Identification of factors predictive of malignancy in patients with atypical biliary brushing results obtained via ERCP.

Biliary brushings obtained during ERCP can have one of three cellular interpretations: benign, malignant, or "atypical." Atypical interpretations usually result in further testing, and may cause controversy over management and increases in cost. We evaluated a large cohort of patients with atypical biliary brushings for analysis and risk stratification. All biliary brushing specimens collected between January 1, 2001 and December 31, 2010 that had an atypical result were included. Hospital electronic records were reviewed for these patients to include: demographics, indication for ERCP, endoscopist/pathologist impressions, serologic testing, stricture site, and information relating to the final clinical diagnosis. Eighty-six patients were included. Totally, 60/86 patients (70%) had malignancies while 26/86 (30%) had no evidence of malignancy during long term follow up. Univariate analysis showed that the risk of malignant outcomes was significantly associated with older age, suspicious/malignant endoscopic impression, pancreatic mass, indications including jaundice and/or dilated bile ducts, stricture location within the common bile duct, PSC, and CA 19-9 levels >300 U/mL. We created a novel scoring system for prediction of malignancy based on clinical and endoscopic factors. We identified parameters that are typically available to the clinician to categorize patients with an "atypical" bile duct brushing results into "high risk" and "lower risk" classifications. Our proposed scoring system would allow such risk stratification to take place.