Biliary Brush Cytology Revisited

Abstract

Purpose: To evaluate the diagnostic yield of biliary brush cytology and the factors affecting positive results in patients with biliary strictures.

Patients and methods: The medical records of all patients who underwent endoscopic retrograde cholangiopancreatography (ERCP) with biliary brush cytology at our institution from November 2004 to December 2013 were reviewed in this retrospective study. The yield of positive brush cytology and the factors affecting positive yield, such as stricture location, age, gender and preprocedure CA 19.9 level were assessed. The final histopathology, diagnosis obtained by other methods, such as endoscopic ultrasound-guided fine-needle aspiration cytology, CT scan, Tru-Cut biopsy and/or clinical/radiological follow-up were used to identify true- and false-positive/negative results. The brush cytology results were divided into 4 main categories: malignant, benign, atypical cells and inadequate.

Results: A total of 1,168 patients underwent ERCP during this 9-year period. Out of these, 142 patients had ERCP and biliary brushings for diagnosis. The mean age of the patients at presentation was 58.7 years (range 23-84 years; 64.8% males). The indication for referral was obstructive jaundice in all patients. Of the 142 patients, 77 (54.2%) had a distal common bile duct (CBD) stricture and 65 (45.8%) had a proximal /complex hilar stricture. The strictures were classified as proximal or distal, based on their relationship with the cystic duct; those below the cystic duct insertion were classified as distal and those above it were considered proximal. The diagnostic yield of brush cytology was 58.5%. The diagnostic yield was higher for proximal than for distal CBD strictures (67 vs. 50%; p = 0.047). It was also higher for females (58 vs. 57.6%; p = 0.94), patients >50 years (60 vs. 50%; p = 0.29) and those with a CA 19.9 level >300 IU/ml (59.4 vs. 55.5%; p = 0.65) but did not reach statistical significance for any of these parameters. Complete follow-up data were available for 96 patients and 46 patients were lost to follow-up. The sensitivity, specificity, positive predictive value and negative predictive value were 65.3, 100, 100 and 27%, respectively. When patients with atypia were included in the group with positive results, the diagnostic yield increased to 65.5% with a diagnostic sensitivity of 68.6%. There were 27 false-negative diagnoses, 10 patients were true-negative and no patients had a false-positive diagnosis.

Conclusion: Biliary brush cytology is a safe and simple initial diagnostic procedure in patients with biliary strictures and can be performed at the time of therapeutic ERCP. If performed correctly and then interpreted by a dedicated cytopathologist, it has a good diagnostic yield and sensitivity. We feel that the low rates of success with this technique reported in some earlier studies have led to a feeling that this is not a particularly useful technique. We recommend that this topic should be revisited, and that the technique should be used more often.