Diagnostic and Prognostic Value of CA19-9 in Pancreatic Tumors: A Retrospective Study

Background: Carbohydrate antigen 19-9 (CA19-9) is a widely used biomarker for pancreatic cancer, but the interpretation of its level is confounded by false positives in patients with cholestasis and false negatives in Lewis antigen-negative patients. The aim of the study was to investigate the diagnostic accuracy of CA19-9 in differentiating malignant from benign pancreatic lesions and its prognostic significance following tumor resection. Methods: We retrospectively reviewed the data of 81 consecutive pancreatic resections performed at Al-Latron Military Hospital, a tertiary center in Amman, Jordan, from January 2016 through June 2024. Preoperative CA19-9 levels, clinical variables, and pathology findings were collected. Diagnostic discriminability was assessed via receiver operating characteristic (ROC) analysis. Multivariable logistic regression was used to assess whether log10(CA19-9) and jaundice could independently predict malignancy. Overall survival (OS) and progression-free survival (PFS) were estimated by Kaplan-Meier survival analysis, and the corresponding curves were compared with log-rank tests, including those of patients stratified a priori by CA19-9 level (<=37 vs. >37 U/mL). Results: CA19-9 values were available for 73/81 patients (59 with malignant lesions, 14 with benign lesions). Preoperative CA19-9 levels were higher in malignant lesions than in benign lesions (median 86 vs. 12 U/mL; p=0.008). The area under the curve (AUC) was 0.82 (95% CI 0.71-0.90), indicating moderate-to-good discriminative ability. A CA19-9 level of 37 U/mL achieved a sensitivity of 62% (95% CI 48%-74%) and a specificity of 89% (95% CI 62%-98%) in discriminating malignant from benign lesions. In the multivariable analysis, log10(CA19-9) independently predicted malignancy (OR 3.97; 95% CI 1.2-13.0; p=0.02), whereas jaundice did not (p=0.37). After a median follow-up of 24 months, the one- and three-year OS rates for patients with malignant lesions were 93.7% and 81.0%, respectively; median OS was not reached, with 12 death events observed during follow-up. Patients with benign lesions had a 100% survival rate at the last follow-up (p=0.03). Among patients with malignant lesions, OS was numerically greater for CA19-9 ≤ 37 vs. >37 U/mL, but the difference was not statistically significant (p=0.15). Among patients with available postoperative measurements, postoperative CA19-9 elevation appeared to accompany early recurrence; however, postoperative sampling was not standardized, and this observation is descriptive only. Conclusions: CA19-9 is a useful diagnostic adjunct and an independent predictor of malignancy after adjustment for jaundice. However, its modest sensitivity at the conventional cutoff limits its use as a stand-alone diagnostic marker. Preoperative CA19-9 may provide exploratory prognostic stratification after resection, although the observed survival difference by CA19-9 group was not statistically significant in this cohort.