Optimizing CA19.9 role in BD-IPMN Surveillance: Enhanced Value of New Thresholds

Aims Monitoring CA19.9 in BD-IPMN surveillance is recommended by international guidelines and elevated levels are considered relative indication for surgery. However, established cut-offs often lead to intensified follow-ups or even surgery without clear benefits. This study aimed to evaluate the role of CA19.9 as a predictive marker of malignant progression in pancreatic cysts and to examine how it is utilized by clinicians in everyday clinical practice.

Methods In this multicentric retrospective observational study, clinical and histopathological data from patients diagnosed with BD-IPMN were collected. Associations between CA19.9 levels and malignant degeneration were evaluated using Chi-square tests and regression analyses. Participants were enrolled if they had at least one serum CA19.9 measurement and a minimum follow-up period of 6 months. Demographic and medical history data were gathered alongside imaging studies, endoscopic evaluations, details of surgical interventions and blood test results. Malignant degeneration was defined as the presence of high-grade dysplasia or invasive neoplasia.

Results Of the initial 333 patients, 185 had at least one recorded CA19.9 assessment. Over a median follow-up of 5 years, 25 patients exhibited elevated CA19.9 levels, of whom four (16%) developed malignancy. In contrast, malignancy occurred in 2 out of 160 (1.3%) patients with normal CA19.9 levels (p=0.001). Notably, 24% of elevated CA19.9 cases normalized with time. ROC curve analysis identified a CA19.9 level of 95 as the optimal threshold, achieving an improved AUC of 0.815 (95%CI 0.592-1), enhanced specificity (0.983 vs 0.916), and comparable accuracy (0.962 vs. 0.908) when compared to the conventional cut-off of 37. Among patients with elevated CA19.9 who developed malignancy, the median time to degeneration was 27 months, earlier than the 42 months of those who encountered degeneration with normal CA19.9 levels. Regression analysis indicated a 2% increased risk of malignancy for each 1 kU/L rise in CA19.9. Despite these findings, clinicians adopted more conservative strategies than those provided by IAP guidelines in 50.9% of cases with elevated CA19.9, compared to only 23.7% of conservative indications when CA19.9 levels were regular (p<0.001).

Conclusions This study supports a higher CA19.9 cut-off of 95 kU/L to optimize surveillance in patients with BD-IPMN. This higher threshold improved predictive value with comparable accuracy, particularly when elevated levels were confirmed in consecutive assessments. Nonetheless, this exam is infrequently requested by clinicians in clinical practice, and in most cases of abnormal CA19.9 levels guideline recommendations are not followed.

Publikationsverlauf

Artikel online veröffentlicht:
27. März 2025