Comprehensive Genomic Analysis Stratified by KRAS Status in Patients with Pancreatic Adenocarcinoma and Its Prognostic Significance

Abstract

KRAS mutations (MUT) are one of the major drivers in pancreatic ductal adenocarcinoma (PDA) with over 90% of patients having alterations. However, the genetic landscape of PDA based on KRAS status is not well studied. The aim of this study is to investigate genomic alterations based on KRAS status and to identify driver mutations in patients with KRAS wild type (WT). Next-generation sequencing with 324 pre-specified genes was performed on patients with histologically confirmed PDA. The landscape of somatic mutations was stratified by KRAS status. Outcomes of interest included overall survival (OS), local failure (LF) following radiotherapy, time to metastasis, and CA 19-9 level. All outcomes were stratified by KRAS status. Genetic alterations exclusive to patients with KRAS WT were analyzed. OS was calculated using the Kaplan-Meier estimates and log-rank test, and LF was measured using cumulative incidence. A multivariate cox-regression analysis (MVA) was performed to identify prognostic factors for survival. Gene ontology analysis of KRAS WT exclusive genes was performed via DAVID (Database for Annotation, Visualization, and Integrated Discovery). A total of 272 patients with metastatic PDA were included. The median age at diagnosis was 65.4 (range, 26.77-83.21) years. The median follow-up was 15.7 (0.06-136.7) months. 91% percent (n = 248) of patients were found to be KRAS MUT. The chi-square test showed that the primary tumor site (p = 0.027) and perineural invasion (p = 0.006) were associated with KRAS status. The median CA 19-9 was 143.6 (15.5-27996) U/ml and 341.15 (0-100000) U/ml for KRAS WT and KRAS MUT, respectively (p = 0.23). The median OS for KRAS WT and MUT was 31.4 (95% CI 25.1-NA) and 13.3 (95% CI 11.3-14.5) months, respectively (p = 0.0008). In patients with metachronous recurrence, the median time to metastasis was 22.4 (1.37-52.97) months and 12.8 (1.8-84.97) months for KRAS WT and KRAS MUT, respectively (p = 0.057). KRAS MUT types including G12 and Q21 were not associated with OS (p = 0.58). On MVA, including KRAS status, age, sex, and metastatic types (de novo vs. metachronous), only KRAS mutation was associated with worse OS [HR 2.64; 95% CI 1.53 to 4.56; p = 0.0004]. In patients treated with radiation, the LF rate at 12 months was 12.5% in patients with KRAS WT and 33.8% in KRAS MUT (p = 0.13). Heatmap analysis identified that RAD50, ALK, BCORL1, BRAF, CDC73, FAM123B, NF2, ERBB4, and ERCC4 were exclusively mutated in patients with KRAS WT. These genes were enriched in pathways associated with tyrosine kinase catalytic domain activity, ubiquitination, and nuclear localization signal. This study identified driver mutations in patients with KRAS WT. KRAS status was associated with pathologic features and disease prognosis after treatment. Further study leveraging more powered cohorts is warranted.