A personalized approach using hypoxia resolution to guide curative-intent radiation dose reduction to 30 Gy: A novel de-escalation paradigm for HPV-associated oropharynx cancers

Sloan-Kettering Inst Can Research, New York NY

Investigators

Linked publications & trials

Abstract

PROJECT ABSTRACT The central theme of this proposal is that a personalized de-escalated chemoradiation (CRT) will maintain efficacy while reducing toxicity versus the standard one-size-fits-all CRT to 70Gy . We completed 4 sequential de-escalation trials for early-stage (T0-2/N1-2c) HPV+ OPC where a hypoxia-specific biomarker, 18F- fluoromisonidazole (FMISO) imaged by positron emission tomography (PET), directed radiation to either 30Gy (FMISO negative scan)[“low-risk”] or to the standard 70Gy (FMISO positive scan). The excellent results led to a phase III FDA approved registration trial for early-stage HPV+ OPC randomizing patients to receive either FMISO PET-directed CRT or standard 70Gy CRT. The results of this trial will provide data for an application for FDA approval of FMISO PET as an imaging biomarker to guide RT for this disease. Given the success in early- stage patients, we postulate that FMISO PET-guided CRT can also personalize treatment for advanced-stage (T3-4/N3/M0) patients. If patients respond to induction chemotherapy with tumors down staged to <T2 & <N3 disease and they have a negative week 2 FMISO PET during CRT, de-escalation to 30Gy can occur without tumor control compromise. Indeed, a completed pilot study demonstrates its feasibility and provides the rationale to initiate a larger phase II validation trial. [Aim 1]. Although FMISO PET has enabled CRT de-escalation, not all centers have access to it. Intra-treatment tumor & microenvironmental changes that occur in response to CRT can be detected through longitudinal imaging with more widely accessible techniques. Therefore, we hypothesize that quantitative imaging biomarkers (QIBs) derived from dynamic contrast-enhanced & diffusion- weighted magnetic resonance imaging, and18F-fluorodeoxyglucose PET can predict a negative week 2 FMISO PET. We propose 30Gy CRT de-escalation by synergizing these QIBs that predict for a negative week 2 FMISO PET with week 2 circulating tumor deoxyribonucleic acid, another useful biomarker that monitors response. [Aim 2] In our trials, early-stage HPV+ OPC patients not eligible for 30Gy CRT (FMISO positive, ~30% of all patients) received 70Gy CRT. Current evidence suggests only 10% of all early-stage HPV+ OPC patients or 1/3 of the FMISO positive cohort require 70Gy (“high risk”). Therefore, de-escalation for the remaining 2/3 of the FMISO positive patients or 20% of all early-stage HPV+ OPC patients (“intermediate risk”) can further reduce toxicity. We will use our novel neural network-based clustered random forest advanced machine learning model that synthesizes all clinical/imaging information to select appropriate intermediate risk patients from the FMISO positive cohort with the goal of further de-escalation. Although FDG PET response, surgery, or induction chemotherapy can select HPV+ OPC patients to receive 45-54Gy, 50-63% of the patients complain of ≥grade 2 mucositis, even with 45Gy when 1.5Gy BID was used. Therefore, once we identified a group of intermediate risk patients, we propose to treat them with 46Gy CRT in standard 2Gy/fraction in a future clinical trial. The goal is to further reduce toxicity without compromise in tumor control compared to the standard 70Gy. [Aim 3]