Safety and Diagnostic Performance of Pancreatic Duct Brushings for Detecting Pancreatic Cancer.

OBJECTIVES: Pancreatic duct (PD) brush cytology (BC) and fluorescence in-situ hybridization (FISH) are viable techniques but underutilized due to limited data on their diagnostic accuracy for pancreatic malignancy and associated risks. This study evaluates the safety and diagnostic performance of BC, FISH, and their combination for detecting pancreatic malignancy. METHODS: A retrospective analysis was conducted on subjects undergoing endoscopic retrograde cholangiopancreatography (ERCP) PD brushings at a tertiary hospital. Data included demographics, procedural details, adverse events, and outcomes. Malignancy was confirmed by EUS-guided biopsy or surgical histology. Diagnostic metrics for BC, FISH, and their combination were calculated. RESULTS: Of the 205 subjects who underwent ERCP PD brushings from 2004 to 2023, 35 (17.1%) were diagnosed with pancreatic malignancy. BC had a sensitivity of 31.4% and specificity of 95.9% (AUC 0.64). Abnormal FISH showed 45.7% sensitivity and 89.4% specificity (AUC 0.68), while polysomy FISH had 28.6% sensitivity and 97.1% specificity (AUC 0.63). Combined BC and abnormal FISH yielded 54.3% sensitivity and 87.7% specificity (AUC 0.71). AEs included post-ERCP pancreatitis in 16 cases (7.8%), bleeding in 2 (1.0%), and micro-perforation in 1 (0.5%). CONCLUSIONS: PD brushing for BC and FISH is generally safe with modest diagnostic accuracy for malignancy. While novel diagnostic tests such as molecular biomarkers in pancreatic fluid for detecting PC are still needed, PD brushings can assist in the detection of PC with low rates of PEP in select subjects.