Precision Radiotherapy: Reduction in Radiation for Oropharyngeal Cancer in the 30 ROC Trial

Abstract

Background: Patients with human papillomavirus-related oropharyngeal cancers have excellent outcomes but experience clinically significant toxicities when treated with standard chemoradiotherapy (70 Gy). We hypothesized that functional imaging could identify patients who could be safely deescalated to 30 Gy of radiotherapy.

Methods: In 19 patients, pre- and intratreatment dynamic fluorine-18-labeled fluoromisonidazole positron emission tomography (PET) was used to assess tumor hypoxia. Patients without hypoxia at baseline or intratreatment received 30 Gy; patients with persistent hypoxia received 70 Gy. Neck dissection was performed at 4 months in deescalated patients to assess pathologic response. Magnetic resonance imaging (weekly), circulating plasma cell-free DNA, RNA-sequencing, and whole-genome sequencing (WGS) were performed to identify potential molecular determinants of response. Samples from an independent prospective study were obtained to reproduce molecular findings. All statistical tests were 2-sided.

Results: Fifteen of 19 patients had no hypoxia on baseline PET or resolution on intratreatment PET and were deescalated to 30 Gy. Of these 15 patients, 11 had a pathologic complete response. Two-year locoregional control and overall survival were 94.4% (95% confidence interval = 84.4% to 100%) and 94.7% (95% confidence interval = 85.2% to 100%), respectively. No acute grade 3 radiation-related toxicities were observed. Microenvironmental features on serial imaging correlated better with pathologic response than tumor burden metrics or circulating plasma cell-free DNA. A WGS-based DNA repair defect was associated with response (P = .02) and was reproduced in an independent cohort (P = .03).

Conclusions: Deescalation of radiotherapy to 30 Gy on the basis of intratreatment hypoxia imaging was feasible, safe, and associated with minimal toxicity. A DNA repair defect identified by WGS was predictive of response. Intratherapy personalization of chemoradiotherapy may facilitate marked deescalation of radiotherapy.

Figure 1.

Precision radiotherapy (RT) and longitudinal imaging. Top) Study schema and results of pre- and intratreatment fluorine-18-labeled fluoromisonidazole positron emission tomography (¹⁸F-FMISO PET) scans (see also Supplementary Figure 1, available online). One patient was removed from the study before his intratreatment scan because of an unrelated medical illness and received the standard 70 Gy of chemo-RT. Bottom) In the illustrated case, intratreatment hypoxia resolved after the 10th fraction of treatment. Intratreatment ¹⁸F-FMISO PET demonstrates resolution of hypoxia. Weekly on-therapy and monthly posttreatment monitoring of therapy with T2-weighted magnetic resonance imaging was performed. Increasing T2-hyperintensity during therapy is consistent with solid tumor turning into fluid, although the anatomic size of the gross nodal disease did not decrease. Posttherapy, the involved lymph node slowly regressed (selected images shown). Posttreatment PET and computed tomography (CT) demonstrated resolution of FDG avid disease, and subsequent neck dissection demonstrated complete pathologic response. *Note, one patient was removed from study due before his intra-treatment scan due to an unrelated medical illness, and was analyzed in standard 70Gy Chemo-rt group.

Figure 2.

Pathologic results and imaging correlates. A, top) Pathologic response (percentage of viable tumor, 100%) at the time of neck dissection is shown in 15 patients who received 30 Gy. Note that 2 of the 4 patients with residual disease had clinically significant abnormal histologic appearance with unclear viability (see also Supplementary Figure 4, available online). Four of the 19 patients who were not deescalated did not undergo a neck dissection, and all remain without evidence of disease. Bottom) Compartmental analysis of hypoxia levels (k₃) in cases, pretherapy, and intratherapy is shown (see also Supplementary Figure 2, available online). Note that tumors with no pretherapy hypoxia did not undergo intratherapy scans, and all of these tumors had a pathologic complete response (pCR) (see 2B). Tumor permeability and perfusion (K^trans on dynamic contrast enhanced-magnetic resonance imaging [MRI]) was numerically higher in patients with a pCR (P = .08; see also Figure 3A). Relative kurtosis (a surrogate for tumor microstructure on diffusion-weighted–MRI) changed more rapidly in patients with CR in the first week intratreatment (P = .01; see also Figure 3C). B) Pathologic status by baseline fluorine-18-labeled fluoromisonidazole (¹⁸F-FMISO) PET scan. Zero of 6 patients with normoxia at baseline had residual disease at the time of neck dissection compared with 4 of 9 patients with baseline hypoxia (P = .1, Fisher’s exact test). C) k₃ values at baseline for patients with pCR tended to be lower, although this was not statistically significant (P = .11, t test). D) For the 9 patients with baseline hypoxia, compartmental analysis revealed a difference in K₁ between those with residual disease and those with complete response (P = .03, t test).

Figure 3.

Weekly magnetic resonance imaging (MRI) changes and pathologic response. A) Pretreatment dynamic contrast enhanced-MRI–derived quantitative K^trans (min⁻¹, a surrogate for tumor perfusion and permeability) demonstrated that K^trans is lower in patients with residual disease (P = .08; Wilcoxon rank sum test). K^trans estimated from a standard Tofts model was incorporated into the shutter speed model. B) Tumor volume, determined from T₂-weighted images, gradually decreased during therapy without a statistically significant difference between those with complete pathologic response and those with residual disease. C) Relative kurtosis (a surrogate for tumor microstructure) changed more rapidly in patients with pathologic complete response immediately during therapy and remained different from those with residual disease (P = .01, P = .08, P = .03, and P = .007 for week 1, 2, 3, and posttherapy, respectively; all Wilcoxon rank sum test). D) Changes in apparent diffusion coefficient (a surrogate for tumor or tissue cellularity) were not statistically significantly different between pathologic response groups until the end of therapy (P = .17, P = .09, P = .69, and P < .001 for week 1, 2, 3, and posttherapy, respectively; all Wilcoxon rank sum test).

Figure 4.

Whole-genome sequencing (WGS) analysis of the Memorial Sloan Kettering Cancer Center (MSKCC) trial. A) Tumor mutation burden (TMB, mutation/Mb), “genome altered” (percentage of the tumor genome not in a diploid state), genes commonly mutated in human papillomavirus (HPV)-related oropharyngeal cancer, and mutational signatures in the 17 tumor-normal pairs subjected to WGS. CD4 results of T-cell score and HPV Class 1 class are derived from RNA-sequencing analysis. HPV Class 1 class is a single-sample Gene Set Enrichment Analysis of a recently described, poor prognosis, HPV-related expression program (41). HPV subtyping of each case is presented in the annotation bar along with whether the virus is integrated into the host genome. B) Clonal reconstructions (left) of a responder (MSK16; Figure 4, B [top]) and nonresponder (MSK20; Figure 4, B [bottom]) to low-dose radiotherapy, with clones shown as numbered circular nodes (Supplementary Methods, available online). Circular cells labeled “N” represent the initial normal cell from which the tumor derived, with the x-axis representing increasing accumulation of mutations. The cellular prevalence of each clone is represented by the height of its corresponding polygon. Along the y-axis, with maximal frequency limited by tumor purity. Circos plots (Figure 4, B, right) of both cases demonstrate differences in mutational signatures and structural events (see main text). Tracks are organized from outside to inside according to legend. C) The proportion of deletions with microhomology is lower among patients who did not respond to low-dose chemoradiotherapy (P = .02, Wilcoxon rank sum test).

Figure 5.

Whole-genome sequencing analysis of the Mayo Clinic cohort. A) Mutation burden (mutation/Mb), “genome altered” (percentage of the tumor genome not in a diploid state), genes commonly mutated in human papillomavirus (HPV)-related oropharyngeal cancer, and mutational signatures in the 19 tumor-normal pairs from the Mayo Clinic Cohort. CD4 T-cell score is derived from immune deconvolution of RNA sequencing. HPV subtyping of each case is presented in the annotation bar along with whether the virus is integrated into the host genome. B) The proportion of deletions with microhomology is lower in patients who did not respond to low-dose postoperative chemo-radiotherapy (P = .03, Wilcoxon rank sum test). NED = no evidence of disease. C) Clonal reconstructions of a responding and nonresponding case from Mayo Clinic, with clones shown as numbered circular nodes (Supplementary Materials, available online). Circular cells labeled “N” represent the initial normal cell from which the tumor derived, with the x-axis representing increasing accumulation of mutations. The cellular prevalence of each clone is represented by the height of its corresponding polygon, along the y-axis, with maximal frequency limited by tumor purity.