Dynamics of the association between circulating levels of miRNA and pancreatic cancer risk through the years prior to pancreatic cancer diagnosis

Early diagnosis of pancreatic ductal adenocarcinoma (PDAC) is critical to improving the overall prognosis. However, biomarkers for early diagnosis are lacking. We used pre-diagnostic plasma samples from 1307 PDAC cases and individually matched cancer-free controls from five prospective cohort studies. Conditional logistic regression was used to assess the association of circulating miRNAs with PDAC risk. Linear regression was conducted to identify miRNAs ratio (miRNA levels of case over their matched control) change associated with the lead time from blood draw to PDAC diagnosis. We found 20 miRNAs significantly associated with PDAC risk (pFDR < .05). Among them, 13 miRNAs showed a significant ratio change within 10 years prior to cancer diagnosis. The case-control difference of miR-155-5p and miR-493-3p increased when blood draw was closer to PDAC diagnosis. While the ratio changes of the remaining 11 miRNAs decreased. Adding 13 miRNAs to known risk factors and CA19-9 significantly improved the prediction performance for imminent PDAC risk, with AUCs of 85%, 77%, and 73% during the time windows of 1, 2, and 3 years prior to PDAC diagnosis. Findings of this large prospective study shield a light on the biology of PDAC and suggest a potential utility of monitoring circulating miRNAs for PDAC risk surveillance among high-risk individuals.