Defining a long-term benefit phenotype to PD-1 inhibitors in recurrent/metastatic head and neck squamous cell carcinoma (R/M HNSCC): A multicentre retrospective study.

6050 Background: Limited data exists on predictors of durable benefit to PD-1 inhibitors in R/M HNSCC. We aimed to identify clinical characteristics associated with long-term benefit from anti-PD-1 therapy in R/M HNSCC. Methods: We reviewed clinical data from patients with R/M HNSCC treated with PD-1 inhibitors between December 2018 and August 2025, at Princess Margaret Cancer Centre (an academic hospital) or within the William Osler Healthy System (a community hospital). Eligible patients received pembrolizumab or nivolumab, either alone or with chemotherapy (for pembrolizumab), and had available tissue for planned biomarker studies. Patients were grouped by progression-free survival (PFS): (A) PFS ≥ 24 months (m), (B) PFS 12-24 m and (C) best response of progressive disease (PD). Best response was determined from treating physician documentation based on clinical and radiographic assessment. PFS was estimated with the Kaplan-Meier method. Univariate analysis used Kruskal-Wallis test for continuous variables and Fisher’s exact or Chi-square tests for categorical variables. Results: A total of 150 patients were included: 31 in group A (PFS ≥ 24 months), 28 in group B (PFS 12-24 months) and 91 in group C (PD as best response). We observed significantly more immune-related adverse events (irAEs) (61% vs 16%, OR 8.0, 95% CI 3.2-19.9, p<0.001) and corticosteroid use (32% vs 8.8%, OR 4.9, 95% CI 1.7-14.1, p= 0.002) in group A compared to group C. There was a trend toward a higher proportion of patients with lung-only metastases in group A compared to C (39% vs 21%, OR 2.4, 95% CI 1.0-5.8, p= 0.052). Similar associations were found for irAEs (p<0.001), corticosteroid use (p= 0.011) and lung-only metastases (p= 0.054) in group B compared with group C. Additionally, we found more active smokers (p=0.010) and ex-smokers (p= 0.036) in group B compared to group C, as well as more active smokers in group B compared to group A (p= 0.042), whereas smoking rates were similar in groups A and C. Interestingly, there was no significant difference in Charlson Comorbidity Index (CCI) scores, albumin levels or neutrophil-to-lymphocyte ratios (NLR) between the three groups. The mean CCI scores were 7.3, 7.3 and 7.1; albumin levels 39.1, 40.1 and 39.7 and NLR 6.8, 6.5 and 8.7 for groups A, B and C respectively. P16 status and PD-L1 CPS were distributed similarly between the groups, although PD-L1 testing was not available in 25% of patients. Conclusions: IrAEs, corticosteroids and lung-only metastases are associated with PFS ≥ 12 m in patients with R/M HNSCC treated with anti-PD-1 agents. Smoking is associated with PFS 12-24 m but not with PFS ≥ 24 m, suggesting a potential role in acquired resistance. Further molecular analysis in this population is underway to identify biomarkers of long-term benefit.