Clinical Outcomes Modeling and Molecular Imaging amos3.aapm.org/abstracts/pdf/124-34900-405535-125572

  • View
    0

  • Download
    0

Embed Size (px)

Text of Clinical Outcomes Modeling and Molecular Imaging...

  • S L I D E 0

    Clinical Outcomes Modeling and Molecular Imaging for Hypofractionated Radiotherapy

    David J. Carlson, Ph.D. Associate Professor Dept. of Therapeutic Radiology Yale University School of Medicine david.j.carlson@yale.edu

    Presented at the 2017 AAPM Spring Clinical Meeting

    Date: Monday, March 20, 2017 from 4:30-5:30 p.m. Location: Hilton New Orleans Riverside, New Orleans, LA Conflict of Interest: None

  • S L I D E 1

    Background and Motivation

    Biologically Guided Radiation Therapy (BGRT) – Systematic method to derive prescription doses that integrate patient-specific

    information about tumor and normal tissue biology – Optimize treatment conditions based on biological objectives

    What are the Big Questions for hypofractionated RT?  To what extent does classical radiobiology apply at high doses?

     Fundamental difference in biology between conventional and SBRT? • Primary mechanism of cell death in fractionated RT is mitotic cell death related to biological

    processing of DSBs (standard view until ~2000) • Could the relevant biological mechanisms differ at high doses

     Are conventional models valid at high doses per fraction?

     Is there an optimal time course for SBRT?

  • S L I D E 2

    The utilization of SBRT is rising

    Primary early-stage NSCLC patients treated with SBRT (U.S. National Cancer Data Base, published in Corso et al. Am. J. Clin. Oncol. 2014)

    Rubio et al., 2013 Reports of Practical Oncology and Radiotherapy; 18: 387–396

  • S L I D E 3

    Why are clinical outcomes so good for SBRT?

    Unique biological mechanisms have been suggested:

    Tumor vasculature damage at high doses – Rapid tumor vascular shutdown due to endothelial cell apoptosis increases tumor hypoxia

    and reduces repair of radiation damage to tumor cells (Fuks and Kolesnick, MSKCC)

    – Vascular damage at high doses produces secondary cell killing (Song, UM)

    Enhanced antitumor immunity at high doses

    A detailed analysis of evidence for and against these mechanisms is in Brown JM, Carlson DJ, Brenner DJ. Int. J. Radiat. Oncol. Biol. Phys. 88: 254–262 (2014).

  • S L I D E 4

    Treatment Planning and Delivery

    • Objective in conventional RT to deliver uniform Rx dose to target volume • Paradigm shift for prescribing dose for SBRT

    1. Target a limited tissue volume, containing gross tumor and margin, with very high doses and hotspots within the target are acceptable  facilitated by advancement in technology of IMRT/IGRT/VMAT

    2. Minimize volume of normal tissue receiving high doses  sharp dose gradients

  • S L I D E 5

    Tumor Control Probability (TCP) Model

     exp ( )TCP N S D  

    Tumor Control Probabilities for intermediate- risk prostate cancer patient group (n = 40) (Levegrun et al 2001)

    Mean Dose (Gy)

    50 60 70 80 90 100

    TC P

    0.0

    0.2

    0.4

    0.6

    0.8

    1.0

    Clinical data from MSKCC TCP Model Fit: N = 4.1 x 106 cells .15 Gy-1

    3.1 Gy

    TCP → relates tumor size and radiation dose to the prob. of tumor control (i.e., no tumor cells survive)

    N = initial # of tumor clonogens

     2exp D DN e       

    Data from: Levegrun et al. IJROBP 2001; 51 (4): 1064–1080

  • S L I D E 6

    Inter-patient variability in radiosensitivity

    Figure from: Keall PJ, Webb S. Optimum parameters in a model for tumour control probability, including interpatient heterogeneity: evaluation of the log- normal distribution. Phys. Med. Biol. 2007; 52: 291302.

     Heterogeneity of human tumour radiation response is well known

     Can account for variation in inter- (and intra-) patient radiosensitivity by assuming that parameter values are normally distributed across the population

     If interpatient heterogeneity is ignored, TCP model generally results in an unrealistically steep dose- response curve

  • S L I D E 7

    How do we move towards hypofractionation?

    2 GyBED 78 Gy 1 130 Gy 3 Gy

         

     

    2/ 4 2 /

    nBEDd n n n

       

           

     

     Rearrange simplified BED equation:

    Number of fractions 0 5 10 15 20 25 30 35 40

    Fr ac

    tio n

    Si ze

    (G y)

    0 2 4 6 8

    10 12 14 16 18 20 22 24 26

    / = 3.0 Gy

    1

    Number of fractions 0 5 10 15 20 25 30 35 40

    Fr ac

    tio n

    Si ze

    (G y)

    0 2 4 6 8

    10 12 14 16 18 20 22 24 26

    / = 3.0 Gy / = 1.5 Gy / = 10 Gy

    1

    King et al. (2009)

    Standard Fractionation

    Kupelian et al. (2005)

    Yeoh et al. (2006)

    BED 1 / dD

       

        

    Isoeffect BED Example for Prostate  Conventional: 39 fractions of 2 Gy (/ = 3 Gy):

  • S L I D E 8

    mins hours days

    DNA repair

    Repopulation

    Reoxygenation & Redistribution

    Factors that alter treatment effectiveness

    Treatment effectiveness

    Treatment duration

    4 R’s of Radiobiology give rise to “dose rate” effects:

    5th R: Intrinsic Radiosensitivity

  • S L I D E 9

    What about tumor hypoxia at high doses?

    Absorbed Dose (Gy) 0 5 10 15 20 25 30 35

    Su rv

    iv in

    g Fr

    ac tio

    n

    10-3

    10-2

    10-1

    100 Fit to aerobic data Fit to hypoxic data Simultaneous fit to aerobic and hypoxic data (OER = 2.96)

    H = A/OER = 0.05 Gy-1 (/)H = (/)A*OER = 13.6 Gy

    A = 0.15 Gy-1 (/)A = 4.6 Gy

    V79 379A Chinese hamster cell survival data from Watts et al. (1986)

    OER = 2.96

     OER values for cell death are relatively constant over a large dose range

    • May actually increase slightly with dose (Wouters and Brown 1997, Nahum et al. 2003)

     Statistically, OER ~ OER • Reasonable assumption for large number

    of in vitro data sets (Carlson et al. 2006)

    Carlson DJ, Stewart RD, Semenenko VA. Effects of oxygen on intrinsic radiation sensitivity - a test of the relationship between aerobic and hypoxic linear-quadratic (LQ) model parameters. Med Phys; 33: 31053115 (2006).

  • S L I D E 10

    Clinical significance of tumor hypoxia

    B. Movsas et al., Urology, 2002D.M. Brizel et al., Radiother. Oncol., 1999

    Prostate cancerHead and neck cancer

    ~90% of solid tumors have median values below normal (40-60 mmHg), half have median values

  • S L I D E 11

    Effects of Hypoxia and Fractionation on Cell Survival

    Carlson DJ, Keall PJ, Loo BW, Chen ZJ, Brown JM. Hypofractionation results in reduced tumor cell kill compared to conventional fractionation for tumors with regions of hypoxia. Int. J. Radiat. Oncol. Biol. Phys. 79: 1188-1195 (2011).

    Number of fractions 0 5 10 15 20 25 30 35 40

    Su rv

    iv in

    g Fr

    ac tio

    n

    10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100

    0 5 10 15 20 25 30 35 40 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100

    Fully oxygenated cells Fully oxygenated cells

    Head and Neck Cancer (/ = 10 Gy)

    Prostate Cancer (/ = 3.0 Gy)

    Dose per fraction (Gy) to yield equivalent tumor BED under normoxic conditions 3.2 2.7 2.4 2.2 2.03.84.97.518.31.75.3 3.9 3.18.6 2.6 2.2 1.923.8

    1 1

     2( ) exp nnoverall A AS S d d d      

    / = 10 Gy / = 3 Gy

    What happens to total cell killing if we include hypoxia?

    Number of fractions 0 5 10 15 20 25 30 35 40

    Su rv

    iv in

    g Fr

    ac tio

    n

    10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100

    fhyp = 0.1 fhyp = 0.2 fhyp = 0.3

    0 5 10 15 20 25 30 35 40 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100

    fhyp = 0.1 fhyp = 0.2 fhyp = 0.3

    Fully oxygenated cells Fully oxygenated cells

    Head and Neck Cancer (/ = 10 Gy)

    Prostate Cancer (/ = 3.0 Gy)

    Dose per fraction (Gy) to yield equivalent tumor BED under normoxic conditions 3.2 2.7 2.4 2.2 2.03.84.97.518.31.75.3 3.9 3.18.6 2.6 2.2 1.923.8

    1 1

      2 2(1 ) ( ) exp / ( ) / ( )hyp nR

    overall hyp hyp hyp A Aa S f S f f r HRF r d HRF r d dr             

  • S L I D E 12

    Hypoxia Imaging Clinical Trial at Yale: Methods

    Day 0 Mon

    FMISO #1 Tue

    Day 2 Wed

    FMISO #2 Thurs

    Day 4 Fri

    FMISO # 3

    SBRT fx #1 of 18Gy

    PET Scan

    PET Scan

    PET Scan

    0 120 150 240 min

    CT CT CT

    180 210

    • IRB-approved protocol to perform serial 18F-fluoromisonidazole (FMISO) PET imaging in early- stage NSCLC cancer patients undergoing SBRT

  • S L I D E 13

    Hypoxia Imaging at Yale: Patient characteristics

    Pt Age Gender Histologic diagnosis Stage (TNM)

    Tumor diameter

    (cm)

    Tumor location

    Ma