Abstract BACKGROUND: Pancreatic cancer is one of the most aggressive cancers with a 5-year survival of less than 10%. Surgery is the only cure for pancreatic cancer but the majority of patients present with metastatic disease at the time of diagnosis. Systemic chemotherapy remains the primary treatment option. Given the insidious clinical presentation and poor treatment response, there is an urgent need to discover a predictive biomarker for appropriate patient selection. We hypothesized that the 4-gene score, which reflects tumor cell proliferation, is both a prognostic and predictive biomarker for pancreatic cancer. METHODS: The 4-gene score is derived from tumor expression of DOK4, HCCS, PGF, and SHCBP1 genes, which were identified based on differential mRNA expression analysis of a human breast cell line, its metastatic variant cells, and clinical outcome data of breast cancer patient cohorts, as previously reported. A total of 954 pancreatic cancer patients were analyzed for both discovery and validation of the 4-gene score from publicly available datasets to investigate the relationship between the score with clinical features such as metastasis, cancer aggressiveness, immune cell infiltration, patient survival, and resectability. RESULTS: We found that the 4-gene score correlation in pancreatic cancer was higher than in breast cancer cohorts, specifically in clinically aggressive parameters such as pathological grade and MKI67 expression. Also, the score in metastatic tumor cohorts was higher than in primary cancer cohorts (p < 0.001). These findings were mirrored in gene set analysis, in which the score high tumors enriched cell proliferation-related gene sets such as E2F targets, G2M checkpoint, MYC targets v1 and v2, mitotic spindle, and metastasis-related gene sets in three cohorts (TCGA, GSE62452, and GSE57495). A high score was associated with increased mutation rates in KRAS and CDKN2A genes (p = 0.006 and 0.044) in the TCGA cohort, as well as enhancement of KRAS, CDKN2A, p53, TGF-β, and E2F pathways in three cohorts. Furthermore, a high score tumor had less infiltration of CD8+ T cells but was associated with high cytolytic activity and IFN-γ scores, which reflect overall cancer immunity. The high score was correlated with the level of drug sensitivity area under the curve of Sorafenib (r = 0.648) in primary cancer, Paclitaxel (r = 0.624) and Doxorubicin (r = 0.575) in metastatic cancer. Finally, a high 4-gene score was significantly associated with worse clinical outcomes in three cohorts (p = 0.044, 0.048, and 0.022), a low rate of R0 resection (p = 0.013), and worse survival after R1/2 resection (p = 0.027). CONCLUSION: The 4-gene score identified poor survival in pancreatic cancer and has potential as a predictive biomarker for R0 resection and treatment response in metastatic pancreatic cancer. Evaluating the degree of the 4-gene score could be a valuable potential prognostic tool. Citation Format: Masanori Oshi, Lan Le, Yoshihisa Tokumaru, Ankit Patel, Ryusei Matsuyama, Itaru Endo, Li Yan, Matthew H.G. Katz, Kazuaki Takabe. A novel 4-gene score predict patient survival as well as pathologically complete (R0) resection in pancreatic cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 404.
Read full abstract