Background
Endoscopic or percutaneous bile duct brushing is often performed as first step to differentiate between benign and malignant biliary strictures. Although brush cytology has a high specificity (95-100%), the sensitivity for detection of malignancy has been reported to be poor (41-67%). This results in repeated diagnostic procedures with potential treatment delay, adverse events, and additional costs. Aim of this study was to evaluate a change in protocol with optimization of obtaining, handling and assessment of brush cytology in patients with suspected perihilar or intrahepatic cholangiocarcinoma (pCCA/iCCA).

Methods
Patients with suspicion of pCCA or iCCA were prospectively included between June 2021 and June 2023. A historical cohort (January 2017-June 2021) was used as control. Difference in protocol between both cohorts are described in the attached table. In both cohorts morphological diagnosis was assessed according to the Papanicolau society of cytopathology. Both the diagnostic category ‘suspicious for malignancy’ and ‘malignant’ were classified as results compatible with malignant disease. Final diagnosis was confirmed by either histological proof of malignancy or in case unavailable, follow-up compatible with malignant disease. Primary endpoint was the sensitivity before and after implementation of the protocol; secondary endpoints were the sensitivity of the individual steps of the modified protocol in the prospective cohort.

Results
In this study, a total of 177 patients were evaluated (62 prospectively and 115 historical controls). The final diagnosis was malignant disease in 166 patients (93.8%). After protocol implementation, the sensitivity raised to 88.3% (95%CI, 76.8-94.8%) versus 50.9% (95%CI, 41.1-60.7%) pre-implementation (difference 37.4%; 95%CI, 23.6-51.2%). Specificity was 100% in both groups (2/2 vs 9/9). Sensitivity of the first brush in the prospective cohort was 78.3% (95%CI, 65.5-87.5%). NGS added value in 3 patients with uncertain results, increasing sensitivity to 83.3% (95%CI, 71.0-91.3%). A second brush was performed in 45 patients; of which one patient benefited from improved diagnostic value. Intraductal biopsies were performed in 34 patients (6 benign, 13 suspicious of malignancy, 15 malignant), leading to a malignant diagnosis in 3 out of 13 patients with false-negative brush cytology.

Conclusion
A modification in the handling of cytopathology, led to a significant improvement in the sensitivity of bile duct brushes to 78% for patients with suspected pCCA or iCCA. Furthermore, adding NGS, use of two brushes, and intraductal biopsies in the initial procedure could further increase sensitivity to 88%.

Mattheus C. Wielenga

Appearances