The spectrum of pathogenic germline variants in pancreatic cancer patients with multiple primary tumors.

Introduction: Approximately 10% of pancreatic ductal adenocarcinomas (PDAC) are considered hereditary due to pathogenic germline variants in cancer predisposition genes. It is crucial to recognize hereditary PDAC to ensure surveillance of affected individuals and their relatives, and to identify the underlying molecular defects for application of targeted therapies. Methods: The study cohort includes 207 patients with PDAC and additional primary tumors in personal history selected from the Dutch Pathology Archive. Germline DNA was analyzed using two customized targeted next-generation sequencing panels including 26 cancer predisposition genes. Identified variants were checked for pathogenicity. Results: The study includes 207 patients diagnosed with PDAC in combination with two other primary tumors (n = 51), breast (n = 51), colorectal (n = 41), prostate (n = 18), ovarian (n = 8), gastric (n = 5), endometrial (n = 4) carcinomas, and melanoma (n = 29). Germline sequencing identified pathogenic or likely pathogenic variants (PV/LPV) in 43/207 (20.8%) of the patients. Per combination, PV/LPV were detected in PDAC patients with ovarian (4/8), gastric (2/5), melanoma (9/29), breast (12/51), prostate (3/18), colorectal (6/41), and two other primary (7/51) cancers. The most frequently affected genes were ATM (14/43), BRCA2 (7/43; one patient with a concurrent PMS2 LPV), CDKN2A (7/43; 5/7 with Dutch founder p16-Leiden PV), and TP53 (3/43). Eight patients carried PV/LPV in BRCA1 (2/43), MSH6 (2/43), biallelic NTHL1 (2/43), and PALB2 (2/43) genes. Four patients each had a single PV/LPV in BRIP1 (1/43), CDH1 (1/43), CHEK2 (1/43), and RAD51C (1/43). Importantly, more than a half of identified PV/LPV (28/43) occurred in double-strand DNA damage repair genes (14 ATM, 7 BRCA2, 2 BRCA1, 2 PALB2, 1 BRIP1, 1 CHEK2, and 1 RAD51C), potentially conferring a sensitivity to PARP inhibitors and platinum-based chemotherapies. Two patients (2/43) had MSH6 PV, predisposing to Lynch syndrome, potentially responsive to immunotherapies. Overall, almost 70% of PV/LPV carriers and about 14.5% of all cases have potentially targetable germline defects. The median age at PDAC in PV/LPV carriers was only slightly younger (64 years, range: 36 – 82) than in non-carriers (66 years, range: 42 – 84), however, the difference was not significant. Conclusions: Up to 20% of patients with PDAC and multiple primary tumors have a hereditary predisposition to cancer onset; the majority harbor PV/LPV in double-strand DNA damage repair genes. Up to 15% of PDAC patients with multiple primaries and up to 70% of PV/LPV carriers have potentially targetable germline defects. Age-based preselection has limited utility to identify hereditary PDAC, however, the presence of specific other malignancies in personal history strongly indicates on underlying genetic predisposition, and should warrant further investigation and genetic screening.