Validation of Immunoscore for prognostic stratification in HPV-associated oropharyngeal cancer: An international multicenter study.

Abstract

6077 Background: Treatment optimization in HPV-associated oropharyngeal cancer (OPSCC) remains challenging due to the limited results of de-escalation trials. Existing patient selection criteria, mainly based on smoking history and TNM classification are insufficient and highlight the urgent need for standardized prognostic biomarkers. Herein, we present the first validation of the Immunoscore (IS) as a prognostic stratification tool in HPV-associated OPSCC. Methods: A cohort of 191 HPV-associated (p16⁺ and HPV DNA/RNA⁺) OPSCC patients treated between 2015–2024 was analyzed, including a French training cohort ( N = 48) and three independent validation cohorts: a French retrospective monocentric ( N = 48), a French prospective multicenter ( N = 50) and a US retrospective multicenter cohort ( N = 45). IS, an IHC-based standardized clinical digital pathology assay, quantifies CD3⁺ and CD8⁺ cell densities in tumor cores and invasive margins of FFPE sections. IS cut-offs were defined using the 25 th percentile of immune cell density in the training cohort and subsequently validated across all cohorts. Associations with disease-free survival (DFS), time to recurrence (TTR), and overall survival (OS) were assessed, along with immune profiling by 3′RNA-seq and sequential immunofluorescence. Results: Median age 65; 80% male; 74% smokers; 66% T1-2; 82% N0-1 (AJCC 8 th ). Treatments included surgery only (9%), radiotherapy ± chemotherapy (27%) and surgery + radiotherapy ± chemotherapy (64%). 52.4% were IS-High ( N = 100) and 47.6% IS-Low ( N = 91). IS-High patients showed significantly improved DFS, consistently across the training and validation cohorts 1-3 (log-rank P = 0.0004, 0.003, 0.006, and 0.001, respectively). Multivariable analysis identified IS-Low as the strongest independent risk factor for DFS (HR 9.27; 95% CI: 4.14-20.76; P < 0.001), outperforming smoking status, T/N stage, and treatment modality. The model combining IS with clinical factors showed higher predictive accuracy for DFS (C-index 0.82) than clinical variables alone (0.70; P < 0.0001). Similar strong prognostic value of IS was observed for TTR (HR 7.64; 95% CI: 3.37-17.33; P < 0.001) and OS (HR 7.26; 95% CI: 2.77-18.99; P < 0.001). IS-High tumors showed enrichment of lymphoid and myeloid immune cell populations, contrasting with immune-poor signatures in IS-Low tumors (all P < 0.05). Conclusions: IS is a robust biomarker that outperforms standard clinical variables in both prognostic and predictive accuracy. The enriched cytotoxic immune infiltrate in IS-High tumors explains favorable outcomes and supports their potential suitability for treatment de-escalation. Prospective validation in future trials is warranted.