How Are We Going to Define RT Targets in 2030? …it will be very different from now! Søren M Bentzen, Ph.D., D.Sc. Departments of Human Oncology; Medical Physics; Biostatistics and Medical Informatics, UW Paul P. Carbone Comprehensive Cancer Center, Madison, Wisconsin, USA bentzen@humonc.wisc.edu /SMB 06/10 The dose-volume trade-off IGRT: The bad old days…. Geographical miss Dose-dependent failure DOSE HOW MUCH? TY CI XI O T VOLUME /SMB WHERE? /SMB Target volume SELECTION in HNSCC Intensity versus volume T2N1 tonsil cancer rS=0.55 2P=0.015 Hong et al. ASTRO (2004) /SMB 06/10 Chemotherapy and target volume Bentzen et al. ASTRO (2005) /SMB Bentzen et al. ASTRO (2005) /SMB Auto-delineation from 18F-FDG PET van Bardwijk et al. IJROBP 68: 771 (2007) /SMB Auto-delineation from 18F-FDG PET Microscopic extension in stage I NSCLC Reduced inter-observer variability with observeredited auto-delineation compared to CT in 33 NSCLC cases (5 observers) van Bardwijk et al. IJROBP 68: 771 (2007) Lower inter-observer variability with contrastenhanced CT than with FDG-PET in 10 HNSCC cases (8 observers) Breen et al. IJROBP 68: 763 (2007) Thanks to Noah Arvidson Reproducibly wrong or reproducibly right? /SMB 06/10 From CTV to RTV GTVPET INF: Isolated nodal failure after chemo-radiation therapy for lung cancer Histology N INF NSCLC 44 1 2% 10% De Ruysscher 2005 LD-SCLC 27 3 11% 29% De Ruysscher 2006 11% van Loon 2010 LC-SCLC 60 2 Crude Upper prop. 95% CL 3% /SMB Grills et al. IJROBP 69: 334, 2007 Lung cancer: Irradiating PET+ nodes only Is involved node irradiation an alternative to elective mediastinal irradiation? Distance from ME Coverage GTV (mm) (%) 1 0 2 0 3 4 4 13 5 25 6 32 7 42 8 52 9 68 10 78 11 84 12 86 13 95 14 97 15 100 Abbreviations: GTV = gross tumor volume; ME = microscopic extensions Real Target Volume Clinical Target Volume Reference /SMB 06/10 Planning Target Volume Bentzen IJROBP 58: 320 (2004) /SMB The radiotherapeutic blind spot DRUGS 20 Gy 35 Gy Cancer stem cells? 70 Gy CT/MRI RT-PCR PET Log # cells in mass /SMB 06/10 /SMB Dirks Nature 444: 687 (2006) Incidence of pN+ nodes in oral cavity SCC Patients undergoing prophylactic radical neck dissection Site No. pts. I II III IV V Tongue 58 14% 19% 16% 3% 0% Floor of mouth 57 16% 12% 7% 2% 0% Gum 52 27% 21% 6% 4% 2% Retromolar trigone 16 19% 12% 6% 6% 0% Cheek 9 44% 11% 0% 0% 0% 192 20% 17% 9% 3% 1% Total http://www.rtog.org/pdf_document/atlasneck_CTV.pdf /SMB Shah et al., Cancer 66: 109 (1990) /SMB 06/10 N0 63% T N+ 37% N0: N(+) or N(-) ? CHART HNSCC Trial QUESTION: Can we select patients for elective irradiation? PROBLEM: Low incidence of events → low statistical power SOLUTION: Analyse molecular risk factors in N+ v N0 & apply to N0 patients ANSWER: Ki-67 pattern, p53 index, p53 staining intensity all associated with N+ Bentzen, Wilson, Saunders (submitted) /SMB Modeled failure-specific rates in HNSCC 2-year distant relapse rate Management of the N0 neck in HNSCC 0.8 309 head & neck SCC 0.6 8% 0.4 0.2 38% 54% 0 0 0.2 0.4 0.6 0.8 2-year local relapse rate Ataman et al. EJC 40: 2738 (2004) Bentzen Sem Rad Onc 8: 75 (2008) /SMB Target volume selection - efficacy:toxicity BUT…. If the benefit:risk is high SUPERSIZE IT! treatment failure after RT is about biology as much as about geometry! If the benefit:risk is low GO LEAN! /SMB 06/10 /SMB 06/10 Dose-painting by numbers Dose-painting by numbers 0.18 0.16 Frequency 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0.0423 0.889 1.73 2.58 3.43 4.27 5.12 5.96 61Cu-ATSM 6.81 SUV Normo-oxic 70 Anoxic 75 80 85 85 Dose in 35 F (Gy) Bentzen Lancet Oncology 6: 112 (2005) /SMB Target selection and delineation 2030 Think for a moment where we were 20 years ago! A number of new imaging and molecular finger-printing tools are at our disposal – the challenge is to demonstrate clinical benefit and cost-effectiveness The gold standard is clinical outcome • Do not develop surrogates of surrogates Choosing to irradiate a given voxel to a given dose involves a (local) trade-off between cost and benefit Imaging & radiation therapy: the perfect match! • But in many cases BIOLOGY > GEOMETRY /SMB 06/10 Bentzen Lancet Oncology 6: 112 (2005) Bowen et al. (in preparation) /SMB