Abstract
OBJECTIVE: Our study examined the association between molecular features and clinical results of pancreatic ductal adenocarcinoma (PDAC) patients, aiming to explore the genomic determinants of the recurrence and prognosis of PDAC after surgical removal. METHODS: This retrospective study analyzed 181 PDAC patients who received pancreatectomy and adjuvant chemotherapy, with 67 patients in the training set. An internal validation set of 48 patients and an external validation set of 66 patients were used to validate the result. Comprehensive genomic profiling was performed on formalin-fixed paraffin-embedded (FFPE) tumor specimens to determine genomic features using the designed cancer-related gene panel based on next-generation sequencing (NGS). RESULTS: Significant differences were identified between the late recurrence (LR) group and early recurrence (ER) group in tumor copy number instability (CNI) levels. Next, the utility of low CNI (the middle and lowest tertile) with regard to predicting LR was confirmed in the training, internal, and external validation sets. Further univariate and multivariate analyses revealed that CNI was an independent predictive and prognostic biomarker, and had higher predictive accuracy for LR than CA19-9 level, pathological stage, tumor size, and age. In addition, CNI combined with lymph node (LN) metastasis status could provide a more accurate model for predicting LR of PDAC. CONCLUSION: We discovered and validated the association between CNI and clinical outcome in 181 patients with resectable PDAC, demonstrating the utility of lower tumor CNI levels as biomarkers of postoperative LR and favorable prognosis. Moreover, the combination of CNI and LN metastasis status elevated the predictive accuracy and illuminated strategies for patient stratification.