Guidance document on delivery, treatment planning, and clinical implementation …lcr.uerj.br/Manual_ABFM/Guidance document on delivery... · 2010. 10. 11. · Guidance document on

Guidance document on delivery, treatment planning, and clinicalimplementation of IMRT: Report of the IMRT subcommitteeof the AAPM radiation therapy committee

Gary A. EzzellMayo Clinic, Scottsdale, Arizona 85259

James M. GalvinThomas Jefferson University Hospital, Philadelphia, Pennsylvania 19019

Daniel LowMallinckrodt Institute of Radiology, St. Louis, Missouri 63101

Jatinder R. Paltaa)

University of Florida, Gainesville, Florida 32610

Isaac RosenUT M.D. Anderson Cancer Center, Houston, Texas 77001

Michael B. SharpePrincess Margaret Hospital, Toronto, Ontario M5G 2M9, Canada

Ping XiaUniversity of California at San Francisco, San Francisco, California 94101

Ying XiaoThomas Jefferson University Hospital, Philadelphia, Pennsylvania 19019

Lei XingStanford University School of Medicine, Stanford, California 94305

Cedric X. YuUniversity of Maryland School of Medicine, Baltimore, Maryland 21201

~Received 27 August 2002; accepted for publication 21 March 2003; published 24 July 2003!

Intensity-modulated radiation therapy~IMRT! represents one of the most significant technical ad-vances in radiation therapy since the advent of the medical linear accelerator. It allows the clinicalimplementation of highly conformal nonconvex dose distributions. This complex but promisingtreatment modality is rapidly proliferating in both academic and community practice settings.However, these advances do not come without a risk. IMRT is not just an add-on to the currentradiation therapy process; it represents a new paradigm that requires the knowledge of multimo-dality imaging, setup uncertainties and internal organ motion, tumor control probabilities, normaltissue complication probabilities, three-dimensional~3-D! dose calculation and optimization, anddynamic beam delivery of nonuniform beam intensities. Therefore, the purpose of this report is toguide and assist the clinical medical physicist in developing and implementing a viable and safeIMRT program. The scope of the IMRT program is quite broad, encompassing multileaf-collimator-based IMRT delivery systems, goal-based inverse treatment planning, and clinical implementationof IMRT with patient-specific quality assurance. This report, while not prescribing specific proce-dures, provides the framework and guidance to allow clinical radiation oncology physicists to makejudicious decisions in implementing a safe and efficient IMRT program in their clinics. ©2003American Association of Physicists in Medicine.@DOI: 10.1118/1.1591194#

Key words: 3-D conformal radiotherapy, intensity-modulated radiation therapy, inverse planning,plan optimization, quality assurance

1
TABLE OF CONTENTS
I. INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . .2090A. Relation of intensity-modulated radiation

therapy~IMRT!, three-dimensionalconformal radiation therapy~3DCRT!, andtraditional practice. . . . . . . . . . . . . . . . . . . . . . . .2090

B. Objectives for this document.. . . . . . . . . . . . . . . 2090C. Organization of this document. . . . . . . . . . . . . . . 2091

2089 Med. Phys. 30 „8…, August 2003 0094-2405 Õ2003Õ30„

II. DELIVERY SYSTEMS FOR IMRT. . . . . . . . . . . . 2091A. General issues of IMRT delivered with MLC. . 209

1. MLC leaf positional accuracy. . . .. . . . . . . . 20912. Linac performance

for small MU delivery. . . . . . . . . . . . . . . . . . 20933. MLC control issues. . . . . . . . . . . . . . . . . . . . 20934. MLC physical characteristics. . . . . . . . . . . . . 2093

B. Additional issues with dynamic IMRT withMLC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2094

20898…Õ2089Õ27Õ$20.00 © 2003 Am. Assoc. Phys. Med.

0

2

mari-ee-nalifiedm’slu-to-sdis-s,

lso,ruc-pro-

ithu-

emsnewustowd

mn-so

on-ro-

ce

2090 Ezzell et al. : Clinical implementation of IMRT 2090

1. MLC leaf positional and leaf speedaccuracy.. . . . . . . . . . . . . . . . . . . . . . . . . . . .2094

2. Other dynamic MLC issues. . . . . . . . . . . . . . 2095C. IMRT with physical attenuators. . . . . . . . . . . . . . 2095D. IMRT with rotating fan beams~tomotherapy!. . 2095

1. Peacock positional accuracy. . . . . . . . . . . . . 20952. Peacock dosimetric measurements. . . . . . . . . 20963. Helical tomotherapy. . . . . . . . . . . . . . . . . . . . 2096

E. IMRT with rotating cone beams~intensity-modulated arc therapy!. . . . . . . . . . . . 2096

F. Leaf sequencing for segmental and dynamicIMRT with MLCs. . . . . . . . . . . . . . . . . . . . . . . . .20961. Sliding window algorithms. . . . . . . . . . . . . . 20962. Areal or reducing algorithms. . . . . . . . . . . . . 2097

III. TREATMENT PLANNING FOR IMRT. . . . . . . . 2098A. Differences between IMRT and conventional

treatment planning: dose calculations andbeam modeling. . . . . . . . . . . . . . . . . . . . . . . . . . .20981. Modeling head scatter, penumbra, and

transmission. . . . . . . . . . . . . . . . . . . . . . . . . .20982. Leaf sequencing and deliverability. . . . . . . . 20983. Heterogeneity corrections.. . . . . . . . . . . . . . . 2098

B. Differences between IMRT and conventionaltreatment planning: Planning algorithms. . . . . . . 2099

C. Differences between IMRT and conventionaltreatment planning: Specific planning issues. . . 211. Dose uniformity vs dose shaping. . . . . . . . . 21002. Target and structure delineation. .. . . . . . . . . 21003. Dose grid. . . . . . . . . . . . . . . . . . . . . . . . . . . .21014. Buildup region. . . . . . . . . . . . . . . . . . . . . . . .21015. Flash and mobile targets. . . . . . . . . . . . . . . . 21016. Margins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21017. Radiobiologic issues.. . . . . . . . . . . . . . . . . . . 21028. Plan evaluation. .. . . . . . . . . . . . . . . . . . . . . .2102

D. Learning how to use the inverse planningsystem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2103

E. Commissioning an IMRT planning system fordosimetric accuracy. . . . . . . . . . . . . . . . . . . . . . .2103

F. QA of individual treatment plans. . . . . . . . . . . . 21051. Independent calculation methods. . . . . . . . . . 21052. Verification measurements. . .. . . . . . . . . . . . 21053. Other plan checks.. . . . . . . . . . . . . . . . . . . . .2106

IV. CLINICAL IMPLEMENTATION OF IMRT. . . . . 2106A. Overview. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2106B. Equipment and space requirements. . . . . . . . . . . 2107

1. Shielding. . . . . . . . . . . . . . . . . . . . . . .. . . . . . 21072. Space planning. . . . . . . . . . . . . . . . . . . . . . . .21073. Equipment. . . . . . . . . . . . . . . . . . . . . . . . . . . .2107

C. Time and personnel requirements. . . . . . . . . . . . 2107D. Changes in treatment planning and treatment

delivery process. . . . . . . . . . . . . . . . . . . . . . . . . .21071. General considerations. . . . . . . . . . . . . . . . . . 21072. Immobilization. . . . . . . . . . . . . . . . . . . . . . . .21073. Image acquisition. . . . . . . . . . . . . . . . . . . . . .21074. Structure segmentation. . . . . . . . . . . . . . . . . . 21085. IMRT treatment planning. . . . . . . . . . . . . . . . 21086. File transfer and management. . . .. . . . . . . . 21087. Plan validation. . . . . . . . . . . . . . . . . . . . . . . .2109

Medical Physics, Vol. 30, No. 8, August 2003

0

8. Position verification. . . . . . . . . . . . . . . . . . . . 21099. IMRT treatment delivery. . . . . . . . . . . . . . . . 2109

E. QA of equipment and individual patienttreatments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2110

F. Staff training and patient education.. . . . . . . . . . 21101. Radiation oncologists. . . . . . . . . . . . . . . . . . . 21102. Radiation oncology physicists. . . . . . . . . . . . 21103. Dosimetrists. . . . . . . . . . . . . . . . . . . . . . . . . . 21114. Radiation therapists. . . . . . . . . . . . . . . . . . . . 21115. Service engineers. . . . . . . . . . . . . . . . . . . . . . 21116. Patient education. . . . . . . . . . . . . . . . . . . . . . 2111

G. Patient scheduling, billing, and charting.. . . . . . 2112H. Overall integration. . . . . . . . . . . . . . . . . . . . . . . . 2112

V. SUMMARY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

I. INTRODUCTION

A. Relation of intensity-modulated radiation therapy„IMRT…, three-dimensional conformal radiationtherapy „3DCRT…, and traditional practice

IMRT is an extension of 3DCRT that uses nonuniforradiation beam intensities that have been determined by vous computer-based optimization techniques. Thrdimensional conformal therapy is a change from traditiopractice in that it uses targets and normal structures identon multiple transverse images, field design based on beaeye view projections, volumetric dose calculations, and vometric plan evaluation tools such as dose–volume hisgrams~DVHs!. IMRT uses all the tools of 3DCRT and addother novel features. IMRT seeks to further shape dosetributions by modulating the intensity of each field. Thunew capabilities of linear accelerators~linacs! and collima-tors must be installed, commissioned, and maintained. Acomputing the needed intensity patterns and machine insttions to create them complicates the treatment planningcess significantly. The computer algorithms associated wIMRT planning must be commissioned for dosimetric accracy. Users must learn how to use inverse planning systto produce and evaluate high quality plans. These aretasks that physicists and other radiation oncology staff maccomplish. Many physicists and their colleagues are nstruggling with the question of ‘‘what do I need to know ando to implement IMRT safely and effectively?’’

B. Objectives for this document

The objectives for this document are~a! to describe in general terms how IMRT differs fro

3DCRT with respect to treatment delivery, treatment planing, and clinical implementation and give referencesreaders can get more details if desired;

~b! to describe how these differences impact commissiing of the treatment planning and delivery systems, and pvide guidance on the commissioning process;

~c! to describe the impact on ongoing quality assuran~QA! and provide guidance on QA practice; and

ao

ellee

utiv

ea

. Il inificuem

soeriagvcuman

idine-dR

to

o

salacrodaaa

rto

uelli

achrise

hatcetic-n-fly.-ryedh-de-

sde-trywero-r-the

ecs.

ngous

hussn-

er-af

en-m-p ae ofve

ewer-slo-tri-y ofontic

nmi-dfinedine


~d! to describe how these processes fit together with eother and provide guidance on the clinical implementationIMRT.

Because of the emerging and rapidly changing natureIMRT, this document cannot be definitive or prescriptivTask group reports and Codes of Practice will eventuaemerge as the field matures. Our intention in this documis to provide guidance during this introductory period. Whave tried to avoid being overly repetitive of other docments, such as the recent report of the IMRT CollaboraWorking Group ~CWG!1 and special issues ofMedicalDosimetry,2,3 with which readers should also be familiar. Whave also consulted with ASTRO representatives whodeveloping recommendations for the clinical use of IMRTshould be recognized that the development of IMRT is stilits infancy and is rapidly evolving. Therefore, many specstatements made within this document are likely to be odated as the new generation of planning and delivery systbecome available.

C. Organization of this document

After this introductory section, this presentation followwith a description in Sec. II of delivery methods used fIMRT and associated commissioning and QA. An undstanding of delivery mechanisms is necessary to apprecsome of the factors that impact IMRT treatment planninSection III on treatment planning follows. That section coers commissioning a planning system for dosimetric acracy, which is inherently related to the delivery mechanisIt also covers learning how to effectively use an inverse plning system. These two sections address objectives~a!-~c!;that is, they explain the differences from 3DCRT and provguidance on commissioning and QA of treatment plannand delivery systems. Finally, in Sec. IV on clinical implmentation we outline the issues that have to be addressethe physicist and other team members in order to bring IMonline, and so we address objective~d!.

II. DELIVERY SYSTEMS FOR IMRT

The difference between 3DCRT and IMRT with respecttreatment delivery is implied in the phraseintensity modula-tion. Three-dimensional conformal therapy uses blocksmultileaf collimators~MLCs! to define fixed field bound-aries. Modulators such as wedges or tissue compensatoroften employed to improve dose homogeneity within the tget. IMRT extends the complexity of the intensity modution to achieve more complex dosimetric aims, such asating dose distributions with concavities. Many methodsachieving this modulation have been proposed and applieclinical practice. One class of techniques holds the bedirection constant during irradiation and indexes the collimtor shape to a fraction of the total prescribed MU for thdirection, thus subjecting any given point in the patient todesired proportion of ‘‘open’’ and ‘‘blocked’’ beam. Anothetechnique uses fixed gantry angles and physical attenuato achieve the modulation. Yet another class of techniqmoves the gantry during the irradiation, indexing the co


chf

of.ynt

-e

ret

t-s

r-te.--.-

eg

byT

r

arer--e-ftom-ta

rss

-

mator shape and gantry angle to the delivered dose. Edelivery technique has its own unique features that giveto different commissioning and QA considerations.

In this section we will emphasize those techniques thave been implemented commercially using MLCs sinthey are the most common and of widest interest to pracing medical physicists. It provides guidance for commissioing and QA for these. Other techniques are described brie

Although IMRT planning and delivery are intimately related, in this section we suggest tests of the IMRT delivesystem using MLC control files that have been developindependently from the IMRT planning system. In this fasion, the causes of dose deviations can be isolated to thelivery or planning system.

In Secs. II A and II B we describe IMRT delivery systemthat use fixed gantry angles and MLCs. In Sec. II C wescribe IMRT delivery systems that make use of fixed ganangles and physical attenuators. In Secs. II D and II Edescribe IMRT delivery systems that make use of gantrytations and MLC. In Sec. II F we provide background infomation on the leaf sequencing algorithms that are used insegmental and dynamic IMRT techniques described in SII A and II B.

A. General issues of IMRT delivered with MLC

The CWG recommends the termsegmental IMRT~SMLC-IMRT! when the collimator shape is constant duriirradiation and changes between irradiations. Synonymterms arestep-and-shootand stop-and-shoot. The gantrydoes not move during irradiation. Each collimator shape tis a subfield~or a segment!. The desired intensity pattern iobtained by the fractional weighted summation of the intesity pattern from all subfields.

1. MLC leaf positional accuracy

In conventional 3DCRT, the MLC defines the outer apture of the beam shape. An uncertainty of 1 to 2 mm in lelocation may be inconsequential to the output and, in geral, to clinical outcome, since the uncertainty is small copared with the aperture size. Segmental IMRT builds ufluence pattern by adding together many segments, somwhich may be quite narrow. Several investigators hashown that, for beam widths of 1 cm, uncertainties of a ftenths of a millimeter in leaf position can cause dose unctainties of several percent.4,5 Furthermore, the beam edgemove to many locations within the treated area, so theircations must be known to high precision so that their conbutions sum accurately. For these reasons, the accuracrelative MLC leaf position must be maintained to a precisiof better than a millimeter. Conventional QA tests for staMLCs are not sufficiently sensitive for this purpose.

A key point for IMRT is that the location of the radiatiofield edge must be well established with respect to the nonal location of the MLC leaf end. For MLCs with roundeleaf ends, there is an offset between the beam edge as deby the light field and that defined by the 50% decrement lof the radiation field.6 This is typically 0.4 to 1.1 mm de-

hoot


FIG. II.1. ~a! MLC test pattern with a 2 cm widestrip. ~b! QA film produced by moving the pattern in 2 cm intervals and irradiating in a step-and-sfashion. The strips should abut at the 50% decrement lines as described in Sec. II A 1. The line on the film shows the location of the scan~c!, which is usedto assess the quality of the matching. This MLC has a rounded leaf end design.

ititerato

ina

anm

sune

t

.a-ettoFoce

nge

i-

ll

am

Cates

mrynk

o-a

estncei-traloningandre-If ah-

pending on the MLC type, beam energy, and location wrespect to the central axis. An offset can also exist wdouble-focused MLCs if the MLC motion deviates from thdesired spherical arc. Users may have the choice of calibing their MLC so that the nominal position correspondsthe light field or radiation field edge.7 ~In practice, calibratingthe MLC nominal position to the light field edge has certaadvantages, especially if it is the standard method usedsupported by the vendor.!

Therefore, the physicist should perform the following.

~a! Measure the offset between the radiation field edgethe nominal leaf position as a function of distance frothe central axis, both positive and negative.~Often, theoffset can be treated as a constant value.!

~b! Create a test sequence that abuts irradiated stripdifferent locations across the field, adjusted to accofor any offset so that the 50% decrement lines supimpose.

~c! Irradiate a film and scan across the match linescheck the uniformity of the dose~Fig. II.1!.

The offset can be measured using the test sequencescribed in~b!–~c! with different values of the offset appliedAlternatively, the full width at half maximum can be mesured for strips of known nominal width to obtain the offsFilms should be obtained at different gantry and collimaangles to check the effect of gravity on the matchlines.MLC systems that employ carriage motion, sequen


hh

t-

nd

d

att

r-

o

de-

.rrs

should be created that test the matchlines over the full raof travel.

Tests of matchline uniformity can detect MLC leaf postion variation to a precision of about 0.2 mm.8,9 More precisecontrol is likely unattainable. This positional variation wiproduce a dose variation of about65% in the matchline andis unlikely to cause significant dose error when many besegments from many angles superimpose.

Another useful test to semi quantitatively check the MLleaf positional accuracy is to film a test sequence that cre1 mm strips at regular intervals.8 A visual inspection candetect improper positioning to a precision of about 0.5 m~Fig. II.2!. Again, such films should be at different gantand collimator angles and over the full range of leaf bamotion.

Physicists must comprehensively check the MLC leaf psitional accuracy during IMRT commissioning and developsubset of checks as part of routine QA. It is prudent to tfrequently at first and reduce the frequency as experiebuilds. In IMRT, unlike conventional treatment, MLC calbration and performance affect dose delivery to the centarget region. This program might include tests that focusmachine performance, such as a daily output check usmultiple narrow-segment tests, films as described above,might also include overall planning and delivery measuments for specific patients, as described in Sec. III F 2.facility moves toward using independent calculation tecniques to check individual patient plans~Sec. III F 1!, then

n

foaes

by.ayike

plusemure

toc-t

nalratetots,or

butat

roltheg-sss.ist

d

be--

im-owU.unt

is-

, so

and

aftor

theleafes-ing

oreoveateate

forLC

o


tests of machine performance will need to be performed oleast a weekly basis.

2. Linac performance for small MU delivery

Depending on the planning and delivery system usedIMRT, segments may be delivered with few or fractionMU. The dose-per-MU constancy should be checkthroughout the range of use for IMRT. Similarly, the flatneand symmetry of the beam should be checked.5,10 Fast filmsuch as Kodak TL can test for flatness and symmetry staity for a few MU, especially if placed on the blocking traSumming several irradiations of small or fractional MU malso be reasonable, since variations at low doses are unlto be clinically important unless they are systematic.

It has been noted that some delivery systems can disdosimetric discrepancies when using very few MU becaof the communication lag between the MLC control systand the linac console.9,11,12 These discrepancies can occwithin the normal range of use for clinical treatments. Th

FIG. II.2. ~a! MLC test pattern with a 1 mm widestrip. ~b! QA film producedby moving the pattern in 2 cm intervals and irradiating in a step-and-shfashion. This MLC has a rounded leaf end design.


at

rlds

il-

ly

aye

y

can affect film QA tests if the number of MU is reducedavoid saturating the film, but this may be mitigated by reduing the MU delivery rate proportionally. The clinical impacof this observation needs further investigation.

3. MLC control issues

Some linac manufacturers~e.g., Siemens! have imple-mented segmental IMRT as an extension of conventiotreatments: each IMRT segment is considered a sepafield. To be efficient, a computer control system is neededset up and verify the potentially large number of segmenbut the process is qualitatively the same for modulatedunmodulated fields. This simplifies the control system,the record/verify overhead limits the number of fields thcan be treated in a given time period. Others~e.g., Elekta andVarian! have developed a dedicated linac and MLC contsystem that directly controls and monitors the indexing ofMLC shape to the delivered MU. This permits more sements to be delivered in a given time at the cost of leopportunity for external verification of individual segmentRegardless of the delivery system, the clinical physicneeds to understand the following:~a! how the MLC is cali-brated,~b! how the MLC leaf position is indexed to MU anwhether fractional MU are permitted,~c! how and to whatprecision the MLC leaf position is measured,~d! what toler-ance applies to the MLC leaf position and whether it cancontrolled,~e! what interlocks check that the MLC leaf position is correct,~f! what verification records or logs are created by the control system,~g! how to respond if the QAchecks show that the calibration has drifted, and~h! how torecover from delivery interruptions.

4. MLC physical characteristics

The transmission characteristics of the MLC are moreportant for IMRT than for 3DCRT because the leaves shadthe treatment area for a large fraction of the delivered MTransmission through the leaf is important, as are the amoand consistency of interleaf leakage.13,14 ~This document ap-plies the term ‘‘leakage’’ for radiation that includes transmsion through materials plus transport through gaps.!. Mostplanning systems require an average transmission valuethe measurement device~film or chamber! should span alarge enough area to adequately sample interleaf leakageintraleaf transmission.

The penumbra of the leaf ends in the direction of letravel should be measured with a high-resolution detecsuch as a film or a diode to permit accurate modeling ofpenumbra by the planning system. Measurement of thepenumbra in the direction perpendicular to leaf travel is prently less of an issue, since most current treatment plannsystems do not model MLC leaf sides and therefore ignthe effect of interdigitations of leaves and tongue-and-groin the dose calculation. However, many reduce or elimininterdigitations in the leaf sequencing step and so mitigthis deficiency.

The available treatment area is less for IMRT thanconventional treatments because IMRT requires that an M

ot

oleththteth

obvenm

itao

thnietehd

II.fosihedt

ca

on

ome

gn-entaf

eausethe

ughn the

Corfor

nmaya-ndA,

e

u-heon-to

on-edatheapsedgu-xisndor

theeedafosso-ss aape

theted.inedgaping

of

fo

lsho

a


leaf traverse the entire field, not simply define an outer bder. Each manufacturer has different specifications forextension, travel across central axis, etc., that affectavailable treatment area. The physicist needs to knowspecifications in order to acceptance test the delivery sysand to test that the planning system correctly handleslimitations.15

The problem of leaf-end transmission is an examplehow MLC design decisions can interact in a way that canvery significant for IMRT. For example, Varian systems harounded leaf ends and the leakage for abutting ends ca20%.58 Furthermore, the control system forces a minimuseparation of 0.5 mm while leaves are moving. Interdigleaf motion is allowed, however, on some MLC systems. FSMLC-IMRT, this means that the leaves need not abut intreated area, but can be moved under a jaw. The plansystem should take advantage of that capability. For Siemunits, on the other hand, the leaf ends are straight and indigital leaf motion is not allowed, so abutments occur in ttreated area. Whether or not problematic leakage occurspends on the accuracy of the leaf positioning. Figureshows a QA film designed to test the effects of abutmentsa Siemens machine. A test pattern was designed to let clopositions of all 27 pairs of MLC leaves move across tIMRT field width while two small openings in the top anthe bottom of the field are used to force the upper jawsremain open. Incomplete abutments will show up as lotions of increased leakage.

B. Additional issues with dynamic IMRT with MLC

The CWG recommends the termdynamic IMRT~DMLC-IMRT! when the collimator shape changes during irradiati

FIG. II.3. ~a! MLC test pattern with all the leaves closed together exceptthe first and last pair, which have a 2 cm opening toforce the upper jaws toremain open.~b! QA film produced by moving the pattern in 2 cm intervaand irradiating in a step-and-shoot fashion. Incomplete abutments will sas strips of increased density. This MLC has a focused leaf end designabutting leaves should close completely.


r-afeeme

fe

be

lrengnsr-

ee-

3r

ng

o-

.

Sliding windowis a synonym~although that term has alsbeen used in the context of segmental MLC to describe soleaf sequencing strategies!. The gantry does not move durinirradiation. However, each pair of leaves in the MLC, defiing a gap, moves unidirectionally, each with an independvelocity as a function of time. Here, the leaf positions, lespeed, delivered MU, and dose rates all interact.

1. MLC leaf positional and leaf speed accuracy

In the DMLC method of IMRT delivery, because of threlatively small gaps between opposed leaves and becmost regions are shielded by leaves most of the time,delivered dose is very sensitive to the transmission throthe leaves and the rounded leaf ends, the leakage betweeleaves, and the magnitude of the extrafocal radiation~headscatter!; these may be of lesser importance in the SMLmethod of IMRT delivery. Therefore, the requirements fMLC leaf positional accuracy are even more stringentdynamic IMRT with MLCs.8,13,16

A key point is that the sensitivity of output to leaf positiodepends on the programmed gap between them, and thisfrequently be smaller for DMLC treatments than for alterntive SMLC treatments. The films suggested in Sec. II A 1 adepicted in Figs. 2.1 and 2.2 could be used for periodic Qand indeed were first suggested in the context of DMLC.8 Avariation of60.2 mm in the gap width can result in a dosvariation of 63% for each clinical DMLC field.13 In addi-tion, the accuracy of DMLC delivery depends on the accracy with which the speed of each leaf is controlled. Tdose rate of the linac is also a related variable, and the ctrol system may vary the leaf speed, dose rate, or bothachieve the desired result. Test patterns should be cstructed to check conditions that are limited by leaf speand dose rate.9,17–19For example, a test pattern could movegap that is 1 cm wide by several centimeters long acrosscentral axis. The gap should travel a fixed distance, perhspanning the maximum field width. Varying the programmMU will cause the dose rate and/or leaf speed to be related. The reading of an ion chamber at the central ashould be directly proportional to the programmed MU, adeviations from that proportionality would be indicators fconcern.

Of course, such a test only checks leaves that coverion chamber position. Film can be used to test leaf spstability for several leaves simultaneously. A specific lepair can be programmed to move a gap of fixed width acrthe field. A fixed gap moving at a uniform rate should prduce a uniform fluence and hence a uniform density acrofilm. ~The fluence and density will also depend on the shof the extended source if a very narrow gap is used.! Bycombining several leaf motion patterns on a single film,stability of the leaves moving at different rates can be tes

The ion chamber and film measurements can be combinto an efficient QA test. The central leaves can scan aacross the ion chamber for a fixed number of MU, produca constancy check. Simultaneously, a film placed upstream

r

wnd

axac

oratr

gaithr-egeas

rn,ceinethmine

haeaif

icafoonceireele

eobaoarreina

li-n.ctoth

tortrol,fana-di-

est.ifi-

d is

-, re-ic-oftheorthermo-is

ter-

-

dicu-li-an-

iseosscen-ce

re-it

ntod

ar-rg

otorfieldctedhe

s of

ode


the chamber can image that gap as well as others off-that are moving at different rates. The density strips, normized to that of the central point, provide additional constaninformation.

As mentioned before, during commissioning the perfmance needs to be checked at different gantry and collimangles. Routine QA will employ a subset of those measuments done during commissioning. The accuracy of theis the critical parameter for accuracy of dose delivery wDMLC. This is impacted by the long-term gradual perfomance degradation of individual drive motors for the leavand stability of the MLC leaf carriage with gravity causinsag and backlash in the MLC carriage and support assblies. A comprehensive QA program for DMLC delivery hbeen recently described.16

2. Other dynamic MLC issues

Most of the considerations listed in Secs. II A 2–II A 4 fosegmental IMRT also apply to dynamic IMRT. In additiothe DMLC control system may have a minimum distanbetween opposing leaves to prevent collisions durmotion.20,21 This minimum gap affects the minimum dosthat can be delivered during a treatment and limitsamount of tissue sparing that can be achieved with dynaIMRT. The physicist should check what that gap is andcorporate a test of its stability into the routine QA of thmachine, especially if the IMRT planning system uses tinformation. For example, a test field could incorporate lpairs with that minimum gap moving across the field at dferent speeds.

C. IMRT with physical attenuators

A number of workers have described the use of physattenuators to accomplish the modulation requiredIMRT.22–25 In these systems, an attenuator must be cstructed for each gantry position employed and then plain the beam for each treatment. The problems of commsioning and maintaining a MLC are replaced by issueslated to material choice, machining accuracy, and placemaccuracy. IMRT delivery with physical attenuators is a viabalternative to IMRT delivery with MLC. In some ways, thphysical attenuators are much simpler and devoid of prlems such as leaf positioning accuracy, interleaf leakageintraleaf transmission, rounded leaf, and tongue-and-groeffect that are intrinsic to MLC systems. However, thereother issues associated with calculating the dose in the pence of a complex metallic filter, such as beam hardenand scatter from the filter, that need to be addressedequately.

D. IMRT with rotating fan beams „tomotherapy …

The first IMRT system to achieve wide commercial appcation was the Peacock developed by Nomos Corporatioslit collimator ~MIMiC ®! is added to a conventional linaand defines a fan beam approximately 20 cm wide and 1cm long. The fan beam irradiates a narrow axial slice of


isl-y

-ore-p

s

m-

g

eic-

tf

-

lr-d

s--nt

-ndvees-gd-

A

4e

patient as the gantry rotates. During the rotation, collimaleaves move in and out of the beam under computer conmodulating the fraction of time that each segment of theis open or blocked. The temporal modulation of the collimtor is indexed to the gantry angle. Several slices are irraated sequentially in order to treat the entire area of interAccurate motion of the couch is necessary to prevent signcant dosimetric errors at the junction between slices anaccomplished using a couch-indexing device~Crane®! fromthe manufacturer.

As an add-on device, the MIMiC® requires special considerations. One is the weight added to the gantry headquiring preliminary testing of gantry balance and isocentrity. Second, the stability of radiation output with rotationthe accelerator should be tested. The alignment ofMIMiC ® collimator with the rotational axis of the acceleratshould also be checked every time that it is attached tomachine. The MIMiC® is not interfaced to the acceleratoand assumes a constant MU delivered per degree of arction. The intensity modulated radiation delivery from thdevice is also not integrated with record-and-verify~R/V!systems, so preparations for recovery from treatment inruptions are necessary.

1. Peacock positional accuracy

Several references26–29describe the key elements in commissioning and QA of the Peacock® system. One is thephysical alignment of MIMiC® collimator on the linac toensure that the device is accurately centered and perpenlar to the axis of gantry rotation. Commissioning the colmator alignment employs superimposed film images at gtry angles of 90° and 270°~Fig. II.427!.

The second element is the determination of the preccouch increment to achieve the best dose uniformity acrthe slice junctions. This latter point is especially crucial sinthe dose can change by 25% per millimeter of misaligment.30,31 The couch is moved between slices a distanequal to the MIMiC® radiation field width projected to theisocenter. The accurate measurement of this width is thesponsibility of the physicist, and the method for measuringis provided by the manufacturer. However, it is the patiethat must move this amount, not only the couch, so gopatient immobilization is required as well. If the couch beings are not operating properly~for example, due to rust ocontaminants!, the couch may bind imperceptibly, causinthe Crane® to twist slightly such that the couch does narrive in the proper location. Only a very small position erris required to cause a measurable dosimetric error in theabutments. A measurement of the abutment can be conduby placing a sheet of radiographic film at the plane of tisocenter and irradiating successive open MIMiC® fields.Uneven couch motion by the Crane® will appear as varyingover- and underlaps between the fields. Periodic checkthe couch motion are necessary. Lowet al.29 describe dailyand weekly QA tests on this delivery system. Also, a methof mitigating the problem by varying the location of thabutment regions has been reported.32,33

ntanu

cain

cade

sjuA

n

ithina

h-ing

-cedvees ofn.

re-that

s,ionsn-rny.ms

lsotor

the

thebyn-ise;

er-ing

afent

ote

themo-

aalan--nd

henC

nscanita-re-

-


2. Peacock dosimetric measurements

As with the MLC systems described earlier, key elemeare to measure the transmission through the collimatorthe penumbra of the leaves. The penumbra must be measwith high spatial resolution~0.2 mm or better!. In addition,the rate at which the binary collimators open and closeaffect the effective output and the physicist should determthis.34

3. Helical tomotherapy

A prototype device that delivers the treatment in a helifashion with simultaneous gantry and couch motion is undevelopment at the University of Wisconsin.35,36 The helicaldelivery has the potential to reduce the dosimetric conquence of errors in couch motion. Because that device isbecoming commercially available, commissioning and Qinformation is limited at this time.

E. IMRT with rotating cone beams „intensity-modulated arc therapy …

Intensity-modulated arc therapy~IMAT ! is a deliverytechnique developed originally at the William BeaumoHospital that may soon be available commercially.37,38 Thismethod combines dynamic motion of the collimator wgantry motion. The MLC shape and gantry position aredexed to the delivered MU. One arc is used to produce e

FIG. II.4. Checkerboard pattern design using MIMiC® leaves and the resulting exposed film from laterals. Reprinted with permission from Sawet al.~Ref. 27!.


sd

red

ne

lr

e-st

t

-ch

intensity level used in the modulation. This is a new tecnique for which there as yet are few references regardcommissioning and QA issues.39

F. Leaf sequencing for segmental and dynamic IMRTwith MLCs

For IMRT delivered with MLCs, leaf sequencing algorithms are needed to translate the intensity patterns produby the planning system into instructions about how to mothe leaves. In general, there are many possible sequencleaf motions that could produce a desired intensity patter40

The search for efficient sequences is an area of ongoingsearch. For example, algorithms have been devisedminimize the number of segments,41–43the number of MU,44

the leaf travel,45 or the delivery time.46–48Additional consid-erations include the smoothness of intensity distribution49

the increments of intensity levels, and the spatial resolutof the intensity map.50–53In general, the number of subfield~segments! calculated by the leaf sequencing algorithm icreases with complexity of intensity pattern, which in tustrongly influences the overall accuracy of IMRT deliverTherefore, it is important that the leaf sequencing algorithminimize the number of subfields~segments! without com-promising the dose conformity. Moreover, algorithms aneed to account for mechanical limitations of the collimaand the need to reduce dosimetric problems such astongue-and-groove effect and the leaf transmission.54–57

In practice, because the leaf sequencing is part ofplanning process, the algorithm employed is determinedthe planning system. For the clinical physicist, commissioing the leaf sequencing algorithm is not a separate exercit is part of commissioning the planning system~seeSec.III E !. Nevertheless, it is important for the physicist to undstand the concepts involved, in part to aid in comparIMRT approaches and choosing between them.

1. Sliding window algorithms

In the sliding window approach to leaf sequencing, a lepair moves from one side to the other across the treatmarea. A point in the patient ‘‘sees’’ the source if it is nblocked by either the leading or trailing leaf. Adjusting threlative motion of the leading and trailing leaves controlsfluence pattern. The basic concept applies whether thetion is continuous during irradiation~dynamic IMRT! or al-ternates with irradiation~segmental IMRT!. Unfortunately,the termsliding windowhas been used in two ways: assynonym for dynamic motion and to signify unidirectionleaf trajectories. We are using it here with the second meing. Figures II.5 and II.6 illustrate the idea of a slidingwindow leaf sequence and its realization in dynamic asegmental modes.~See alsoFigs. 5 and 6 in the CWG report1

and Fig. 2 in Chuiet al.58!Conceptually, each leaf pair is considered separately w

constructing the pattern of motions. However, practical MLlimitations require modifications to account for interactiobetween neighboring leaves. Sliding window approachesbe constructed to accommodate leaf extension, interdigtion, and tongue-and-groove constraints. Interdigitation

x

f

-


FIG. II.5. The leaf trajectory of oppos-ing leaves as a function of dose indefor dynamic MLC delivery ~DMLC-IMRT!. A nonzero slope indicates leamotion during irradiation;~b! is the in-tensity map. Reprinted with permission from Xia and Verhey~Ref. 59!.

ase

nrinigth

nofsthxesehlid-

sele

veis

nroweg-

ingndents

esat-g-

fers to the end of a trailing leaf extending past the end ofadjacent leading leaf. Such a pattern is more likely to caucollision and is forbidden for some MLCs~seeFig. 2 of Xiaand Verhey59!. The tongue-and-groove effect refers to an uderdose that occurs in a junction region between neighboleaves if the tongue on one leaf extends beyond its nebor’s groove and later the situation is reversed withgroove extending beyond the tongue~seeFig. 1 of Xia andVerhey59!. This is attributed to the design of the MLC iwhich the sides of each leaf have steps or some kindtongue-and-groove arrangement to reduce the transmisbetween leaves. The width of the step is small, usually oforder of 1 mm, and as a result is ignored when planning fifields. However, this can cause a problem when MLC is ufor IMRT or to provide internal blocking. Incorporating succonstraints complicates the motion; however, in general, sing window algorithms effectively minimize the total number of MU required for treatment at the cost of an increanumber of segments~i.e., subfields for SMLC or contropoints for DMLC!.42 In practice, these algorithms may b


na

-g

h-e

aionedd

-

d

more efficient for delivery systems that can quickly mofrom segment to segment and in which treatment timelimited by physical leaf motion.

2. Areal or reducing algorithms

Areal and reducing algorithms allow bi-directional motioand consider the entire intensity pattern instead of eachindependently. These algorithms reduce the number of sments required at the cost of increased total MU. Addinterleaf motion constraints to deal with interdigitation atongue-and-groove effects increases the number of segmby about 20% to 35%.41,45 In practice, these leaf sequencmay be more efficient for delivery systems in which trement time is limited by the overhead in moving from sement to segment.

-t

-

FIG. II.6. The leaf trajectory as a function of dose index for step-and-shooMLC delivery ~SMLC-IMRT!; ~b! isthe intensity map. Reprinted with permission from Xia and Verhey~Ref.59!.

eninio

gh-

rs

RT

ts

lle-llnsc

neeu

iothnin-iran

fo%ogetheino

orinc

eah

ers,n

end

s.beer

rsely.

nsi-e.nsaf-in-

foroseal-nal

d onny

butan-

thera-

hassed

g-veent

he

o-n,finaler-a-

ionthe

fteraylan.imi-ccu-

forns.

ntvi-


III. TREATMENT PLANNING FOR IMRT

In this section we provide guidance related to treatmplanning issues to clinical physicists who anticipate settup an IMRT program. The specific purposes of this sectare the following:

~a! to describe the IMRT treatment planning process, hilighting areas that differ from ‘‘conventional’’ treatment planning~Secs. III A–III C!;

~b! to describe a process for learning how to apply inveplanning to particular clinical cases~Sec. III D!;

~c! to describe an approach to commissioning an IMplanning system for dosimetric accuracy~Sec. III E!;and

~d! to describe approaches to QA for individual patientreatment plans and treatment delivery~Sec. III F!.

A. Differences between IMRT and conventionaltreatment planning: dose calculations andbeam modeling

1. Modeling head scatter, penumbra, andtransmission

IMRT doses are calculated by dividing beams into smasections, calledbeamlets, that have varying intensities. Because the dimensions of the beamlets may be too smaestablish electronic equilibrium within them, calculatiobased on corrections to broad-beam data will not suffiSome method of integrating pencil beams or dose kermust be used,60–64 or Monte Carlo techniques must bapplied.65–68The small collimator openings also make accrate head-scatter modeling important.69,70

For conventional fields, issues such as transmissthrough collimators and penumbra affect the results atedges of and outside beams and so have reduced cliimportance. For IMRT delivered with MLCs, beamlet intesities are varied by moving the MLC leaves through theradiated field; therefore, accurately modeling penumbratransmission for the MLC leaves is critical.71–73 For ex-ample, a typical five-field prostate treatment plannedIMRT blocks a point within the prostate for more than 60of the MU, and leaf transmission typically contributes 4%the total dose. Since IMRT fields have multiple beam edthroughout the target volume, the dosimetric accuracy ofplan is dependent on the fidelity of the penumbra represtation. Special care must be taken during commissionwhen measuring these characteristics. Experience has shthat the penumbra should be measured with film, diode,very small chamber. A beam model based on scans obtawith a chamber having an inner diameter larger than 0.3may not produce accurate IMRT calculations. For this rson, special care must be taken during commissioning wmeasuring these parameters.

2. Leaf sequencing and deliverability

Inverse planning systems must determine a patternbeamlet intensities for each field and translate it to delivinstructions for the system being used. For MLC systemleaf-sequencing algorithm determines the MLC moveme


tgn

-

e

’

r

to

e.ls

-

ne

cal

-d

r

fsen-gwnaedm-

en

ofyats

to best replicate the desired patterns~seeSec. II F!. Param-eters such as collimator transmission, leaf shape at theand sides~rounded-end and tongue-and-groove effects!, andphysical limitations to motion all affect the delivered doseSome idealized intensity patterns may not, in fact,deliverable.74 For example, leaf transmission sets a lowbound on the minimum deliverable intensity.

Different systems handle the interplay between inveplanning, leaf sequencing, and dose calculation different

~a! Some systems first determine a set of beamlet inteties that, if delivered, would give the desired dosDose calculations during the inverse planning iteratioare for idealized beamlets. Subsequently, a lesequencing algorithm is used to create the deliverystructions. This algorithm incorporates correctionstransmission, penumbra, etc., so that the delivered dclosely resembles that which had been previously cculated, but no calculation is done based on the fidelivery sequence.

~b! Some systems append a final dose calculation basethe actual delivery sequence, in order to reduce adifference between what is planned and delivered,possibly obscuring the connection between the plning parameters and the final result.

~c! Some systems incorporate full dose calculations forproposed leaf sequences into all or some of the itetions of the inverse planner, thus ensuring that whatbeen planned can be delivered, at the cost of increacalculation time.

~d! Some systems permit weight optimization of the sements of actual delivery sequence to further improthe dose conformation and its adherence to treatmplanning objectives.

The manner in which this interplay is handled affects taccuracy of dose calculation and the speed of planning.

Note that some IMRT systems may use different algrithms during optimization than for a final dose calculatioin order to accelerate the process. The accuracy of thecalculation is most important, but the accuracy of the intmediate method may influence the quality of the optimiztion results. For example, if the optimization dose calculatover- or underestimates penumbra or scatter dose, thendose distribution returned by the optimizer may change athe final calculation, producing suboptimal results. It mnot be clear to the user what to change to improve the pThe physicist needs to know the approach used and its ltations. There is usually a tradeoff between speed and aracy, and the commissioning process~Sec. III E! should iden-tify any weaknesses.

3. Heterogeneity corrections

Heterogeneity corrections may be more importantIMRT than for conventional treatments, for several reaso

~a! IMRT treatments often incorporate more and differebeam directions than are used conventionally, so pre

nomsx

uanf

ltsr-cfo

Rd

io

froT

tie

ve

omrar-ngin’s

rte

Foiethano

hecgeb

os

llythnla

d

n-ed-to-

entsruc-ch

amof

etsryon.

res,in-eetsedra-

orhaveg

thee–sentseorn

cor-s,

sult.el-ndcialat

athatoseti-rchimaler-aan

t to. Aec.

s tolyely

c-nis-


ous clinical experience with uncorrected doses maytranslate well. Heterogeneities that affect some bealets more than others may give rise to localized dodifferences that are different from those previously eperienced.

~b! IMRT is used to escalate doses to targets and/or reddoses to critical organs. DVHs are used to evaluate~frequently! prescribe treatments. The reliability oclinical experience with DVH prescriptions and resuwill be significantly compromised if heterogeneity corections are not used, in particular, for body sites suas lung in which the corrections are clearly neededaccurate results.

To summarize, practitioners need to understand that IMoften uses higher prescribed doses, larger fraction sizes,ferent beam arrangements, and/or different dose distributthan conventional treatments. Clinical experience withheterogeneous/homogeneous conversion factor derivedconventional treatment planning may be irrelevant to IMRespecially in the lung.

Facilities that presently do not correct for heterogeneiwill face certain new tasks.

~a! Determine the conversion from CT number to relatielectron density for the imagers used.

~b! Check the planning system results using heterogenephantoms. Simple slab geometry using solid phantowith air cavities or cork inclusions has been used tditionally to check low-density effects. Anthropomophic phantoms are another possibility, typically usiTLD for point dose measurement. Some simple testby each clinic is needed to validate the institutionimplementation of the heterogeneity correction.

~c! Plan how to handle contrast agents or streaking afacts that may assign undesired CT numbers to voxand inappropriately influence the dose calculations.example, many planning systems allow bulk densitto be assigned to specified regions, replacingtroublesome areas. Also, plans should be run withwithout the corrections to determine the magnitudeany effects.

~d! Decide which types of plans need corrections. TCWG report recommends that heterogeneity corrtions be used; however, it may well be that heteroneity corrections are necessary for lung treatmentsare less necessary for prostate treatments and evendesirable if contrast material or rectal gas causes dmetric artifacts.

B. Differences between IMRT and conventionaltreatment planning: Planning algorithms

Simple IMRT planning can be accomplished by manuaadding subfields with various weights and evaluatingdose distribution. In each iteration of the process, the plandecides what changes to make to revise the design. The pning process is not automated and is sometimes calledfor-ward planning. This method typically produces a limite


t-

e-

ced

hr

Tif-nsam,

s

uss-

g

i-lsr

sedf

e--utun-i-

eern-

number of subfields and is a natural evolution of 3-D coformal planning. A number of publications have describsuccessful methods.75–79 The method lends itself to ‘‘stepand-shoot’’ delivery techniques. This approach can be aumated to various degrees, both for designing the segmbased on beam’s-eye-view projections of targets and sttures and determining the relative weight to give easegment.80,81

Another approach to IMRT planning breaks each beinto many small beamlets and determines the intensityeach.82–86 Having a large number of segments or beamlmakes the problem of determining individual intensities vecomplex and requires computerized methods for solutiThis process has come to be calledinverse planning. Theplanner specifies beam directions~or arc angles!, target dosegoals, and dose constraints or goals for sensitive structuand then an automated optimization algorithm calculatestensity patterns that create a dose distribution that best mthe prescription.~In the general literature of optimization, thterm constraintsrefers to limits that cannot be violated, anthe term goals denotes desired objectives. In these pagraphs, we use the termobjectivesto indicate both goals andconstraints.! If the planner wishes to change the result, heshe alters the objectives and reoptimizes. Some systemslimited ability to modify the intensity patterns by deletinsegments.

In inverse planning, the user specifies objectives fordose distribution using single dose value, a few dosvolume points, or fully flexible DVHs. Importance factormay be used to change the relative weight given to differobjectives.87 Internally, the planning system represents theobjectives in a cost function, which must be maximizedminimized by an optimization algorithm. The cost functionumerically attempts to represent the tradeoffs that are inporated into clinical judgment. By changing the objectivethe user alters the cost function and so influences the re

Many investigators have worked on the problem of devoping clinically successful inverse planning algorithms, athe literature is rapidly expanding, as are the commerimplementations. It is important to realize, however, th‘‘inverse planning’’ and ‘‘optimization’’ do not guaranteegood solution. The planner may set up dose objectivesare impossible to achieve or, conversely, that are so lothat the optimizer is not guided in a useful direction. Opmization algorithms are mostly heuristic local minima seaschemes that do not always guarantee that a globally optsolution to the problem as stated will be identified, and ctainly cannot guarantee clinical optimality. In general,treatment planner often needs several trials before findingacceptable solution, and it may not be easy to know whachange in order to push the solution in a desired directionprocess for developing that knowledge is suggested in SIII D. The success of an inverse planning system dependa large extent on offering a cost function that effectiverepresents clinical concerns and that a user can intuitivregulate.

Optimization algorithms used to minimize the cost funtions can be classified into two broad categories: determi

/on

ndi-c

oin

sblhotaedod

thonithis

cathey

i

tioa.nthaystiain

ari

eaisimeliplenoeiefew

nspese

the

ticalti-e

ateseningheifor-

t are

anbenaltion

phanossning,lds,lin-so

tanti-

x-in

ec-ced

tis-are

lds. Theec-ts ofnt.ituentre-alusererse

on-liedTon-sarys.ues


tic and stochastic.1 Deterministic methodsmove from oneproposed solution to the next using computed first andsecond derivatives of the cost function. The direction asize of each step~i.e., which beamlet intensities change aby how much! depend on the computed gradients. Minimzation can be relatively fast but cannot escape from a lominimum.

Stochastic methodsmove from one proposed solution tthe next by randomly changing beamlet intensities accordto some scheme. Because disadvantageous changesometimes allowed, escape from local minima is possiSuch methods are slower than the gradient descent metmentioned previously because the optimizer spends a lotime evaluating and rejecting random moves. Simulatednealing is one stochastic technique that has been adaptIMRT. In practice, stochastic and gradient descent methcan be combined.

The possible existence of local minima depends onform of the cost function and objectives. If the cost functidepends only on simple linear or quadratic functions wone goal dose per structure, then local minima do not exDose–volume objectives can cause local minima,88 and localminima can exist if the cost function depends on biologimodels in which different dose distributions can result insame complication or control possibilities. Similarly, thcan exist if the number and orientation of treatment fieldsa parameter to be optimized.

Since most inverse planning systems permit~or require!dose–volume objectives, then it appears that the soluspace for many clinical problems will have local minimThere is no difficulty if they are clinically equivalent, but, igeneral, it is not at all clear how a planner might know thagiven solution is the best solution. Planners have the clenge of discovering ways to force the inverse planning stems into different parts of solution space by changing iniconditions, such as by rearranging beam order, changingtial beam weights, or changing initial fluence patterns.

C. Differences between IMRT and conventionaltreatment planning: Specific planning issues

1. Dose uniformity vs dose shaping

Target dose inhomogeneity has been claimed to beunavoidable consequence of IMRT. This is not necessatrue and is a consequence of the characteristics of someIMRT planning systems and their applications. If IMRTdirected to produce a uniform dose to the target as its prgoal, then it should be able to accomplish that, effectivreplacing wedges and tissue compensators. In princIMRT should never do worse than conventional treatmtechniques, for the former has more flexibility or degreesfreedom.89 On the other hand, if IMRT is used to producdose distributions with concave shapes and/or steep gradnear critical organs, then target dose uniformity may sufTo create a complex dose distribution, IMRT casts shadowith some beamlets and balances them with higher inteties from other beamlets. Because the balancing is notfect, localized dose variations within the target and el


rd

al

gare

e.ds

ofn-

tos

e

t.

le

s

n

al--li-

nlyrly

eye,tf

ntsr.si-r--

where can be expected. In general, one should expectdose inhomogeneity in the target to increase as~a! the re-quired dose difference between target and adjacent cristructure increases;~b! the distance between target and crical structure decreases;~c! the concavity of the required dosdistribution increases, and~d! the number of available beamdirections decreases.

As part of the commissioning process, a user can evaluthe performance of the optimization with respect to theexpectations. As noted above, a successful inverse planalgorithm should allow a user intuitive means to control tbalance between the competing goals of target dose unmity and low dose outside the target.

2. Target and structure delineation

There are issues in target and structure delineation thaspecific to inverse planning for IMRT.

Inverse planning puts more responsibility on the clinicito carefully delineate what is to be treated and what is toavoided. For example, in conventional radiotherapy, regiotreatments can be designed by drawing ports on simulafilms that encompass the gross target and the draining lymnodes. To treat the same region with IMRT, the clinicimust contour the nodal regions explicitly as well as the grdisease and assign the desired doses. With inverse planthe physician designates targets instead of designing fieso careful and accurate contouring is essential. The deeated structures should be consistent from slice to slicethat structures are smooth in three dimensions. It is importo review the outlined structures prior to beginning optimzation.

In addition to designating targets, the clinician must eplicitly define all volumes that should be kept below certadoses. If an important structure is not identified and objtives set for it, then unacceptably high doses may be plathere by the inverse planning system.

Treating with novel beam arrangements may put newsues at risk. For example, many head and neck patientsconventionally treated with parallel-opposed lateral fiethat are reduced to deliver boost doses to gross diseasespinal cord is blocked after approximately 40 Gy, and eltron fields then boost the posterior neck nodes. Some parthe oral cavity may be blocked throughout the treatmeApplying IMRT with five to nine axial beams may makepossible to spare much of the parotids and reduce subseqxerostomia. However, parts of the anterior mucosa that pviously were totally spared would now be within severfields, as would tissue posterior to the spine. Unless theestablishes dose objectives for these regions, the invplanner may give undesired dose there~Fig. III.1!.

In general, all areas of potential interest should be ctoured so that DVHs can be evaluated and objectives appif needed. It is very important to recognize that in an IMRplan extremely, high-dose areas can show up in uncstrained normal tissue. Therefore, it may become necesto define ‘‘normal tissue’’ objectives to avoid such problemWhen assigning dose–volume objectives to normal tiss

-.-es

-


FIG. III.1. The use of avoidance structures to limit doses in inverse plansThe left panel shows the beam directions used, and the right shows thstructures, including avoidance areathat would not be contoured in conventional planning.

alcbees

itip

enreuathd

thc

erupteitiothplngtmla

e.n

enb

lcuonan

em-

al-cialnd

sbesedlan-ner

for

pert.’’

nifi-m-ibu-hethenrly,t ishpro-oflly

for

esng

getal-

redre-

and when evaluating DVHs for the plan, it is important ththe entire organ is contoured and included in the dose calation volume. If not, the planner and physician mustaware of the fact and adjust their plan evaluation procaccordingly.

3. Dose grid

As is generally the case with 3DCRT, the size of the crcal organs and the expected dose gradients near them imthe choice of the resolution of the dose grid used. OftIMRT is used in situations in which high gradients aneeded, and the dose grid may have to be finer than us

The dose grid also should be finer than the size ofbeamlets or incident fluence map so that the effects of molation are adequately sampled.

4. Buildup region

Care must be taken when target volumes are drawn withe buildup region. First, calculated doses are often inacrate and lower than delivered doses. Second, the invplanning algorithm will see the low doses in the buildregion as underdosing the target and will increase the insities of the corresponding beamlets. Those high intensmay well degrade the overall plan quality, likely causing hspots in the target or elsewhere. It may not be obvious touser that the hot spots are a consequence of the inversening engine fighting with the buildup effect instead of bei‘‘unavoidable with IMRT.’’ This issue is especially importanfor planning systems that expand the clinical target volu~CTV! by defined margins in three dimensions and then pto the expanded planning target volume~PTV!. Even if theCTV is well within the buildup region, the PTV may not bUnless the user inspects the PTV on each slice, this maybe detected.

Of course, if the target really is in the buildup region, ththe dosimetric problem is also real and is better solvedadding bolus than by relying on the accuracy of dose calations in the buildup region. It is better to put the bolusfor scanning so that it is accurately represented in the pl


tu-

s

-act,

l.eu-

inu-se

n-este

an-

en

ot

y-

.

5. Flash and mobile targets

Inverse planning for targets such as the breast is problatic. Conventional plans add beam margins in air~flash! toaccount for daily changes in shape, but inverse planninggorithms only treat defined targets. At present, commerplanning systems do not offer reliable heuristics to expathe beams to accommodate these needs.

For breast IMRT, both the flash and buildup problempresent significant difficulties; therefore, that site shouldconsidered with caution. Most published studies have umanually created segments or university-based inverse pning systems where additional control by the human planis possible.90–93

Respiratory motion can also cause more problemsIMRT treatments than for conventional treatments.94,95 Anyplan evaluation must consider how the plan shown on pafor a static image might be different in the living patienSome IMRT planning systems produce relatively ‘‘noisyintensity maps; that is, adjacent beamlets may have sigcantly different intensities. The summation of all these bealets on a static image may produce an acceptable distrtion. But if respiratory motion moves tissues during ttreatment over distances comparable to the beamlet size,deviations in delivered dose may be substantial. Similatomotherapy with slit collimators presumes that the patiena rigid body that can be indexed longitudinally with higaccuracy. Studies have shown that positioning errors canduce dose gradients of 25% for each millimetermisalignment.31 Physicians and physicists must realisticaassess these potential errors when selecting patientsIMRT, especially for sites in the abdomen and thorax.

6. Margins

Deciding what margins to apply is a question for all typof conformal radiotherapy, but IMRT and inverse plannicreate additional issues.

Planning systems often offer means for expanding tarcontours in three dimensions, often with six independent vues ~anterior, posterior, medial, lateral, superior, inferio!.However, it may be difficult to encode more complicatinstructions, such as avoiding intersections with other

thH

seoteT

an

eeveof

-

natth

fn

di

otsio

etennthTVettna

uilia

hetod

ithmopTaese

nd

uf-yFor

arewseh-of

theage

gHin-

t of

thesat-

at-

land?stict-lt?

e

behas

ro-

er-fer-

an


gions or boundaries. An experienced planner can handledeficiency by designing the beams appropriately. If the DVfor a brain tumor PTV shows low doses, and those low doare seen to be outside the skull, the planner can decide nworry about them. An inverse planning algorithm cannot dcide to ignore certain parts of a PTV. In such cases, the Pmust be explicitly drawn instead of produced by the expsion tools.

More generally, the ability of IMRT to produce rapid dosfalloff outside a target makes the assessment of marginsmore important.96–99 Where gradients are high, the consquence of localization errors is large, as for retreatmentparaspinal tumor. Hence the need to combine the abilityperform IMRT with excellent localization tools if high precision radiotherapy is the goal.

Planning systems differ in how they expand targets anormal structures and how the expansion regions are trein the inverse planning. Users need to understand whetargets can expand into structures~and vice versa!, whetherregions can overlap, whether priorities can be assignedoptimization, how doses are reported in expansion regioetc.

7. Radiobiologic issues

IMRT plans can have radiobiologic consequences thatfer from conventional plans.100–102 Conventionally, patientsare treated with a consistent dose per fraction. To give mdose to gross disease, field sizes are reduced and boosgiven at the same dose per fraction. Clinical experience wthis system has established the prescription doses. WhenIMRT plan is used from the beginning of treatment, targthat are to get different total doses also receive differdoses per fraction.77,103For example, a head and neck patieto receive 66 Gy to the base of tongue and 50 Gy toposterior neck nodes would receive 2 Gy/fraction to the Gand 1.5 Gy/fraction to the nodes. The 50 Gy would be givin 33 fractions instead of the typical 25. The target dosethe nodes might have to be increased in order to havesame radiobiologic effect as 50 Gy in 25 fractions. Coversely, the lower doses per fraction may improve the sping of normal tissues.104 One could also use multiple IMRTplans in a regional-treatment-plus-boost fashion, therebying a consistent dose per fraction, but this requires the abto sum distributions and the user to apportion dose gobetween plans.

These effects are reduced if IMRT is only used for tboost portion of the treatment, but the ability of IMRTproduce unconventional dose distributions is compromiseonly used for a part of the treatment.

Target doses are often less uniform with IMRT than wthe conventional treatment. The clinical consequencesdepend on whether the target is bulky disease or microscinclusions in normal mucosa. Initial reports comparing IMRto conventional treatments indicate that acute reactionsless for prostate treatments but more for head and ntreatments.103,105,106Physician training needs to include theanticipated changes from conventional practice.


is

sto-V-

en-a

to

deder

ors,

f-

reare

thne

st

te

nohe-r-

s-tyls

if

ayic

reck

8. Plan evaluation

IMRT treatment plans need to be evaluated carefully asomewhat differently than other plans.

Inspecting and comparing DVHs are useful, but not sficient, since DVHs have no spatial information. IMRT macreate hot spots or cold spots in unexpected locations.example, in 3-D conformal treatments in which beamsdefined using beam’s eye views, the user typically knothat the CTV is well within every field, and so a low-dostail on a DVH for the PTV reflects penumbra at the peripery. With IMRT, those low doses may occur in the centerthe CTV, with a different effect on tumor control~Fig. III.2!.Conversely, localized high doses may occur well outsidetarget. Planners need to inspect the isodoses on each imslice. At a minimum, it is very important that the planninsystem reports the global hot spot, and it is better if the DVfor all nontarget or nonsegmented tissue is available forspection.

Plan evaluation for IMRT should include an assessmenthe potential problems and pitfalls outlined below.

~a! Is the dose uniformity in the target acceptable? Arestated plan goals for hot spots and target coverageisfied?

~b! Are the stated plan goals for normal-tissue sparing sisfied?

~c! Were organs contoured in their entirety? Are the pgoals appropriate for the fraction of organ contoure

~d! Are the margins and dose gradients safe given realiexpectations for setup reproducibility? Might geomeric miss of the target or overdose to a structure resu

~e! Will patient or organ intrafraction motion during thtreatment compromise the accuracy?

~f! Are there high doses in the buildup region that mayinaccurate or an indication that the inverse plannerstruggled to ‘‘fix’’ low doses there?

~g! Have inhomogeneity corrections been applied apppriately?

~h! How does this plan compare with a conventional altnative? What regions are being treated or spared difently compared with traditional methods?

FIG. III.2. The unwanted appearance of a localized cold spot frominverse-planning system.

a

lv

re

io

aninthinnmm

ethonstA

onlaeoos

iminne-

dan-

oal

ra

m

a

lyhanonng

hao

th

esto

in-y.

ant

be--

Be-mem-gebe°lyc-

n-arengpar-ed-

r-mar-

lish

aven-

m

de-theasicFor

nten-Af-

in

or-ameasyy to

for-owed


~i! Is the increased whole body dose with IMRTconcern?107

~j! Are there unusual beam orientations that might invocollision with or shadowing by the treatment table?

~k! Are there low intensity segments that could bemoved without compromising plan quality?

This list is not exhaustive but serves to illustrate the cautand skepticism that should be brought to bear.

D. Learning how to use the inverse planning system

Learning how to use a particular system’s inverse plning tools to best advantage can be a significant undertakThe previous sections have outlined some of the issuesmay be challenging for a new user. More fundamentally,verse planning requires learning a new set of skills. Ochallenge is getting a feel for how to adjust the plan paraeters~prescription, goals, constraints, priorities, beam geoetry, and so forth! in order to shift a dose distribution in thdesired direction. The user needs to learn how muchresults of optimization change with changes in available ctrol parameters. A second challenge is developing realiexpectations for what can be accomplished with IMRT.common problem is asking for an impossible distributiand therefore getting poor results. In such a situation, reing the objectives may produce a better plan. The user neto learn how to express the clinical objectives using the toavailable in the planning system and then to adjust thparameters to steer the plan.

New users should expect to spend considerable tlearning how to apply IMRT to the body sites of interesttheir institution. Each new site should be regarded as acommissioning effort, with implications for imaging, immobilization, setup verification, etc., as well as planning~seeSec. IV!. Setting aside overall clinical implementation anconcentrating on planning issues, developing an IMRT plning procedure for a clinical site~e.g., prostate or head-andneck with parotid sparing! consists of several steps.

~a! Determine conventions for contouring targets and nmal tissues. For example, will the rectum or rectal wbe contoured, and over what length?

~b! Decide what margins should apply and what dose gdients are appropriate.

~c! Decide what dose–volume limits define the minimucharacteristics of anacceptableplan, both for targetsand normal tissues. RTOG protocol H-0022 for orophryngeal cancer~http://www.rtog.org! provides a goodexample. This is a nontrivial exercise but absolutenecessary. Evaluating hot spots may be especially clenging since the DVHs of these plans often have lohigh-dose tails. Is the maximum reported dose a ccern, given that it may be a single voxel? Is reviewithe dose to a minimum volume, perhaps 1 cm3, morerealistic?

~d! Once the criteria for acceptability are set, decide waspects are to be optimized. For example, the gmight be to minimize the dose to the hottest 30% of


e

-

n

-g.at-e--

e-

ic

x-dslse

e

w

-

r-l

-

-

l-g-

tale

rectum while maintaining the prostate CTV doswithin certain ranges. Conversely, the goal might bemaximize the dose in the prostate CTV while maintaing the dose to the hottest 10% of the rectum to 75 GIt is useful to decide on one parameter to hold constfor all the subsequent comparisons.

~e! Having determined how to evaluate the plans, thengin to try different combinations of the planning parameters to find those that produce good results.cause the range of possibilities is huge, sosystematic approach is needed. One might fix the nuber and orientation of beams to some relatively larnumber so that the beam selection is not likely tolimiting plan quality ~e.g., nine coaxial beams at 40increments!. Then, for fixed target doses, gradualtighten the normal-tissue objectives. After the objetives are finalized, try different beam combinations.

~f! Compare the results with a manually planned, 3-D coformal alternative. Carefully assess what volumesbeing treated that were not before. What is beispared that was not before? Does improved tissue sing justify nonuniform target doses? Are the increascost and complexity justified by real dosimetric improvement? When comparing IMRT with 3-D confomal plans, it is crucial to make sure that the probledefinition is consistent, e.g., the same contours, mgins, and criteria for acceptability.

~g! Repeat the process for a number of patients to estaba robust methodology.

Some studies have reported specific protocols that hproved useful for particular body sites and particular planing systems.108,109

E. Commissioning an IMRT planning systemfor dosimetric accuracy

Dosimetric commissioning of an IMRT planning systeshould follow a systematic sequence.17,110,111Many of thesetests require that the system allow the user to specify asired intensity pattern and apply it to a phantom so thatresulting doses can be measured and confirmed. The bscheme is to advance from simple to more complex tests.example, start with single beams on a simple, flat~i.e., geo-metric! phantom with controlled intensity patterns. Whethose are validated, then progress to using controlled insity patterns for multiple beams on the simple phantom.ter that, apply multiple beams treating hypothetical targetsthe flat phantom. Finally~if possible! progress to testingmultiple beams treating hypothetical targets in anthropomphic phantoms. The goals are, first, to determine if the beparameters are accurate using simple situations that areto evaluate and, second, to determine the level of accuracexpect in clinical situations.

In this discussion we assume that the required input inmation has been given for beam modeling and focus on hto test the resulting calculations. In Sec. II A 4 we discuss


FIG. III.3. Examples of user-controlled intensity shapes used for commissioning tests.

T

he,

Ctivanth

neo

leth

setie

rmle

g

nspla

te

di-sen-ns.

a

try,iftsrly

ir-x-

nts,an-s.ing

flatth

et/

particular concerns in obtaining the initial data for IMRmodeling.

The primary dosimetry tools are water-equivalent or otplastic phantom~s!, ionization chamber, electrometer, filmand a film scanning system. Note that if the phantom isscanned with the ionization chamber in place, the sensivolume can be outlined as a region of interest in the plThe mean dose to this region as reported by the plan canbe directly compared with the measured dose.

Cylindrically symmetric chambers are preferable to plaparallel chambers for multiple beam irradiation becausetheir axial symmetry. Small-volume chambers are best unthe dose gradients can be kept low over the size ofchamber.112 Film that can be irradiated to a typical daily dois preferred over faster films in order to remove uncertaincaused by MU scaling.

~a! For a series of open fields on the flat phantom, confithat the central axis depth dose and off-axis profimatch expected values.

~b! For a series of simple intensity patterns, e.g., wedpyramid, or well@Figs. III.3~a!–III.3~c!#, measure thedose per MU at multiple points in low gradient regiowith an ion chamber. Measure dose profiles at multilocations and directions with film. Create patterns thhave systematic changes in intensity levels. As no


r

Te.en

-f

sse

s

s

e,

etd

above, careful attention to agreement along high graent edges at this point can uncover penumbra repretation problems that might cascade in full patient plaA random distribution@Figs. III.3~d! and III.4# helps todetermine the level of accuracy one might see inpatient treatment.

~c! Apply a simple modulated shape to plans using gancollimator, and couch angles and translational shand confirm that these geometric motions are propeimplemented and understood.

~d! Apply a simple intensity pattern to multiple beamsradiating the flat phantom at different angles. For eample, create a 10310 cm array of high intensitybeamlets with a central 535 cm section with reducedintensity @Fig. III.3~e!#. Irradiate the flat phantom withfive to seven axial beams at equal angular incremeeach having that intensity pattern. This tests the plning and delivery for a summation of simple fieldVary the central section intensities to test the plannand delivery over a range of conditions..

~e! Design a series of tests of idealized targets in thephantom to be treated with multiple fields. Start wisimple targets requiring little modulation~such as asphere! and progress to more complicated targcritical organ combinations that require more~such as a

s
FIG. III.4. The dose profile measured with film across one line of a random intensity pattern (plan5dotted, film5solid), showing some systematic differencein low intensity regions.

aly

n

e

pla

nsitynt

aln

blen

s

eeh

amam

oo

erts

n

tivicae

ant

alcuhar-a

qunod

e

rlyt ishree

e-

s ofch.n,eeenicistchndI A.

on.a-

ormas

d toand

so-o-ndan-redte-

re-dent

eirt

n

theonede-tion

iontchns

ousosee-ceds

af


C shape surrounding a critical organ or a cylindricshell surrounding a critical organ, with progressivetighter objectives for the organ!. As before, measurethe dose in a low-gradient region with the chamber athe dose distribution in multiple planes with film.

~f! Evaluate dose calculation accuracy in the presencheterogeneities using a simple geometry.

~g! As need and resources permit, test simple and comtargets in heterogeneous and anthropomorphic phtoms.

It is difficult at this time to give specific recommendatiofor dosimetric accuracy of IMRT plans given the complexof the plans and the measurement problems. A statemethe TG-53 report113 deserves repeating:

Here, we will not provide a table of recommended vues, since it is clear that what is achievable with okind of planning system may be quite unachievawith another. It is the responsibility of the radiatiooncology physicist to determine

1. the accuracy of the institution’s particular RTP sytem for a range of clinical situations; and

2. how that expectation of accuracy must be modifito account for any particular clinical situation, thkinds of treatment plans that are created, and otaspects of the local situation.

There is a developing consensus, however, that ion chber measurements in low gradient areas of single [email protected]., seeFigs. III.3~a!–III.3~c! and III. 3~e!# should agreewith the plans to the same accuracy as is achieved with cventional treatments, i.e., on the order of 2% to 3%. Fmore complex irradiations typical of patient treatments, this a developing consensus that ion chamber measuremenhigh dose, low gradient regions should agree with the plawithin 3% to 4%.

Attention must be paid to high dose regions representaof targets and low dose regions representative of critstructures. A goal of commissioning is to develop an undstanding of the dosimetric uncertainties so that clinical plcan be meaningfully evaluated, especially with respectcritical structures. It is true that IMRT plans may have locized dose gradients that make measurement more diffibut these may be more problematic for individual beams tfor the combination of all. It may also be difficult to detemine if differences between measurement and calculationcaused by a planning, delivery, or measurement techniFor this reason, the delivery system should be commissioseparately from and before the planning system. The cstruction of good commissioning tests is a challenge ansubject for ongoing research and development.

F. QA of individual treatment plans

A primary difficulty with designing QA tests for indi-vidual patient plans is not knowing all the likely failur


l

d

of

exn-

in

-e

-

d

er

-s

n-rein

to

el

r-so-lt,n

ree.

edn-a

modes for this new modality. Concern and caution are cleaindicated. In developing a comprehensive QA system, iuseful to separate the complete treatment process into tsequential elements:~a! Dose and MU calculation;~b! infor-mation transfer from planning system to R/V system to dlivery system; and~c! dose delivery.

Each step in the process has its own potential sourceerror, and the physicist should develop checks for eaThese checks will involve some combination of inspectiocalculation, and experiment. In the following sections wdescribe some of the possible techniques that have bused; each has its strengths and limitations. Each physand facility will need to balance patient-specific tests, suas described in this section, with standardized MLC alinac performance tests, such as described in Sec. IPatient-specific verification measurements test many~but notall! aspects of planning and delivery in a combined fashiPatient-specific calculations combined with frequent mchine QA represent another approach. The latter is the nfor conventional treatments and may become so for IMRTthe field evolves. In these early stages, physicists neecarefully assess the overall structure of their QA testsfrequencies.

1. Independent calculation methods

Independent calculation methods to verify MU and ablute doses are becoming available for IMRT plans. Algrithms have been reported that take MLC delivery files acalculate doses that can be compared with the IMRT plning system’s prediction. Some methods calculate deliveintensities from the delivery files and then apply sector ingration or other techniques to approximate the dose.114–119

Some facilities have eliminated most point-dose measuments after developing and commissioning such indepensystems, but that commissioning task is a large one.

‘‘Independent’’ dose calculation methods that derive thinput information from the planning system files will nocatch errors in that input information~such as a plan done othe wrong patient or the with the wrong treatment unit! orerrors in transferring data from the planning system toR/V and treatment systems. To give the most confidence,should use output from the R/V system as input to the inpendent calculator, along with necessary patient informasuch as source-to-skin distances.

As mentioned in Sec. II A 1, independent dose calculatmethods based on pre-treatment information will not caerrors in the treatment delivery. Patient-specific calculationeed to be part of a larger QA process that includes rigortesting of the delivery system. In principle, independent dcalculations could include information derived from the dlivery itself, such as from electronic portal imaging devi~EPID! measurements or MLC log files, but such methoare still under development.

2. Verification measurements

Verification measurements are commonly made of‘‘phantom plan’’ or ‘‘hybrid plan.’’ This technique consists o

foCTfin

se

ecctceealo

hauria

ans,lmthtere

oncthtengrioin

hoapthno

cme

reagtte

a

atin

a.

rT

te

ispots

ionreisto

geslan

theheh adis-lan

iser-

l inthen-

isweid-

.onri-RT

laceuttors,sid-top-RTnd

ysi-tionurening.-

re-

ts


applying the MLC segments, leaf trajectories and MUeach field, derived from the final patient calculation, to astudy of a standard phantom and then recalculating thedeliverable dose distribution in the phantom.120,121The phan-tom is then irradiated according to this plan and the domeasured using ion chambers, film,122–124 or otherdetectors.125–127 The results of the measurement are thcompared to the predicted dose to the phantom. The logi‘‘phantom plan’’ methodology is that it verifies the corretranscription of IMRT delivery parameters, leaf sequenand MU calculation. As mentioned above, there is a devoping consensus that a reasonable action level for ion chber measurements of such phantom plans in high dose,gradient regions is 3% to 4%, with the understanding tsmall fields and localized gradients may cause additionalcertainties in some cases. Film is generally used to vevisually the dose distribution on at least one plane, at lequalitatively. Note that measurements on a single axial pllikely will be sensitive to the motions of a few leaf pairperhaps only one. It should also be noted that not all fiscanning systems can track optical density accurately inpresence of high gradients. Therefore, the scanning sysmust be validated for accuracy for quantitative measuments.

It is important to realize that some errors in input datacalculations will not be caught by using phantom plans, sithe dose distribution in the phantom is not expected to besame as in the patient. For example, the planning sysmight ‘‘see’’ the CT couch as part of the patient, addiseveral centimeters of radiological depth to the postefields and inappropriately increasing those intensities. Andependent calculation using the correct depths would sthe error. However, a phantom plan would apply these inpropriate intensities to the phantom. Measurements inphantom that confirmed this new dose calculation woulduncover the error. Phantom measurements test the doseculation and delivery mechanism, but do not check soassumptions used in the planning process. Measuremthat test the dose in the actual patient would be preferseveral groups are working on using electronic portal iming devices to perform QA measurements using transmidose through the patient.128–132

It is useful to have phantoms that reasonably approximthe body site in question. Examples could be a 30330315cm rectangular phantom for the trunk and 15315315 cmrectangular phantom for the head. The routine use of a phtom that is not equivalent in size must be validated by tesat least once against a more appropriate phantom. Morethropomorphic phantoms are also commercially available

3. Other plan checks

TG-40133 and TG-53113 both have recommendations fochecks of individual plans that certainly apply to IMRplans, but again there are additional concerns. Becauseverse planning systems, not planners, design the beam in


r

al

s

nof

,l-m-wt

n-fyste

em-

reem

r-w-etal-entsd;-d

te

n-gn-

in-n-

sities, it is important to check that the target coverageadequate in three dimensions and that the hot and cold sare well understood.

Similarly, inverse-planning systems may have the optof shifting the isocenter from an original setup point befotreatment. Clearly, recognizing and verifying such a shiftcrucial. A helpful method to check for these situations iscompare digitally reconstructed radiographs~DRRs! fromthe plan, with target volumes superimposed, to portal imaof the treatment. Since the DRR is generated from the pdata, correspondence to the actual patient as seen onportal image confirms that the virtual model aligns with treal world. Clearly, high quality DRRs are needed for succomparison to be trustworthy. The plan evaluation issuescussed in Sec. III C 8 should also be considered during pchecks.

In summary, commissioning an IMRT planning systema challenging project that must be undertaken with an undstanding of the dosimetric and clinical concerns. Our goathis section has been to provide a framework on whichphysician and clinical physicist can build a plan for that udertaking.

IV. CLINICAL IMPLEMENTATION OF IMRT

A. Overview

Work needed to implement IMRT includes all thatneeded to implement 3DCRT and more. In this sectionwill concentrate on the additional aspects and provide guance related to issues of clinical implementation of IMRT

Each facility should designate an IMRT implementatiteam to think through the implications in advance and peodically update procedures as lessons are learned. For IMto truly produce a benefit, various resources must be in pand all persons involved in IMRT, not only physicists balso physicians, dosimetrists, therapists and administramust be properly trained before the actual treatment. Coneration should be given not only to bringing the modalitythe clinic, but also to keeping it running smoothly and keeing pace with upgrades and future enhancement in IMtechnology. Furthermore, IMRT is an integrated system, acareful thought should be given to every technical and phcal component and treatment step. The overall integrashould also consider human involvement in the procedand address the issues related to staff education and trai

The clinical implementation of IMRT includes the following aspects:

~a! Equipment and space requirements~Sec. IV B!;~b! time and personnel requirements including their

sponsibilities~Sec. IV C!;~c! changes in treatment planning practice~Secs. IV D 1–

IV D 5!;~d! changes in treatment delivery practice~Secs. IV D 6–

IV D 9!;~e! QA of equipment and individual patient treatmen

~Sec. IV E!;~f! staff training and patient education~Sec. IV F!;

ice

tstaerga

reuro

ed

oucsBnf

ra

iomob

sp-ss-Rd

eoi-m

Rupexngm

a-

toe for

affT

m-ean-ningandp-seteat-heandss

ofics

i-ese

wn

t-to

gtongng,eing,

tom-for

hised

er.

entleartab-nt


~g! changes in scheduling, billing, and charting pract~Sec. IV G!;

~h! overall integration~Sec. IV H!.

In the following we will offer guidance on these aspecof IMRT, suggesting questions that the clinical implemention team will need to ask and providing potential answwhere possible. The goal is to provide a framework to ornize the task of bringing IMRT into the clinic.

B. Equipment and space requirements

1. Shielding

IMRT treatments require about a factor of 2 to 10 moMU than conventional treatments, so room shielding shobe reevaluated.134,135The MU are about 2 to 4 times more fothe MLC-based IMRT treatments. For sequential tomtherapy delivery, up to 10-fold greater MU may be needdepending on number of rotations involved.134,135 Primarybarriers are not usually affected, although use factors shbe assessed because IMRT treatments typically use armany more gantry angles than conventional treatments.cause the enhanced workload affects the leakage compoof radiation reaching secondary barriers, shielding designthese barriers must be evaluated.

2. Space planning

Extra space may be needed for additional computer wostations, especially if IMRT planning is to be done ondedicated system. Space may also be needed for additequipment, such as add-on collimators, dosimetry phantofilm scanner, and instrumentation, as well as patient immbilization devices. Space for additional personnel mayrequired.

3. Equipment

It may be necessary to upgrade existing acceleratorprovide IMRT functionality, such as adding an MLC, ugrading an existing MLC to dynamic capability, or purchaing special add-on collimators. Similarly, existing R/V sytems may need to be upgraded to accommodate IMtreatments. Computer networks may need to be enlargeimproved to permit the needed file transfers.

Additional dosimetry equipment including small volumdetectors may be needed for the commissioning and ongQA of IMRT. It is important to have an efficient film scanning system to accomplish these tasks. Additional phantomay also be needed.

IMRT planning capabilities must be provided, either asstand-alone IMRT planning system or as an add-on IMmodule to a conventional planning system. Many issues mbe considered. For example, stand-alone systems mayvide more resources for computation time and/or morepertise with regard to IMRT planning. Conventional plannisystems may allow more easily the combination and/or coparison of non-IMRT and IMRT plans. However, if this c


-s-

ld

-,

ldor

e-entor

k-

nals,-e

to

-

Tor

ng

s

aTstro--

-

pability is not provided, it certainly is useful to be ablecontour on one system and have those contours availablboth IMRT and non-IMRT planning.

C. Time and personnel requirements

It is essential to anticipate the number of additional stthat will be needed to implement and maintain an IMRprogram.

Sufficient time and resources must be allocated to coplete all the tasks involved in clinical implementation. Thphysics staff will need to complete comprehensive and qutitative measurements to assure that the treatment planand treatment delivery systems are accurate. Physicianstreatment planners will need to learn a very different aproach to planning. The implementation team will need toup and test the processes used for individual patient trments. QA procedures will have to be modified. Many of tstaff—physicians, physicists, dosimetrists, therapists,engineers—will need special training. It is important to strethat these tasks will likely require an initial investmentseveral person–months of work on the part of the physstaff and other members of the implementation team.

After the initial implementation effort, the ongoing QAactivities will increase for both the IMRT systems and indvidual patient treatments. In other sections we describe thactivities in detail.

D. Changes in treatment planning and treatmentdelivery process

1. General considerations

The details of IMRT treatment will differ from institutionto institution, but the general IMRT treatment process shoin Fig. IV.1 will serve to frame the discussion.

2. Immobilization

Because of the highly conformal nature of IMRT treament, new immobilization techniques may be necessarysafely use the technology,96,136,137 such as supplementinthermoplastic masks with bite block fixation. Techniquesreduce or follow internal organ motion, such as by usiultrasound localization of the prostate or respiratory gatimay be desired.138,139All these new procedures will impostheir own burdens with respect to procedure design, trainand validation. If not already known, it may be necessarystudy the reproducibility that can be achieved with the imobilization system in order to establish realistic marginsplanning.140–142 Electronic Portal Imaging Devices~EPID!and implanted fiducial markers can provide a big help in tarea. Generally, the patients will be immobilized and markas close as possible to the anticipated treatment isocent

3. Image acquisition

At an early stage in the process, the goals of treatmshould be discussed carefully with the planner so that a cunderstanding of the imaging and planning needs is eslished. As for 3-D conformal treatments, a CT for treatme

d


FIG. IV.1. The overall process of IMRT planning andelivery.

othri

merenhiedha

fotoTh

n

euohea

anthgeDoe

etidthacbn

in

n-of

ethedesbelan-by

tobe

esa

.orsonb-w

epro-ch-l. Inatfi-

ndent

nceexlannualntlyoft-

at-to

toeensk,

planning will be performed with the patient in treatment psition with the immobilization device. Clinics may find thathey need to obtain more slices at a finer spacing thanbeen the norm previously. For inverse planning systems ding a 1 cm MLC, slice spacing of no more than 0.5 cshould be used, and finer spacing may be needed to genDRRs of sufficient quality. This is especially important whusing an inverse planning system that may call for a sfrom the original alignment point, and in any case one neto verify that the isocenter in the plan corresponds to tused for treatment.

The range of slice acquisition may also be expandedorder to permit the use of nonaxial beams. For example,isocenters above the base of the sphenoid sinus, the promay be to acquire slices through the top of the head.acquisition of enough CT slices~fine slice thickness! may benecessary to produce DRRs of high quality and defianatomy adequately.

Highly conformal treatments, especially when designwith inverse planning, require target and normal tissue strtures to be identified very accurately. Hence, the use of ctrast agents for the CT and registration of images from otmodalities, such as MRI or PET, are often needed and mrepresent a change in typical practice.

4. Structure segmentation

Structure segmentation is one of the most importantcrucial steps of the IMRT procedure. The success ofIMRT procedure is closely tied to the accuracy of the tarvolume and critical structure delineation. As with all 3-planning, contouring targets and normal structures is labintensive for physicians and planners. With IMRT more dmand is placed on the physicians to define structures in dand with rigor. For example, implementing a new parotsparing protocol for head and neck patients would requireparotids and at-risk nodal volumes to be defined on eaxial slice, with due consideration for margins. This canmore difficult than defining conventional lateral fields osimulator films to treat the nodal volumes, hence requirmore of the physician’s time.


-

adv-

ate

ftst

inr

cole

e

dc-n-ry

det

r--ail-eh

e

g

5. IMRT treatment planning

The differences between planning for IMRT and for coventional treatments are discussed in Sec. III. In termsclinical implementation, a key point is to allow time for thphysicians and planners to develop their skills in usingsystem. Inverse planning, in particular, requires new moof thinking: physicians need to quantitatively prescridose–volume limits that define an acceptable plan, and pners need to learn how to improve the dose distributionmodifying unfamiliar input parameters. Clinics will needdevelop tools to aid these tasks. Special forms shouldimplemented for recording the desired clinical objectiv@Sec. III D, ~b!–~d!#, the planning parameters entered, andcomparison of the plan results with the clinical objectives

Note that it is not certain that IMRT plans will be superito alternative 3DCRT plans. For a specific site, a compariof 3DCRT plans and IMRT plans may be obtained from pulished literature, showing the benefit of IMRT. Even so, neusers need to demonstrate to themselves that they can rduce the essential characteristics of IMRT treatment teniques that the published literature has shown beneficiaany event, practitioners should not utilize IMRT plans thare inferior to the treatments currently employed even ifnancially advantageous.

6. File transfer and management

When an IMRT plan has been satisfactorily computed aapproved by the physician, one can generate the treatmcontrol files. For MLC systems, these include leaf sequefiles for each gantry angle. Since IMRT involves complbeam shapes and control files, the digital capability for ptransfer is essential to avoid possible mistakes during matransfer. Depending on the individual clinic’s informatiosystem, the files can be transferred by floppy disk or directransferred to the R/V server through data exchange sware.

Since information transfer is a common source of trement error, the clinical implementation team will needanswer many important questions. The therapist will needbe able to verify every day that the appropriate file has bselected for each field or arc. If the files are on a floppy di


FIG. IV.2. An example of DRR andportal image used for the IMRT iso-center and field shape verification.

theethR

heTefo

rotury

ors

ethe

trreb

bu

lcun-itenithceve

nt-ld be

u-esrsen-the

forLCcor-

iblesedften

arrtalkly

nyac-of-

on,illndim-

anom-nd

how will the disks be stored and labeled so that choosingwrong one is unlikely? Will patient and field identifiers bdisplayed so that they can be checked? Will a double chof that selection be required? Will it be documented? Ifdepartment has an R/V system that fully supports the IMtreatments, then many of these problems are eliminated~andreplaced by the need to verify the initial programming of tR/V system!. If the R/V system does not fully support IMRtreatments, can it still verify some parameters, such asergy and MU? Does it have to be bypassed or turned offthe IMRT treatments? If so, how might that affect other pcesses, such as electronic record-keeping or charge cap

To expedite IMRT delivery, an autosequencing delivesystem is sometimes used. Such delivery systems~in differ-ent forms! are currently available from all major acceleratvendors. ‘‘Dry runs’’ to test for collisions or other problemshould be a part of routine plan validation.

7. Plan validation

The goal of IMRT plan validation is to verify that thcorrect dose and dose distribution will be delivered topatient. One needs to check that the plan has been propcomputed and that the leaf sequence files and treatmenrameters charted and/or stored in the R/V server are coand will be executable. Items that need to be validated,fore the first treatment, include MU~or absolute dose to apoint!, MLC leaf sequences or fluence maps, dose distrition, and collision avoidance.

Note that the details of what is to be measured or calated for dosimetric validation will be tailored to each cliic’s needs and may change with experience. However,important to emphasize that new users will need to spmuch more time validating IMRT plans than is common wconventional treatments. Direct measurements will be nesary until independent dose calculation methods are deoped and validated.


e

ckeT

n-r

-re?

erly

pa-ct

e-

-

-

isd

s-l-

8. Position verification

Clearly, position verification is an important part of plavalidation. The most critical point is to verify that the treament isocenter matches the planned isocenter. This shouaccomplished by comparing orthogonal films taken at simlation, DRRs from the planning system, and portal imagfrom the treatment unit. As mentioned above, an inveplanning system may call for a shift from the original aligment point, so it is crucial to compare the isocenter onDRRs with the setup films.

Wherever possible, portal images should be obtainedthe fields used for treatment, and it is useful to have the Mfield boundary as apertures for the ports and compare toresponding DRRs from the planning system~Fig. IV.2!. De-pending on the imaging system available, it may be possto obtain a portal image of the modulated field superimpoon the patient’s regional anatomy, but such images are ohard to interpret.

If IMRT is to be applied to highly precise treatments necritical structures, then the frequency of on-treatment poimaging may need to be evaluated. As a minimum, weeportal imaging is necessary.

In general, the implementation team must consider achanges in the portal imaging process, such as how toquire the bounding MLC shape, how to verify the positiona slit collimator, or how to operate an electronic portal imaging system in the presence of dynamic fields. In additiuse of daily target localization tools, such as ultrasound, wimpact the need for and interpretation of portal images amay add the need to acquire, review, and archive otherages.

9. IMRT treatment delivery

IMRT treatments often take more time to deliver ththeir conventional counterparts due to their increased cplexity. They require larger numbers of MU and fields a

seter-ngtr

redatth

ingr-b

sdndkngnilo-reis

so

n

snan.

cecity

on

Trsr-ilyacewhinth

eni

in

llye.

aryOnthemi-s inthehm.ts inive.-ro-ofif

.arealseg-er ises-inoodin-

ormi-ess

ra-e-lresnt

getldca-s.RT.

yon-ectalg ofer-do-si-ent

acta-

oset iseseic


are more likely to use oblique gantry angles than are uconventionally. Because maximum field sizes maybe limifor IMRT because of limitations on leaf and/or jaw ovetravel, IMRT treatments are more likely to require abuttitwo or more adjacent treatment fields for a single ganangle. Experience has shown that, in head and neck tment, the treatment time ratio between an IMRT plan anconventional 3DCRT plan is about 1.5 to 2.5. For prosttreatment, the time ratio is about 1 to 2, depending ondelivery system.

Foresight and training with respect to patient positionwill be needed to avoid problems with collisions or interfeence by patient support systems. ‘‘Dry run’’ tests mayuseful.

E. QA of equipment and individual patient treatments

In general, the QA of IMRT consists of three main taskcommissioning and testing of the treatment planning andlivery systems, routine QA of the delivery system, apatient-specific validation of treatment plans. The first tasmainly concerned with the integrity of the inverse planniand IMRT delivery system. The second aspect is concerwith the normal operation of the delivery system and winvolve additions to the daily, monthly, and annual QA prtocols. The third task is to ensure an accurate and safe tment of a patient. It is important to emphasize that IMRTa rapidly evolving modality and the QA program must alevolve to handle new issues that arise.

F. Staff training and patient education

Like any other radiation therapy modality, IMRT is aintegrated process, and staff training and education areimportant part of the clinical implementation of IMRT. It imuch more complex and less intuitive than conventio3DCRT. Experience gained by the staff in 3-D treatmeplanning and delivery is helpful but not sufficient for IMRTThere are significant differences between the two that nesitate additional specialized training. IMRT is often assoated with sharp dose gradients, increased heterogeneidose within the target volume, low MU efficiency~muchlarger number of MU compared with conventional radiatitherapy for the same prescribed dose!, and complex motionof MLCs. It is imperative that each member of the IMRteam understands the implications of each of these factouse this technology safely and effectively. IMRT is so diffeent from traditional radiation therapy that it can be easconsidered as a special procedure necessitating didtraining for key members before they implement this nmodality in their clinics. The training curriculum for eacIMRT team member must include all of the critical stepsthe IMRT process. Patient education in the nuances ofnew treatment modality is also essential.

1. Radiation oncologists

IMRT represents a significant departure from the currparadigm used in radiation oncology. Treatment planningconventional radiation therapy is accomplished in a very


dd

yat-aee

e

:e-

is

edl

at-

an

lt

s--of

to

tic

is

tn-

tuitive manner by optimizing the weights of strategicaplaced radiation portals that conform to the target volumPlanning solutions are often well understood and do not vmuch from patient to patient for a particular disease site.the other hand, the IMRT planning process starts withdefinition of treatment goal and objectives. The dose optization is completely computer controlled, and its succesachieving the clinical goals is very much dependent onset of parameters used as input to the computer algoritLearning how to adjust the parameters to steer the resulthe desired direction is complex and sometimes nonintuitTherefore, it is difficult to identify an optimal solution without having a complete understanding of the optimization pcess and its limitations. There is a significant potentialtreating a patient with a suboptimal IMRT treatment planthe radiation oncologist lacks the training in this process

One of the basic uses of IMRT is to treat tumors thateither in close proximity to or surrounded by critical normstructures, and this presents two challenges. One is toment the structures precisely and accurately, and the othto choose appropriate planning margins judiciously. It issential that the radiation oncologists are well-trainedimage-guided treatment planning and that they have a gunderstanding of treatment planning and delivery uncertaties.

Unlike conventional radiation therapy, the gross tumand regions of subclinical disease are often treated concotantly to different doses per fraction in IMRT. Moreover, thdose distribution in the target volume is often much lehomogeneous in an IMRT plan. It is important that thediation oncologists critically evaluate differential dosfractionation schedules for IMRT in light of their clinicaexperience with conventional radiation therapy. This requian understanding of the biologically effective equivaledose concepts and tissue tolerance doses.

Radiation oncologists who did not have the chance totraining in the IMRT process during their residency shouconsider attending special workshops conducted by ademic institutions that have active clinical IMRT programSome private companies have also started courses in IM

2. Radiation oncology physicists

IMRT is much more challenging for radiation oncologphysicists than conventional radiation therapy. Radiationcology physicists have a much more significant and dirrole in IMRT planning and delivery than in conventionradiation therapy. It requires an advanced understandinmathematical principles of dose optimization, computcontrolled delivery systems, and issues that relate to thesimetry of small and complex shaped radiation fields. Phycists also need to have a better understanding of treatmsetup, planning and delivery uncertainties, and their impon patients treated with IMRT. Treatment planning optimiztion for IMRT is based on dose–volume objectives and dlimits for critical structures and target tissues. Therefore, iimportant that radiation oncology physicists understand thconcepts and have a good familiarity with tomograph

in

isonith

dmapnis

tiopits,elss

iheia

ofdo

th

nt

m.o

rsheaprticitiaelb

inFoestiodreoyfilleacth

t-encant.n-ndent

to.resithust

ip-ardnlyancel-ispro-nty-thatms.lateIt iseral

ofler-t beandthet be

toen-ectsgi-ents

are

ralimeza-ndsntn-ol-tialringin-


anatomy. They must understand the implications of busytensity patterns~with large peaks and valleys! on treatmentdelivery accuracy and efficiency. QA testing for IMRTmuch more complex than it is for conventional radiatitherapy. It is imperative that each physicist involved wIMRT should have special training in the whole process.

3. Dosimetrists

The dosimetrists have the particularly difficult task of ajusting to IMRT planning. IMRT planning uses a paradigthat they are not used to in conventional radiation therplanning. Compared with treatment planning for convetional radiation therapy, the emphasis in IMRT planningmore on selecting the most appropriate dose optimizaparameters. Less importance is assigned to beam shaplacement and weight optimization in IMRT. Like physicisdosimetrists must understand the implications of dosvolume objectives on optimized dose distributions. They aneed to understand, at least conceptually, the implicationtreatment setup, planning, and delivery uncertaintiesIMRT. The best source of training for a dosimetrist is tfacility’s radiation oncologist and physicist who have spectraining in the use of IMRT.

4. Radiation therapists

Implementing IMRT requires the active involvementthe radiation treatment therapists. They should be involvethe design and testing of treatment procedures. It is imptant to set aside sufficient time for that participation andrelated training.

If the IMRT delivery involves specialized equipme~e.g., an add-on collimating device!, then there will be theneed to train the therapists in its use and storage. Theyalso have responsibilities for basic maintenance and QA

Therapists will have to be trained to use any new immbilization or localization systems.

However IMRT is delivered, be it with special collimatoor existing MLCs, therapists will need to be trained in tnew procedures. Carrying out mock procedures with phtoms needs to be part of the process of testing the newcedures. Delivery details that escape the physicist’s nomay be important to the therapists. For example, the infield shape for an IMRT treatment may obscure the light fior the crosshair, requiring that the patient be positionedfore the MLC is programmed.

Therapists must be provided with the means of knowthat the treatment they are about to deliver is correct.conventional treatments with blocks or static MLC shapthey can compare the field on the patient to the simulafilm, DRR, or other plan data. For IMRT, the initial fielshape may show only a narrow segment or be closed entiFor IMRT treatments, the analog to the physical blockstatic MLC file is the dynamic IMRT file. The physicist mawell have validated the intensity map produced by eachbefore treatment, but every day the therapist must be abverify that the appropriate file has been selected for efield or arc.~These issues were discussed previously in


-

-

y-

nng,

–oofn

l

inr-e

ay

-

n-o-el

de-

gr,n

ly.r

etohe

section on file transfer and management.! Given the com-plexity of IMRT treatments, it is clearly best for the treament delivery to be fully monitored by an R/V system. Evin that case, therapists will need to be trained so theyverify for themselves that the R/V programming is correc

Therapists will need to be shown how to respond to uplanned events. They need to know how to interrupt arestart a treatment, how to recover from a partial treatmthat requires the console to be reprogrammed, and howrecognize and act on new error messages and interlocks

Therapists will need to be trained on any new procedurelated to portal imaging and to new daily QA tests. As wany QA procedure, clear instructions and action levels mbe provided.

5. Service engineers

Reliable performance of all aspects of the delivery equment used for IMRT is essential. Compared with standtreatment techniques, it can be much more difficult to clearecover from an interruption in dose delivery afterintensity-modulated treatment has started. Therefore, acerators with a poor history of reliability are not suited for thtype of treatment, and expanded preventive maintenancegrams are extremely important. This is particularly importafor the MLC component of the overall system. Intensitmodulated dose delivery places demands on the MLCfar exceed the criteria used for the design of these systeWhen the standard MLC systems were designed in the1980s, IMRT was not anticipated as a routine treatment.now evident that some implementations can require sevhundred field changes per patient, or many thousandsfields per treatment day. This situation can lead to acceated component failure, and special QA procedures musadopted to guarantee proper calibration of leaf positionto avoid treatment interruptions. With the assistance ofmedical physicist, preventive maintenance programs musexamined to determine that they are properly designedaddress the special needs of IMRT. Additionally, servicegineers must have a good working knowledge of the aspof the treatment unit that are unique to IMRT. Service enneers need to understand that small changes or adjustmto a MLC can affect the machine output13 for IMRT deliveryand should confer with the physicist whenever changesmade.

6. Patient education

Patients treated with IMRT should be informed of seveissues. They need to be given realistic estimates of the trequired for each treatment, a description of the immobilition method used, and delivery system motions and southey will experience. A description of the goal of treatmeand potential side effects may differ from that given for coventional radiotherapy. These will be site- and protocspecific. If IMRT is used to escalate doses, then the potenfor acute or chronic sequelae may increase. Parotid-spaprotocols may decrease the incidence of xerostomia but

s

ioteis

eth

athalao

adb

nse

e-clo

eruatoo

inggdn

ianoisetibis

ficehr

as-

yfor-

ub-es-nic

p,st,’’

rt

’’

de-

thoot

a

.LC

af

F.ed

in--

nd-

ing’

ectsle-

ics.

g

ur-af

s-ra-r.

lied

-,’’

am


crease acute mucositis, especially if target doses are lesmogeneous than with conventional treatments.

Another issue is the need to manage patients’ expectatfor IMRT. Patients may come with the desire to be treawith this new, highly advertised modality, whether or not itadvantageous or appropriate for their condition. Patients~andtheir families! also converse together, and some may qution why their experience differs from that described by oers.

G. Patient scheduling, billing, and charting

IMRT treatments may take longer than conventional trements. They may also be implemented on only some oftreatment machines. New immobilization techniques malso be introduced simultaneously and would impact simution and treatment times. New imaging studies may bedered. Staff responsible for scheduling will need to bevised of new scheduling requirements. They shouldconsulted early in the implementation process so that coquences of those changes can be anticipated and adjustmmade.

Implementing IMRT offers new opportunities and requirments for billing and requires careful attention to complianissues. Administrators and other staff will need to deveefficient tools for billing and documentation.

The implementation team will need to consider needchanges in charting procedures. This could relate to insttions for treatment delivery, documentation of daily trement with many complex fields, documentation of QA prcedures, and dose summaries that adequately describe dvolume goals and results.

H. Overall integration

In this section we have stressed the importance of usthe combined expertise of an implementation team. Althouthe physics staff will carry much of the burden of installinand commissioning an IMRT system, ultimate successpends on the active support and involvement of physiciadosimetrists, therapists, and administrators.

V. SUMMARY

This document provides guidance to the practicing radtion oncology physicists in treatment delivery, treatmeplanning, and clinical implementation of IMRT. Becausethe emerging and rapidly changing nature of IMRT, thdocument cannot be definitive or prescriptive at this timHowever, several task group reports and a code of pracwill eventually emerge as the field matures. The IMRT Sucommittee of the AAPM Radiation Therapy Committeecurrently working on a document that will provide specirecommendations on the tolerance limits and action levfor different IMRT tests that are described in this report. Tintention here is to provide guidance during this introductoperiod.


ho-

nsd

s--

t-ey-

r--ee-nts

ep

dc---se–

gh

e-s,

-tf

.ce-

lsey

ACKNOWLEDGMENTS

The IMRT Subcommittee gratefully acknowledges thesistance of the Radiation Therapy Committee members~Jef-frey F. Williamson, Ph.D., Chairman!, who have contributedsignificantly to the quality of this report. We particularlthank John Gibbons, Ph.D., and Michael Herman, Ph.D.,a critical review of this document and making helpful comments and suggestions.

a!Address all correspondence to Jatinder R. Palta, Chairman IMRT Scommittee: Department of Radiation Oncology, P.O. Box 100385, Gainville Florida 32610. Phone: 352-265-8217; fax: 352-265-8417; electromail: [email protected] Modulated Radiation Therapy Collaborative Working Grou‘‘Intensity-modulated radiotherapy: Current status and issues of intereInt. J. Radiat. Oncol., Biol., Phys.51, 880–914~2001!.

2C. B. Saw, K. M. Ayyangar, and C. A. Enke, ‘‘MIMiC-based IMRT-PaI,’’ Med. Dosim. 26, 1 ~2001!.

3C. B. Saw, K. M. Ayyangar, and C. A. Enke, ‘‘MLC-based IMRT-Part II,Med. Dosim.26, 111–112~2001!.

4J. H. Kung and G. T. Chen, ‘‘Intensity modulated radiotherapy doselivery error from radiation field offset inaccuracy,’’ Med. Phys.27, 1617–1622 ~2000!.

5M. B. Sharpe, M. B. Miller, D. Yan, and J. W. Wong, ‘‘Monitor unisettings for intensity modulated beams delivered using a step-and-sapproach,’’ Med. Phys.27, 2719–2725~2000!.

6A. L. Boyer and S. Li, ‘‘Geometrical analysis of light-field position ofmultileaf collimator with curved ends,’’ Med. Phys.24, 757–762~1997!.

7M. N. Graves, A. V. Thompson, D. L. Martel, D. L. McShan, and B. AFraass, ‘‘Calibration and quality assurance for rounded leaf-end Msystems,’’ Med. Phys.28, 2227–2233~2001!.

8C. S. Chui, S. Spirou, and T. LoSasso, ‘‘Testing of dynamic multilecollimation,’’ Med. Phys.23, 635–641~1996!.

9D. A. Low, J. W. Sohn, E. E. Klein, J. Markman, S. Mutic, and J.Dempsey, ‘‘Characterization of a commercial multileaf collimator usfor intensity modulated radiation therapy,’’ Med. Phys.28, 752–756~2001!.

10C. W. Cheng and I. J. Das, ‘‘Comparison of beam characteristics intensity modulated radiation therapy~IMRT! and those under normal treatment condition,’’ Med. Phys.29, 226–230~2002!.

11G. A. Ezzell and S. Chungbin, ‘‘The overshoot phenomenon in step-ashoot IMRT delivery,’’ J. Appl. Clin. Med. Phys.2, 138–148~2001!.

12P. Xia, C. F. Chuang, and L. J. Verhey, ‘‘Communication and samplrate limitations in IMRT delivery with a dynamic multileaf collimator,’Med. Phys.29, 412–423~2001!.

13T. LoSasso, C. S. Chui, and C. C. Ling, ‘‘Physical and dosimetric aspof a multileaf collimation system used in the dynamic mode for impmenting intensity modulated radiotherapy,’’ Med. Phys.25, 1919–1927~1998!.

14E. E. Klein and D. A. Low, ‘‘Interleaf leakage for 5 and 10 mm dynammultileaf collimation systems incorporating patient motion,’’ Med. Phy28, 1703–1710~2001!.

15Q. Wu, M. Arnfield, S. Tong, Y. Wu, and R. Mohan, ‘‘Dynamic splittinof large intensity-modulated fields,’’ Phys. Med. Biol.45, 1731–1740~2000!.

16T. LoSasso, C. S. Chui, and C. C. Ling, ‘‘Comprehensive quality assance for the delivery of intensity modulated radiotherapy with a multilecollimator used in the dynamic mode,’’ Med. Phys.28, 2209–2219~2001!.

17M. Essers, M. de Langen, M. L. Dirkx, and B. J. Heijmen, ‘‘Commisioning of a commercially available system for intensity-modulateddiotherapy dose delivery with dynamic multileaf collimation,’’ RadiotheOncol.60, 215–224~2001!.

18M. L. Dirkx and B. J. Heijmen, ‘‘Testing of the stability of intensitymodulated beams generated with dynamic multileaf collimation, appto the MM50 racetrack microtron,’’ Med. Phys.27, 2701–2707~2000!.

19G. J. Budgell, J. H. Mott, P. C. Williams, and K. J. Brown, ‘‘Requirements for leaf position accuracy for dynamic multileaf collimationPhys. Med. Biol.45, 1211–1227~2000!.

20H. G. Kuterdem and P. S. Cho, ‘‘Leaf sequencing with secondary be

te

gei

ofCa

m,ia-

ri-

ton

keion

ke

n

A.p

A

. Aa-

i,s

ys

a-l.,

t oed

rd,’’

osu-

l-

d,ed.

afar

onr,’

med

celli-

ryco

li-

gs.

g

dnd

,’’

g,cy

h

-

ofra-

tedl.

hed.

ed.

ed.,

n

ofol-ys.

io-

-

,ity

ed

ra-d.

ingys.

icy-

e

te


blocking under leaf positioning constraints for continuously modularadiotherapy beams,’’ Med. Phys.28, 894–902~2001!.

21G. J. Budgell, J. R. Sykes, and J. M. Wilkinson, ‘‘Rectangular edsynchronization for intensity modulated radiation therapy with dynammultileaf collimation,’’ Phys. Med. Biol.43, 2769–2784~1998!.

22W. U. Laub, A. Bakai, and F. Nusslin, ‘‘Intensity modulated irradiationa thorax phantom: Comparisons between measurements, Montecalculations and pencil beam calculations,’’ Phys. Med. Biol.46, 1695–1706 ~2001!.

23J. Meyer, J. A. Mills, O. C. Haas, E. M. Parvin, and K. J. Burnha‘‘Some limitations in the practical delivery of intensity modulated radtion therapy,’’ Br. J. Radiol.73, 854–863~2000!.

24H. Thompson, M. D. Evans, and B. G. Fallone, ‘‘Accuracy of numecally produced compensators,’’ Med. Dosim.24, 49–52~1999!.

25S. B. Jiang and K. M. Ayyangar, ‘‘On compensator design for phobeam intensity-modulated conformal therapy,’’ Med. Phys.25, 668–675~1998!.

26C. B. Saw, K. M. Ayyangar, R. B. Thompson, W. Zhen, and C. A. En‘‘Commissioning of Peacock System for intensity-modulated radiattherapy,’’ Med. Dosim.26, 55–64~2001!.

27C. B. Saw, K. M. Ayyangar, W. Zhen, R. B. Thompson, and C. A. En‘‘Quality assurance for the Peacock system,’’ Med. Dosim.26, 83–90~2001!.

28B. J. Salter, ‘‘NOMOS Peacock IMRT utilizing the Beak post collimatiodevice,’’ Med. Dosim.26, 37–45~2001!.

29D. A. Low, K. S. Chao, S. Mutic, R. L. Gerber, C. A. Perez, and J.Purdy, ‘‘Quality assurance of serial tomotherapy for head and necktient treatments,’’ Int. J. Radiat. Oncol., Biol., Phys.42, 681–692~1998!.

30D. A. Low, S. Mutic, J. F. Dempsey, J. Markman, K. S. Chao, and J.Purdy, ‘‘Abutment dosimetry for serial tomotherapy,’’ Med. Dosim.26,79–82~2001!.

31D. A. Low, S. Mutic, J. F. Dempsey, J. Markman, S. M. Goddu, and JPurdy, ‘‘Abutment regional dosimetry for serial tomotherapy,’’ Int. J. Rdiat. Oncol., Biol., Phys.45, 193–203~1999!.

32L. B. Leybovich, N. Dogan, A. Sethi, M. J. Krasin, and B. Emam‘‘Improvement of tomographic intensity modulated radiotherapy dodistributions using periodic shifting of arc abutment regions,’’ Med. Ph27, 1610–1616~2000!.

33N. Dogan, L. B. Leybovich, A. Sethi, M. Krasin, and B. Emami, ‘‘Amodified method of planning and delivery for dynamic multileaf collimtor intensity-modulated radiation therapy,’’ Int. J. Radiat. Oncol., BioPhys.47, 241–245~2000!.

34J. S. Tsai, M. J. Rivard, and M. J. Engler, ‘‘Dependence of linac outputhe switch rate of an intensity-modulated tomotherapy collimator,’’ MPhys.27, 2215–2225~2000!.

35T. R. Mackie, J. Balog, K. Ruchala, D. Shepard, S. Aldridge, E. FitchaP. Reckwerdt, G. Olivera, T. McNutt, and M. Mehta, ‘‘TomotherapySemin Radiat. Oncol.9, 108–117~1999!.

36J. P. Balog, T. R. Mackie, P. Reckwerdt, M. Glass, and L. Angel‘‘Characterization of the output for helical delivery of intensity modlated slit beams,’’ Med. Phys.26, 55–64~1999!.

37C. X. Yu, ‘‘Intensity-modulated arc therapy with dynamic multileaf colimation: An alternative to tomotherapy,’’ Phys. Med. Biol.40, 1435–1449 ~1995!.

38L. Ma, C. X. Yu, M. Earl, T. Holmes, M. Sarfaraz, X. A. Li, D. SheparP. Amin, S. DiBiase, M. Suntharalingam, and C. Mansfield, ‘‘Optimizintensity-modulated arc therapy for prostate cancer treatment,’’ IntCancer96, 379–384~2001!.

39C. R. Ramsey, K. M. Spencer, R. Alhakeem, and A. L. Oliver, ‘‘Leposition error during conformal dynamic arc and intensity modulatedtreatments,’’ Med. Phys.28, 67–72~2001!.

40S. Webb, ‘‘Configuration options for intensity-modulated radiatitherapy using multiple static fields shaped by a multileaf collimatoPhys. Med. Biol.43, 241–260~1998!.

41P. Xia and L. J. Verhey, ‘‘Multileaf collimator leaf sequencing algorithfor intensity modulated beams with multiple static segments,’’ MPhys.25, 1424–1434~1998!.

42M. Langer, V. Thai, and L. Papiez, ‘‘Improved leaf sequencing redusegments or monitor units needed to deliver IMRT using multileaf comators,’’ Med. Phys.28, 2450–2458~2001!.

43J. Dai and Y. Zhu, ‘‘Minimizing the number of segments in a delivesequence for intensity-modulated radiation therapy with a multileaflimator,’’ Med. Phys.28, 2113–2120~2001!.


d

c

rlo

,

,

a-

.

.

e.

n.

,

,

J.

c

’

.

s

l-

44L. Ma, A. L. Boyer, L. Xing, and C. M. Ma, ‘‘An optimized leaf-settingalgorithm for beam intensity modulation using dynamic multileaf colmators,’’ Phys. Med. Biol.43, 1629–1643~1998!.

45Y. Wu, D. Yan, M. B. Sharpe, B. Miller, and J. W. Wong, ‘‘Implementinmultiple static field delivery for intensity modulated beams,’’ Med. Phy28, 2188–2197~2001!.

46R. A. Siochi, ‘‘Minimizing static intensity modulation delivery time usinan intensity solid paradigm,’’ Int. J. Radiat. Oncol., Biol., Phys.43, 671–680 ~1999!.

47L. D. Potter, S. X. Chang, T. J. Cullip, and A. C. Siochi, ‘‘A quality anefficiency analysis of the IMFAST™ segmentation algorithm in head aneck ‘‘step and shoot’’ IMRT treatments,’’ Med. Phys.29, 275–283~2002!.

48S. M. Crooks, L. F. McAven, D. F. Robinson, and L. Xing, ‘‘Minimizingdelivery time and monitor units in static IMRT by leaf-sequencingPhys. Med. Biol.47, 3105–3116~2002!.

49S. V. Spirou, N. Fournier-Bidoz, J. Yang, C. S. Chui, and C. C. Lin‘‘Smoothing intensity-modulated beam profiles to improve the efficienof delivery,’’ Med. Phys.28, 2105–2112~2001!.

50A. L. Boyer and C. X. Yu, ‘‘Intensity-modulated radiation therapy witdynamic multileaf collimators,’’ Semin Radiat. Oncol.9, 48–59~1999!.

51W. Que, ‘‘Comparison of algorithms for multileaf collimator field segmentation,’’ Med. Phys.26, 2390–2396~1999!.

52A. W. Beavis, P. S. Ganney, V. J. Whitton, and L. Xing, ‘‘Optimizationthe step-and-shoot leaf sequence for delivery of intensity modulateddiation therapy using a variable division scheme,’’ Phys. Med. Biol.46,2457–2465~2001!.

53G. J. Budgell, ‘‘Temporal resolution requirements for intensity modularadiation therapy delivered by multileaf collimators,’’ Phys. Med. Bio44, 1581–1596~1999!.

54J. R. Sykes and P. C. Williams, ‘‘An experimental investigation of ttongue and groove effect for the Philips multileaf collimator,’’ Phys. MeBiol. 43, 3157–3165~1998!.

55J. Deng, T. Pawlicki, Y. Chen, J. Li, S. B. Jiang, and C. M. Ma, ‘‘ThMLC tongue-and-groove effect on IMRT dose distributions,’’ Phys. MeBiol. 46, 1039–1060~2001!.

56L. Ma, A. L. Boyer, C. M. Ma, and L. Xing, ‘‘Synchronizing dynamicmultileaf collimators for producing two-dimensional intensity-modulatfields with minimum beam delivery time,’’ Int. J. Radiat. Oncol., BiolPhys.44, 1147–1154~1999!.

57J. P. van Santvoort and B. J. Heijmen, ‘‘Dynamic multileaf collimatiowithout ‘‘tongue-and-groove’’ underdosage effects,’’ Phys. Med. Biol.41,2091–2105~1996!.

58C. S. Chui, M. F. Chan, E. Yorke, S. Spirou, and C. C. Ling, ‘‘Deliveryintensity-modulated radiation therapy with a conventional multileaf climator: Comparison of dynamic and segmental methods,’’ Med. Ph28, 2441–2449~2001!.

59P. Xia and L. J. Verhey, ‘‘Delivery systems of intensity-modulated radtherapy using conventional multileaf collimators,’’ Med. Dosim26, 169–177 ~2001!.

60D. A. Low and S. Mutic, ‘‘A commercial IMRT treatment-planning dosecalculation algorithm,’’ Int. J. Radiat. Oncol., Biol., Phys.41, 933–937~1998!.

61A. Boyer, L. Xing, C. M. Ma, B. Curran, R. Hill, A. Kania, and A. Bieier‘‘Theoretical considerations of monitor unit calculations for intensmodulated beam treatment planning,’’ Med. Phys.26, 187–195~1999!.

62M. H. Phillips, K. M. Singer, and A. R. Hounsell, ‘‘A macropencil beammodel: Clinical implementation for conformal and intensity modulatradiation therapy,’’ Phys. Med. Biol.44, 1067–1088~1999!.

63J. V. Siebers, S. Tong, M. Lauterbach, Q. Wu, and R. Mohan, ‘‘Acceletion of dose calculations for intensity-modulated radiotherapy,’’ MePhys.28, 903–910~2001!.

64J. V. Siebers, M. Lauterbach, S. Tong, Q. Wu, and R. Mohan, ‘‘Reducdose calculation time for accurate iterative IMRT planning,’’ Med. Ph29, 231–237~2002!.

65H. H. Liu, F. Verhaegen, and L. Dong, ‘‘A method of simulating dynammultileaf collimators using Monte Carlo techniques for intensitmodulated radiation therapy,’’ Phys. Med. Biol.46, 2283–2298~2001!.

66X. A. Li, L. Ma, S. Naqvi, R. Shih, and C. Yu, ‘‘Monte Carlo dosverification for intensity-modulated arc therapy,’’ Phys. Med. Biol.46,2269–2282~2001!.

67P. J. Keall, J. V. Siebers, M. Arnfield, J. O. Kim, and R. Mohan, ‘‘Mon

iol

rsee

gas

el

ofthys

n

T:u

in

-d

W.ionap

chle-In

, WHof. J

myraiol

, Coner

evt.

onor

r-ol.

ith

ty

no

n

b-

na

Gam

.-al

, S.ofatic

‘‘Anetri-

ns.

e

tyn-

S.e-du-

.feck

,’’

i-ity-

diat.

b-

nuni-

,

-ten-

B.oseate

dalate

M.ar-d

-

epy,’’

ntdio-

C.ity-

an,

.


Carlo dose calculations for dynamic IMRT treatments,’’ Phys. Med. B46, 929–941~2001!.

68R. Jeraj and P. Keall, ‘‘The effect of statistical uncertainty on invetreatment planning based on Monte Carlo dose calculation,’’ Phys. MBiol. 45, 3601–3613~2000!.

69S. A. Naqvi, M. Sarfaraz, T. Holmes, C. X. Yu, and X. A. Li, ‘‘Analysincollimator structure effects in head-scatter calculations for IMRT clfields using scatter raytracing,’’ Phys. Med. Biol.46, 2009–2028~2001!.

70Y. Yang, L. Xing, A. Boyer, Y. Song, and Y. Hu, ‘‘A three-source modfor the calculation of head scatter factors,’’ Med. Phys.29, 2024–2033~2002!.

71R. Mohan, M. Arnfield, S. Tong, Q. Wu, and J. Siebers, ‘‘The impactfluctuations in intensity patterns on the number of monitor units andquality and accuracy of intensity modulated radiotherapy,’’ Med. Ph27, 1226–1237~2000!.

72Y. Chen, A. L. Boyer, and C. M. Ma, ‘‘Calculation of x-ray transmissiothrough a multileaf collimator,’’ Med. Phys.27, 1717–1726~2000!.

73J. D. Azcona, R. A. Siochi, and I. Azinovic, ‘‘Quality assurance in IMRImportance of the transmission through the jaws for an accurate calction of absolute doses and relative distributions,’’ Med. Phys.29, 269–274 ~2002!.

74P. S. Cho and R. J. 2. Marks, ‘‘Hardware-sensitive optimization fortensity modulated radiotherapy,’’ Phys. Med. Biol.45, 429–440~2000!.

75Y. Xiao, J. Galvin, M. Hossain, and R. Valicenti, ‘‘An optimized forwardplanning technique for intensity modulated radiation therapy,’’ MePhys.27, 2093–2099~2000!.

76B. Fraass, M. L. Kessler, D. L. McShan, L. H. Marsh, B. A. Watson,J. Dusseau, A. Eisbruch, H. M. Sandler, and A. S. Lichter, ‘‘Optimizatand clinical use of multisegment intensity-modulated radiation therfor high-dose conformal therapy,’’ Semin Radiat. Oncol.9, 60–77~1999!.

77B. Pickett, E. Vigneault, J. Kurhanewicz, L. Verhey, and M. Roa‘‘Static field intensity modulation to treat a dominant intra-prostaticsion to 90 Gy compared to seven field 3-dimensional radiotherapy,’’J. Radiat. Oncol., Biol., Phys.44, 921–929~1999!.

78F. Claus, W. De Gersem, C. De Wagter, R. Van Severen, I. VanhoutteDuthoy, V. Remouchamps, B. Van Duyse, L. Vakaet, M. Lemmerling,Vermeersch, and W. De Neve, ‘‘An implementation strategy for IMRTethmoid sinus cancer with bilateral sparing of the optic pathways,’’ IntRadiat. Oncol., Biol., Phys.51, 318–331~2001!.

79W. De Gersem, F. Claus, C. De Wagter, and W. De Neve, ‘‘An anatobased beam segmentation tool for intensity-modulated radiation theand its application to head-and-neck cancer,’’ Int. J. Radiat. Oncol., BPhys.51, 849–859~2001!.

80F. Claus, W. De Gersem, I. Vanhoutte, W. Duthoy, V. RemouchampsDe Wagter, and W. De Neve, ‘‘Evaluation of a leaf position optimizatitool for intensity modulated radiation therapy of head and neck cancRadiother. Oncol.61, 281–286~2001!.

81W. De Gersem, F. Claus, C. De Wagter, B. Van Duyse, and W. De N‘‘Leaf position optimization for step-and-shoot IMRT,’’ Int. J. RadiaOncol., Biol., Phys.51, 1371–1388~2001!.

82M. Carol, ‘‘An automatic 3-D treatment planning and implementatisystem for optimized conformal radiation therapy by the NOMOS Cporation,’’ Int. J. Radiat. Oncol., Biol., Phys.23, 1081~1992! ~abstract!.

83A. L. Boyer, P. Geis, W. Grant, and M. Carol, ‘‘Modulated beam confomal therapy for head and neck tumors,’’ Int. J. Radiat. Oncol., BiPhys.39, 227–236~1997!.

84S. V. Spirou and C. S. Chui, ‘‘A gradient inverse planning algorithm wdose–volume constraints,’’ Med. Phys.25, 321–333~1998!.

85Q. Wu and R. Mohan, ‘‘Algorithms and functionality of an intensimodulated radiotherapy organization system,’’ Med. Phys.27, 701–711~2000!.

86L. Xing, J. G. Li, A. Pugachev, Q. T. Le, and A. L. Boyer, ‘‘Estimatiotheory and model parameter selection for therapeutic treatment plantimization,’’ Med. Phys.26, 2348–2358~1999!.

87L. Xing, J. G. Li, S. Donaldson, Q. T. Le, and A. L. Boyer, ‘‘Optimizatioof importance factors in inverse planning,’’ Phys. Med. Biol.44, 2525–2536 ~1999!.

88J. O. Deasy, ‘‘Multiple local minima in radiotherapy optimization prolems with dose–volume constraints,’’ Med. Phys.24, 1157–1161~1997!.

89C. S. Chui and S. V. Spirou, ‘‘Inverse planning algorithms for exterbeam radiation therapy,’’ Med. Dosim.26, 189–197~2001!.

90L. Hong, M. Hunt, C. Chui, S. Spirou, K. Forster, H. Lee, J. Vahalom,J. Kutcher, and B. McCormick, ‘‘Intensity-modulated tangential be


.

d.

s

e.

la-

-

.

y

,

t.

..

.

-py.,

.

,’’

e,

-

,

p-

l

.

irradiation of the intact breast,’’ Int. J. Radiat. Oncol., Biol., Phys.44,1155–1164~1999!.

91L. L. Kestin, M. B. Sharpe, R. C. Frazier, F. A. Vicini, D. Yan, R. CMatter, A. A. Martinez, and J. W. Wong, ‘‘Intensity modulation to improve dose uniformity with tangential breast radiotherapy: Initial clinicexperience,’’ Int. J. Radiat. Oncol., Biol., Phys.48, 1559–1568~2000!.

92P. M. Evans, E. M. Donovan, M. Partridge, P. J. Childs, D. J. ConveryEagle, V. N. Hansen, B. L. Suter, and J. R. Yarnold, ‘‘The deliveryintensity-modulated radiotherapy to the breast using multiple stfields,’’ Radiother. Oncol.57, 79–89~2000!.

93B. van Asselen, C. P. Raaijmakers, P. Hofman, and J. J. Lagendijk,improved breast irradiation technique using three-dimensional geomcal information and intensity modulation,’’ Radiother. Oncol.58, 341–347 ~2001!.

94C. X. Yu, D. A. Jaffray, and J. W. Wong, ‘‘The effects of intra-fractioorgan motion on the delivery of dynamic intensity modulation,’’ PhyMed. Biol. 43, 91–104~1998!.

95P. J. Keall, V. R. Kini, S. S. Vedam, and R. Mohan, ‘‘Motion adaptivx-ray therapy: A feasibility study,’’ Phys. Med. Biol.46, 1–10~2001!.

96D. A. Jaffray, D. Yan, and J. W. Wong, ‘‘Managing geometric uncertainin conformal intensity-modulated radiation therapy,’’ Semin Radiat. Ocol. 9, 4–19~1999!.

97L. Xing, Z. Lin, S. S. Donaldson, Q. T. Le, D. Tate, D. R. Goffinet,Wolden, L. Ma, and A. L. Boyen, ‘‘Dosimetric effects of patient displacment and collimator and gantry angle misalignment on intensity molated radiation therapy,’’ Radiother. Oncol.56, 97–108~2000!.

98M. A. Manning, Q. Wu, R. M. Cardinale, R. Mohan, A. D. Lauve, B. DKavanagh, M. M. Morris, and R. K. Schmidt-Ullrich, ‘‘The effect osetup uncertainty on normal tissue sparing with IMRT for head-and-ncancer,’’ Int. J. Radiat. Oncol., Biol., Phys.51, 1400–1409~2001!.

99J. G. Li and L. Xing, ‘‘Inverse planning incorporating organ motionMed. Phys.27, 1573–1578~2000!.

100R. Mohan, Q. Wu, M. Manning, and R. Schmidt-Ullrich, ‘‘Radiobiologcal considerations in the design of fractionation strategies for intensmodulated radiation therapy of head and neck cancers,’’ Int. J. RaOncol., Biol., Phys.46, 619–630~2000!.

101A. Brahme, ‘‘Optimized radiation therapy based on radiobiological ojectives,’’ Semin Radiat. Oncol.9, 35–47~1999!.

102Q. Wu, R. Mohan, A. Niemierko, and R. Schmidt-Ullrich, ‘‘Optimizatioof intensity-modulated radiotherapy plans based on the equivalentform dose,’’ Int. J. Radiat. Oncol., Biol., Phys.52, 224–235~2002!.

103E. B. Butler, B. S. Teh, W. H. Grant III, B. M. Uhl, R. B. KuppersmithJ. K. Chiu, D. T. Donovan, and S. Y. Woo, ‘‘SMART~simultaneousmodulated accelerated radiation therapy! boost: A new accelerated fractionation schedule for the treatment of head and neck cancer with insity modulated radiotherapy,’’ Int. J. Radiat. Oncol., Biol., Phys.45,21–32~1999!.

104L. J. Bos, E. M. Damen, R. W. de Boer, B. Mijnheer, D. L. McShan,A. Fraass, M. L. Kessler, and J. V. Lebesque, ‘‘Reduction of rectal dby integration of the boost in the large-field treatment plan for prostirradiation,’’ Int. J. Radiat. Oncol., Biol., Phys.52, 254–265~2002!.

105H. K. Shu, T. T. Lee, E. Vigneauly, P. Xia, B. Pickett, T. L. Phillips, anM. Roach, ‘‘Toxicity following high-dose three-dimensional conformand intensity-modulated radiation therapy for clinically localized prostcancer,’’ Urology57, 102–107~2001!.

106M. J. Zelefsky, Z. Fuks, L. Happersett, H. J. Lee, C. C. Ling, C.Burman, M. Hunt, T. Wolfe, E. S. Venkatraman, A. Jackson, M. Skwchuk, and S. A. Leibel, ‘‘Clinical experience with intensity modulateaccelerated radiation therapy~IMRT! in prostate cancer,’’ Radiother. Oncol. 55, 241–249~2000!.

107D. Followill, P. Geis, and A. Boyer, ‘‘Estimates of whole body dosequivalent produced by beam intensity modulated conformal theraInt. J. Radiat. Oncol., Biol., Phys.38, 667–672~1997!.

108G. A. Ezzell, S. E. Schild, and W. W. Wong, ‘‘Development of a treatmeplanning protocol for prostate treatments using intensity modulated ratherapy,’’ J. Appl. Clin. Med. Phys.2, 59–68~2001!.

109J. Stein, R. Mohan, X. H. Wang, T. Bortfield, Q. Wu, K. Preiser, C.Ling, and W. Schlegel, ‘‘Number and orientations of beams in intensmodulated radiation treatments,’’ Med. Phys.24, 149–160~1997!.

110X. Wang, S. Spirou, T. LoSasso, J. Stein, C. S. Chui, and B. Moh‘‘Dosimetric verification of intensity-modulated fields,’’ Med. Phys.23,317–327~1996!.

111L. Xing, Y. Curran, R. Hill, T. Holmes, L. Ma, K. M. Forster, and A. L

n-

ini

anondio

n-im

a

nlit

r,

tio

tte

ity

Tdia

ms.

dioe,

a

ity

ar-D

48te

-

sid-nic

d-a

ndA

ed.

ivey,’’

tede,’’

n,andf

ods

ms,’’

r,y

‘Anmo-

ks,rnalgy

o-

A.adnt of

r

,ed

ofiol.,


Boyer, ‘‘Dosimetric verification of a commercial inverse treatment planing system,’’ Phys. Med. Biol.44, 463–478~1999!.

112C. Martens, C. De Wagter, and W. De Neve, ‘‘The value of the PinPoion chamber for characterization of small field segments usedintensity-modulated radiotherapy,’’ Phys. Med. Biol.45, 2519–2530~2000!.

113B. Fraass, K. Doppke, M. Hunt, G. Kutcher, G. Starkschall, R. Stern,J. Van Dyke, ‘‘American Association of Physicists in Medicine RadiatiTherapy Committee Task Group 53: Quality assurance for clinical ratherapy treatment planning,’’ Med. Phys.25, 1773–1829~1998!.

114K. M. Ayyangar, C. B. Saw, B. Shen, C. A. Enke, and P. S. Nizin, ‘‘Idependent dose calculations for the PEACOCK system,’’ Med. Dos26, 29–35~2001!.

115Y. Watanabe, ‘‘Point dose calculations using an analytical pencil bekernel for IMRT plan checking,’’ Phys. Med. Biol.46, 1031–1038~2001!.

116J. H. Kung, G. T. Chen, and F. K. Kuchnir, ‘‘A monitor unit verificatiocalculation in intensity modulated radiotherapy as a dosimetry quaassurance,’’ Med. Phys.27, 2226–2230~2000!.

117C. M. Ma, T. Pawlicki, S. B. Jiang, J. S. Li, J. Deng, E. Mok, A. KapuL. Xing, L. Ma, and A. L. Boyer, ‘‘Monte Carlo verification of IMRTdose distributions from a commercial treatment planning optimizasystem,’’ Phys. Med. Biol.45, 2483–2495~2000!.

118L. Xing, Y. Chen, G. Luxton, J. G. Li, and A. L. Boyer, ‘‘Monitor unicalculation for an intensity modulated photon field by a simple scatsummation algorithm,’’ Phys. Med. Biol.45, N1–7 ~2000!.

119L. Xing and J. G. Li, ‘‘Computer verification of fluence maps in intensmodulated radiation therapy,’’ Med. Phys.27, 2084–2092~2000!.

120D. A. Low, R. L. Gerber, S. Mutic, and J. A. Purdy, ‘‘Phantoms for IMRdose distribution measurement and treatment verification,’’ Int. J. RaOncol., Biol., Phys.40, 1231–1235~1998!.

121B. Paliwal, W. Tome, S. Richardson, and T. R. Makie, ‘‘A spiral phantofor IMRT and tomotherapy treatment delivery verification,’’ Med. Phy27, 2503–2507~2000!.

122A. J. Olch, ‘‘Dosimetric performance of an enhanced dose range ragraphic film for intensity-modulated radiation therapy quality assurancMed. Phys.29, 2159–2168~2002!.

123S. Chungbin and G. Ezzell, ‘‘Streamlined IMRT patient verification mesurements using Kodak Extended Dose Range film~EDR2!,’’ Med. Phys.28, 1208~2001! ~abstract!.

124S. J. Ju, Y. C. Ahn, S. J. Huh, and I. J. Yeo, ‘‘Film dosimetry for intensmodulated radiation therapy: Dosimetric evaluation,’’ Med. Phys.29,351–355~2002!.

125D. A. Low, J. F. Dempsey, R. Venkatesan, S. Mutic, J. Markman, E. MHaacke, and J. A. Purdy, ‘‘Evaluation of polymer gels and MRI as a 3dosimeter for intensity-modulated radiation therapy,’’ Med. Phys.26,1542–1551~1999!.

126C. Martens, C. De Wagter, and W. De Neve, ‘‘The value of the LAlinear ion chamber array for characterization of intensity-modulabeams,’’ Phys. Med. Biol.46, 1131–1148~2001!.

127J. S. Li, A. L. Boyer, and C. M. Ma, ‘‘Verification of IMRT dose distributions using a water beam imaging system,’’ Med. Phys.28, 2466–2474~2001!.


tn

d

-

.

m

y

n

r-

t.

-’’

-

k

d

128P. M. Evans, M. Partridge, and J. R. Symonds-Tayler, ‘‘Sampling conerations for intensity modulated radiotherapy verification using electroportal imaging,’’ Med. Phys.28, 543–552~2001!.

129H. Keller, M. K. Fix, L. Liistro, and P. Ruegsegger, ‘‘Theoretical consierations to the verification of dynamic multileaf collimated fields withSLIC-type portal image detector,’’ Phys. Med. Biol.45, 2531–2545~2000!.

130M. Partridge, P. M. Evans, M. van Herk, L. S. Ploeger, G. J. Budgell, aH. V. James, ‘‘Leaf position verification during dynamic beam delivery:comparison of three applications using electronic portal imaging,’’ MPhys.27, 1601–1609~2000!.

131J. Chang, G. S. Mageras, C. S. Chui, C. C. Ling, and W. Lutz, ‘‘Relatprofile and dose verification of intensity-modulated radiation therapInt. J. Radiat. Oncol., Biol., Phys.47, 231–240~2000!.

132A. J. Curtin-Savard and E. B. Podgorsak, ‘‘Verification of segmenbeam delivery using a commercial electronic portal imaging devicMed. Phys.26, 737–742~1999!.

133G. J. Kutcher, L. Coia, M. Gillin, W. F. Hanson, S. Leibel, R. J. MortoJ. R. Palta, J. A. Purdy, L. E. Reinstein, G. K. Svensson, M. Weller,L. Wingfield, ‘‘Comprehensive QA for radiation oncology: Report oAAPM Radiation Therapy Committee Task Group 40,’’ Med. Phys.21,581–618~1994!.

134J. E. Rodgers, ‘‘Radiation therapy vault shielding calculational methwhen IMRT and TBI procedures contribute,’’ J. Appl. Clin. Med. Phys.2,157–164~2001!.

135S. Mutic, D. A. Low, E. E. Klein, J. F. Dempsey, and J. A. Purdy, ‘‘Rooshielding for intensity-modulated radiation therapy treatment facilitieInt. J. Radiat. Oncol., Biol., Phys.50, 239–246~2001!.

136C. B. Saw, R. Yakoob, C. A. Enke, T. P. Lau, and K. M. Ayyanga‘‘Immobilization devices for intensity-modulated radiation therap~IMRT!,’’ Med. Dosim. 26, 71–77~2001!.

137B. J. Salter, J. M. Hevezi, A. Sadeghi, M. Fuss, and T. S. Herman, ‘oblique arc capable patient positioning system for sequential totherapy,’’ Med. Phys.28, 2475–2488~2001!.

138J. Lattanzi, S. McNeeley, A. Hanlon, T. S. Schultheiss, and G. E. Han‘‘Ultrasound-based stereotactic guidance of precision conformal extebeam radiation therapy in clinically localized prostate cancer,’’ Urolo55, 73–78~2000!.

139J. E. McGary, B. S. Teh, E. B. Butler, and W. Grant III, ‘‘Prostate immbilization using a rectal balloon,’’ J. Appl. Clin. Med. Phys.3, 6–11~2002!.

140B. J. Salter, M. Fuss, D. G. Vollmer, A. Sadeghi, C. A. Bogaev, D.Cheek, T. S. Herman, and J. M. Hevezi, ‘‘The TALON removable heframe system for stereotactic radiosurgery/radiotherapy: Measuremethe repositioning accuracy,’’ Int. J. Radiat. Oncol., Biol., Phys.51, 555–562 ~2001!.

141J. E. McGary and W. Grant III, ‘‘A clinical evaluation of setup errors foa prostate immobilization system,’’ J. Appl. Clin. Med. Phys.1, 138–147~2000!.

142J. S. Tsai, M. J. Engler, M. N. Ling, J. K. Wu, B. Kramer, T. Dipetrilloand D. E. Wazer, ‘‘A non-invasive immobilization system and relatquality assurance for dynamic intensity modulated radiation therapyintracranial and head and neck disease,’’ Int. J. Radiat. Oncol., BPhys.43, 455–467~1999!.

Documents

Guidance document on delivery, treatment planning, and clinical implementation …lcr.uerj.br/Manual_ABFM/Guidance document on delivery... · 2010. 10. 11. · Guidance document on