Diagnostic and prognostic applications of blood cadherin-17 (CA17) and CA19.9 on pancreatic cancer.

Abstract

e16459 Background: The asymptomatic nature of pancreatic cancer (PC) poses significant challenges for effective initial evaluation of suspected cases. Consequently, most patients are diagnosed at stage IV, resulting in poor survival. CA19.9 is the only biomarker for PC detection, but its performance is suboptimal, underscoring the need for better alternatives. Cadherin-17 (CA17) expression is regulated in healthy individuals but overexpressed in gastrointestinal cancers. We have developed an immunoassay using in-house patented antibodies to detect CA17 in blood, offering potential utility for PC screening and prognostic assessment. Methods: The study included plasma samples from 156 and 30 asymptomatic donors serving as normal controls for CA17 and CA19.9 analysis. 28 PC patients were included, with paired pre- and post-treatment specimens available for CA17 analysis in all 28 cases and for CA19.9 analysis in 27 cases. Plasma CA17 levels were measured using the CA17 assay, and CA19.9 measured with a commercial ELISA kit. Results: CA17 expression was restricted in controls but significantly elevated in stage I/II PC, returning to lower thresholds in stage III/IV (mean: 1.51 ng/mL, 6.39 ng/mL, 3.46 ng/mL, respectively, p < 0.0001). No differences were observed in CA19.9 expression between controls and stage I/II PC, though a significant increase occurred in stage III/IV (mean: 44.5 U/mL, 471 U/mL, 15,581 U/mL, respectively). By ROC analysis, the CA17 assay showed an AUC of 0.92 with 85.7% sensitivity and 87.2% specificity for detecting stage I/II PC, outperforming CA19.9 (54.6% sensitivity and 90% specificity). CA19.9 alone detected only 54% of patients across all PC stages (13/24), but combining CA17 identified 6/11 additional patients missed by CA19.9, improving the detection rate by 25%. Stage III/IV PC patients were stratified into median-high (MH) and median-low (ML) groups based on CA17 and CA19.9 levels for 1-year overall survival (OS) analysis. Patients with concurrent MH-CA17 and ML-CA19.9 in both pre- and post-treatment statuses exhibited significantly better OS than those with ML-CA17 and MH-CA19.9 (median survival: 12.5 vs. 6 months, p = 0.0469). These data suggest the CA17/CA19.9 combination is valuable for prognostic stratification in PC. Conclusions: CA17 outperforms CA19 as a biomarker for PC screening, and the combination of these biomarkers further enhances the overall detection rate of PC while also providing predictive value for disease prognosis.