The tumor inflammation signature (TIS) is associated with anti-PD-1 treatment benefit in the CERTIM pan-cancer cohort

Background The 18-gene tumor inflammation signature (TIS) is a clinical research assay that enriches for clinical benefit to immune checkpoint blockade. We evaluated its ability to predict clinical benefit of immunotherapy in cancer patients treated with PD-1 checkpoint inhibitors in routine clinical care. Methods The CERTIM cohort is a prospective cohort which includes patients receiving immune checkpoint inhibitors in Cochin University hospital. RNA extracted from 58 archival formalin fixed paraffin embedded tumor blocks (including 38 lung cancers, 5 melanomas, 10 renal carcinomas, 4 urothelial carcinomas and 1 colon carcinoma) was hybridized to a beta version of the NanoString (R) PanCancer IO360 (TM) CodeSet using nCounter (R) technology. Gene expression signatures were correlated with tumor responses (by RECIST criteria) and overall survival. PD-L1 immunostaining on tumor cells was assessed in 37 non-small cell lung cancer (NSCLC) samples and tumor mutational burden (TMB) measured by whole exome sequencing in 19 of these. Results TIS scores were significantly associated with complete or partial response to anti-PD-1 treatment in the whole cohort (odds ratio = 2.64, 95% CI [1.4; 6.0], p = 0.008), as well as in the NSCLC population (odds ratio = 3.27, 95% CI [1.2; 11.6], p = 0.03). Patients whose tumor had a high TIS score (upper tertile) showed prolonged overall survival compared to patients whose tumor had lower TIS scores, both in the whole cohort (hazard ratio = 0.37, 95% CI [0.18, 0.76], p = 0.005) and in the NSCLC population (hazard ratio = 0.36, 95% CI [0.14, 0.90], p = 0.02). In the latter, the TIS score was independent from either PD-L1 staining on tumor cells (spearman coefficient 0.2) and TMB (spearman coefficient - 0.2). Conclusions These results indicate that validated gene expression assay measuring the level of tumor microenvironment inflammation such as TIS, are accurate and independent predictive biomarkers and can be easily implemented in the clinical practice.