Improved predictability of pancreatic ductal adenocarcinoma diagnosis using a blood immune cell biomarker panel developed from bulk mRNA sequencing and single-cell RNA-sequencing

Abstract

Abstract Background Pancreatic ductal adenocarcinoma (PDAC) remains a devastating cancer due to its poor survival rate, early detection, and resectability. This study aimed to determine the peripheral blood mononuclear cell (PBMC) immune biomarkers in patients with PDAC and investigate the PDAC-specific peripheral blood biomarker panel and validate its clinical performance. Methods In this prospective, blinded, case–control study, a biomarker panel formula was generated using a development cohort—including healthy controls, patients at high risk of PDAC, and patients with benign pancreatic disease, PDAC, or other gastrointestinal malignancies—and its diagnostic performance was verified using a validation cohort, including patients with ≥ 1 lesion suspected as PDAC on computed tomography (CT). Results RNA-sequencing of PBMCs from patients with PDAC identified three novel immune cell markers, IL-7R, PLD4, and ID3, as specific markers for PDAC. Regarding the diagnostic performance of the regression formula for the three biomarker panels, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 84.0%, 78.8%, 47.2%, 95.6%, and 79.8%, respectively. Based on the formula scores for the biomarker panel, the false-negative rate (FNR) of the biomarkers was 8% (95% confidence interval [CI] 3.0–13.0), which was significantly lower than that based on CT in the validation cohort (29.2%, 95% CI 20.8–37.6). Conclusions The regression formula constructed using three PBMC biomarkers is an inexpensive, rapid, and convenient method that shows clinically useful performance for the diagnosis of PDAC. It aids diagnoses and differential diagnoses of PDAC from pancreatic disease by lowering the FNR compared to CT. Clinical trial registration Clinical Research Information Service, KCT0004614 (08 January 2020).