Abstract 5682: Histomorphological and molecular immune classification of head and neck cancer and its relation to prognosis

Abstract

Abstract Recently, nivolumab and pembrolizumab have been approved as second-line treatment for recurrent and/or metastatic squamous cell carcinomas of the head and neck (SCCHN). However, while nivolumab improved response rate and overall survival in the CheckMate-141 trail and led to long-lasting responses, the overall response rate of 13% was limited. Therefore, biomarkers to select patients for immune therapy and to analyze and overcome mechanisms of resistance are urgently needed. To gain inside into the immunological status of SCCHN, we evaluated digital slides (HE stain) and molecular data of 528 primary tumors from TCGA. Tumor infiltrating lymphocytes (TILs) were evaluated separately intra-epithelial and in tumor stroma using a semi-quantitative method. The content of T cells, CD8+ T cells, cytotoxic lymphocytes, B lineage cells, NK cells, monocytic lineage cells, myeloid dendritic cells, neutrophils, fibroblasts and endothelial cells was estimated from RNA-seq data using the MCP-counter method. The T cell level was classified as high and low using a median cutoff point. As both intra-epithelial TILs and stromal TILs followed bimodal distributions, cutoff points were introduced leading to the flowing classification: Intra-epithelial TILs were present in 61% and absent in the remaining 39% of tumors. Levels of stromal TILs were high in 52% and low in the remaining 48% of tumors. Intra-epithelial and stromal TILs correlated highly significant (p=1.1E-06), but there were considerable proportions of tumors with absent intra-epithelial and high stromal TILs (13%) as well was with present intra-epithelial and low stromal TILs (21%). Intra-epithelial TILs correlated significantly with HPV status, but not with PD-L1 mRNA. Stromal TILs correlated significantly with PD-L1 mRNA, but not with HPV status. In univariate Kaplan-Meier analysis, intra-epithelial TILs correlated significantly with better DFS (p=0.009), but not with OS, while high stromal TILs correlated borderline significantly with better OS (p=0.071), but not with DFS. Tumors with many T cells had both significantly better DFS (p=0.018) and better OS (p<0.0001). PD-L1 mRNA expression correlated significantly with T cell content (R=0.45), but did not correlate significantly with DFS and OS. However, in a bivariate analysis of DFS including PD-L1 mRNA and T cells, PD-L1 was a significant negative prognostic maker, (HR per doubling=1.13, 1.01-1.25), while T cells were a strong positive prognostic maker (HR per doubling=0.74, 0.64-0.68). These results remained significant in bivariate analysis of OS and in a multivariate analysis including correction for clinical-pathological patients and tumors characteristics. The histopathological and molecular features describing the immune status of SCCHN investigated here should be further analyzed for their relevance in immune therapy patients. Citation Format: Jan Budczies, Mohamed Badr, Ingeborg Tinhofer, Carsten Denkert, Korinna Jöhrens. Histomorphological and molecular immune classification of head and neck cancer and its relation to prognosis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5682.