Increased Diagnostic Yield of Brushings for Biliary Strictures Using New Cytology Brush Design

Abstract

Introduction: The sensitivity of brush cytology for biliary strictures has typically been low, usually 30-60%. There have been multiple studies using various techniques to improve cellular yield and increase diagnostic accuracy. We compared the cellular yield and diagnostic accuracy using a new cytology brush (n=16) versus standard biliary brushings (n=16) in a total of 32 patients who underwent ERCP with brushings for evaluation of a biliary stricture for malignancy. Methods: We performed retrospective chart reviews of 16 consecutive ERCPs with brushings performed for the cytologic evaluation of a biliary stricture for malignancy using the new cytology brush between January 2016 and February 2017 at our institution. Our control cohort was 16 consecutive ERCP cases performed for the same indication directly preceding the availability of the new cytology brush (September 2015 to December 2015). All ERCPs were performed by the same operator (DGA). All cytology slides were pulled and reviewed by a board certified cytopathologist. The cytology preparation for each case was comprised of a single liquid-based Papanicolaou stained slide. The pathologist was blinded to the brushing device utilized for each case. Each slide was first assessed to determine a representative 20x microscopic field (accounting for approximately 1/25th of the total slide area). Once the field was selected, the area was carefully evaluated for the epithelial cellularity where large groups (>50 cells), medium groups (6-49 cells), small groups (2-5 cells) and single cells were quantified. Results: Please refer to Tables 1 and 2.Table: Table. Diagnostic Accuracy for New Cytology Brush Versus Standard Biliary BrushingsTable: Table. Average Number of Cell Clusters Per 20x Field for New Cytology Brush Versus Standard Biliary BrushingsConclusion: As demonstrated in Table 1, there were no statistically significant differences with regards to diagnostic accuracy for the new cytology brush versus standard biliary brushing. The biliary brushing cases performed using the new cytology brush demonstrated a significantly increased number of total cell clusters per representative 20x field compared with cases using the standard brush (mean 24.6 versus 14.4, p=0.03). This trend continued when assessing large (> 50 cells) clusters (mean 5.8 versus 3.3, p=0.02) and medium (6-49 cells) clusters (11.1 versus 5.8, p=0.03). There was not a statistically significant difference with cell yield for small (2-5 cells) clusters and single cells between the two groups. Although the new cytology brush increased the amount of tissue obtained on most parameters, it did not improve diagnostic yield.