Abstract 5877: Genetic analysis of metachronous pancreatic cancers

[Introduction] Early detection of pancreatic cancer is a key to curable surgery. However, many patients are diagnosed in an advanced stage and even in those cases who had a surgical resection of early stage tumors, metachronous pancreatic cancer in residual pancreas prevents cure. Thus, it is important to control metachronous tumors to improve clinical outcomes, whose pathogenesis is poorly understood. In this study, we aimed to reveal the origin of metachronous pancreatic cancers using an unbiased detection of somatic mutations in primary and metachronous cancers as well as adjacent precursor lesions. [Methods] Serially obtained formalin-fixed paraffin-embedded surgical specimens from 8 patient who had undergone curative surgery for an early stage pancreatic cancer were subjected to laser microdissection for the enrichment of tumor and PanIN lesions, from which DNA was extracted and analyzed for somatic mutations of each lesion by whole-exome sequencing (WES) with matched normal DNA. Based on shared and private mutations across different samples, we interrogated history of clonal evolution of these lesions. [Results] Pathology for primary cancer were margin-negative in all patients. The median interval between the initial and second surgery was 29.8 months (23 - 64.4 months). We analyzed a median of two cancer and one PanIN samples (0-5), from which we identified a median number of 45(range: 26-69) and 20 (14-42) somatic mutations using WES, respectively. None of the patients have known pathogenic germline variants. All samples had one or more driver mutations. In each patient, all driver mutations and more than half of passenger mutations were shared between primary and metachronous cancer samples, suggesting that those metachronous tumors are evolutionally closely related to the primary cancer. Negative pathology of the margins at the initial surgery suggested that those metachronous lesions originated from distant dissemination or metastasis, rather than contiguous, intraductal invasion. By contrast, none of the mutations other than hotspot KRAS mutations were shared between precursor lesions and cancer samples. Despite multiple independent clones in the precancerous lesion, all metachronous tumors originated from the primary lesions in this study. [Conclusions] Our study suggests that even early pancreatic cancer might be disseminated within the pancreas and contribute to metachronous cancers, suggesting the importance of close monitoring of the recur