Abstract 4546: The tumor inflammation signature is predictive of anti-PD1 treatment benefit in the CERTIM pan-cancer cohort

Purpose of the study : The 18-gene Tumor Inflammation Signature (TIS) (Ayers et al, J Clin Invest 2017; 127:2930) is a clinical research assay that enriches for clinical response to immune checkpoint blockade. In this work, the performance of the TIS was evaluated on standard tumor specimens for its ability to predict clinical benefit of immunotherapy in a cohort of cancer patients treated with PD1 checkpoint inhibitors in routine clinical care. Experimental procedures : Consecutive metastatic cancer patients treated with anti-PD1 in the outpatient monocentric CERTIM cohort with available formalin fixed paraffin embedded tumor specimen were selected for analysis. RNA were extracted from tissue samples with High Pure FFPET RNA Isolation Kit (Roche) and 25-100 ng RNA was hybridized to a pilot version of the NanoString® PanCancer Immuno-oncology 360 Panel. The association between individual gene expression (GE), TIS and other GE signatures with outcome (response to therapy and overall survival) was assessed with logistic resgression model, LogRank tests and Cox model. Results : A total of 69 patients with non-small cell lung carcinoma (NSCLC, n=42), small cell lung carcinoma (n=2), melanoma (n=9), renal cell carcinoma (n=10), urothelial carcinoma (n=5) or colorectal carcinoma (n=1), who had received nivolumab (n=59) or pembrolizumab (n=10) were analyzed. All samples but one SCLC could be successfully analyzed, demonstrating the feasibility of the testing even in small biopsy samples with little available RNA (>30 ng). In the whole cohort, clinical response (PR or CR versus SD or progression) was significantly associated with TIS score [OR=2.002, 95%CI=(1.123,3.977), p =0.029] as well as with the individual expression of PDL1, CTLA4, IDO1 or GE scores characterizing Th1, cytotoxic, NK cells infiltration, immuno-proteasome or antigen processing machinery. Receiver operating characteristics for response status demonstrated the high discriminatory ability of the TIS signature in this unselected series of patients (area under the ROC curve 0.71). Patients with a high TIS score (> 9.39) showed a longer OS (median OS 20.2 months) compared to those with lower scores (median OS 5.6 months, p value = 0.0017). A high TIS score was also significantly associated with clinical response when the analysis was limited to the NSCLC (n=42, [OR=3.156, 95%CI=(1.168,11.455), p-value=0.042] - or showed a strong trend in lung adenocarcinoma (n=29 [OR=11.251, 95%CI=(0.962,131.534), p-value=0.054] patients. Conclusion : Due to costs and toxicity, it is mandatory to identify the patients who benefit of immune checkpoint inhibitors. Here we demonstrate that the Tumor Inflammation Signature, that reflects key processes involved in anti-tumor immune response, may be used in routine samples to predict clinical benefit and prolonged survival of cancer patients. Citation Format: diane damotte, jennifer arrondeau, pascaline boudou-rouquette, audrey lupo, jerome biton, hanane ouakrim, Marco Alifano, Gervais Claire, Audrey Bellesoeur, Nora Kramkimel, Angeline Duche, Franck Letourneur, Xing Ren, Pascale Morel, Alessandra Cesano, Sarah Warren, Francois Goldwasser, Karen Leroy. The tumor inflammation signature is predictive of anti-PD1 treatment benefit in the CERTIM pan-cancer cohort [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 4546.