Correlation of degree of tumor immune infiltration and insertion-and-deletion (indel) burden with outcome on programmed death 1 (PD1) therapy in advanced renal cell cancer (RCC).

Martin H. Voss, Jacqueline L. Buros, Matthew D. Hellmann, Mark W. Ball, A. Ari Hakimi, Diana Miao, Claire A. Margolis, Christine E. Horak, Megan Wind‐Rotolo, Guillermo de Velasco,

Journal of Clinical Oncology · 2018 · 21 citations · 0 references

Concepts

Abstract

4518 Background: Mutation load and neoantigen burden predict benefit from anti-PD1 therapy across several cancers but are generally low in RCC. Pan-cancer comparisons highlight RCC for its enriched tumor immune infiltrate and a high proportion of indels, which may be immunogenic. We investigated these phenomena in anti-PD1 treated patients (pts) with comparison to pts receiving tyrosine kinase inhibitors (TKI). Methods: Whole exome sequencing (WES) was done for 77 tumors from 2 anti-PD1 treated pt cohorts; 25 pts provided RNAseq data. Overall mutation burden, indels, and predicted neoantigens were quantified. ESTIMATE and CIBERSORT algorithms were used to infer abundances of immune infiltrate and detailed immunophenotypes from RNAseq. We investigated correlation of molecular features with overall survival (OS) and durable clinical benefit (DCB = progression free survival > 6mo). A cohort of TKI-treated pts with no prior anti-PD1, previously analyzed as part of the TCGA, was used for comparison (35 with WES, 33 with RNAseq). Results: Higher overall immune infiltration correlated favorably with DCB in the PD1 cohort (stratified Mann Whitney test; p = 0.05) but not in TKI treated pts (p = 0.84). As the two most abundant infiltrating cell types, M2 macrophages (p < 0.001) but not CD8+ T cells (p = 0.22) associated with DCB from anti-PD1 therapy; neither associated with outcomes in TKI pts (p = 0.15, p = 0.76). Expectedly, mutation counts were low (median 68; inter-quartile range 45-93). While neither mutation nor neoantigen burden impacted on OS with anti-PD1 therapy (stratified Cox Proportional HR 1.02; p = 0.6; HR 1.01; p = 0.69, respectively), association with OS was significant for frameshift indel count (HR 0.85; p = 0.006). Correlation was similar but did not reach significance in TKI pts (HR 0.87; p = 0.07). Conclusions: In anti-PD1 treated RCC pts, baseline tumor immune infiltration, particularly by M2 macrophages, associated favorably with outcome. In contrast to other cancers, mutation and neoantigen load did not affect treatment outcomes; yet higher indel counts appeared to confer superior OS. Such associations were not seen for TKI treated pts.