Identification of Nectin Family Interactive Gene Panel and Stratification of Clinical Outcomes in Patients with Pancreatic Cancer

Abstract

Background: Although patient-risk stratification is important for selecting individualized treatment for pancreatic ductal adenocarcinoma (PDAC), predicting the oncologic outcomes after surgery remains a challenge. In this study, we identified a nectin family gene panel (NFGP) that can accurately stratify oncologic outcomes in patients with PDAC.

Study design: Comprehensive analysis of the expression of 9 nectin family genes identified the NFGP, which was assessed for predictive performance in 2 independent public cohorts (The Cancer Genome Atlas [TCGA] n = 176; International Cancer Genome Consortium [ICGC] n = 89). It was subsequently trained and validated for the in-house training cohort without neo-adjuvant therapy (NAT, n = 213) and the validation cohort with NAT (n = 307).

Results: Using the Cox regression model, NFGP derived from 9 nectin family genes accurately stratified overall survival (OS) in TCGA (p = 0.038) and ICGC (p = 0.005). We subsequently optimized NFGP, which robustly discriminated postoperative prognosis, OS (p = 0.014) and relapse-free survival ([RFS] p = 0.006) in the training cohort. The NFGP was successfully validated in an independent validation cohort (OS: p < 0.001; RFS: p = 0.004). Multivariate analysis demonstrated the NFGP was an independent prognostic factor for OS and RFS in the training (p = 0.028 and 0.008, respectively) and validation (p < 0.001 and 0.013, respectively) cohorts. The subcohort analyses showed that the predictive performance of NFGP is applicable to the patients' subcohort according to resectability or adjuvant therapy status. Additionally, a combination model of NFGP score and CA19-9 level emerged with improved accuracy for predicting prognosis.

Conclusions: This study established the predictive significance of NFGP for oncologic outcomes after surgery in PDAC. Our data demonstrate its clinical impact as a potent biomarker for optimal patient selection for individualized treatment strategies.