Circulating tumor DNA-based genomic landscape of early-onset pancreatic cancer.

Abstract

588 Background: The incidence of gastrointestinal and obesity related cancers, including pancreatic cancer, is increasing in individuals <50 years old. Early-onset pancreatic cancer (EOPC; age<50 years old at diagnosis) is characterized by higher incidence in males, enrichment for KRAS wild-type tumors with targetable genomic alterations (GAs), and improved outcomes compared to average-onset pancreatic cancer (AOPC; age>50 years old at diagnosis). Little is known about the genomic correlates underlying these clinical differences. In this study, we sought to characterize the circulating tumor DNA (ctDNA) landscape in EOPC compared to AOPC. Methods: We analyzed ctDNA samples from a total of 8548 patients (EOPC n=488 [age<30yo n=13, 30-39yo n=87, 40-49yo n=388]; AOPC n=8060) collected prospectively between December 2017 to March 2021 using a 73 gene next generation sequencing panel (Guardant360). Statistical analyses were performed using Fisher’s exact test. Results: Of the 488 EOPC patients, 261 (53%) were male and 227 (47%) were female. Median age was 45yo in EOPC and 69yo in AOPC. Contrary to prior reports from tissue-based sequencing studies, EOPC patients were more likely to harbor KRAS alterations (p≤0.0001), specifically KRAS G12V (p≤0.0072) and KRAS amplifications (p≤0.0001). 1.24% of pancreatic cancer patients harbored KRAS G12C mutations; 0.05% were in EOPC patients. The most common currently-targetable GAs identified in EOPC were PIK3CA (10%), BRCA1/2 (10%), EGFR (9%), BRAF (7%), MET (7%), ATM (7%), FGFR1/2 (7%), CDK6 (5%), and ERBB2 (4%). EOPC patients had higher proportion of targetable GAs in BRCA2 (p≤0.041), MET (p≤0.005), and PIK3CA (p≤0.0023) compared to AOPC. Conversely, AOPC patients had higher proportion of GAs in TP53 (p≤0.0001) and ATM (p≤0.0001). MSI-high and TMB-high were detected in 0.07% and 0.06% of EOPC patients respectively, while 0.6% and 1.7% AOPC patients were MSI-high and TMB-high. Gene fusions were detected in 0.3% pancreatic cancer patients, predominantly in FGFR2 (0.2%) and FGFR3 (0.02%). Potential germline variants were detected in both EOPC and AOPC patients, most commonly in BRCA2 (54%). Females displayed higher proportion of TP53 (p<0.0001), EGFR (p<0.0029), CDKN2A (p<0.0001), BRCA2 (p<0.0336) and ATM (p<0.0001) mutations compared to males. Gender was predictive of the pattern of GAs in EOPC with MET (p≤0.0027) , ARID1A (p≤0.0376) , and BRAF (p≤0.0034) alterations enriched in males, and PIK3CA (p≤0.0017) alterations enriched in females. Conclusions: This study represents the first large-scale blood-based ctDNA genomic profiling of EOPC. Based on this age and gender-stratified molecular characterization of pancreatic cancer, EOPC is a distinct molecular entity compared to AOPC. Identification of multiple targetable GAs may improve patient outcomes in EOPC, especially when a tissue biopsy is not feasible or sufficient for comprehensive genomic profiling.