical applicability of the results.

52

Figure 2.1 A to D: Cranial to caudal axial T2w FSE, T2* GRE and T1w images superior

(A) to inferior (C and D) the bifurcation of the abdominal aorta. (Red arrow denotes the

division of the aorta in the right and left common iliac arteries).

53

2.2.6 Lymph node nomenclature and segmentation

Following administration of ferumoxtran-10, lymph nodes were detected by a decrease

in signal on the dual gradient-echo images. Lymph nodes were classified into anatomic

groups according to the following definitions, as illustrated in Figure 2.1:

• Distal para-aortic: Adjacent to the aorta and IVC from the inferior mesenteric

artery to the bifurcation of these vessels.

• Sub-aortic/presacral: Inferior to the sacral promontory over the mid-portion of the

sacrum, medial to the sacral foramina and posterior to the peritoneum and the

posterior rectal surface.

• Common iliac: Adjacent to the common iliac vessels from the aortic bifurcation to

the branching of the internal iliac artery.

• Internal iliac: Along the medial aspect of the internal iliac vessels.

• External iliac: Adjacent to the external iliac vessels inferior to the origin of the

deep inferior epigastric vessels, extending anterior to the iliopsoas muscle and

posterior to include the obturator lymph nodes within the obturator fossa.

All visible lymph nodes in these regions were contoured regardless of size and whether

or not they were judged to contain metastatic disease.

2.2.7 Musculoskeletal boundary nomenclature and segmentation

Musculoskeletal landmarks were identified for the sub-aortic/presacral and pelvic

sidewall lymph nodes. The sub-aortic/presacral landmark was the anterior aspect of the

fifth lumbar vertebrae and sacrum from the bifurcation of the aorta and IVC superiorly to

54

the inferior aspect of S3, and laterally to the sacral foramina. The pelvic sidewall

landmark was the ilium/ischium and internal obturator muscle from the bifurcation of the

common iliac vessels superiorly to the femoral canal inferiorly.

2.2.8 Measure of Proximity of Lymph Nodes to Vascular Segments

To generate a complete three-dimensional spatial representation of the lymphatic tissue

in each nodal segment, individual lymph nodes were first divided into small subunits

measuring 0.5 (x-axis) x 0.5 (y-axis) x 3 (z-axis) mm in size using an automated

algorithm (Figure 2.2). Each subunit represented 0.75 mm3 of lymphatic tissue.

Distances in 3-dimensions from the center of each nodal subunit to the closest vascular

edge (arterial or venous) were then calculated and grouped by vascular segment. Sub-

aortic/presacral lymph nodes and iliac lymph nodes along the lateral pelvic sidewall

were also localized in relation to musculoskeletal landmarks in the same manner.

Histograms were generated for each vascular segment to illustrate the spatial

distribution of nodal tissue in relation to the vessels or bony landmarks. Summary

statistics were derived from the histograms to determine the vascular expansions

required in each segment to encompass 50%, 80%, 90%, 99% and 100% of the

lymphatic tissue.

55

Figure 2.2: A three-dimensional surface rendering of the contoured pelvic lymph nodes

identified using ferumoxtran-10 in relation to the pelvic vasculature for a typical patient.

Arteries are indicated in red, veins in blue and lymph nodes in green.

Distal para-aortic

Common iliac

Sub-aortic/presacral

Internal iliac

External iliac

56

Figure 2.3: A three-dimensional surface rendering of the contoured pelvic lymph nodes

(a) and a schematic representation (b) of the method used to determine the spatial

distribution of lymphatic tissue. The distance from the center of each lymph node

volume unit (green) to the closest artery (red) and vein (blue) was calculated in three

dimensions.

57

2.3 Results

2.3.1 Lymph node frequency, size and location

A total of 1670 lymph nodes defined by 5663 separate nodal contours were detected in

the 55 patients. The median number of lymph nodes was 30 with a range of 5 to 62

(mean 30 and standard deviation 12.8). For those lymph nodes below the chosen

radiographic size criterion for metastasis (>10 mm), the median short axis size was 4

mm, with a range between 2 and 10 mm (mean 3.7 mm and standard deviation 1.2

mm). The median total nodal volume was 6650.0 mm3 (760.6-31946.9 mm3) and mean

7532.0 mm3 (standard deviation 5569.9 mm3). Table 2.1 summarizes the frequency

distribution of lymph nodes in each of the vascular segments. Nodes were relatively

uniformly distributed in the common iliac, external iliac and internal iliac regions, but

were less frequently detected in the sub-aortic/presacral region as well as the distal

para-aortic. There were no significant differences in nodal frequency, size or volume

between the left and right sides of the pelvis or between males and females.

58

Table 2.1: Pelvic lymph node frequency and distribution

Number of nodes

(Median and range) Vascular segment

Total Left Right

Distal vena cava and

para-aortic 1 (0-16)

Common iliac 7 (0-20) 3 (0-10) 3 (0-10)

External iliac 11 (1-19) 5 (1-10) 5 (0-10)

Internal iliac 5 (0-15) 1 (0-8) 3 (0-7)

Sub-aortic/presacral 2 (0-6)

59

The main purpose of this study was to define lymph node target volumes for

radiotherapy. It is unlikely, based on previous reports, that MR image enhancement with

ferumoxtran-10 will allow lymph node metastases <3 mm in size to be detected

reliably122, 124. Therefore, radiotherapy target volumes will need to encompass all lymph

nodes and not just those with imaging-evidence of metastatic disease. In addition,

lymph nodes that were enlarged and at high risk of harboring metastatic disease by

commonly accepted CT and MR size criteria (short axis diameter >10 mm) were

excluded from the histogram analysis to avoid biasing the results. This resulted in the

removal of 7 lymph nodes from the analysis in 4 of the study patients (4 left internal

iliac, 1 right common iliac, 1 right external iliac and 1 right external iliac) and represents

less than 0.5% of the nodes identified.

2.3.2 Quantitative distance histogram analysis of lymph node volume

Lymph nodes in each patient were analyzed as a function of distance from the closest

vessel. Figure 2.3 depicts representative histograms from a typical patient for each of

the vascular regions. The small branches of the internal iliac vasculature in the pre-

sacral region were not consistently visualized on MR. Therefore, the sub-

aortic/presacral lymphatic tissue was related to the closest aspect of the anterior

sacrum rather than the nearest visible vessel to avoid biasing the results towards

unusually long distances. For the same reason, obturator and medial internal iliac

lymphatic tissue along the pelvic sidewall was related to the pelvic sidewall as

previously described (see Materials and Methods).

60

The spatial distribution of nodal tissue for the 55 eligible patients is summarized in Table

2.1. On average, 90% of the lymph nodes were detected within 8.7 mm of the nearest

vessel in the distal para-aortic region, 7.3 mm in the common iliac region, 10.1 mm in

the external iliac region and 12.1 mm in the internal iliac region. However, there was

marked variability among patients. While the distribution of the nodal tissue along the

blood vessels was noted to be anisotropic and varied along the length of each relevant

vascular segment, isotropic (uniform) margins of vascular expansion were assessed

with the intent of generating a practicable method of 3D nodal CTV delineation.

The small vessels associated with the sub-aortic/presacral and obturator lymph nodes

were not consistently visualized and readily identifiable musculoskeletal landmarks were

used instead. In the sub-aortic/presacral nodal region, the mean distance anterior to the

sacrum to encompass 90% of the nodal tissue was 8.2 mm and, in the obturator region,

the mean distance from the pelvic sidewall to encompass 90% of the nodal tissue was

7.9 mm.

61

Figure 2.4: Histograms for a typical patient illustrating the spatial distribution of nodal

tissue in relation to the closest edge of the nearest artery or vein in each anatomic nodal

region: a) distal para-aortic; b) common iliac; c) external iliac; d) internal iliac; e) pelvic

side-wall f) the sub-aortic/presacral. Note the different Y-axis scales, reflecting

differences in total lymphatic nodal volume from one region to another.

62

2.3.3 Nodal CTV for pelvic radiotherapy

The main objective of this study was to define clinically practicable rules for determining

a lymph node CTV for patients receiving pelvic radiotherapy. The spatial distributions of

lymph node tissue in relation to vascular or musculoskeletal landmarks summarized in

Tables 2.2 and 2.3 suggest that different margins of expansion around these anatomic

surrogates may be optimal in different anatomic regions. The rules for nodal CTV

delineation were derived using 3D margins of expansion sufficient to cover 90% of the

nodal tissue in 90% of patients. This definition was chosen as a compromise between

nodal coverage and excessive normal tissue irradiation. The required margins of

expansion around the distal para-aortic, common iliac, external iliac and internal iliac

vessels were 12 mm, 10 mm, 9 mm and 10 mm respectively. In addition, a 22 mm

medial expansion from the lateral pelvic sidewall, contiguous anteriorly and posteriorly

with the expansions of the external and internal iliac vessels respectively, was required

to encompass the obturator and medial internal iliac lymph nodes. A 12 mm expansion

was required anterior to the sacrum from S1-S3, contiguous laterally with the expansion

of the iliac vessels, to encompass the sub-aortic/presacral nodes. Figure 2.4 illustrates

the use of these guidelines to devise a nodal CTV for pelvic radiotherapy. The anatomic

expansions were trimmed to exclude bone, muscle and fascia, as well as bowel and

bladder, because the space occupied by any of these tissues has a very low probability

of also containing nodal tissue either at the time of imaging or at later times (data not

shown).

63

Table 2.2: Spatial distribution of lymph node tissue in relation to the closest vessel in

each vascular segment

Distance of lymphatic tissue from nearest vessel

(Mean for the entire cohort and range in individual patients, mm)

Percentile 50% 80% 90% 95% 99% 100%

Distal para-aortic 4.8

(2.4-11.5)

7.6

(3.2-22.4)

8. 7

(3.6-23. 2)

9.4

(3.7-23.9)

10.4

(4.0-24.6)

11.1

(4.0-24.8)

Common iliac 4.1

(2.1-10.5)

6.2

(2.8-18.8)

7.3

(3.1-20.9)

8.2

(3.3-21.9)

9.6

(3.5-23.1)

10.4

(3.8-23.9)

External iliac 5.2

(2.4-11.5)

8.3

(3.1-14.7)

10.1

(3.4-17.4)

11.5

(3.7-19.4)

14.0

(4.2-22.9)

15.8

(4.2-26.9)

Internal iliac 7.0

(2.8-20.3)

10.0

(3.3-27.2)

12.1

(3.4-29.8)

13.2

(3.4-30.6)

14.7

(3.4-31.8)

15.0

(3.4-32.4)

Sub-aortic/pre-

sacral*

15.5

(3.6-31.1)

19.9

(5.1-36.9)

20.5

(5.9-37.7)

20.8

(6.4-38.1)

21.2

(6.7-38.6)

21.3

(6.9-38.6)

Obturator*

Lymphatic tissue in each anatomic region was localized in relation to the closet artery or

vein.

64

Table 2.3: Spatial distribution of lymph node tissue in relation to the closest vessel in

each vascular segment and musculoskeletal land markers

Distance of lymphatic tissue from nearest vessel or musculoskeletal land

marker

(Mean for the entire cohort and range in individual patients, mm)

Percentile 50% 80% 90% 95% 99% 100%

Distal para-aortic 4.6

(2.4-11.5)

7.0

(3.2-17.7)

8.2

(3.6-19. 8)

8. 8

(3.7-20.9)

9.9

(4.0-22.4)

10.5

(4.0-23.8)

Common iliac 4.0

(2.1-10.4)

6.1

(2.8-18.8)

7.2

(3.1-20.9)

8.2

(3.2-21.9)

9.6

(3.5-23.1)

10.3

(3.8-23.9)

External iliac 3.9

(2.4-5.5)

5.5

(3.1-7.8)

6.6

(3.4-10.1)

7.6

(3.7-12.3)

9.4

(4.2-15.0)

10.9

(4.2-17.8)

Internal iliac 4.6

(2.8-8.4)

6.1

(3.3-11.0)

6.9

(3.4-13.5)

7.6

(3.4-14.5)

8.5

(3.4-15.6)

8.9

(3.4-16.6)

Sub-aortic/pre-

sacral*

6.0

(3.0-10.3)

7.8

(3.5-15.1)

8.2

(3.8-16.5)

8.5

(4.2-17.2)

8.8

(4.2-17.8)

8.9

(4.2-18.2)

Obturator* 4.6

(2.6-7.2)

7.1

(3.4-12.2)

7.9

(3.6-18.9)

9.8

(3.9-19.8)

11.8

(4.3-21.1)

13.3

(4.3-22.7)

* Lymphatic tissue in each anatomic region was localized in relation to the closet

artery or vein except in the sub-aortic/pre-sacral and obturator regions, where it was

related to musculoskeletal land markers (see text).

65

2.4 Discussion

The lymphatic system is a prominent route of spread for many malignancies. Current

imaging techniques are limited in their ability to reliably detect micro-metastases in

lymph nodes. Therefore, radiation treatment of apparently normal lymph nodes is often

advocated as a means of improving local disease control and improving survival in

patients with endometrial, cervix, bladder, prostate and gastrointestinal cancer. The

current evolution towards the use of high-precision IMRT to treat pelvic lymph nodes at

risk of harboring occult disease demands accurate knowledge of the location of these

nodes for target volume definition. This study provides a comprehensive description of

the spatial distribution of pelvic lymph nodes derived from MR imaging and the use of a

lymph node contrast agent in 55 patients. The results of this study demonstrate that: 1.

ferumoxtran-10 subjectively enhances the capability of MR to detect pelvic lymph

nodes, whether involved by cancer or not; 2. for CTV definition, appropriate margins of

expansion around major pelvic vessels or musculoskeletal nodal surrogates vary

according to anatomic regions; and 3. the margins of expansion that are required

around these anatomic surrogates to reliably encompass high-risk nodes are larger than

advocated in previous studies169. A population-based model of pelvic lymphatic

topography has been presented, along with a set of clinically-relevant rules for nodal

target volume definition.

This study, in agreement with others, demonstrated that pelvic lymph nodes are

generally situated close to vessels3. This was particularly evident in the distal para-

aortic, common iliac and external iliac regions, as illustrated in Figure 2.3 for a

66

representative patient and summarized for the entire cohort in Tables 2.2 and 2.3.

However, nodal tissue in the sub-aortic/presacral and obturator regions was found at

greater distance from the major arteries and veins. These nodes are associated with

smaller branches of the iliac vessels that are too small to be visualized reliably using

MR. Therefore, the major vessels alone do not provide sufficient surrogate information

for lymph node CTV definition, and it was necessary to incorporate bony landmarks into

our population-based model of nodal topography. The model suggests that 90% of the

nodal tissue in 90% of patients will be encompassed by symmetrical three-dimensional

expansions of the distal para-aortic (12 mm), common iliac (10 mm), external iliac (9

mm) and internal iliac (10 mm) vessels, drawn in continuity with a 12 mm expansion

anterior to the sacrum and a 22 mm expansion medial to the pelvic sidewall, as shown

in Figure 2.4. The union of the major artery and vein contours should be used to form

the nodal CTV, and the resultant volume trimmed to exclude bone, muscle, fascia,

bladder and bowel, consistent with other recommendations. While population-based

models facilitate treatment planning, it is important to carefully review all relevant

imaging for each patient and adapt the nodal CTV contours to encompass overt

lymphadenopathy or nodal outliers.

There are at least two other descriptions of pelvic lymph node topography for

radiotherapy treatment planning derived from MR imaging of ferumoxtran-10169, 170. Shih

et al. mapped the center of mass of lymph nodes metastases in 18 men with prostate

cancer to a standardized three-dimensional model using anatomic landmarks170. A 2 cm

uniform margin of expansion was found to encompass 94.5%

67

(a) (b)

(c) (d)

(e) (f)

68

Figure 2.4: Pelvic nodal CTV at six cranio-caudal levels in the pelvis developed using

our population-based model of lymph node tomography. The vascular, sacral and pelvic

sidewall expansions were trimmed to exclude bone, muscle, fascia, bladder and bowel.

(a) Distal para-aortic, (b) Bifurcation of the aorta and inferior vena cava, (c) Inferior to

the bifurcation of the common iliac vessels, (d) Mid-pelvis (e) Cranial to the acetabulum,

and (f) Mid-acetabulum.

69

of the lymph nodes thought to be at risk of harboring disease, assuming a fixed nodal

diameter of 1 cm. This margin of expansion probably is excessive based on our analysis

of the spatial distribution of lymphatic tissue in 55 patients, and would unnecessarily

limit the utility of IMRT to spare adjacent normal tissues. Taylor et al., in a series of 20

women with gynecologic tumors, found a 7 mm vessel expansion to be optimal provided

that the contours were manually modified using specified rules to include potential nodal

locations at greater distances from major arteries and veins169. Despite the apparent

difference in the proposed margins of expansion between this study and ours, the

derived nodal CTV’s are actually quite similar: The manual contour modifications that

are implicit in the Taylor et al. CTV definition, in effect, further expand their 7 mm vessel

margins so that the final volumes are relatively coincident. Our approach may be less

time consuming, and less prone to error, because less revision is required for individual

cases.

The greatest limitation of this and previous studies is the ability of MR imaging with

ferumoxtran-10 to identify small lymph nodes <3 mm in size, and microscopic lymph

node metastases124, 174. The nodes comprising the presacral and internal iliac groups

are likely underrepresented by virtue of their size. Taylor et al. reported the frequency of

lymph node contours and a summation of consecutive contours through successive

axial images is required to determine the lymph node numbers identified for each lymph

node group specified169. Within their 20 patient series, a mean of 0.4 (0-2) and 7.2 (1-

22) lymph node contours were identified for the presacral and internal iliac lymph node

groups respectively. In comparison, a median 2 (0-6) and 5 (0-15) presacral/subaortic

and internal iliac lymph nodes were identified in our 55 patient series. Recognizing that

70

the presacral lymph node group in our series included subaortic lymph nodes and that

nodes within each group may lie across more than one adjacent axial plane, there

would appear to be an equivalent or greater number of lymph nodes identified in our

series with a larger sample size. Our model for nodal CTV definition was derived from

the spatial distribution of visible lymph nodes, on the assumption that these would be

representative of all nodes at risk of harboring disease. It is possible that there are small

lymph nodes at greater distances from the major vessels that were not visualized using

this approach. However, these are likely to represent a small proportion of the total

nodal volume in each anatomic region and would not substantially alter the proposed

margins of expansion, which were defined to encompass 90% of the lymphatic tissue in

90% of patients. Ferumoxtran-10, while lowering the MR threshold for the detection of

subclinical disease, does not enhance the sensitivity enough to reliably detect

microscopic metastases174. This currently limits the utility of this approach for defining

patient-specific radiotherapy volumes that target only regions of known tumor. However,

innovative cellular MR techniques using magnetic nanoparticles have been shown in

preclinical studies to detect as few as 100 malignant cells in lymph nodes, and hold

great promise for the future175.

71

2.6 Conclusion

In conclusion, the use of MR lymphography provides an objective, three-dimensional

spatial description of relevant pelvic lymph nodes in relation to easily visualized

anatomic landmarks. Radial 3D margins of expansion around the major pelvic vessels,

with the inclusion of separate margins for the sacral and medial external and internal

iliac lymph nodes, can be used to develop a population-based nodal CTV for

radiotherapy treatment planning. The use of this model in clinical practice will assure a

high probability of encompassing most of the nodal tissue at risk of harboring

metastases in most patients, while minimizing the dose to normal tissues.

72

Chapter 3. Inguinal Lymph Node Topography for Radiotherapy

Treatment Planning from Contrast-Enhanced MR Imaging

3.1 Introduction

Primary and adjuvant inguinal radiotherapy is employed in the management of

genitourinary, gynecologic and gastrointestinal malignancies. Intensity modulated

radiotherapy may permit the delivery of an escalated dose to the lymph nodes at risk of

occult metastatic spread while potentially reducing the probability of toxicity such as

lymphedema and fibrosis of normal tissues.

Adequate delineation of a nodal clinical target volume (CTV) for inguinal lymph node

irradiation is predicated upon sound knowledge of lymphatic anatomy. Unfortunately,

the location and depth of the inguinal lymph nodes has not been well defined and

inguinal nodal radiotherapy volumes are traditionally related to anatomic surrogates

such as bony landmarks or the femoral vasculature176, 177. A potential inability to

accurately define and encompass the inguinal lymphatic CTV may result in the delivery

of an inadequate dose and a geographic miss resulting in a failure to eradicate clinically

evident or subclinical disease178. Koh et al. demonstrated the importance of adequately

defining the depth of the inguinal lymph nodes in a re-analysis of patients with inguinal

nodal recurrence following treatment for vulvar carcinoma179. The use of an absolute

depth to represent the position of the inguinal nodes without regard to an individual

patient’s weight and body habitus resulted in a failure to adequately encompass the

lymph node CTV within the high dose radiation volume and subsequent high risk of

73

inguinal node failures. From that analysis, an average patient would require a

prescription depth of 6.0 cm to adequately encompass the inguinal lymphatics. However

as demonstrated by the significant range observed (minimum mean 6.1 cm and

maximum mean 17.3 cm), this depth would need to be modified for a given individual to

respect patient specific anatomic factors. A poor understanding of inguinal lymph node

target location relative to other anatomic landmarks affected this study178, 180, 181. A

related follow-up study analyzing 31 women with cervical, vaginal or vulvar carcinomas

revealed the mean depth of the most superficial inguinal lymph node to be 2.6 cm and

the mean depth of the deepest to be 5.8 cm181. These depths were dependent on body

habitus and further support the need for a patient specific plan as opposed to an

arbitrary prescription depth.

CT based treatment planning has the ability to identify the vessels within an individual

patient avoiding the potential for over or under estimating the depth of the femoral

vessels. However, CT lacks sufficient soft tissue resolution to depict the small clinically

normal lymph nodes of interest and little is known about the precise anatomic location of

the inguinal lymph nodes relative to the femoral vessels and other adjacent anatomic

structures. While knowledge of nodal topography is still incomplete, it has been known

for some time that the spatial relationship between the lymphatics and the vasculature is

one of high anatomic reproducibility and fidelity7, 19, 55. Wang et al. plotted the

topographic distribution of clinically palpable gross inguinal lymphadenopathy on a

representative fluoroscopic simulation film to derive field boundaries to encompass the

inguinal lymphatics176. While this may be suitable for two-dimensional planning, it does

not adequately convey sufficient information to generate an acceptable patient specific

74

three-dimensional radiotherapy treatment plan. For conformal and intensity modulated

radiotherapy no validated rationale for the generation of an inguinal nodal clinical target

volume is yet available.

The objective of this study was to develop a robust and generalizable CTV definition for

inguinal lymph node irradiation based on an isotropic expansion of vascular anatomy to

encompass nodal frequency and topography using MR lymphography with ferumoxtran-

10 enhanced magnetic resonance images of the lower pelvic and inguinal region. Based

on an analysis of this data, we provide a practical working CTV definition that can be

applied to patients with gynecologic, genitourinary or gastrointestinal tumors who are

deemed to be at high risk of harboring occult inguinal lymph node metastases.

3.2 Materials and Methods

This was a single center pilot study eligible to patients with histologically confirmed

endometrial, cervix, prostate (cT1-cT3) or bladder cancer with no distant metastases.

To ensure uniformity of the sample and to control for artifacts and lymphadenopathy, all

subjects were screened to exclude: men with histologically confirmed carcinoma of the

prostate and a PSA < or = 4.0 or > 100; evidence of distant metastasis (M1); radical

surgery or prior cryosurgery for carcinoma of the prostate, previous radiation, hormonal

manipulation or chemotherapy; patients with biopsy-proven lymph node involvement;

women with histologically confirmed carcinoma of the endometrium or cervix who have

undergone recent pelvic surgery with the exclusion of colposcopy, dilatation and

curettage or biopsy; men or women with histologically confirmed carcinoma of the

75

bladder who have undergone recent pelvic surgery with the exclusion of TURBT;

previous or concurrent cancers other than superficial basal or squamous cell skin

carcinoma unless disease free for at least 5 years; or a hip prosthesis. MR imaging

remained uniform with no changes in equipment or technique over the duration of the

study period.

3.2.1 Patient characteristics

From August 2004 to June 2006, 30 patients were prospectively enrolled in this study

with biopsy proven primary carcinoma of the prostate (20), bladder (5), endometrium (3)

or cervix (2). Patients who had undergone prior chemotherapy, radiotherapy or

pelvic/inguinal lymph node dissection or hernia surgery were excluded. None of the

patients analyzed for this study had clinically or radiographically detectable inguinal

nodal metastases. The median age was 64 years (range 39-82). There were 25 men

and 5 women. All patients were imaged prior to undergoing therapy for their cancer. The

local Research Ethics Board approved the study and all of the patients who participated

provided written informed consent.

3.2.2 MR imaging

MR imaging was carried out using a 1.5 T magnet (GE EXCITE 1.5 T, Waukesha, WI,

USA) with a four-channel pelvic phased-array coil. Patients were instructed to take

nothing by mouth four hours prior to the study.

Axial imaging sequences were obtained at 3 mm intervals through the pelvis before and

24-36 hours after contrast administration. Pulse sequences used consisted of a T2w

76

FSE (TR/TE 6500/80; field of view 24-28 cm), and a T2*w dual echo GRE (TR/TE1/TE2

1400/14/21ms; field of view 26-28 cm). Motion artifacts from peristalsis were minimized

with the administration of hyoscine-N-butylbromide (Buscopan) 30 mg or glucagon 1mg

i.v. or i.m. prior to each of the MR imaging sessions.

3.2.3 Ferumoxtran-10

Ferumoxtran-10 (Combidex, AMAG Pharmaceuticals; Lexington, MA) was reconstituted

from a lyophilized powder. A dose of 2.6 mg Fe/kg was dissolved in 100 ml of normal

saline (0.9% NaCl) and infused intravenously.

3.2.4 Image analysis

Image analysis was performed on archived DICOM images using a PACS radiology

workstation (OsiriX image processing software, Version 2.7.5; OsiriX Foundation,

Geneva, Switzerland)182 and 3D modeling and image processing software (3D-

DOCTOR®: Able Software Corp., Lexington, MA). A single observer (R.D.) manually

delineated the lymph nodes, inguinal and pelvic vasculature, as well as the skin and

musculoskeletal landmarks. Lymph node size evaluation was undertaken using the T2w

FSE image series as opposed to the T2*w dual echo GRE which is subject to USPIO

induced susceptibility artifacts. To verify the identification and anatomic accuracy of the

manual operator (R.D.) defined regions of interest (vascular segments and lymph

nodes) an expert abdominal-pelvic radiologist (M.H.) reviewed 10 (33%) of the image

datasets. No noted variation between the observers was apparent with regards to either

nodal identification and delineation or vessel segment identification. Points defining the

edge of each distinct vascular and nodal element were identified in the x- and y- planes

for each axial slice (z-plane)183.

77

3.2.5 Vascular segmentation

All image sets were reviewed with partitioning of the image data to define the vascular

and lymphatic structures of interest. Manual tracing of the vessel walls was undertaken

using continuous sampling with vertices recorded along the path length. A closed spline

curve was then generated by connection of the vertices. Object specific boundaries on

each axial slice were integrated through the series to define a three-dimensional

volume. All points (x, y, and z) at the boundary edge and those pixels enclosed by these

points were considered to be part of the object. Those points and pixels not part of a

target structure were considered to be background. Window width and window level

were standardized to ensure optimization of contrast and brightness.

Segmentation of the inguinal arterial and venous system was performed in a cranio-

caudal direction beginning from the external iliac artery and vein at the level where the

inguinal ligament traversed the medial aspect of the artery. At bifurcations, the first

section that did not contain any part of the "mother vessel" was described as the

beginning of the "daughter vessel ". The T2w FSE, T2* GRE and T1w image series

were reviewed for each patient to identify the points of bifurcation. Segmentation of the

venous system was performed in an analogous manner. The vessels of interest

(femoral artery and vein, superficial circumflex iliac vein, great saphenous vein,

superficial external pudendal vein and the superficial epigastric vein) were delineated

from their origin at the point of the passage of the external iliac artery and vein under

the inguinal ligament superiorly and distally to the inferior border of Scarpa’s triangle

defined by the adductor longus and sartorius muscles on both the right and left sides7.

78

Vessel topology, course, segment length, and bifurcation location varies within and

between individuals. Bifurcations of the iliac vessels on the right and left sides do

demonstrate varying degrees of spatial asymmetry. Similarly, the vessel bifurcation

angle may occur through a broad range. Vessel courses, lengths, curvatures, and

branching angles will all impact upon object appreciation and segmentation accuracy.

Given that this limitation exists within current clinical radiotherapy treatment planning, its

presence within this study will not limit the clinical applicability of the results.

3.2.6 Lymph node nomenclature and segmentation

After administration of ferumoxtran-10, lymph nodes were identified by a decrease in

signal on the gradient-echo images, which was more marked at the longer TE.

Localization was aided by combined review with T2 images. The visualized lymph

nodes were identified and divided into right and left sides as illustrated in Figure 3.1. All

of the lymph nodes lying within anatomic volume defined by the inguinal ligament, the

adductor longus muscle and the sartorius muscle, were segmented as separate

structures.

Prior to analyzing the patient data, our approach was validated using the

photomicrographs of the adult human male (Visible Human Male – VHM) from the

Visible Human Project, National Library of Medicine20. The VHM dataset is comprised of

high resolution, 1 mm, anatomic, axial images, and allows lymph nodes as small as 1 or

2 mm in size to be identified reliably. Lymph nodes in this size range may not be seen

even with ferumoxtran-10 enhanced MR images of the study subjects, and the VHM

was therefore used to assure that the nodal distribution maps and the nodal CTV

79

definition derived from the patient MR image sets were truly representative of the nodal

anatomy.

3.2.7 Quantitative distance histogram analysis of lymph node volume

The x-y-z coordinates of the segmented structures (vascular and lymphatic) were used

to compute their spatial relationships. The boundary descriptors for the pelvic

vasculature and lymphatics were compiled in one data file for each study subject and

analyzed. Each lymph node in every image set was divided into nodal volume elements

measuring 0.5 (x-axis) x 0.5 (y-axis) x 3 (z-axis) mm in dimension using an automated

algorithm in order to generate a complete 3D spatial representation of all of the

lymphatic tissue. Paired 3D distances were determined from the centre of each nodal

volume element (representing 0.75 mm3 of lymphatic tissue) to the closest vascular

edge (arterial or venous) for each vascular segment across all axial planes (Figure 3.2).

80

Figure 3.1: (A) A three-dimensional image rendering of the contoured lymph nodes,

vasculature and bony anatomy derived from the VHM with a representative image

adjacent. (B) three-dimensional rendering of pelvic and inguinal vasculature (1 inguinal

ligament, 2 sartorius muscle, 3 adductor longus muscle, 4 femoral artery and vein, 5

superficial circumflex iliac vein, 6 great saphenous vein, 7 inferior epigastric vein and

the 8 external pudendal vein) and (C) the corresponding lymph nodes.

81

Anisotropic and isotropic histograms describing the distribution of distances from the

centre of each nodal volume element to the nearest vessel edge for varying bin sizes

were generated for each vascular segment to illustrate the spatial distribution of nodal

tissue in relation to the associated vessels. Scatter plot displays of the nodal volume

elements were plotted on a Cartesian coordinate system where the intersection of the

abscissa and ordinate (origin) was defined as the collapse surface of the vessels

comprising the vascular segments of interest. Summary statistics were derived from the

histograms to determine the vascular expansions required in each vascular segment to

encompass 50%, 80%, 90%, 95%, 99% and 100% of the lymphatic tissue. From this,

uniform (isotropic) vascular margins of expansion were generated. These margins were

then applied to the vessels to produce an inguinal nodal CTV delimited by the adjacent

musculature, body of the pubis and skin.

3.3 Results

On each of the 1 mm axial macrotomal sections of the Visible Human Male dataset, the

inguinofemoral region bounded by the inguinal ligament superiorly, medially, and

laterally by the adductor longus and sartorius muscles were delineated. One hundred

and sixty one sections (16.1 cm) encompassed this volume. On the right, 8 discrete

inguinal lymph nodes were identified with a range in axial size of 3 to 16 mm (mean 6)

and on the left, 17 separate lymph nodes were identified with a range in greatest axial

size of 2 to 14 mm (mean 5) (Figure 3.1). The total volume of nodal tissue on the right

was 2690.3 mm3 and 3114 mm3 on the left.

82

(A)

(B)

Closest distanceof each nodal volumesegment to an arteryor vein

(A)

(B)

Closest distanceof each nodal volumesegment to an arteryor vein

Figure 3.2: (A) A three-dimensional surface rendering of the contoured left inguinal

lymph nodes with a schematic representation of the method used to determine the

spatial distribution of lymphatic tissue inset. (B) The distance from the center of each

lymph node volume unit (green) to the closest artery (red) and vein (blue) was

calculated in three dimensions.

83

Following delineation of the lymph nodes in the VHM, the right and left femoral artery

and vein, superficial circumflex iliac vein, great saphenous vein, superficial external

pudendal vein and the inferior epigastric vein were identified and contoured. The

distances from each vessel to the lymph nodes in three-dimensions were analyzed

individually for a given vessel and in combination with the other vessels to determine

which vessel(s) were most closely related to the adjacent lymph nodes (Table 3.1). The

femoral artery and vein, superficial circumflex iliac vein and the great saphenous vein

had the most clinically relevant associations to the adjacent lymph nodes while the

superficial external pudendal vein and the inferior epigastric vein had the least relevant

ones (data not shown).

(A)

84

Table 3.1: Vessel combinations and the spatial distribution of lymph node tissue in

relation to the closest vascular segment

Distance of lymphatic tissue in mm from nearest vessel to

encompass a given percentage of the identified lymph node (LN)

volume (Vol.)

Vessels 50% of

LN Vol.

80% of

LN Vol.

90% of

LN Vol.

95% of

LN Vol.

99% of

LN Vol.

100% of

LN Vol.

F AV 13.9 17.1 18.8 20.4 25.0 28.0

F AV

GS V

11.1 15.3 17.6 20.1 25.0 28.0

F AV

SCI V

9.5 15.1 16.9 18.2 20.0 22.8

F AV

GS V

SCI V

7.4 10.8 12.7 14.3 16.6 22.3

F AV: femoral artery and vein; GS V: Great Saphenous Vein; SCI V: Superficial

Circumflex Iliac Vein.

85

The high resolution (4096 x 2048 x 24 bits) color images of the axial anatomic images of

the VHM afford an opportunity for a detailed study of lymphovascular anatomy20.

Current clinically available MR and CT imaging is not able to provide this level of

anatomic detail and resolution. To approximate what may be imaged in current clinical

care, those lymph node contours, which were 3 mm or less in diameter were censored

and the remaining lymph nodes were analyzed to yield distances of 15.2, 18.0, 19.6,

21.4, 25.2 and 27.7 mm to encompass 50%, 80%, 90%, 95%, 99% and 100% of the

inguinal lymph nodes. Censoring those lymph nodes, which were greater than 3 mm in

diameter, resulted in distances of 11.7, 14.7, 16.6, 18.3, 23.7 and 28.0 mm. The

analysis of this dataset demonstrates the feasibility of using identifiable vascular

surrogates to define nodal positions and those lymph nodes which may be at the

threshold of detection or fall below it do not appear to adversely impact upon the lymph

node to vessel distance data obtained. Those nodes that are larger than 3 mm in

diameter (detectable with the current spatial resolution of clinically available MR

scanners) are related closely enough to those less than or equal to 3 mm so that the

smaller nodes would be encompassed by the values defined using the larger nodes.

Based upon the findings from the VHM analysis, the ferumoxtran-10 data were

analyzed (Figure 3.3). For the thirty patients accrued from August 2004 to June 2006, a

mean of 8 and median of 8 inguinal lymph nodes were identified within the volume

encompassing Scarpa’s triangle on both the right and left side (Range: 3-16 right and 2-

13 left) with a mean and median depth of 26.7 and 26.4 mm (9.4-49.5) on the right and

27.7 and 27.3 mm (8.1-52.9) on the left. There was no significant difference in the

number and distribution of inguinal lymph nodes between the left and right sides. The

86

mean and median short axis diameters of the lymph nodes were 5.7 and 6 mm (range

3-12). Tables 3.2 to 3.4 depict the distance of the nodal volume elements from the

closest vessel edge using both isotropic and anisotropic analyses. The effect of the

inclusion of the superficial circumflex iliac vein upon the distribution of the nodal volume

elements in relationship to the surface of the nearest vessel is illustrated in Figure 3.4.

With the inclusion of the superficial circumflex iliac vein, a reduction in the isotropic

margin of vascular expansion is achieved by better capturing the true anatomical spatial

relationships of the lymph nodes to the corresponding vessel. Using only the femoral

artery and vein, the nodal tissue is distributed in a hemispheric fashion anterior to the

vessels with 90% of this tissue lying within 27.7 mm of the vessels. The congruence in

the distribution on the scatter plot with the adjacent anatomic rendering demonstrates

the absence of visible lymph nodes posterior to the femoral vessels (Figure 3.4).

87

Figure 3.3: Histograms derived from the Visible Human Male illustrating the spatial

proximity of nodal tissue in relation to the closest edge of the nearest artery or vein for:

a) the femoral artery and vein; b) femoral artery and vein and great saphenous vein; c)

femoral artery and vein and superficial circumflex iliac vein and d) femoral artery and

vein, great saphenous vein and superficial circumflex iliac vein.

88

Table 3.2: Spatial distribution of lymph node tissue in relation to the closest vascular segment (Femoral Artery and Vein) derived from the 30 patient series

Distance of lymphatic tissue in mm from nearest vessel to encompass a given percentage of the identified lymph node (LN) volume (Vol. )

Percentile 50% of LN Vol.

80% of LN Vol.

90% of LN Vol.

95% of LN Vol.

99% of LN Vol.

100% of LN Vol.

10th 9.7 12.8 14.3 15.3 17.1 17.9

20th 10.2 13.8 15.2 16.1 18.2 19.3

30th 10.8 14.2 16.1 18.4 20.1 21.4

40th 12.1 15.2 18.2 21.3 23.1 24.6

50th 13.5 17.3 20.4 21.7 24.7 26.2

60th 14.2 18.8 21.1 24.2 25.8 26.9

70th 15.4 20.1 22.3 25.3 27.5 29.2

80th 16.9 22.0 23.7 29.6 31.6 34.3

90th 18.8 23.7 27.7 35.4 37.8 39.5

100th 21.4 27.1 38.5 41.3 44.0 45.3

89

Table 3.3: Spatial distribution of lymph node tissue in relation to the closest vascular segment (Femoral Artery and Vein and Superficial Circumflex Iliac Vein) derived from the 30 patient series

Distance of lymphatic tissue in mm from nearest vessel to encompass a given percentage of the identified lymph node (LN) volume (Vol. )

Percentile 50% of LN Vol.

80% of LN Vol.

90% of LN Vol.

95% of LN Vol.

99% of LN Vol.

100% of LN Vol.

10th 6.4 9.4 10.9 12.2 14.1 14.8

20th 6.8 9.9 11.9 14.4 16.2 17.5

30th 7.8 11.7 13.4 15.3 17.9 19.0

40th 8.2 12.7 14.9 17.1 19.3 20.7

50th 8.7 13.7 16.0 17.6 21.1 23.3

60th 9.1 14.8 17.1 20.0 22.2 24.3

70th 10.2 15.6 18.5 20.6 23.8 26.0

80th 10.8 17.0 19.2 21.9 24.8 26.6

90th 11.7 17.8 20.5 22.5 26.9 28.7

100th 13.7 18.9 21.0 22.7 37.7 40.4

90

Table 3.4: Anisotropic analysis of the spatial distribution of lymph node tissue in relation to the closest vascular segment derived from the 30 patient series

Distance of lymphatic tissue in mm from nearest vessel to encompass a 90% of the identified lymph node (LN) volume (Vol. ) in 90% of patients

Vessels ANT MED ANT ANT LAT POST LAT POST POST MED

F AV 21.1 24.3 32.1 0 0 10.1

F AV

SCI V

15.3 20.2 19.7 7.9 8.7 13.1

F AV: femoral artery and vein; SCI V: Superficial Circumflex Iliac Vein.

91

Inclusion of the superficial circumflex iliac vein, which does have lymph nodes related to

its posterior aspect, results in a significant change in the distribution of the lymph node

volume elements around the vessels’ surfaces. While the bulk of the elements are still

situated antero-laterally, points are now apparent posterior to the surface effecting a

partial shift to a more circumferential distribution closer to the vessel surface (coordinate

(0,0) on the scatter plot). The radial space around the vessel surfaces was divided into

six equal parts and analysis of the distribution of the lymph node volume elements was

undertaken to determine the anisotropic distribution of the lymph nodes. Table 3.4

depicts the margin of expansion to encompass 90% of the lymph nodes in 90% of the

patients for each sextant.

92

Figure 3.4: (A). A frontal view of a three-dimensional surface rendering for a typical

patient with the left femoral artery and vein and inguinal lymph nodes delineated and (B)

the corresponding scatter plot distribution of the lymph node volume elements in relation

to the point collapse surface of the femoral artery and vein. (C) A frontal view with the

left femoral artery and vein and superficial circumflex iliac vein delineated and (D). the

corresponding scatter plot distribution of the lymph node volume elements in relation to

the point collapse surface of these vascular segments. Note the antero-radial

distribution of the lymph node volume elements in relationship to the femoral vessels

and the effect of the inclusion of the superficial circumflex iliac vein on the distribution.

93

3.3.1 Nodal CTV for adjuvant inguinal radiotherapy

The main objective of this study was to define a set of practical clinical rules for

determining a lymph node CTV in relation to easily visualized anatomic surrogates of

lymph node position. The margins of expansion to define the inguinal nodal CTV were

based on coverage of 90% of the nodal tissue in 90% of patients and represent a

compromise between nodal coverage and the avoidance of excessive normal tissue

irradiation (Figure 3.5)183.

An inguinal nodal CTV may be delineated by:

1. Segmenting the femoral artery and vein within Scarpa’s triangle;

2. Segmenting the superficial circumflex iliac vein within Scarpa’s triangle if a more

refined nodal CTV with greater potential sparing of adjacent tissues is desired;

3. Segment the adjacent anatomic land markers defining Scarpa’s triangle;

4. Expand the vessel contours by 3 cm if only the femoral vessels are used or 2 cm

if both the femoral and circumflex vessels are used;

5. Revise or delimit the expanded volume to exclude the adjacent musculature and

those contours extending inferior to the region within which the sartorius muscle

crosses the adductor longus muscle; and

6. Visually inspect the volumes to ensure inclusion of any potential enlarged lymph

nodes or outliers.

94

Figure 3.5: Inguinal nodal clinical target volume at twelve cranio-caudal levels in the

groin developed using our population-based model of lymph node topography. The

vascular expansions were trimmed to exclude bone, muscle, and skin. Legend:

Femoral arteries – red; Femoral vein – dark blue; Superficial circumflex iliac vein –

light blue; Sartorius muscle – brown; Adductor longus muscle – khaki and lymph

nodes – green.

95

3.4 Discussion

In spite of improvements in anatomic imaging and the use of novel radiologic contrast

agents, our knowledge of lymphatic anatomy and ability to identify lymph nodes in vivo

requires further improvement to more accurately identify occult nodal metastases and

delineate patient specific patterns of loco-regional spread. The inguinal lymphatics are

sites of potential spread for some gynecologic, genitourinary and gastrointestinal

malignancies. With the use of existing anatomic and surgical data to provide parameters

for a model, imaging information can be employed to generate a three-dimensional

probability atlas for nodal location in relation to adjacent vessels and other anatomic

structures. This model provides a means of defining a class solution for radiotherapy

treatment volume delineation.

The inguinal lymph nodes are situated inferior to the inguinal ligament and are divided

into superficial and deep components. The superior superficial inguinal nodes are

distributed along the line of the inguinal ligament and lie anterior to the femoral fascia

and deep to the subcutaneous fascia while the superficial inferior nodes are situated

vertically along the proximal aspect of the great saphenous vein prior to its union with

the femoral vein at the fossa ovalis (Figure 3.1). The number of superficial nodes varies

with a range of 4 to 13 reported in the literature7. The deep inguinal lymph nodes are

situated posterior to the fascia lata within the fossa ovalis (fossa ileopectinea) and lie

medial to the femoral vein within the femoral canal. Typically 1-3 lymph nodes are

identified within this region7, 49.

96

Non-lymphatic anatomic structures, which can be related to the boundaries of the

inguinal region and to the inguinal nodes, are of benefit in defining and refining a nodal

CTV. In an attempt to minimize the morbidity of inguinal lymph node dissections,

several studies have assessed the inferior, lateral, and superior extents of the deep and

superficial lymph nodes184, 185. Borgno et al. reviewed the distribution of the deep

inguinal lymphatics and noted that 33% of these nodes were situated inferior to the

junction of the saphenous vein with the femoral vein and all of the nodes fell within the

boundaries defined by the fossa ovalis185. No lymph nodes were identified inferior to the

most distal margin of the fossa ovalis. Additionally, no deep inguinal lymph nodes were

identified lateral to the femoral artery. The superficial inguinal lymphatics have been

studied in order to limit the lateral extent of the surgical field required for an inguinal

lymph node dissection186. These nodes lie within the superior aspect of Scarpa’s

triangle and are closely related to the superficial circumflex iliac vasculature. As a

consequence of this close relationship, the lymph nodes are not typically found lateral to

the medial aspect of the sartorius muscle superior to the point at which it crosses the

lateral aspect of the inguinal ligament (Figures 3. 1 and 3. 2)184, 186, 187. Similarly, the

posterior most extension of the inguinal lymph nodes have been shown not to extend

beyond the point at which the sartorius muscle crosses the adductor longus muscle in

the anatomic position (Figures 3. 1 and 3. 5).

The smallest short axis distance of the lymph nodes delineated in the VHM dataset was

2 mm. While this is smaller than the slice thickness (3 mm) used in this study and raises

concerns about identifying these lymph nodes using MRI, the close relationship of these

lymph nodes to the adjacent femoral vessels would result in them being encompassed

in a nodal CTV generated using a margin of expansion derived on the basis of larger

97

visible nodes around these vessels. The adjacent lymph nodes which are of a larger

size than the slice thickness used and are apparent on the MRI images can help to

define the outer boundaries for an anatomic envelope to encompass the inguinal nodal

basin within which the smaller (<3 mm) lymph nodes may reside.

The lymphatic system arises embryologically from the venous system with a strong

anatomic and spatial relationship between the two. The topographic distribution of the

lymph nodes is influenced by the adjacent nonvascular anatomy, which influences the

spatial distribution of the lymph nodes in relation to the adjacent vessels. While isotropic

margins of vascular expansion afford a sufficiently accurate and clinically practical

means of lymph node CTV delineation, they are an over-simplification. The anisotropic

distribution reflects the impact of the musculature, bony and integumentary anatomy on

the location of the lymph nodes and shows their close association with the vessels. A

dynamic adaptive margin of expansion to account for variations in nodal distribution

through a consecutive series of axial planes would be required to fully exploit these

anatomic relationships for clinical use. An adequate approximation for this can be

obtained with an isotropic margin of expansion limited by adjacent structures. By

delimiting the isotropic nodal expansion to avoid inclusion of the muscles surrounding

the inguinal region the superior to inferior changes can be accounted for. With the aid of

magnetic resonance lymphography to facilitate identification of the lymph nodes, the

spatial relationship to the adjacent vessels has been determined and the expansions

around the vessels necessary to encompass a specified percentage of these nodes

have been quantified (Tables 3.1 and 3.2). This knowledge may be useful in

radiographic staging and in surgical as well as radiotherapy treatment planning to

facilitate nodal localization.

98

The use of conformal and intensity modulated radiotherapy enables sparing of adjacent

normal tissues. This potential can be further optimized by using the adjacent

musculature to limit an isotropic margin of expansion to produce a nodal CTV that

encompasses all of the inguinal lymph nodes while eliminating the inclusion of adjacent

normal structures within the CTV (Figure 3.5). While further studies are required to

better delineate nodal patterns of failure, the nodal CTV generated using a modified 2

cm isotropic margin of expansion around the femoral artery, vein and the superficial

circumflex vein provides a volume, which is consistent with that proposed by others

while facilitating the potential for normal tissue avoidance 176, 188.

3.5 Conclusion

In conclusion, the use of MR lymphography has facilitated the construction of an

objective, three-dimensional spatial description of relevant inguinal lymph nodes in

relation to easily visualized anatomic landmarks. Radial 3D margins of expansion

around the femoral artery, vein and the superficial circumflex iliac vein were used to

develop a population-based inguinal nodal CTV for radiotherapy treatment planning.

The use of this model in clinical practice will assure a high probability of encompassing

most of the nodal tissue at risk of harboring metastases in most patients, while

minimizing the dose to normal tissues

99

Chapter 4. General Discussion

Prior to the development of lymphography, no clinically practical means of lymph node

delineation for radiotherapy treatment was available. With the entry of oil-based

lymphography into clinical practice, lymphograms were used to identify lymph nodes at

high risk of harboring metastatic disease, which advanced the treatment of

genitourinary, gynecologic, gastrointestinal and other malignancies156, 163, 189-200.

However, the sensitivity of clinical lymphography for the detection of microscopic

disease is markedly inferior to surgical dissection and histologic examination. Non-

contrast and to a significant extent contrast-enhanced x-ray radiographs of this era were

unable to fully visualize relevant soft tissue anatomy including the pelvic organs, the

vasculature as well as the lymphatics. Therefore, target localization for radiotherapy

treatment planning was incomplete and imprecise, necessitating that large volumes of

normal tissues be treated to avoid missing regions of disease. In the absence of a

clinically feasible means for accurate vascular or lymphatic identification, the wealth of

anatomic information obtained up to this point in time could not be adequately translated

to the two-dimensional treatment planning radiographs nor could it be leveraged in the

routine provision of clinical care.

The development and use of non-contrast and contrast enhanced CT and MRI and the

availability of multi-planar cross-sectional imaging has enabled three-dimensional

radiotherapy treatment planning. Multi-slice imaging with CT and MRI are increasingly

able to generate three-dimensional visualizations of greater and greater anatomic detail.

While there has been an improvement in the level of resolution that can be appreciated

100

with these modalities, they remain sub-optimal in the identification of small lymph nodes

in the pelvis and groin as well as in the detection of microscopic lymph node

metastases201, 202. Functional imaging with PET scanning has the potential to improve

the detection of clinically occult lymph node metastases to some degree but remains

insensitive to micrometastatic disease203. With further improvements in each of these

modalities, the identification of anatomical features below the current limits of resolution

will become a reality and rival the detailed anatomic information that can be obtained

through surgical dissection. Similarly, the continued development of novel

lymphotrophic contrast agents augurs a new era of lymphatic target definition. Systems

of anatomic knowledge management will become increasingly important to ensure that

the emerging information is utilized to its maximal potential for radiotherapy treatment

planning as well as other endeavors. Chapters 2 and 3 outline the most comprehensive

quantitative assessment of the pelvic and inguinal lymphatics to date. An overview of

the methods that have been used in the past to record and communicate anatomic

information provides a foundation for new approaches to translating these findings into

clinical practice.

The origins of anatomical study can be traced back to ancient Egypt. The Ebers

Papyrus of 1550 BC provides a written record of ancient Egyptian medicine that

provides a description of the heart and its vessels204, 205. This represents one of, if not

the first, recorded symbolic representations of human anatomy. Following this period,

the later works of Hippocrates (460-377 BC) and Galen (129-200/217 AD) document the

medical knowledge within ancient Greek and Roman society in both symbolic and

pictorial forms206, 207. Both Hippocrates and Galen have had a profound and lasting

101

influence on Western medical thought. Roman law, during the time of Galen, forbade

human dissection and anatomic investigations were primarily limited to the study of non-

human primates, pigs and other mammals. Galen’s works would later be supplanted by

those of Andreas Vesalius (December 31, 1514 - October 15, 1564). Vesalius, unlike

Galen, faced no similar constraints and undertook careful and methodical observational

studies of human anatomy, which resulted in the publication of De humani corporis

fabrica208. Galenic anatomy, because of the limitations during that period, contained

many inaccuracies. Through a fastidious focus on detail and iterative refinements,

Vesalius was able to refute the errors in Galenic teachings and raise the standard for

medical works with the creation of his anatomic atlas207.

This initial atlas and subsequent ones provided a means of communicating knowledge

through pictorial representations as opposed to text alone. Accurate visual depictions of

human anatomy can convey the classification of spatial objects, their structural

networks (topological boundaries), the relevant part-of networks (component parts of

objects of interest) and the spatial associations between these objects and adjacent

related or unrelated objects209. Figure 3.1C depicts a three-dimensional rendering of the

arterial and venous vasculature of interest in addition to the lymphatics and relevant

musculoskeletal anatomy that readily conveys the relationships and topographic

boundaries among these structures.. The component parts are similarly appreciable in

the vasculature and its bifurcations. Anatomy atlases can facilitate the generation of

mental constructs through a representation of reality that acts to record and convey

spatial and structural complexity 210-215 and are fundamental cornerstones of

understanding and teaching. Continued revision, the incorporation of new knowledge

102

and improvements in the manner of its representation help these atlases to be better

utilized for research and clinical care delivery.

Paper, derived from the Greek for papyrus - the medium used by the ancient Egyptians,

affords primarily a two-dimensional means to record anatomical information204, 205. Early

three-dimensional anatomic representations were undertaken as evidenced by the wax

models created under the supervision of Paolo Mascagni216-218. While of high anatomic

fidelity, they were limited as a means of conveying the individual anatomic vagaries that

are present within a population and were both labor and cost intensive. More recently,

advances in computer based medical imaging and image analysis have provided a

means of developing three-dimensional digital anatomic models. These virtual atlases

have been generated for a variety of anatomic sites based on representative individuals,

as well as for populations219-222.

Neuroanatomy and functional mapping have utilized atlases with defined coordinate

systems using data derived from a carefully studied individual221. Relating features

known for a “representative” individual to a particular subject of interest is undertaken

via application of a defined set of proportionality rules using a common reference space.

For anatomic structures that do not demonstrate significant inter-subject variability, and

for which any variability that is present can be addressed through the application of

proportionality rules, the use of one single average subject is appropriate. However, a

single average subject poorly represents anatomical structures with inherent random

variability. The USPIO studies described in this work as well as the work published by

103

Taylor et al. demonstrate sufficient random variability in pelvic lymphatic and vascular

anatomy to argue against the use of a single individual or small sample169, 183.

In representing structures with a large degree of inter-subject variability, such as the

pelvic and inguinal lymph nodes, a mathematical model that describes the spatial

distributions and topographic relationships derived from a large population is more apt

to provide a representative reflection of the structure of interest and facilitate

differentiation of sub-populations222. Such atlases have been both probabilistic and

statistically based. Probabilistic atlases provide an “average” depiction of some feature

of interest in reference to a standard space or co-ordinate system for a collection of

individuals. Statistical atlases depict the variability within a population and facilitate the

evaluation of factors that account for the anatomic variability within the population and

allow the shape properties of anatomic structures and the spaces between them to be

explored223. The papers by Taylor et al., Shih et al. and Dinniwell et al. demonstrate to

varying degrees the potential utility of such an approach169, 170, 183. No one study

however fully exploits the potential of either a probabilistic or a statistical lymphatic

anatomic atlas.

From a broader perspective, there is also ongoing work to develop a standardized

representation of anatomic ontology to create the Digital Anatomist Symbolic

Knowledge Base215. This work focused on the symbolic representations and spatial

relationships of anatomic structures to produce an adaptable and integrated knowledge

base of systemic and regional anatomy that can be queried at varying levels of physical

scale215. Together a model of anatomic variability and of complex relationships

104

expressed through a standardized ontology would facilitate knowledge management,

the incorporation of future data and the provision of clinical care.

At present, no such model exists, with the topographic lymph node atlas developed by

Qatarneh et al. representing the closest approximation to a whole body atlas224. Unlike

the model developed by Talairach, the model based on the Visible Human Male lacks

any means of scaling the anatomic properties or features to other individuals, which

curtails its utility225. The transition from film to digital image processing in radiology has

facilitated computer based quantitative image analysis. The introduction of the DICOM

(Digital Imaging and Communications in Medicine) standard for storing digital imaging

information allows large cross-sectional anatomic image series to be rendered and

viewed in two- or three-dimensions for volumetric analysis. Two prior studies and those

presented in Chapters 2 and 3 use the digital images from USPIO-enhanced MR

imaging to derive lymph node clinical target volumes for radiotherapy treatment

planning of pelvic malignancies169, 170, 183.

The model described by Shih et al. is comprised of 18 prostate cancer patients (10

newly diagnosed and 8 with persistent or recurrent disease)170. Within the group were

69 metastatic lymph nodes with a range of 1 to 21 (median 2 mean 3.8) identified per

patient. Fourteen of the 69 lymph nodes were situated in a para-aortic distribution, and

the remaining 55 were in the pelvis. The majority of the metastatic lymph nodes (40 of

69) were ≤ 1 cm in size. Using the 69 metastatic lymph nodes, a model was generated

with the center point of each lymph node plotted in spatial relation to the adjacent

anatomy. The center point was expanded for all nodal centers by a uniform 5 mm

105

margin to yield a sphere 1 cm in diameter, which was taken to be representative of the

node. The distance from the surface of each representative nodal sphere to the surface

the adjacent blood vessels was then assessed to determine that 94.5% of lymph nodes

were encompassed with a 2.0 cm margin of expansion around sections of the common,

external and internal iliac vessels.This study has several limitations. Only “metastatic”

lymph nodes were included in the analysis and the topologic distribution of

radiographically normal lymph nodes at risk of harboring occult micrometastatic disease

was not addressed, thereby limiting this as a useful guide for defining target volumes for

adjuvant radiotherapy. All lymph nodes were assumed to be adequately represented as

1 cm spheres, which appears to run counter to earlier works examining the short-axis

dimensions of the retroperitoneal, pelvic and inguinal lymph nodes226-229.

Grubnic et al., using a sample of 12 patients who had undergone lymphangiographic,

MRI and CT studies, determined the 95th percentile maximum sort axis dimension of

pelvic lymph nodes to be: 4 mm for the common iliac and obturator, 5 mm for the

external and internal iliac 5 mm and 6 mm for the hypogastric227. Within the

retroperitoneum, the 95th percentile maximum short axis dimension was 3 mm for the

paracaval and inter-aortocaval nodes, 4 mm for the posterior caval nodes, and 5 mm for

the low left para-aortic nodes227. Normalizing all of the nodes to 1.0 cm for all nodal

stations regardless of the underlying anatomy, could have resulted in additional

adjacent normal tissue being unnecessarily encompassed in the vascular expansion

used to form the nCTV170. These potential problems have limited the clinical acceptance

of this model.

106

A related study by Taylor et al. examined 20 patients, all of whom were newly

diagnosed with endometrial or cervical carcinomas169. None of the patients was

reported to have pathological lymph node metastases. A total of 1216 nodal contours

were generated, ranging from 30-101 (median 58) in individual patients. Of these, 627

were from the external iliac region, 303 from the obturator region, 144 from the internal

iliac region and 135 from the common iliac region (under represented as only 13/20

patients were imaged with an adequate field of view). The nodal contours had a mean

short axis of 3.6 mm with a standard deviation of 1.6 mm (1.1 to 12.1 mm). Only seven

contours (0.6%) were greater than 10 mm. Following identification and delineation of the

lymph nodes, varying margins of expansion were applied to the adjacent vasculature

and the percentage lymph node coverage was calculated. It was determined that a 7

mm margin encompassed 88% of the adjacent lymph nodes. With the addition of a 10

mm presacral margin anterior to S1/S2 and an 18 mm margin along the lateral pelvic

sidewall between the external iliac volume anteriorly and the internal iliac volume

posteriorly, all demonstrable lymph node tissue could be encompassed. In a follow-up

validation cohort using a series of 10 patients with 741 identifiable nodal contours,

application of these margins resulted in inclusion of 737 of the contours. A 17 mm

sidewall margin was found to be adequate rather than the 18 mm margin proposed

originally. Despite limitations in the field of view and the absence of data concerning

nodal frequency as opposed to contours, this study is useful for its assessment of the

distribution of lymph nodes in relation to the adjacent vessels.

The model proposed by Taylor et al. represents an improvement over the one by Shih

et al. in better reflecting the differences between the distinct vascular segments and the

107

importance of considering adjacent anatomy when delineation nodal target volumes169,

170. However, difficulties arise from the manner in which the lymph node data were

reported, and this limits comparison with the existing published literature. Qaterneh, in

an analysis of the high resolution National Institute of Health Visible Human Male data

set, identified 6 internal iliac, 14 obturator, 1 promontorial, 3 sacral and 6 common iliac

lymph nodes224. Grubnic et al. reported mean nodal numbers: of 11.9 (4-18) for the

common iliac, 10.4 (1-16) for the external iliac, 4.8 (1-9) for the obturator, 5.1 (0-13) for

the internal iliac and 1.8 (0-4) for the hypogastric227. The long-axes of a significant

number of these nodes extended across more than one axial plane. However, instead

of lymph nodes, Taylor et al. reported their data using nodal contours and nodal

frequency for each site was not specified. Similarly, the short axis dimension was not

reported in a manner consistent with the published literature230-232. By convention, there

can only be one short-axis dimension per lymph node. However, in reporting nodal

contours, a median short axis distance of 3.6 mm with a standard deviation of 1.6 and a

range of 1.1 to 12.1 was described169. Lymph nodes and nodal contours were used

interchangeably in this study, which might have introduced a left skew to the distribution

by repeatedly sampling multiple “short-axes” for a given lymph node. They cannot then

be meaningfully compared to other reported series. Unlike Shih et al., the volumes

proposed by Taylor et al. did incorporate the influence of adjacent musculoskeletal and

visceral anatomy with modifications that effect a transition from an isotropic to an

anisotropic distribution169, 170. This helps to ensure adequacy of lymph node coverage

while facilitating normal tissue avoidance.

108

Both of these studies have been incorporated into expert consensus nodal volume

definitions. A Radiation Therapy Oncology Group Genitourinary (RTOG GU) expert

panel was assembled to review material from lymphographic studies and the results of

extended surgical and sentinel lymph node dissections, in addition to the studies by

Shih and Taylor233, 234. It was concluded that the patterns of prostatic lymphatic drainage

were primarily to the internal iliac, external iliac and sacral lymphatic nodal basins. The

recommendation was for a 7 mm uniform vascular expansion with pelvic sidewall strips

of unspecified width and a 10 mm strip anterior to S1-S3 to encompass the presacral

space. The RTOG Gynecologic Oncology (GYN) expert panel reviewed source material

from anatomical and surgical atlases, surgical texts, imaging references, as well as

IMRT and patient experiences to derive their target volumes235. Similar to the GU

guidelines, they recommended a 7 mm vascular expansion with a 15 mm strip anterior

to the sacrum. A further margin to unite the external and internal iliac contours along the

lateral pelvic sidewalls was recommended but not quantified. The RTOG

Gastrointestinal (GI) group, having identified an educational need for nodal clinical

target volume delineation, formed a consensus panel and each member was asked to

delineate the clinical target volume for a representative case. Using these target

volumes, 95% confidence level consensus contours were developed236. From this

analysis, it was concluded that a 7 to 8 mm margin of expansion around iliac vessels

(≥10 mm anterolaterally) and a 10 mm strip anterior to the sacrum should be used. No

specific margin of expansion was recommended for the femoral vessels. The inguinal

lymph node target volume was to be delineated as a compartment and not based on

vascular expansion. Table 4.1 summarizes the recommendation from prior studies and

the consensus definitions.

109

Table 4.1: Comparison of recommendations for nodal CTV delineation guidelines

Shih170 Taylor*169 RTOG

GU

RTOG

GYN

RTOG

GI

Common Iliac AV 20 mm 7 mm 7 mm 7 mm

External Iliac AV 20 mm 7 mm

(18 mm

Distal

Ant/Lat)

7 mm

(10 mm

Ant)

7 mm 7-8 mm

Internal Iliac AV 20 mm 7 mm 7 mm 7 mm 7-8 mm

Obturator

(Lateral pelvic sidewall)

18 mm Not

specified

?7 mm

7 mm Not

specified

Presacral space 10 mm 10 mm 15 mm ≥10 mm

*7 mm expansion encompassed 88% of lymph node contours so modification recommended

110

In the work presented in Chapters 2 and 3, unlike Shih et al., all gross nodes were

censored and no attempt was made to generate representative nodes in their place,

since the main objective was to develop clinically relevant target volumes for adjuvant

radiotherapy183. Given the small number of such nodes, it is not likely that this had any

meaningful impact on the results. As with Taylor et al., the volumes were modified to

reflect differences in each nodal vascular segment and the adjacent anatomy169, 170. In

contradistinction to Taylor et al., who determined the isotropic vascular expansion

required to include a set amount of the adjacent lymph node tissue and then modified it

to capture the remainder, the method outlined in Chapters 2 and 3 aims to capture all

lymph nodes and then modify the volume obtained to exclude adjacent non-lymphatic

anatomy. Despite the apparent difference in the proposed margins of expansion

between Taylor’s study and ours, the derived nodal CTV’s are actually quite similar. The

manual contour modifications that are implicit in the Taylor et al. CTV definition, in

effect, further expand their 7 mm vessel margins so that the final volumes are relatively

coincident169, 170. Our approach may be less time consuming and less prone to error

because less revision is required for individual cases.

The greatest limitation of our work as well as previous studies is the inability of MR

imaging with lymphotrophic USPIO’s like ferumoxtran-10 to identify small lymph nodes

<3 mm in size, and microscopic lymph node metastases. The nodes comprising the

presacral and internal iliac groups are likely underrepresented in all of these studies by

virtue of their size. Taylor et al. reported the frequency of lymph node contours and a

summation of contours through successive axial images would be required to determine

the number of lymph nodes in each nodal group169. Within their 20 patient series, a

111

mean of 0.4 (0-2) and 7.2 (1-22) lymph node contours were identified for the presacral

and internal iliac lymph node groups respectively. In comparison, a median 2 (0-6) and

5 (0-15) presacral/subaortic and internal iliac lymph nodes were identified in our 55

patient series. Recognizing that the presacral lymph node group in our series included

subaortic lymph nodes and that nodes within each group may lie across more than one

adjacent axial plane, there would appear to be an equivalent or greater number of

lymph nodes identified in our series with a larger sample size.

Our model for nodal CTV definition was derived from the spatial distribution of visible

lymph nodes, on the assumption that these would be representative of all nodes at risk

of harboring disease. It is possible that there are small lymph nodes at greater distances

from the major vessels that were not visualized using this approach. The smallest short

axis dimension of the lymph nodes delineated in the high resolution Visible Human Male

dataset was 2 mm. While this is smaller than the slice thickness (3 mm) used in this

study and raises concerns about identifying these lymph nodes using MRI, the close

relationship of these lymph nodes to the adjacent vessels would result in them being

encompassed in a nodal CTV generated using a margin of expansion derived from

larger visible nodes. The adjacent lymph nodes that are of a larger size than the slice

thickness used, and are apparent on the MRI images, can help to define the outer

boundaries for an anatomic envelope to encompass the nodal basin within which

smaller (<3 mm) lymph nodes may reside. However, these are likely to represent a

small proportion of the total nodal volume in each anatomic region and would not

substantially alter the proposed margins of expansion, which were defined to

encompass 90% of the lymphatic tissue in 90% of patients. Ferumoxtran-10, while

112

lowering the MR threshold for the detection of subclinical disease, does not enhance the

sensitivity enough to reliably detect microscopic metastases. This currently limits the

utility of this approach for defining patient-specific radiotherapy volumes that target only

regions of known tumor.

All of the inherent limitations in our work and other previous studies of pelvic lymph

node target volume definition (adequacy of anatomic coverage, knowledge of patterns

of spread, patient specific nCTV delineation) could be mitigated in whole or in part with

the use of imaging techniques and contrast agents that are more sensitive and specific

for the detection of small lymph nodes and microscopic lymph node metastases.

Unfortunately, there is no lymph node contrast agent currently available for clinical use.

In the 55 patients analyzed to determine pelvic nodal target volumes (Chapter 2) and

the 30 patients that formed the basis for the inguinal target volume definition (Chapter

3), it has been demonstrated that a population based class solution can be derived for

isotropic margins of vascular expansion. While the model would benefit from a larger

sample size to ensure sufficient power for clinical use, it is instructive in its

demonstration of the potential of isotropic margins. These margins are capable of

encompassing the lymph nodes of interest while reducing the volume of normal tissue

included in the nCTV. A reduction in the volume of normal tissue within the high dose

nCTV may help lead to a reduction in the acute and late toxicities from pelvic

radiotherapy.

A further incremental improvement might be realized with the use of non-uniform

margins of expansion or the addition of further vascular segments upon which margins

113

of expansion may be applied. Within each vascular segment, characteristic distributions

of the lymph nodes are apparent, as illustrated in Figure 4.1. The common iliac lymph

nodes appear to be closely related to the posterior aspect of the adjacent vessels while

the external and internal iliac lymph nodes are situated more circumferentially and at

greater distances from the vessels. These differences primarily arise because of the

relationship of the vessels and nodes to the adjacent musculoskeletal anatomy,

including bone and muscle. While anisotropic expansion is a potential refinement on

isotropic margins, still further improvement could be obtained with continuously variable

patient-specific margins around vessels based on more sensitive and specific nodal

imaging techniques.

The guidelines derived for pelvic and inguinal nCTV's in Chapters 2 and 3 are based

upon the paired relationship between regional lymph nodes and adjacent vascular

segments. Within the femoral triangle, lie the femoral artery and vein in addition to the

great saphenous, superficial circumflex iliac, superficial epigastric, external pudendal as

well as the lateral and medial accessory saphenous veins (Figure 1.8 and 1.9). In

exploring the spatial distribution between the lymph nodes and vascular segments, the

impact upon the histogram plot of the addition of vascular segments to the femoral

artery and vein is readily apparent (Figure 2.3). Using the femoral artery and vein alone,

a larger margin of expansion is required to encompass the lymph nodes and the centre

of balance appears to be about 15 mm. With the addition of new vascular segments, the

balance shifts left with an associated reduction in the margin of expansion required to

encompass an equivalent volume of nodal tissue. In Figure 2.3, the skewness of the

histogram shifts left with the incorporation of the great saphenous and superficial

114

circumflex iliac veins along with a slight reduction in the range. Those lymph nodes

(lateral superficial and inferior superficial) which had been encompassed by an over

expansion of the femoral artery and vein are now paired with their corresponding artery

or vein. Were it technically possible to identify, delineate and add the remaining

vascular segments, the histogram plot would assume a normal distribution with a

balance less than 10 mm and a narrow range. In an effort to ensure an accurate and

efficient means of nCTV delineation, those vessels which are difficult to appreciate in

the absence of vascular contrast or additional imaging modalities have been omitted.

Were there a clinically practical means of defining further vascular segments such as

the branches arising from the internal iliac artery and the veins and arteries within

femoral triangle, a re-analysis of the data presented in Chapters 2 and 3 could be

undertaken to generate a population based statistical library of vascular segment

expansions that could be applied to individual named vascular arterial or venous

branches. The long intravascular circulatory time of the USPIO does enable MR

arteriograms and venograms to be obtained from which those smaller vascular

branches could be reliably segmented. This would yield an iterative refinement in the

data presented in Chapters 2 and 3 in addition to a reduction in the individual vascular

margins of expansion.

A similar effect is noted in the histogram shown in Figure 2.3. In delineating only the

proximal internal iliac vessels cephalad to their origin and extending the object

delineations along the course of the superior gluteal vessels, the remaining branches of

the internal iliacs are omitted from the analysis (Figure 1.8 and 1.9). This is also the

case for the middle sacral artery. Through the delineation of anatomic land markers for

115

Figure 4.1: Axial MR images for: (A) the common iliac vessels and common iliac lymph

node (red arrow); (C) the external iliac vessels and external iliac lymph node (white

arrow) and (E) the internal iliac vessels and an internal iliac lymph node (blue arrow)

and the corresponding scatter plot distributions of the lymph node volume elements in

relation to the surface of these vascular segments. Note: the red dot (0,0) denotes the

point defining the vessel surface.

116

the lateral pelvic sidewall and the distal lumbar and anterior sacral surface, those nodes

which would require an over expansion of the common, external and internal iliac

vessels to assure their inclusion (data not shown). With the use of the land markers,

their histogram plots remain with a narrower range and a balance more reflective of the

underlying anatomy. Were those vessels that had been omitted to be delineated and

included in the analysis, the anatomic land markers for the subaortic/sacral and lateral

pelvic sidewall would confer no additional gain and could be eliminated..

117

Chapter 5. Conclusions and Future Directions

In conclusion, isotropic margins can be derived for the pelvic and inguinal vasculature.

The application of these margins in the context of adjuvant IMRT for pelvic

malignancies should enable consistency in lymph node clinical target volume

delineation while facilitating normal tissue avoidance. Derivation of anisotropic margins

of expansion may enable additional sparing of normal tissues while providing similar

lymph node coverage to isotropic margins. Analysis of a larger sample size and

validation would help to better power this model and assess its utility for clinical use.

Studies of IMRT in the treatment of pelvic malignancies have demonstrated reduced

volumes of the small bowel, rectum, bladder and bone marrow in the high dose regions,

with a resultant decrease in acute and chronic gastrointestinal toxicities132, 157, 158, 237-240.

Application of the modified isotropic or the anisotropic margins in a prospective clinical

study with IMRT would help to determine any further reduction in toxicities related to the

reduced volume of normal tissue irradiated using these nCTV margins.

At present, the limitations to patient specific USPIO imaging include: the absence of a

clinically available agent, cost of the MRI and radiology reporting time (mean time to

report one image series by an expert observer is 80 minutes with pre- and post-contrast

imaging required), and the learning curve of the reporting radiologist174. USPIO currently

is not available in North America outside of a clinical trial. The models presented here

as well as those reported by others represent means of using a population based model

for the radiotherapy treatment planning of an individual patient169, 170, 183. These nascent

118

anatomic models parallel the efforts that have recently been undertaken in other

domains to develop an animated four-dimensional model of the human body using

morphologic, functional and molecular information within a unified representation220.

The core elements of this model include: a 3-dimensional digital atlas of the human

body, ontology-based metadata support and biomedical datasets220. This will allow

spatio-temporal modeling of complex disease processes and is sufficiently evolved to

demonstrate changes over time reflecting progression of underlying pathology. In the

context of the body of work discussed here, a depiction of the spatial topology of the

lymph node distributions within the pelvis and groin as well as the pattern of local,

regional and distant metastatic spread both within and external to the lymphatics could

be incorporated into such a model. From a lymphatic atlas, composite organ systems

could then be added and integrated into a unified and dynamic digital construct. Into this

form, medical images and spatio-temporal data could be incorporated and manipulated

in four-dimensions over a range of scales from the subcellular to the that of the entire

organism..

A standard three-dimensional anatomic form affords a framework to which additional

morphologic information can be added. For example, lymphatic vessels cannot be

adequately imaged using currently available cross-sectional imaging approaches.

However, with the use of microsurgical techniques and computed tomographic

lymphangiography, the relevant efferent and afferent lymphatic vessels can be

detected241. This new information could be incorporated into a standardized three-

dimensional anatomic model to refine currently available information about the micro-

regional distribution of lymphatic vessels and lymph nodes.

119

Image data with USPIO is limited by MRI resolution. Additional complementary pulse

sequences, such as diffusion-weighted MRI (DW-MRI), may enhance the diagnostic

accuracy. DW-MRI is a non-invasive imaging technique that examines the structure of

a biologic tissue and facilitates the assessment and measurement of tissue

microstructure (Figure 5.1). The random thermal motion of water molecules, Brownian

motion, results in tissue dependent signal attenuation. This allows tissues to be

evaluated on the basis of differences in water-proton mobility within and between

anatomic structures. DW-MRI has been investigated for a variety of pathologic

conditions such as acute stroke, multiple sclerosis, abscesses and tumors242-247. More

recently, it has been used to serially monitor solid tumor response to chemotherapy

over time and identify nodal metastases248, 249. Changes in the mobility of water

molecules in a tissue afford a non-invasive means of further characterizing a given

tissue.

120

Figure 5.1: (Left) T2-w Fast Spin Echo MR image of the pelvis at the level of the

acetabulum with ( ) denoting a left lateral superficial inguinal lymph node. (Right) DWI

MR image at the level of the acetabulum with ( ) denoting a left lateral superficial

inguinal lymph node.

121

A recent report by Thoeny et al. examined the potential of DW-MRI to facilitate reader

reporting of USPIO-enhanced MR imaging studies of normal sized lymph nodes in 20

patients with bladder or prostate cancer174. The authors evaluated the potential

diagnostic synergism between USPIO and DW-MRI using a 3-T MRI with pre- and post-

USPIO imaging series. At the completion of imaging, all patients underwent an

extended pelvic lymphadenectomy and resection of the primary tumor. The addition of

DW-MRI failed to improve the overall diagnostic accuracy for lymph node metastases.

However, DW-MRI did appear to facilitate nodal identification, as the reporting time for

the combined arm was less than that for the USPIO arm (mean of 13 minutes vs 80

minutes respectively, p<0.0001).. Eight percent (2 of 26) metastatic lymph nodes were

missed by imaging. These nodes had dimensions of 1.0 x 0.2 mm and 0.7 x 0.4 mm.

This highlights an important limitation of commonly-used MR technology and pulse

sequences in relation to the detection of micrometastatic disease.

The diagnostic accuracy of CT or MR for the detection of lymph node metastases may

be improved by integrating these anatomic imaging modalities with functional and

molecular PET or SPECT imaging. This creates the potential for synchronized

acquisition of morphologic, functional and molecular information250. While historically of

limited resolution, enhancements such as pinhole SPECT have enabled the detection of

discrete nodal metastases in sites such as the axilla251. As further improvements are

made in camera design with greater intrinsic spatial resolution, the potential for

additional discrete nodal metastasis detection will increase.

122

Additional improvement in nodal detection can be obtained by increasing the inherent

contrast between the different tissues of interest (e.g. lymph node versus adjacent non-

lymphatic anatomy) with the administration of a contrast agent. Radiographic

lymphography has recently re-emerged for axillary lymphatic mapping with the

interstitial administration of commercially-available water-soluble CT contrast media into

the breast89, 241, 252-254. Unlike the current sentinel lymph node technique, 3D-CT

lymphography provides a permanent record of the route and pattern of lymphatic

drainage from the peri-tumoral injection site within the breast through to the first, second

and third order draining lymphatics. Optical fluorophores, quantum dots, have also

been used to identify the sentinel lymph nodes in animal models. They consist of

nanometer sized (5 to 20 nm) cadmium-selenium or cadmium-tellurium semi-conductor

crystals255. Quantum dots emit energy over a narrow wavelength that is determined by

their size and shape. Therefore, using quantum dots with different characteristics allows

an assessment of routes of transport in tissue. As with 3D-CT breast lymphography, the

use of quantum dots to image lymph node drainage basins is currently limited to

interstitial peri-tumoral administration as they are not taken up by the lymphatics

following intravenous infusion255, 256. Other limitations include the potential for toxicity

from the semi-conductor core elements [261] and the production of free radicals from

the optical excitation energy leading to DNA damage and cell death257. While the unique

physical properties of quantum dots give them utility in the study of animal model

systems, the current toxicities and lack of lymphotrophism are impediments to clinical

translation.

123

Adenoviruses are known to be lymphotrophic258. This may be a consequence of their

size, negative charge or some additional biological feature. Recombinant adenoviruses

have been created to act as gene delivery vectors. Imaging reporter genes have been

successfully inserted into prostate-specific adenoviral vectors and detected following

administration of the virus in an animal model259. Unfortunately, in spite of the known

lymphotrophism, only a small fraction (0.1%) of the adenoviral dose is detectable in

lymph nodes following interstitial administration. 258. This along with the potential for

incomplete gene transfer may limit its sensitivity and subsequent clinical utility.

The overall quality of all of the different imaging modalities is improving and with this

progress, the potential for individual lymph node identification and segmentation will

become more accessible in the clinic. Using high-resolution isotropic voxels and

segmentation algorithms, one can volumetrically identify the lymph nodes of interest

within a given anatomic site. Currently available commercial software allows automated

segmentation and analysis of lymph nodes following manual identification of seed points

to guide this process260. The nodes may then by analyzed in two- and or three-

dimensional visualizations using accepted means, such as those specified by the

response evaluation criteria in solid tumors231, 261. This potential has also been explored

by others 262, 263 and can be improved upon by subdividing regions, as was presented in

Chapters 2 and 3 for inguinal and pelvic lymph nodes. Region specific delimiters for the

head and neck, axilla, thorax, abdomen, pelvis and groin would enable further

refinement of these methods.

124

In summary, this study has highlighted the importance of high resolution lymph node

imaging with an appropriate contrast agent in the current era of high precision

radiotherapy. USPIO offer several advantages for the detection of lymph nodes and

lymph node metastases that can aid in radiotherapy target volume definition.. contrast

agents are not currently available for routine clinical use. However, their demonstrated

utility will help to ensure that they (or other newly emerging nodal imaging agents) will

ventually find a place in the diagnosis and treatment of cancer, leading to greater

individualization of care. In the absence of widely available nodal contrast agents for

patient-specific treatment planning, the use of class solution nodal clinical target

volumes based on contrast-enhanced high resolution nodal imaging as presented here

will form the foundation for current patient care and the development of innovative

treatment strategies in the future.

125

References

1. Vondenhoff MF, van de Pavert SA, Dillard ME, et al. Lymph sacs are not

required for the initiation of lymph node formation. Development. Jan 2009;136(1):29-34.

2. De Roo T. Atlas of Iymphography. Leiden: J.B. Lippincott; 1975. 3. Viamonte M, Ruttimann A. Atlas of Lymphography. Stuttgart: Thieme; 1980. 4. Foldi M, Foldi E, eds. Foeldi's Textbook of Lymphology: For Physicians and

Lymphedema Therapists 2nd ed. Munich: Elsevier; 2006. 5. Kubik S. The anatomy of the lymphatic system. In: Musshoff K, ed. Recent

Results in Cancer Research Vol 46. New York Heidelberg Berlin: Springer; 1974. 6. Kubik S. Anatomie des Lymphgefafisystems. In: Frommhold W, Gerhardt P, eds.

Klinisch-radiologisches Seminar. Erkrankungen des Lymphsystems. Stuttgart: Thieme; 1981.

7. Rouviere H. Anatomy of the Human Lymphatic System: A Compendium Translated From the Original ‘Anatomie des Lymphatiques de l'Homme’ and Rearranged for the Use of Students and Practitioners by M.J. Tobias. Ann Arbor, MI; 1938.

8. Farrell WJ. Lymphangiographic demonstration of lymphovenous communication after radiotherapy in Hodgkin's disease. Radiology. Oct 1966;87(4):630-634.

9. Hidden G, Menard P, Zorn JY. Lymphaticovenous communications. Role of the lymph nodes. Anat Clin. 1985;7(2):83-91.

10. Modi S, Stanton AW, Mortimer PS, Levick JR. Clinical assessment of human lymph flow using removal rate constants of interstitial macromolecules: a critical review of lymphoscintigraphy. Lymphat Res Biol. 2007;5(3):183-202.

11. Roxin T, Bujar H. Lymphographic visualization of lymphaticovenous communications and their significance in malignant hemolymphopathies. Lymphology. Sep 1970;3(3):127-135.

12. Threefoot S, Kossover MF, Aiken DW. RADIOISOTOPIC DETECTION OF LYMPHATICOVENOUS COMMUNICATIONS IN LIVING ANIMALS. J Lab Clin Med. Apr 1965;65:688-697.

13. Threefoot SA. Gross and microscopic anatomy of the lymphatic vessels and lymphaticovenous communications. Cancer Chemother Rep. Jan 1968;52(1):1-20.

14. Threefoot SA. The clinical siginifcance of lymphaticovenous communications. Ann Intern Med. Jun 1970;72(6):957-958.

15. Threefoot SA, Kent WT, Hatchett BF. Lymphaticovenous and lymphaticolymphatic communications demonstrated by plastic corrosion models of rats and by postmortem lymphangiography in man. J Lab Clin Med. Jan 1963;61:9-22.

16. Threefoot SA, Kossover MF. Lymphaticovenous communications in man. Arch Intern Med. Feb 1966;117(2):213-223.

17. Threefoot SA, Kossover MF, Kent WT, Hatchett BF, Pearson JE, Jr., Cabrera-Gil C. Factors stimulating function of lymphaticovenous communications. Angiology. Nov 1967;18(11):682-698.

126

18. Wolfel DA. Lymphaticovenous communications; a clinical reality. Am J Roentgenol Radium Ther Nucl Med. Nov 1965;95(3):766-768.

19. Mascagni P. Vasorum lymphaticorum corporis humani: historia et ichnographia. Siena; 1787.

20. Ackerman MJ, Spitzer VM, Scherzinger AL, Whitlock DG. The Visible Human data set: an image resource for anatomical visualization. Medinfo. 1995;8 Pt 2:1195-1198.

21. Lopez-Cano M, Rodriguez-Navarro J, Rodriguez-Baeza A, Armengol-Carrasco M, Susin A. A real-time dynamic 3D model of the human inguinal region for surgical education. Comput Biol Med. Sep 2007;37(9):1321-1326.

22. Mathers CD, Shibuya K, Boschi-Pinto C, Lopez AD, Murray CJ. Global and regional estimates of cancer mortality and incidence by site: I. Application of regional cancer survival model to estimate cancer mortality distribution by site. BMC Cancer. Dec 26 2002;2:36.

23. Murray CJ, Lopez AD. Mortality by cause for eight regions of the world: Global Burden of Disease Study. Lancet. May 3 1997;349(9061):1269-1276.

24. Shibuya K, Mathers CD, Boschi-Pinto C, Lopez AD, Murray CJ. Global and regional estimates of cancer mortality and incidence by site: II. Results for the global burden of disease 2000. BMC Cancer. Dec 26 2002;2:37.

25. Bakkum-Gamez JN, Gonzalez-Bosquet J, Laack NN, Mariani A, Dowdy SC. Current issues in the management of endometrial cancer. Mayo Clin Proc. Jan 2008;83(1):97-112.

26. Benedetti Panici P, Basile S, Maneschi F, et al. Systematic pelvic lymphadenectomy vs. no lymphadenectomy in early-stage endometrial carcinoma: randomized clinical trial. J Natl Cancer Inst. Dec 3 2008;100(23):1707-1716.

27. Breyer BN, Greene KL, Dall'Era MA, Davies BJ, Kane CJ. Pelvic lymphadenectomy in prostate cancer. Prostate Cancer Prostatic Dis. 2008;11(4):320-324.

28. Briganti A, Blute ML, Eastham JH, et al. Pelvic lymph node dissection in prostate cancer. Eur Urol. Jun 2009;55(6):1251-1265.

29. Burkhard FC, Studer UE. Regional lymph node staging in prostate cancer: prognostic and therapeutic implications. Surg Oncol. Sep 2009;18(3):213-218.

30. Buscarini M, Josephson DY, Stein JP. Lymphadenectomy in bladder cancer: a review. Urol Int. 2007;79(3):191-199.

31. Chang SJ, Kim WY, Yoon JH, Yoo SC, Chang KH, Ryu HS. Para-aortic lymphadenectomy improves survival in patients with intermediate to high-risk endometrial carcinoma. Acta Obstet Gynecol Scand. 2008;87(12):1361-1369.

32. Da Pozzo LF, Cozzarini C, Briganti A, et al. Long-term follow-up of patients with prostate cancer and nodal metastases treated by pelvic lymphadenectomy and radical prostatectomy: the positive impact of adjuvant radiotherapy. Eur Urol. May 2009;55(5):1003-1011.

33. de Vries RR, Nieuwenhuijzen JA, Meinhardt W, Bais EM, Horenblas S. Long-term survival after combined modality treatment in metastatic bladder cancer patients presenting with supra-regional tumor positive lymph nodes only. Eur J Surg Oncol. Apr 2009;35(4):352-355.

127

34. Ferrandina G, Fanfani F, Ludovisi M, et al. Is frozen section analysis of pelvic lymph nodes accurate in locally advanced cervical cancer patients administered preoperative chemoradiation? Gynecol Oncol. Jan 2008;108(1):220-225.

35. Fujimoto T, Fukuda J, Tanaka T. Role of complete para-aortic lymphadenectomy in endometrial cancer. Curr Opin Obstet Gynecol. Feb 2009;21(1):10-14.

36. Hosaka M, Watari H, Takeda M, et al. Treatment of cervical cancer with adjuvant chemotherapy versus adjuvant radiotherapy after radical hysterectomy and systematic lymphadenectomy. J Obstet Gynaecol Res. Aug 2008;34(4):552-556.

37. Karl A, Carroll PR, Gschwend JE, et al. The impact of lymphadenectomy and lymph node metastasis on the outcomes of radical cystectomy for bladder cancer. Eur Urol. Apr 2009;55(4):826-835.

38. Kitchener H, Swart AM, Qian Q, Amos C, Parmar MK. Efficacy of systematic pelvic lymphadenectomy in endometrial cancer (MRC ASTEC trial): a randomised study. Lancet. Jan 10 2009;373(9658):125-136.

39. Klopp AH, Jhingran A, Ramondetta L, Lu K, Gershenson DM, Eifel PJ. Node-positive adenocarcinoma of the endometrium: outcome and patterns of recurrence with and without external beam irradiation. Gynecol Oncol. Oct 2009;115(1):6-11.

40. Lerner SP. The role and extent of pelvic lymphadenectomy in the management of patients with invasive urothelial carcinoma. Curr Treat Options Oncol. Aug 2009;10(3-4):267-274.

41. Marana HR, de Andrade JM, Dos Reis FJ, et al. Impact of surgical staging in locally advanced cervical cancer and subsequent chemotherapy. J Surg Oncol. Nov 1 2009;100(6):505-510.

42. Ordon M, Nam RK. Lymph node assessment and lymphadenectomy in prostate cancer. J Surg Oncol. Mar 15 2009;99(4):215-224.

43. Sonpavde G, Goldman BH, Speights VO, et al. Quality of pathologic response and surgery correlate with survival for patients with completely resected bladder cancer after neoadjuvant chemotherapy. Cancer. Sep 15 2009;115(18):4104-4109.

44. Spiess PE, Hernandez MS, Pettaway CA. Contemporary inguinal lymph node dissection: minimizing complications. World J Urol. Apr 2009;27(2):205-212.

45. Vanni AJ, Moinzadeh A. Lymphadenectomy in minimally invasive urologic oncology. Curr Opin Urol. Mar 2008;18(2):163-166.

46. Wagner M, Sokoloff M, Daneshmand S. The role of pelvic lymphadenectomy for prostate cancer--therapeutic? J Urol. Feb 2008;179(2):408-413.

47. Wang D, Lawton C. Pelvic lymph node irradiation for prostate cancer: who, why, and when? Semin Radiat Oncol. Jan 2008;18(1):35-40.

48. Winter A, Wawroschek F. Lymphadenectomy in prostate cancer. Radio-guided lymph node mapping: an adequate staging method. Front Radiat Ther Oncol. 2008;41:58-67.

49. Fuchs W, Davidson J, Fischer H. Lymphography in Cancer. New York: Spriner-Vertag; 1969.

50. Lengele B, Nyssen-Behets C, Scalliet P. Anatomical bases for the radiological delineation of lymph node areas. Upper limbs, chest and abdomen. Radiother Oncol. Sep 2007;84(3):335-347.

128

51. Lengele B, Scalliet P. Anatomical bases for the radiological delineation of lymph node areas. Part III: Pelvis and lower limbs. Radiother Oncol. Jul 2009;92(1):22-33.

52. Kinmonth J. The Lymphatics: Diseases, Lymphography and Surgery. London: Edward Arnold; 1972.

53. Gregoire V, Scalliet P, Ang K, eds. Clinical Target Volumes in Conformal and Intensity Modulated Radiation Therapy: A Clinical Guide to Cancer Treatment. Heidelberg: Springer; 2004.

54. Reiffentstuhl G. The Lymphatics of the Female Genital Organs. Philadelphia: J. B. Lippincott Co.; 1964.

55. Jossifow G. Das Lymphgefäsystem des Menschen. Jena: Gustav Fischer; 1930. 56. Levenback C, van der Zee A, Coleman R, eds. Clinical lymphatic mapping in

gynecologic cancers. London: Taylor & Francis; 2004. 57. Escobar-Prieto A, Gonzalez G, Templeton AW, Cooper BR, Palacios E.

Lymphatic channel obstruction. Patterns of altered flow dynamics. Am J Roentgenol Radium Ther Nucl Med. Oct 1971;113(2):366-375.

58. Heymans O, Fallais C, Hustinx R. Intratissular lymphaticovenous anastomoses demonstrated by perioperative intramuscular injection of 99mTC-colloids. Lymphat Res Biol. Spring 2006;4(1):29-33.

59. Brossner C, Ringhofer H, Hernady T, Kuber W, Madersbacher S, Pycha A. Lymphatic drainage of prostatic transition and peripheral zones visualized on a three-dimensional workstation. Urology. Feb 2001;57(2):389-393.

60. Brossner C, Ringhofer H, Schatzl G, Madersbacher S, Powischer G, Kuber W. Sacral distribution of prostatic lymph nodes visualized on spiral computed tomography with three-dimensional reconstruction. BJU Int. Jan 2002;89(1):44-47.

61. Cellini N, Luzi S, Mantini G, et al. Lymphatic drainage and CTV in carcinoma of the prostate. Rays. Jul-Sep 2003;28(3):337-341.

62. Heidenreich A, Varga Z, Von Knobloch R. Extended pelvic lymphadenectomy in patients undergoing radical prostatectomy: high incidence of lymph node metastasis. J Urol. Apr 2002;167(4):1681-1686.

63. Park JM, Charnsangavej C, Yoshimitsu K, Herron DH, Robinson TJ, Wallace S. Pathways of nodal metastasis from pelvic tumors: CT demonstration. Radiographics. Nov 1994;14(6):1309-1321.

64. Raghavaiah NV, Jordan WP, Jr. Prostatic lymphography. J Urol. Feb 1979;121(2):178-181.

65. Ravery V, Chopin DK, Abbou CC. [Surgical anatomy of the lymphatic drainage of the bladder]. Ann Urol (Paris). 1993;27(1):9-11.

66. Sherif A, De La Torre M, Malmstrom PU, et al. Lymphatic mapping and detection of sentinel nodes in patients with bladder cancer. J Urol. Sep 2001;166(3):812-815.

67. Stein JP. Lymphadenectomy in bladder cancer: how high is "high enough"? Urol Oncol. 2006 Jul-Aug 2006;24(4):349-355. 68. Lerner SP. The role and extent of pelvic lymphadenectomy in the management of

patients with invasive urothelial carcinoma. JCurr Treat Options Oncol. Aug 2009; 10(3-4);267-74.

129

69. Coleman RL, Frumovitz M, Levenback CF. Current perspectives on lymphatic mapping in carcinomas of the uterine corpus and cervix. J Natl Compr Canc Netw. May 2006;4(5):471-478.

70. Eichner E, Goldberg I, Bove ER. In vivo studies with direct sky blue of the lymphatic drainage of the internal genitals of women. Am J Obstet Gynecol. Jun 1954;67(6):1277-1287.

71. Kocisova M, Munka V, Fabianova B. Lymph drainage of the human uterus. Folia Morphol (Praha). 1983;31(2):128-130.

72. Pandit-Taskar N, Gemignani ML, Lyall A, Larson SM, Barakat RR, Abu Rustum NR. Single photon emission computed tomography SPECT-CT improves sentinel node detection and localization in cervical and uterine malignancy. Gynecol Oncol. Apr;117(1):59-64.

73. Ercoli A, Delmas V, Iannone V, et al. The lymphatic drainage of the uterine cervix in adult fresh cadavers: Anatomy and surgical implications. Eur J Surg Oncol. Jul 21 2009.

74. Garriba AP, Restaino A, Selvaggi L, Ettorre GC, Ferreri R, Sabatini R. Direct lymphadenography of the cervix: personal data. Eur J Gynaecol Oncol. 1983;4(3):234-236.

75. Hubalewska A, Sowa-Staszczak A, Huszno B, et al. Use of Tc99m-nanocolloid for sentinel nodes identification in cervical cancer. Nucl Med Rev Cent East Eur. 2003;6(2):127-130.

76. Perrone AM, Casadio P, Formelli G, et al. Cervical and hysteroscopic injection for identification of sentinel lymph node in endometrial cancer. Gynecol Oncol. 2008 Oct;111(1):62-67.

77. Carvalho R, Rodriguez A, Pereria S. La mise en evidence pa le radiographique du system chez le vivant. Ann d'anat path. 1931;8:193.

78. Funaoka S. Abt D'Anat Inst D'Kaiseri, Univ Kyoto Ser D HFT. 1930;1. 79. Malek P, Kolc J, Belan A. Lymphography of the deep lymphatic system of the

thigh. Acta radiol. Jun 1959;51(6):422-428. 80. Ahuja AT, Ying M. Sonographic evaluation of cervical lymph nodes. AJR Am J

Roentgenol. May 2005;184(5):1691-1699. 81. Ganeshalingam S, Koh DM. Nodal staging. Cancer Imaging. 2009;9:104-111. 82. Hovels AM, Heesakkers RA, Adang EM, et al. The diagnostic accuracy of CT

and MRI in the staging of pelvic lymph nodes in patients with prostate cancer: a meta-analysis. Clin Radiol. Apr 2008;63(4):387-395.

83. Kim J, Lee J, Chang E, et al. Selective Sentinel Node Plus Additional Non-Sentinel Node Biopsy Based on an FDG-PET/CT Scan in Early Breast Cancer Patients: Single Institutional Experience. World J Surg. May 2009;33(5):943-949.

84. Kwee RM, Kwee TC. Imaging in assessing lymph node status in gastric cancer. Gastric Cancer. 2009;12(1):6-22.

85. Mattei A, Fuechsel FG, Bhatta Dhar N, et al. The template of the primary lymphatic landing sites of the prostate should be revisited: results of a multimodality mapping study. Eur Urol. Jan 2008;53(1):118-125.

86. Selman TJ, Mann C, Zamora J, Appleyard TL, Khan K. Diagnostic accuracy of tests for lymph node status in primary cervical cancer: a systematic review and meta-analysis. CMAJ. Mar 25 2008;178(7):855-862.

130

87. Lee KC, Moon WK, Chung JW, et al. Assessment of lymph node metastases by contrast-enhanced MR imaging in a head and neck cancer model. Korean J Radiol. Jan-Feb 2007;8(1):9-14.

88. Pandharipande PV, Mora JT, Uppot RN, et al. Lymphotropic nanoparticle-enhanced MRI for independent prediction of lymph node malignancy: a logistic regression model. AJR Am J Roentgenol. Sep 2009;193(3):W230-237.

89. Yamamoto S, Maeda N, Tamesa M, Nagashima Y, Suga K, Oka M. Sentinel Lymph Node Detection in Breast Cancer Patients by Real-Time Virtual Sonography Constructed With Three-Dimensional Computed Tomography-Lymphography. Breast J. Oct 12 2009.

90. Hsu WH, Hsu PK, Wang SJ, et al. Positron emission tomography-computed tomography in predicting locoregional invasion in esophageal squamous cell carcinoma. Ann Thorac Surg. May 2009;87(5):1564-1568.

91. Hu Q, Wang W, Zhong X, et al. Dual-time-point FDG PET for the evaluation of locoregional lymph nodes in thoracic esophageal squamous cell cancer. Eur J Radiol. May 2009;70(2):320-324.

92. Johansen J, Buus S, Loft A, et al. Prospective study of 18FDG-PET in the detection and management of patients with lymph node metastases to the neck from an unknown primary tumor. Results from the DAHANCA-13 study. Head Neck. Apr 2008;30(4):471-478.

93. Jover R, Lourido D, Gonzalez C, Rojo A, Gorospe L, Alfonso JM. Role of PET/CT in the evaluation of cervical cancer. Gynecol Oncol. Sep 2008;110(3 Suppl 2):S55-59.

94. Kernstine KH, Stanford W, Mullan BF, et al. PET, CT, and MRI with Combidex for mediastinal staging in non-small cell lung carcinoma. Ann Thorac Surg. Sep 1999;68(3):1022-1028.

95. Kim SK, Choi HJ, Park SY, et al. Additional value of MR/PET fusion compared with PET/CT in the detection of lymph node metastases in cervical cancer patients. Eur J Cancer. Aug 2009;45(12):2103-2109.

96. Kitajima K, Murakami K, Yamasaki E, et al. Performance of integrated FDG-PET/contrast-enhanced CT in the diagnosis of recurrent uterine cancer: comparison with PET and enhanced CT. Eur J Nucl Med Mol Imaging. Mar 2009;36(3):362-372.

97. Kitajima K, Murakami K, Yamasaki E, et al. Accuracy of 18F-FDG PET/CT in detecting pelvic and paraaortic lymph node metastasis in patients with endometrial cancer. AJR Am J Roentgenol. Jun 2008;190(6):1652-1658.

98. Kitajima K, Murakami K, Yamasaki E, Kaji Y, Sugimura K. Accuracy of integrated FDG-PET/contrast-enhanced CT in detecting pelvic and paraaortic lymph node metastasis in patients with uterine cancer. Eur Radiol. Jun 2009;19(6):1529-1536.

99. Liao CT, Chang JT, Wang HM, et al. Preoperative [18F]fluorodeoxyglucose positron emission tomography standardized uptake value of neck lymph nodes predicts neck cancer control and survival rates in patients with oral cavity squamous cell carcinoma and pathologically positive lymph nodes. Int J Radiat Oncol Biol Phys. Jul 15 2009;74(4):1054-1061.

100. Miele E, Spinelli GP, Tomao F, et al. Positron Emission Tomography (PET) radiotracers in oncology--utility of 18F-Fluoro-deoxy-glucose (FDG)-PET in the

131

management of patients with non-small-cell lung cancer (NSCLC). J Exp Clin Cancer Res. 2008;27:52.

101. Molina M, Avisar E. Use of PET probe in surgical oncology. J Surg Oncol. Apr 1 2008;97(5):369-371.

102. Morikawa M, Demura Y, Ishizaki T, et al. The effectiveness of 18F-FDG PET/CT combined with STIR MRI for diagnosing nodal involvement in the thorax. J Nucl Med. Jan 2009;50(1):81-87.

103. Mueller-Lisse UG, Scher B, Scherr MK, Seitz M. Functional imaging in penile cancer: PET/computed tomography, MRI, and sentinel lymph node biopsy. Curr Opin Urol. Jan 2008;18(1):105-110.

104. Nakagawa T, Yamada M, Suzuki Y. 18F-FDG uptake in reactive neck lymph nodes of oral cancer: relationship to lymphoid follicles. J Nucl Med. Jul 2008;49(7):1053-1059.

105. Nakamoto Y, Tamai K, Saga T, et al. Clinical value of image fusion from MR and PET in patients with head and neck cancer. Mol Imaging Biol. Jan-Feb 2009;11(1):46-53.

106. Nishiyama Y, Yamamoto Y, Kimura N, Ishikawa S, Sasakawa Y, Ohkawa M. Dual-time-point FDG-PET for evaluation of lymph node metastasis in patients with non-small-cell lung cancer. Ann Nucl Med. May 2008;22(4):245-250.

107. Okada M, Murakami T, Kumano S, et al. Integrated FDG-PET/CT compared with intravenous contrast-enhanced CT for evaluation of metastatic regional lymph nodes in patients with resectable early stage esophageal cancer. Ann Nucl Med. Jan 2009;23(1):73-80.

108. Rembielak A, Price P. The role of PET in target localization for radiotherapy treatment planning. Onkologie. Feb 2008;31(1-2):57-62.

109. Yoon DY, Hwang HS, Chang SK, et al. CT, MR, US,18F-FDG PET/CT, and their combined use for the assessment of cervical lymph node metastases in squamous cell carcinoma of the head and neck. Eur Radiol. Mar 2009;19(3):634-642.

110. Ueda S, Tsuda H, Asakawa H, et al. Clinicopathological and prognostic relevance of uptake level using 18F-fluorodeoxyglucose positron emission tomography/computed tomography fusion imaging (18F-FDG PET/CT) in primary breast cancer. Jpn J Clin Oncol. Apr 2008;38(4):250-258.

111. Garcia C, Flamen P. Role of positron emission tomography in the management of head and neck cancer in the molecular therapy era. Curr Opin Oncol. May 2008;20(3):275-279.

112. Ferumoxtran-10: AMI 227, AMI 227cn, AMI 27, BMS 180549, Combidex, G 53425, Sinerem. Drugs R D. 2003;4(1):52-54.

113. Anzai Y, McLachlan S, Morris M, Saxton R, Lufkin RB. Dextran-coated superparamagnetic iron oxide, an MR contrast agent for assessing lymph nodes in the head and neck. AJNR Am J Neuroradiol. Jan 1994;15(1):87-94.

114. Anzai Y, Piccoli CW, Outwater EK, et al. Evaluation of neck and body metastases to nodes with ferumoxtran 10-enhanced MR imaging: phase III safety and efficacy study. Radiology. Sep 2003;228(3):777-788.

115. Anzai Y, Prince MR. Iron oxide-enhanced MR lymphography: the evaluation of cervical lymph node metastases in head and neck cancer. J Magn Reson Imaging. Jan-Feb 1997;7(1):75-81.

132

116. Barentsz JO, Futterer JJ, Takahashi S. Use of ultrasmall superparamagnetic iron oxide in lymph node MR imaging in prostate cancer patients. Eur J Radiol. Sep 2007;63(3):369-372.

117. Bellin MF, Lebleu L, Meric JB. Evaluation of retroperitoneal and pelvic lymph node metastases with MRI and MR lymphangiography. Abdom Imaging. Mar-Apr 2003;28(2):155-163.

118. Bellin MF, Roy C, Kinkel K, et al. Lymph node metastases: safety and effectiveness of MR imaging with ultrasmall superparamagnetic iron oxide particles--initial clinical experience. Radiology. Jun 1998;207(3):799-808.

119. Bordat C, Sich M, Rety F, et al. Distribution of iron oxide nanoparticles in rat lymph nodes studied using electron energy loss spectroscopy (EELS) and electron spectroscopic imaging (ESI). J Magn Reson Imaging. Sep 2000;12(3):505-509.

120. Chambon C, Clement O, Le Blanche A, Schouman-Claeys E, Frija G. Superparamagnetic iron oxides as positive MR contrast agents: in vitro and in vivo evidence. Magn Reson Imaging. 1993;11(4):509-519.

121. Choi SH, Moon WK, Hong JH, et al. Lymph node metastasis: ultrasmall superparamagnetic iron oxide-enhanced MR imaging versus PET/CT in a rabbit model. Radiology. Jan 2007;242(1):137-143.

122. Harisinghani MG, Barentsz J, Hahn PF, et al. Noninvasive detection of clinically occult lymph-node metastases in prostate cancer. N Engl J Med. Jun 19 2003;348(25):2491-2499.

123. Li W, Tutton S, Vu AT, et al. First-pass contrast-enhanced magnetic resonance angiography in humans using ferumoxytol, a novel ultrasmall superparamagnetic iron oxide (USPIO)-based blood pool agent. J Magn Reson Imaging. Jan 2005;21(1):46-52.

124. Li Z. Efficay and Safety of Combidex (NDA 21-115). http://www.fda.gov/OHRMS/DOCKETS/ac/05/slides/2005-4095S1_02_01-FDA-Li.ppt.

Accessed November 24th, 2010. 125. Bermudez RS, Izaguirre A, Roach M, 3rd. State-of-the-art radiotherapy in the

management of clinically localized prostate carcinoma. Future Oncol. Feb 2007;3(1):103-111.

126. Roach M, 3rd. Targeting pelvic lymph nodes in men with intermediate- and high-risk prostate cancer, and confusion about the results of the randomized trials. J Clin Oncol. Aug 1 2008;26(22):3816-3817; author reply 3817-3818.

127. Eifel PJ, Winter K, Morris M, et al. Pelvic irradiation with concurrent chemotherapy versus pelvic and para-aortic irradiation for high-risk cervical cancer: an update of radiation therapy oncology group trial (RTOG) 90-01. J Clin Oncol. Mar 1 2004;22(5):872-880.

128. Grigsby PW. Postoperative irradiation in cervical cancer: prognostic factors and outcome. Radiat Med. Mar-Apr 2004;22(2):106-110.

129. Grigsby PW, Heydon K, Mutch DG, Kim RY, Eifel P. Long-term follow-up of RTOG 92-10: cervical cancer with positive para-aortic lymph nodes. Int J Radiat Oncol Biol Phys. Nov 15 2001;51(4):982-987.

130. Grigsby PW, Perez CA, Chao KS, Herzog T, Mutch DG, Rader J. Radiation therapy for carcinoma of the cervix with biopsy-proven positive para-aortic lymph nodes. Int J Radiat Oncol Biol Phys. Mar 1 2001;49(3):733-738.

133

131. Pearcey R, Brundage M, Drouin P, et al. Phase III trial comparing radical radiotherapy with and without cisplatin chemotherapy in patients with advanced squamous cell cancer of the cervix. J Clin Oncol. Feb 15 2002;20(4):966-972.

132. Portelance L, Chao KS, Grigsby PW, Bennet H, Low D. Intensity-modulated radiation therapy (IMRT) reduces small bowel, rectum, and bladder doses in patients with cervical cancer receiving pelvic and para-aortic irradiation. Int J Radiat Oncol Biol Phys. Sep 1 2001;51(1):261-266.

133. Rushdan MN, Tay EH, Khoo-Tan HS, et al. Tailoring the field and indication of adjuvant pelvic radiation for patients with FIGO stage Ib lymph nodes-negative cervical carcinoma following radical surgery based on the GOG score--a pilot study. Ann Acad Med Singapore. Jul 2004;33(4):467-472.

134. Stryker JA, Mortel R. Survival following extended field irradiation in carcinoma of cervix metastatic to para-aortic lymph nodes. Gynecol Oncol. Dec 2000;79(3):399-405.

135. Aizer AA, Yu JB, McKeon AM, Decker RH, Colberg JW, Peschel RE. Whole pelvic radiotherapy versus prostate only radiotherapy in the management of locally advanced or aggressive prostate adenocarcinoma. Int J Radiat Oncol Biol Phys. Dec 1 2009;75(5):1344-1349.

136. Barbera L, Thomas G. Management of early and locally advanced cervical cancer. Semin Oncol. Apr 2009;36(2):155-169.

137. Boran N, Akdag D, Halici F, et al. A retrospective analysis of the diameter of metastatic lymph nodes in apparently early stage endometrial cancer. Tumori. Sep-Oct 2008;94(5):681-685.

138. Cummings BJ. The Role of Radiation Therapy With 5-Fluorouracil in Anal Cancer. Semin Radiat Oncol. Oct 1997;7(4):306-312.

139. Cummings BJ. Current management of anal canal cancer. Semin Oncol. Dec 2005;32(6 Suppl 9):S123-128.

140. Cummings BJ, Keane TJ, O'Sullivan B, Wong CS, Catton CN. Epidermoid anal cancer: treatment by radiation alone or by radiation and 5-fluorouracil with and without mitomycin C. Int J Radiat Oncol Biol Phys. Oct 1991;21(5):1115-1125.

141. George L, Bladou F, Bardou VJ, et al. Clinical outcome in patients with locally advanced bladder carcinoma treated with conservative multimodality therapy. Urology. Sep 2004;64(3):488-493.

142. Gerard JP, Chapet O, Samiei F, et al. Management of inguinal lymph node metastases in patients with carcinoma of the anal canal: experience in a series of 270 patients treated in Lyon and review of the literature. Cancer. Jul 1 2001;92(1):77-84.

143. Kaufman DS. Challenges in the treatment of bladder cancer. Ann Oncol. May 2006;17 Suppl 5:v106-112.

144. Kong A, Powell M, Blake P. The role of postoperative radiotherapy in carcinoma of the endometrium. Clin Oncol (R Coll Radiol). Aug 2008;20(6):457-462.

145. Monk BJ, Tewari KS, Koh WJ. Multimodality therapy for locally advanced cervical carcinoma: state of the art and future directions. J Clin Oncol. Jul 10 2007;25(20):2952-2965.

146. Moore DH. Chemotherapy and radiation therapy in the treatment of squamous cell carcinoma of the vulva: Are two therapies better than one? Gynecol Oncol. Jun 2009;113(3):379-383.

134

147. Pagliaro LC, Crook J. Multimodality therapy in penile cancer: when and which treatments? World J Urol. Apr 2009;27(2):221-225.

148. Rogers LJ, Howard B, Van Wijk L, et al. Chemoradiation in advanced vulval carcinoma. Int J Gynecol Cancer. May 2009;19(4):745-751.

149. Tans L, Ansink AC, van Rooij PH, Kleijnen C, Mens JW. The Role of Chemo-Radiotherapy in the Management of Locally Advanced Carcinoma of the Vulva: Single Institutional Experience and Review of Literature. Am J Clin Oncol. Jan 15 2010.

150. Tzioras S, Pavlidis N, Paraskevaidis E, Ioannidis JP. Effects of different chemotherapy regimens on survival for advanced cervical cancer: systematic review and meta-analysis. Cancer Treat Rev. Feb 2007;33(1):24-38.

151. Creutzberg CL, van Putten WL, Koper PC, et al. The morbidity of treatment for patients with Stage I endometrial cancer: results from a randomized trial. Int J Radiat Oncol Biol Phys. Dec 1 2001;51(5):1246-1255.

152. Kirwan JM, Symonds P, Green JA, Tierney J, Collingwood M, Williams CJ. A systematic review of acute and late toxicity of concomitant chemoradiation for cervical cancer. Radiother Oncol. Sep 2003;68(3):217-226.

153. Perez CA, Grigsby PW, Lockett MA, Chao KS, Williamson J. Radiation therapy morbidity in carcinoma of the uterine cervix: dosimetric and clinical correlation. Int J Radiat Oncol Biol Phys. Jul 1 1999;44(4):855-866.

154. Dinter DJ, Hofheinz RD, Hartel M, Kaehler GF, Neff W, Diehl SJ. Preoperative staging of rectal tumors: comparison of endorectal ultrasound, hydro-CT, and high-resolution endorectal MRI. Onkologie. May 2008;31(5):230-235.

155. Malmstrom PU. Lymph node staging in prostatic carcinoma revisited. Acta Oncol. 2005;44(6):593-598.

156. Chao KS, Lin M. Lymphangiogram-assisted lymph node target delineation for patients with gynecologic malignancies. Int J Radiat Oncol Biol Phys. Nov 15 2002;54(4):1147-1152.

157. Mundt AJ, Lujan AE, Rotmensch J, et al. Intensity-modulated whole pelvic radiotherapy in women with gynecologic malignancies. Int J Radiat Oncol Biol Phys. Apr 1 2002;52(5):1330-1337.

158. Mundt AJ, Roeske JC, Lujan AE. Intensity-modulated radiation therapy in gynecologic malignancies. Med Dosim. Summer 2002;27(2):131-136.

159. Salama JK, Roeske JC, Mehta N, Mundt AJ. Intensity-modulated radiation therapy in gynecologic malignancies. Curr Treat Options Oncol. Apr 2004;5(2):97-108.

160. Finlay MH, Ackerman I, Tirona RG, Hamilton P, Barbera L, Thomas G. Use of CT simulation for treatment of cervical cancer to assess the adequacy of lymph node coverage of conventional pelvic fields based on bony landmarks. Int J Radiat Oncol Biol Phys. Jan 1 2006;64(1):205-209.

161. Bonin SR, Lanciano RM, Corn BW, Hogan WM, Hartz WH, Hanks GE. Bony landmarks are not an adequate substitute for lymphangiography in defining pelvic lymph node location for the treatment of cervical cancer with radiotherapy. Int J Radiat Oncol Biol Phys. Jan 1 1996;34(1):167-172.

162. McAlpine J, Schlaerth JB, Lim P, Chen D, Eisenkop SM, Spirtos NM. Radiation fields in gynecologic oncology: correlation of soft tissue (surgical) to radiologic landmarks. Gynecol Oncol. Jan 2004;92(1):25-30.

135

163. Pendlebury SC, Cahill S, Crandon AJ, Bull CA. Role of bipedal lymphangiogram in radiation treatment planning for cervix cancer. Int J Radiat Oncol Biol Phys. Nov 15 1993;27(4):959-962.

164. Bipat S, Glas AS, van der Velden J, Zwinderman AH, Bossuyt PM, Stoker J. Computed tomography and magnetic resonance imaging in staging of uterine cervical carcinoma: a systematic review. Gynecol Oncol. Oct 2003;91(1):59-66.

165. Golder WA. Lymph node diagnosis in oncologic imaging: a dilemma still waiting to be solved. Onkologie. Apr 2004;27(2):194-199.

166. Hacker A, Jeschke S, Leeb K, et al. Detection of pelvic lymph node metastases in patients with clinically localized prostate cancer: comparison of [18F]fluorocholine positron emission tomography-computerized tomography and laparoscopic radioisotope guided sentinel lymph node dissection. J Urol. Nov 2006;176(5):2014-2018; discussion 2018-2019.

167. Park JY, Kim EN, Kim DY, et al. Comparison of the validity of magnetic resonance imaging and positron emission tomography/computed tomography in the preoperative evaluation of patients with uterine corpus cancer. Gynecol Oncol. Mar 2008;108(3):486-492.

168. Taylor A, Rockall AG, Powell ME. An atlas of the pelvic lymph node regions to aid radiotherapy target volume definition. Clin Oncol (R Coll Radiol). Sep 2007;19(7):542-550.

169. Taylor A, Rockall AG, Reznek RH, Powell ME. Mapping pelvic lymph nodes: guidelines for delineation in intensity-modulated radiotherapy. Int J Radiat Oncol Biol Phys. Dec 1 2005;63(5):1604-1612.

170. Shih HA, Harisinghani M, Zietman AL, Wolfgang JA, Saksena M, Weissleder R. Mapping of nodal disease in locally advanced prostate cancer: rethinking the clinical target volume for pelvic nodal irradiation based on vascular rather than bony anatomy. Int J Radiat Oncol Biol Phys. Nov 15 2005;63(4):1262-1269.

171. Tabatabaei S, Harisinghani M, McDougal WS. Regional lymph node staging using lymphotropic nanoparticle enhanced magnetic resonance imaging with ferumoxtran-10 in patients with penile cancer. J Urol. Sep 2005;174(3):923-927; discussion 927.

172. Harada T, Tanigawa N, Matsuki M, Nohara T, Narabayashi I. Evaluation of lymph node metastases of breast cancer using ultrasmall superparamagnetic iron oxide-enhanced magnetic resonance imaging. Eur J Radiol. Sep 2007;63(3):401-407.

173. Curvo-Semedo L, Diniz M, Migueis J, et al. USPIO-enhanced magnetic resonance imaging for nodal staging in patients with head and neck cancer. J Magn Reson Imaging. Jul 2006;24(1):123-131.

174. Thoeny HC, Triantafyllou M, Birkhaeuser FD, et al. Combined ultrasmall superparamagnetic particles of iron oxide-enhanced and diffusion-weighted magnetic resonance imaging reliably detect pelvic lymph node metastases in normal-sized nodes of bladder and prostate cancer patients. Eur Urol. Apr 2009;55(4):761-769.

175. Foster PJ, Dunn EA, Karl KE, et al. Cellular magnetic resonance imaging: in vivo imaging of melanoma cells in lymph nodes of mice. Neoplasia. Mar 2008;10(3):207-216.

176. Wang CJ, Chin YY, Leung SW, Chen HC, Sun LM, Fang FM. Topographic distribution of inguinal lymph nodes metastasis: significance in determination of

136

treatment margin for elective inguinal lymph nodes irradiation of low pelvic tumors. Int J Radiat Oncol Biol Phys. Apr 1 1996;35(1):133-136.

177. Agurio A, Filippone A, Sanpaolo P. Pelvic Lymph Nodes. . In: Ausili Cefaro GP CAG, D.; Vinciguerra, A., editor. , ed. A Guide for Delineation of Lymph Nodal Clinical Target Volume in Radiation Therapy: Springer; 2008:154-160.

178. Stehman FB, Bundy BN, Thomas G, et al. Groin dissection versus groin radiation in carcinoma of the vulva: a Gynecologic Oncology Group study. Int J Radiat Oncol Biol Phys. 1992;24(2):389-396.

179. Koh WJ, Chiu M, Stelzer KJ, et al. Femoral vessel depth and the implications for groin node radiation. Int J Radiat Oncol Biol Phys. Nov 15 1993;27(4):969-974.

180. McCall AR, Olson MC, Potkul RK. The variation of inguinal lymph node depth in adult women and its importance in planning elective irradiation for vulvar cancer. Cancer. May 1 1995;75(9):2286-2288.

181. Kalidas H. Influence of inguinal node anatomy on radiation therapy techniques. Med Dosim. Winter 1995;20(4):295-300.

182. Rosset A, Spadola L, Ratib O. OsiriX: an open-source software for navigating in multidimensional DICOM images. J Digit Imaging. Sep 2004;17(3):205-216.

183. Dinniwell R, Chan P, Czarnota G, et al. Pelvic lymph node topography for radiotherapy treatment planning from ferumoxtran-10 contrast-enhanced magnetic resonance imaging. Int J Radiat Oncol Biol Phys. Jul 1 2009;74(3):844-851.

184. Micheletti L, Levi AC, Bogliatto F, Preti M, Massobrio M. Rationale and definition of the lateral extension of the inguinal lymphadenectomy for vulvar cancer derived from an embryological and anatomical study. J Surg Oncol. Sep 2002;81(1):19-24.

185. Borgno G, Micheletti L, Barbero M, et al. Topographic distribution of groin lymph nodes. A study of 50 female cadavers. J Reprod Med. Dec 1990;35(12):1127-1129.

186. Nicklin JL, Hacker NF, Heintze SW, Van Eijkeren M, Durham NJ. An anatomical study of inguinal lymph node topography and clinical implications for the surgical management of vulval cancer. Int J Gynecol Cancer. Mar 1995;5(2):128-133.

187. Micheletti L, Levi AC, Bogliatto F. Anatomosurgical implications derived from an embryological study of the Scarpa's triangle with particular reference to groin lymphadenectomy. Gynecol Oncol. Sep 1998;70(3):358-364.

188. Dinniwell R, Chan P, Kim J, Milosevic M, Haider M. Inguinal nodal clinical target volume delineation: Validation of a class solution for precision radiotherapy. Radiother Oncol. 2008;88(Supp 1):S28.

189. Fernando IN, Moskovic E, Fryatt I, Glaholm J, Blake PR. Is there still a role for lymphography in the management of early stage carcinoma of the cervix? Br J Radiol. Nov 1994;67(803):1052-1056.

190. Hanks GE, Bagshaw MA. Megavoltage radiation therapy and lymphangiography in ovarian cancer. Radiology. Sep 1969;93(3):649-654.

191. Hliniakowa I, Skalska-Vorbrodt J, Godawski J. [Value of lymphography in establishing areas of irradiation]. Pol Przegl Radiol Med Nukl. May-Jun 1972;36(3):391-394.

192. Johnson DW, Cox RS, Howes AE, Castellino RA, Martinez A. The use of para-aortic radiation therapy based on lymphangiogram interpretation in uterine cervical carcinoma. Gynecol Oncol. Dec 1983;16(3):326-333.

137

193. Keinert K, Schumann E, Kirsch M. [The use of lymphography in dose planning and localization method in radiation therapy (author's transl)]. Radiobiol Radiother (Berl). 1973;14(4):337-344.

194. Lameer C. Lymphography in the intensive radiological treatment of Hodgkin's disease. Radiol Clin Biol. 1969;38(1):60-66.

195. Macdonald JS. The value of lymphography to the radiotherapist. Clin Radiol. Oct 1969;20(4):447-452.

196. Mendenhall NP, Holland KW, Sombeck MD. The role of lymphangiography in designing fields for elective pelvic node irradiation in Hodgkin's disease. Int J Radiat Oncol Biol Phys. Nov 15 1994;30(4):993-995.

197. van den Berg HA, Olofsen-van Acht MJ, van Santvoort JP, Seven M, Quint S, Levendag PC. Definition and validation of a reference target volume in early stage node-positive cervical carcinoma, based on lymphangiograms and CT-scans. Radiother Oncol. Feb 2000;54(2):163-170.

198. Weisenburger TH, Juillard G. Axillary lymphangiograms in radiation therapy of lymphomas. Preliminary report. Radiology. Nov 1974;113(2):463-465.

199. Weisenburger TH, Juillard GJ. Upper extremity lymphangiography in the radiation therapy of lymphomas and carcinoma of the breast. Radiology. Jan 1977;122(1):227-230.

200. Zunino S, Rosato O, Lucino S, Jauregui E, Rossi L, Venencia D. Anatomic study of the pelvis in carcinoma of the uterine cervix as related to the box technique. Int J Radiat Oncol Biol Phys. Apr 1 1999;44(1):53-59.

201. Arii K, Takifuji K, Yokoyama S, et al. Preoperative evaluation of pelvic lateral lymph node of patients with lower rectal cancer: comparison study of MR imaging and CT in 53 patients. Langenbecks Arch Surg. Sep 2006;391(5):449-454.

202. Bipat S, Glas AS, Slors FJ, Zwinderman AH, Bossuyt PM, Stoker J. Rectal cancer: local staging and assessment of lymph node involvement with endoluminal US, CT, and MR imaging--a meta-analysis. Radiology. Sep 2004;232(3):773-783.

203. von Schulthess GK. Positron emission tomography versus positron emission tomography/computed tomography: from "unclear" to "new-clear" medicine. Mol Imaging Biol. Jul-Aug 2004;6(4):183-187.

204. Frey EF. The earliest Egyptian medical texts. Bookman. Nov-Dec 1984;11(6):3-7. 205. Frey EF. The earliest medical texts. Clio Med. 1985;20(1-4):79-90. 206. Cruse JM. History of medicine: the metamorphosis of scientific medicine in the

ever-present past. Am J Med Sci. Sep 1999;318(3):171-180. 207. Sakai T. Historical evolution of anatomical terminology from ancient to modern.

Anat Sci Int. Jun 2007;82(2):65-81. 208. Joutsivuo T. [Vesalius and De humani corporis fabrica: Galen's errors and the

change of anatomy in the sixteenth century]. Hippokrates (Helsinki). 1997:98-112.

209. Neal PJ, Shapiro LG, Rosse C. The digital anatomist structural abstraction: a scheme for the spatial description of anatomical entities. Proc AMIA Symp. 1998:423-427.

210. Agoncillo AV, Mejino JL, Jr., Rosse C. Influence of the Digital Anatomist Foundational Model on traditional representations of anatomical concepts. Proc AMIA Symp. 1999:2-6.

138

211. Kalet IJ, Wu J, Lease M, Austin-Seymour MM, Brinkley JF, Rosse C. Anatomical information in radiation treatment planning. Proc AMIA Symp. 1999:291-295.

212. Kim S, Brinkley JF, Rosse C. Design features of on-line anatomy information resources: a comparison with the Digital Anatomist. Proc AMIA Symp. 1999:560-564.

213. Mejino JL, Jr., Rosse C. Conceptualization of anatomical spatial entities in the Digital Anatomist Foundational Model. Proc AMIA Symp. 1999:112-116.

214. Rosse C. Anatomy atlases. Clin Anat. 1999;12(4):293-299. 215. Rosse C, Shapiro LG, Brinkley JF. The digital anatomist foundational model:

principles for defining and structuring its concept domain. Proc AMIA Symp. 1998:820-824.

216. Hilloowala R. Illustrations from the Wellcome Institute Library. The origin of the Wellcome anatomical waxes. Albinus and the Florentine collection at La Specola. Med Hist. Oct 1984;28(4):432-437.

217. Hilloowala R, Renahan J. XVIII century anatomical models at La Specola, Florence. Anat Anz. 1985;159(1-5):141-158.

218. Holubar K. The anatomical wax preparations in the Josephinum in Vienna, Austria (on the improvement of surgical education 200 years ago). Arch Surg. Apr 1991;126(4):421-422.

219. Turinsky AL, Fanea E, Trinh Q, et al. Integration of genomic and medical data into a 3D atlas of human anatomy. Stud Health Technol Inform. 2008;132:526-531.

220. Turinsky AL, Fanea E, Trinh Q, et al. CAVEman: Standardized anatomical context for biomedical data mapping. Anat Sci Educ. Jan 2008;1(1):10-18.

221. Talairach J, Tournoux P. Co-planar Stereotaxic Atlas of the Human Brain: 3-Dimensional Proportional System - an Approach to Cerebral Imaging. New York: Thieme Medical Publishers; 1988.

222. Hill DLG, Hajnal JV, Rueckert D, Smith SM, Hartkens T, McLeish K. A Dynamic Brain Atlas, in Medical Image Computing and Computer-Assisted Intervention — MICCAI 2002. In: Dohi T, Kikinis R, eds: Springer; 2002:532-539.

223. Subburaj K, Ravi B, Agarwal M. Automated identification of anatomical landmarks on 3D bone models reconstructed from CT scan images. Comput Med Imaging Graph. Jul 2009;33(5):359-368.

224. Qatarneh SM, Kiricuta IC, Brahme A, Tiede U, Lind BK. Three-dimensional atlas of lymph node topography based on the visible human data set. Anat Rec B New Anat. May 2006;289(3):98-111.

225. Nowinski WL. The cerefy brain atlases: continuous enhancement of the electronic talairach-tournoux brain atlas. Neuroinformatics. 2005;3(4):293-300.

226. Dorfman RE, Alpern MB, Gross BH, Sandler MA. Upper abdominal lymph nodes: criteria for normal size determined with CT. Radiology. Aug 1991;180(2):319-322.

227. Grubnic S, Vinnicombe SJ, Norman AR, Husband JE. MR evaluation of normal retroperitoneal and pelvic lymph nodes. Clin Radiol. Mar 2002;57(3):193-200; discussion 201-194.

228. Vinnicombe SJ, Norman AR, Nicolson V, Husband JE. Normal pelvic lymph nodes: evaluation with CT after bipedal lymphangiography. Radiology. Feb 1995;194(2):349-355.

139

229. Grey AC, Carrington BM, Hulse PA, Swindell R, Yates W. Magnetic resonance appearance of normal inguinal nodes. Clin Radiol. Feb 2000;55(2):124-130.

230. Steinkamp HJ, Cornehl M, Hosten N, Pegios W, Vogl T, Felix R. Cervical lymphadenopathy: ratio of long- to short-axis diameter as a predictor of malignancy. Br J Radiol. Mar 1995;68(807):266-270.

231. Schwartz LH, Bogaerts J, Ford R, et al. Evaluation of lymph nodes with RECIST 1.1. Eur J Cancer. Jan 2009;45(2):261-267.

232. Matsuoka H, Nakamura A, Masaki T, et al. Optimal diagnostic criteria for lateral pelvic lymph node metastasis in rectal carcinoma. Anticancer Res. Sep-Oct 2007;27(5B):3529-3533.

233. Lawton CA, Michalski J, El-Naqa I, et al. RTOG GU Radiation oncology specialists reach consensus on pelvic lymph node volumes for high-risk prostate cancer. Int J Radiat Oncol Biol Phys. Jun 1 2009;74(2):383-387.

234. Lawton CA, Michalski J, El-Naqa I, et al. Variation in the definition of clinical target volumes for pelvic nodal conformal radiation therapy for prostate cancer. Int J Radiat Oncol Biol Phys. Jun 1 2009;74(2):377-382.

235. Small WJ, Mell L, Anderson P, et al. Consensus guidelines for delineation of clinical target volume for intensity-modulated pelvic radiotherapy in postoperative treatment of endometrial and cervical cancer. International Journal of Radiation Oncology, Biology, Physics. 2008 Jun 1 2008;71(2):428-434.

236. Myerson RJ, Garofalo MC, El Naqa I, et al. Elective clinical target volumes for conformal therapy in anorectal cancer: a radiation therapy oncology group consensus panel contouring atlas. Int J Radiat Oncol Biol Phys. Jul 1 2009;74(3):824-830.

237. Lujan AE, Mundt AJ, Yamada SD, Rotmensch J, Roeske JC. Intensity-modulated radiotherapy as a means of reducing dose to bone marrow in gynecologic patients receiving whole pelvic radiotherapy. Int J Radiat Oncol Biol Phys. Oct 1 2003;57(2):516-521.

238. Mundt AJ, Roeske JC, Lujan AE, et al. Initial clinical experience with intensity-modulated whole-pelvis radiation therapy in women with gynecologic malignancies. Gynecol Oncol. Sep 2001;82(3):456-463.

239. Nutting CM, Convery DJ, Cosgrove VP, et al. Reduction of small and large bowel irradiation using an optimized intensity-modulated pelvic radiotherapy technique in patients with prostate cancer. Int J Radiat Oncol Biol Phys. Oct 1 2000;48(3):649-656.

240. Roeske JC, Lujan A, Rotmensch J, Waggoner SE, Yamada D, Mundt AJ. Intensity-modulated whole pelvic radiation therapy in patients with gynecologic malignancies. Int J Radiat Oncol Biol Phys. Dec 1 2000;48(5):1613-1621.

241. Pan WR, Rozen WM, Stella DL, Ashton MW. A three-dimensional analysis of the lymphatics of a bilateral breast specimen: a human cadaveric study. Clin Breast Cancer. May 2009;9(2):86-91.

242. Rivers CS, Wardlaw JM. What has diffusion imaging in animals told us about diffusion imaging in patients with ischaemic stroke? Cerebrovasc Dis. 2005;19(5):328-336.

243. Rivers CS, Wardlaw JM, Armitage PA, et al. Persistent infarct hyperintensity on diffusion-weighted imaging late after stroke indicates heterogeneous, delayed, infarct evolution. Stroke. Jun 2006;37(6):1418-1423.

140

244. Rivers CS, Wardlaw JM, Armitage PA, et al. Do acute diffusion- and perfusion-weighted MRI lesions identify final infarct volume in ischemic stroke? Stroke. Jan 2006;37(1):98-104.

245. Rivers CS, Wardlaw JM, Armitage PA, Bastin ME, Hand PJ, Dennis MS. Acute ischemic stroke lesion measurement on diffusion-weighted imaging--important considerations in designing acute stroke trials with magnetic resonance imaging. J Stroke Cerebrovasc Dis. Mar-Apr 2007;16(2):64-70.

246. Benedict RH, Bruce J, Dwyer MG, et al. Diffusion-weighted imaging predicts cognitive impairment in multiple sclerosis. Mult Scler. Jul 2007;13(6):722-730.

247. Lichy MP, Aschoff P, Plathow C, et al. Tumor detection by diffusion-weighted MRI and ADC-mapping--initial clinical experiences in comparison to PET-CT. Invest Radiol. Sep 2007;42(9):605-613.

248. Sharma U, Danishad KK, Seenu V, Jagannathan NR. Longitudinal study of the assessment by MRI and diffusion-weighted imaging of tumor response in patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy. NMR Biomed. Jan 2009;22(1):104-113.

249. Dirix P, Vandecaveye V, De Keyzer F, et al. Diffusion-weighted MRI for nodal staging of head and neck squamous cell carcinoma: impact on radiotherapy planning. Int J Radiat Oncol Biol Phys. Mar 1;76(3):761-766.

250. Bockisch A, Freudenberg LS, Schmidt D, Kuwert T. Hybrid imaging by SPECT/CT and PET/CT: proven outcomes in cancer imaging. Semin Nucl Med. Jul 2009;39(4):276-289.

251. Spanu A, Chessa F, Sanna D, et al. Breast cancer axillary lymph node metastasis detection by a high-resolution dedicated breast camera: a comparative study with SPECT and pinhole SPECT. Cancer Biother Radiopharm. Dec 2007;22(6):799-811.

252. Takahashi M, Sasa M, Hirose C, et al. Clinical efficacy and problems with CT lymphography in identifying the sentinel node in breast cancer. World J Surg Oncol. 2008;6:57.

253. Ue H, Ogawa Y, Kariya S, et al. Sentinel lymph node detection using computed tomography lymphography is accurate after neoadjuvant chemotherapy for breast cancer. Oncol Rep. Jun 2007;17(6):1475-1480.

254. Yamashita K, Shimizu K. Evaluation of sentinel lymph node metastasis alone guided by three-dimensional computed tomographic lymphography in video-assisted breast surgery. Surg Endosc. Mar 2009;23(3):633-640.

255. Kobayashi H, Ogawa M, Kosaka N, Choyke PL, Urano Y. Multicolor imaging of lymphatic function with two nanomaterials: quantum dot-labeled cancer cells and dendrimer-based optical agents. Nanomed. Jun 2009;4(4):411-419.

256. Kosaka N, Ogawa M, Sato N, Choyke PL, Kobayashi H. In vivo real-time, multicolor, quantum dot lymphatic imaging. J Invest Dermatol. Dec 2009;129(12):2818-2822.

257. Hild WA, Breunig M, Goepferich A. Quantum dots - nano-sized probes for the exploration of cellular and intracellular targeting. Eur J Pharm Biopharm. Feb 2008;68(2):153-168.

258. Pouliot F, Johnson M, Wu L. Non-invasive molecular imaging of prostate cancer lymph node metastasis. Trends Mol Med. Jun 2009;15(6):254-262.

259. Iyer M, Wu L, Carey M, Wang Y, Smallwood A, Gambhir SS. Two-step transcriptional amplification as a method for imaging reporter gene expression

141

using weak promoters. Proc Natl Acad Sci U S A. Dec 4 2001;98(25):14595-14600.

260. http://www.definiens.com/definiens-lymphexpert_186_7_50.html. Accessed November 22nd, 2009.

261. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. Jan 2009;45(2):228-247.

262. Noble JH, Dawant BM, Warren FM, Labadie RF. Automatic identification and 3D rendering of temporal bone anatomy. Otol Neurotol. Jun 2009;30(4):436-442.

263. Handels H, Hacker S. A framework for representation and visualization of 3D shape variability of organs in an interactive anatomical atlas. Methods Inf Med. 2009;48(3):272-281.