Zevallos, Jose

Abstract Purpose: Relapse is a major cause of failure in human papillomavirus–independent head and neck squamous cell carcinoma. Clinicopathologic criteria for adjuvant treatment remain imprecise and have not changed for decades. We investigated whether ctDNA in lymphatic exudate collected via surgical drains (“lymph”) 24 hours after surgery identified molecular residual disease (MRD) and compared its performance with time-matched plasma. Experimental Design: Using an ultra-sensitive tumor-informed sequencing approach, tumor variants were called in lymph and plasma to classify patients as ctDNA positive or ctDNA negative, trained in an initial cohort of 36 patients, and replicated in an independent cohort of 37 patients. Progression-free survival was compared in ctDNA(+) versus ctDNA(–) patients. Results: Lymph identified MRD in two independent multi-site cohorts (initial cohort sensitivity = 76% and specificity = 63%; log-rank P = 0.01; replication cohort sensitivity = 65% and specificity = 70%; log-rank P = 0.04). Lymph performance was enhanced in locoregional relapse (sensitivity = 78% and specificity = 67%; log-rank P = 0.0004) and generalized to early-stage patients. Analysis of matched plasma collected at this early timepoint was not predictive of recurrence (sensitivity = 35% and specificity = 72%; log-rank P = 0.7). In patients with intermediate-risk pathology, lymph ctDNA was associated with recurrence (sensitivity = 88% and specificity = 67%; log-rank P = 0.0008), suggesting an opportunity for improved stratification of patients who may benefit from additional adjuvant treatment. Conclusions: Postoperative lymph is a novel, proximal, and early source of MRD with the potential to introduce more precision into adjuvant therapy decision-making and improve outcomes, especially for patients with intermediate-risk human papillomavirus–independent head and neck squamous cell carcinoma.

Abstract Purpose: NFE2L2/KEAP1/CUL3 mutations have been validated for radioresistance in cell-based assays and animal models. However, clinical validation of these biomarkers has been challenging because of multimodality treatment regimens. This study aims to investigate the association between NFE2L2/KEAP1/CUL3 mutations and patient outcomes, including local failure, locoregional failure, disease-free survival (DFS), and overall survival, using samples from a phase III trial in which patients were treated with radiation monotherapy at two controlled doses. Patients and Methods: We investigated NFE2L2/KEAP1/CUL3 mutations in 250 randomized patients with T2N0 glottic squamous cell carcinoma receiving definitive radiotherapy in the NRG/RTOG 9512 trial. A total of 119 patients had available biospecimens that were subjected to amplicon-based next-generation sequencing to assess for the presence of NFE2L2/KEAP1/CUL3 mutations without regard to outcomes. Mutations in NFE2L2/KEAP1/CUL3 were assessed blinded to clinical outcomes. Cox models (two-sided α = 0.05) were used to evaluate the association with clinical outcomes, performed by an independent statistical team. Results: Nineteen of 119 patients (16.0%) had NFE2L2/KEAP1/CUL3 mutations. Patient, treatment, and tumor characteristics were similar between those with and without mutations. Patients with mutation compared with those without had significantly more local failure [HR = 3.50; 95% confidence interval (CI), 1.56–7.89; P = 0.0025] and locoregional failure (HR = 3.80; 95% CI, 1.80–8.03; P = 0.0005). DFS was significantly worse for the mutated compared with the nonmutated group in the first 2 years (HR = 2.88; 95% CI, 1.46–5.66; P = 0.0022). The median DFS was shorter in the mutation group (10.3 months) versus those with intact NFE2L2/KEAP1/CUL3 (4.2 years). Conclusions: NFE2L2/KEAP1/CUL3 mutations may predict radiation treatment failure in T2N0 glottic cancer. See related commentary by Rao, p. 1563