The role of proinflammatory immune (IFN-gamma) gene signature in predicting prognosis and response to chemotherapy in uterine carcinoma.

e14209 Background: Immunotherapy is promising option given low toxicity and potential durable response. In mismatch repair proficient endometrial and ovarian cancers, the reported response rate is ranging from 10-15% in recurrent setting. We need to better identify subset of patients who benefits from immunotherapy. Multigene immune signatures represent a robust means of capturing a complex, T cell–inflamed phenotype necessary for the clinical activity of PD-1–/PD-L1–directed monoclonal antibodies. IFN-γ is a key cytokine produced by activated T cells, as well as natural killer (NK) and NK T cells, in the tumor microenvironment. An immune-related IFN-γ 18-gene profile was derived through a cross-validated penalized regression modeling strategy to predict response to anti-PD1 therapy across 9 different tumor types. We want to test if this gene panel can also predict cancer outcome and response to chemotherapy. Methods: We used whole transcriptome sequencing of RNA matched tumor-normal samples from 38 high stage (Stage III and IV) uterine serous cancer patients. All patients received chemotherapy with platinum and taxanes. IFN-18 gene expression score was calculated by averaging the normalized and log transformed individual gene read counts. The optimized score cut off was selected to best separating the progression free survival. Then the cut off score was tested in The Cancer Genomic Atlas (TCGA) uterine and ovarian cancer RNAseq datasets. Results: The IFN score of 2.46 was determined based on 18-gene expression derived from 38 high-stage uterine serous cancer samples. Average age was 67 years (range: 56-82 years). Uterine serous cancer is known to be MSI stable. Patients with score higher than 2.46 showed significantly longer progression free survival (PFS – 57.6 months vs 15months, p = 0.002) and longer overall survival (73.1 months vs 51.1 months), not statistically significant given our small sample size, p = 0.13) compared to the patients with score lower than 2.46. Then this IFN based gene signature was then applied to TCGA 541 uterine cancer samples with RNAseq data. Similarly, this signature predicted significant improvement in both progression-free survival (p = 0.001) and overall survival (p = 0.005). Interestingly, this score cannot separate outcome for TCGA ovarian cancer cohort. Conclusions: Immune-related IFN-γ gene signature predicted prognosis and response to chemotherapy. We plan to assess if this signature will predict endometrial cancer patients who benefits from anti-PD1 therapy.