Abstract A085: Urinary microRNA Profiling Identifies Early Pancreatic Cancer Among Individuals at Increased Risk
Tomoya Kawase, Koji Yoshida, Yasutaka Kato, Shogo Baba, Tadatoshi Kawasaki, Hiroshi Kurahara, Hideyuki Oi, Shunsuke Kondo, Mao Okada, Tomoyuki Satake, Yukiko Shimada. Igawa, Tatsuya Yoshida, Junji Kita, Kazuya Kinoshita, Masaya Yokoyama, Johji Imura, Atsushi Satomura, Takayama Kazuya, Motoki MIkami, Yumi Nishiyama, Mika Mizunuma, Yuki IchikawaAbstract
Background:
Early detection of pancreatic ductal adenocarcinoma (PDAC) remains a critical unmet need because population-wide screening is not currently feasible and conventional biomarkers lack sufficient sensitivity for early-stage disease. We previously identified urinary microRNAs (miRNAs) associated with early pancreatic cancer using a platform for the isolation and analysis of extracellular vesicle-derived miRNAs from urine. Here, we investigated whether urinary miRNA signatures could distinguish patients with pancreatic cancer, including early-stage disease, from individuals at increased risk for PDAC without pancreatic cancer.
Methods:
Urine samples from 253 participants were retrospectively analyzed, including 144 patients with pancreatic cancer and 109 high-risk individuals (HRIs) without pancreatic cancer who were evaluated at our institution and affiliated centers between 2019 and 2024. The pancreatic cancer cohort included invasive PDAC (n=138), HG-PanIN (n=1), intraductal papillary mucinous carcinoma (IPMC; n=3), adenosquamous carcinoma (n=1), and unclassifiable carcinoma (n=1), with clinical stages ranging from 0 to IV. The HRI cohort included individuals with main pancreatic duct dilatation, type 2 diabetes mellitus, chronic pancreatitis, pancreatic cysts or intraductal papillary mucinous neoplasms, and/or a family history of PDAC. Urine samples were processed using an extracellular vesicle lysis-based protocol, followed by miRNA extraction and next-generation sequencing. A LightGBM gradient-boosting model was developed to discriminate pancreatic cancer/HG-PanIN from HRIs and to identify informative miRNA features.
Results:
Based on our previous analysis, 16 urinary miRNAs showing significant differential expression in stage IA pancreatic cancer were selected as candidate biomarkers. The study cohort included patients across the full spectrum of pancreatic cancer progression, including stage 0 (HG-PanIN) and stage IA disease. Using urinary miRNA profiles, the LightGBM classifier distinguished pancreatic cancer/HG-PanIN from individuals at increased risk for PDAC with a sensitivity of 0.80 and specificity of 0.80. The areas under the receiver operating characteristic curve were 0.888 in the training cohort and 0.889 in the independent test cohort, demonstrating consistent diagnostic performance.
Conclusions:
Urinary miRNA profiling demonstrated promising performance for distinguishing pancreatic cancer, including stage 0 and early-stage disease, from individuals at increased risk for PDAC. A urinary miRNA signature derived from stage IA pancreatic cancer may complement risk-based enrichment strategies and facilitate the identification of individuals who should undergo intensive pancreatic imaging. Prospective validation in independent, longitudinal cohorts is warranted to establish its clinical utility for the early detection of pancreatic cancer.
Citation Format:
Tomoya Kawase, Koji Yoshida, Yasutaka Kato, Shogo Baba, Tadatoshi Kawasaki, Hiroshi Kurahara, Hideyuki Oi, Shunsuke Kondo, Mao Okada, Tomoyuki Satake, Yukiko Shimada. Igawa, Tatsuya Yoshida, Junji Kita, Kazuya Kinoshita, Masaya Yokoyama, Johji Imura, Atsushi Satomura, Takayama Kazuya, Motoki MIkami, Yumi Nishiyama, Mika Mizunuma, Yuki Ichikawa. Urinary microRNA Profiling Identifies Early Pancreatic Cancer Among Individuals at Increased Risk [abstract]. In: Proceedings of the AACR Conference on Pancreatic Cancer: New Frontiers in Biology and Therapeutic Development; 2026 Sep 25-28; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(18_Suppl_2):Abstract nr A085.