Tumor Hypoxia on 18F-fluoromisonidazole Positron Emission Tomography and Distant Metastasis From Head and Neck Squamous Cell Carcinoma.

Importance: Given high rates of locoregional control after definitive management of head and neck squamous cell carcinoma (HNSCC), better methods are needed to project distant metastasis (DM) risk. Tumor hypoxia on 18F-fluoromisonidazole (FMISO) positron emission tomography (PET) is associated with locoregional failure, but data demonstrating an association with DM are limited. Objective: To determine whether tumor hypoxia on FMISO PET is associated with DM risk after chemoradiotherapy (CRT) for HNSCC. Design, Setting, and Participants: This cohort study assessed patients with HNSCC enrolled in 2 prospective clinical trials at a single academic referral center from 2004 to 2021 in which participants received FMISO PET before and during CRT. Data analysis occurred from May 2023 to May 2024. Exposures: FMISO PET scans before and 1 to 2 weeks after starting CRT were evaluated for tumor hypoxia by nuclear medicine physicians. Main Outcomes and Measures: The primary outcome was DM, defined as biopsy-proven HNSCC outside the primary site and regional lymph nodes. Time to DM was modeled with competing risk regression, with death as a competing risk. Overall survival (OS) was assessed secondarily and modeled with Cox regression. Results: Among 281 patients (median [range] age at CRT, 58.7 [25.5-85.6] years; 251 male [89.3%]) included in this study, 242 (86.1%) had oropharyngeal primary cancer, and 266 (94.7%) had human papillomavirus-positive disease. Of all patients, 217 (77.2%) had T stage 1 or 2, and 231 patients (82.2%) had N stage 2b or less. De-escalated 30 Gy CRT was delivered to 144 patients (51.2%), and the remainder received standard 70 Gy CRT. On FMISO PET examination, 73 patients (26.0%) had hypoxia-negative disease before CRT, 138 patients (49.1%) had hypoxia-positive disease before CRT and then hypoxia-negative disease during CRT, and 70 patients (24.9%) persistently had hypoxia-positive disease before and during CRT. At a median (IQR) 58 (46-91) months of follow-up, 12 DM events and 22 deaths were observed. Persistent intratreatment hypoxia was associated with increased DM risk (hazard ratio, 3.51; 95% CI, 1.05-11.79; P = .04) and worse OS (hazard ratio, 2.66; 95% CI, 1.14-6.19; P = .02). No patients with hypoxia-negative disease before CRT experienced DM. Conclusions and Relevance: In this cohort study using pooled analysis of prospective nonrandomized clinical trials incorporating FMISO PET in the definitive management of HNSCC, persistent intratreatment hypoxia was associated with increased risk of DM and worse OS. Conversely, all patients with hypoxia-negative disease before treatment remained free of DM. These findings suggest that pretreatment and intratreatment FMISO PET results may serve as biomarkers for DM risk and aid in identifying candidates for escalated therapeutic strategies.