The presence of a biliary stricture is a medical challenge. ERCP brush cytology (BR) and endo-biliary forceps biopsies (BX) are commonly used tissue sampling techniques. BR is favoured for its simplicity; however, its diagnostic sensitivity has historically been low. Its utility is limited by poor specimen quality and equivocal results are common (1). BX, while technically more demanding, offers superior histological detail and higher diagnostic yield (2). Given the implications of missed or delayed diagnoses, it is crucial to establish a standardised, evidence-based approach to tissue sampling during ERCP.

We aimed to compare the diagnostic utility of ERCP brushings versus endo-biliary biopsies in patients with confirmed malignant biliary obstruction.

We retrospectively reviewed all patients who underwent ERCP for investigation of biliary strictures during a three-year period (08/2021–07/2024) in our department (a referral centre). We acquired all relevant data (demographics, histological diagnosis, method of diagnosis, dates) through patient records.

Diagnostic yield was defined as a positive cytology or histology confirming malignancy. Results were analysed in three categories (Negative, Positive, Suspicious). Agreement analysis (kappa values) was performed in all patients who underwent both BR and BX. 

Total number of patients investigated for biliary stricture was 208. A confirmed diagnosis of malignancy was established in 110 patients (53.4%) through ERCP brushings (BR), ERCP biopsies (BX), or further investigations. BR was performed in 110 patients, with the addition of BX in 55/110. BR was positive in 33/110 (30%) and negative in 77/110 (70%) (including suspicious 28/110; 25.9%), which required another modality to confirm diagnosis.

BX were positive in 34/55 (61.8%) and negative in 21/55 (38.2%) (including suspicious 8/55; 14.5%). All suspicious cases for BX and BR were ultimately proven malignant. There were no cases where BR was positive but BX was negative. Among the 55 patients who had both tests, BR was positive in 13/55 (23.7%) compared with 34/55 (61.8%) for BX. No complications were noted as a direct result of BX in addition to BR.

Concordance analysis:

Overall agreement between BR and BX was low (43.6%), with only fair concordance (Cohen’s kappa 0.20–0.21). Even when results were merged, agreement improved only modestly (58.2%). A key finding was that BR frequently missed malignancy: 29% of patients with negative brushings had positive biopsies, and 36.4% had positive or suspicious biopsy results, demonstrating the limited reliability of brushings as a standalone test.

The data suggests BR cytology has limited sensitivity as a standalone diagnostic tool in malignant biliary strictures, and biliary biopsies have almost 3 times the diagnostic yield. This is significant as delayed diagnosis often leads to delay in definitive treatment and delays referral for subsequent diagnostic modalities pending results. The low kappa values (0.20–0.21) indicate only fair agreement between BR and BX, reinforcing that the two methods do not perform similarly and that BR frequently under-detects malignancy compared with BX.

In this cohort of patients, brushings provided low diagnostic yield whether performed alone or with BX (30% and 23.7% in respective cohorts). Endo-biliary BX have a higher diagnostic yield (61.8%). The data also suggests that brushings add little incremental value and could be omitted unless biopsies are technically challenging. The poor concordance between the two modalities (Cohen’s kappa 0.20–0.21) further demonstrates that brushings frequently miss malignancy captured by biopsies. Biopsy should be considered the standard of practice, with brushings reserved as an adjunctive option rather than a primary diagnostic tool.