Brush cytology allows sampling of broad mucosal areas. It has various gastrointestinal indications such as malignancy, Barrett’s esophagus and Helicobacter pylori infection. During endoscopy, the brush bristles exit a protective sheath. After sampling, the bristles retract back into the sheath. While few studies have explored the role of the plastic sheath on cellularity, sheath-influenced loss of cellularity is observed in the literature.[1, 2] As diagnostic accuracy depends on cellularity yield, this study investigated whether not retracting the bristles into the sheath is feasible and results in better cellular yield.

Patients with Barrett esophagus undergoing standard-of-care endoscopy were included at the Antwerp University Hospital. Endoscopy was performed by two experienced gastroenterologists. Brush samples were taken with either 'standard-of-care' brush retraction (=control) or 'unsheathed' (=intervention). Samples were stored in Cytolyt cell medium and directly transported to the department of Pathology for cytological evaluation and formalin-fixation for cellblock. Cellularity results were defined as ‘insufficient’, ‘low’, ‘moderate’ and ‘high’. [3] Cellularity results were bundled in two groups ‘insufficient-to-low’ and ‘moderate-to-high’ in order to facilitate statistical analysis using Fisher’s exact test at 95% confidence interval. Diagnostic accuracy was evaluated whether the brush cytology could confirm diagnosis of Barrett esophagus, irrespective of dysplasia grade.

Seventeen patients with Barrett esophagus were included during a one-year period for a total of 32 brush samples. When brush retraction was performed, cellularity was insufficient-to-low in 47.4% (9 brushes) and moderate-to-high in 52.6% (10 brushes). When the unsheathed technique was used, cellularity was moderate-to-high in 100% (13 brushes). There were no brushes that yielded insufficient-to-low cellularity. When comparing the two techniques, there is more cellular yield when not retracting (p = 0.0042).

Diagnostic accuracy was 52.6% (10/19) when brush retraction was performed, while diagnostic accuracy increased to 61.5% (8/13) with the unsheathed technique. This result was not statistically significant (p-value 0.72).

The Fisher exact test statistic value is 0.0042. The result is significant at p < .05. 

 The Fisher exact test statistic value is 0.7249. The result is not significant at p < .05.

This study points towards better cellular yield by not retracting the brush into the sheath during endoscopic brush sampling. A larger sample size is needed to evaluate diagnostic benefit and extrapolation for other indications. As part of the Horizon Europe Project ENDEAVOR, endoscopic brush samples will be taken from multiple international sites between abstract submission and presentation to evaluate result replicability and performance in a population with Barrett esophagus high-grade dysplasia and T1 adenocarcinoma.