Wide-Area Transepithelial Sampling with 3D analysis (WATS3D) is widely used as an adjunct to random forceps biopsies (RFB) during Barrett’s esophagus (BE) surveillance endoscopies. In a previous randomized trial (WATS-EURO1) comparing WATS with RFB in 172 patients in 17 BE expert centers in Europe, we found that inadequate cellularity of WATS3D samples prohibited a reliable diagnosis in 13% of cases, despite that procedures were performed by experienced BE endoscopists who were instructed on the brush technique and with in-room training on the sample preservation.

The WATS-EURO2 study was initiated after the WATS-EURO1 study to study the relevance of WATS3D-positive-RFB-negative findings during follow-up and aimed to enroll 200 BE patients with low grade dysplasia, high-grade dysplasia or early cancer and 200 non-dysplastic BE patients with a follow-up of 3 years and without ablation for all WATS3D-positive-RFB-negative cases. Based on the sample inadequacy rates in the WATS-EURO1, the WATS-EURO2 was separated into three chronologically distinct phases : (1) Conventional Preservation Phase: after brush collection, a portion of the sample was applied as a dry smear on a glass slide, followed by the addition of preservation fluid for transport and processing; (2) Improved Preservation Phase: no dry smear was obtained and the entire brush sample was preserved in a transportation medium before transportation and processing; and (3) Post-training Phase: the same preservation method was used, but endoscopists received additional training instructions, video tutorial, and individual feedback on their sample adequacy. Sample adequacy was evaluated across each phase, with subgroup analyses by endoscopist. Specimens were categorized and assessed at the CDx Diagnostics laboratory, blinded to case indications and RFB outcomes. Samples were classified as inadequate if either the smear or cell block showed poor preservation or low cellularity. 

Seven BE expert endoscopists contributed to the study. Sample adequacy did not improve in the Improved Preservation Phase compared to the Conventional Preservation Phase (126/218 cases (57.8%) vs. 113/173 cases (65.3%), p=0.15). In the Post-Training Phase adequacy rates rose to 299/326 (91.7%; p< 0.001, compared to the preceding phases; Figure 1). All 7 endoscopists significantly improved their adequacy rates with final adequacy rates between 73% and 100% (median: 92%; Figure 2). Despite the overall improvement post-training, inadequate samples persisted in 0 to 27% (median 8%) of cases. 

Despite that all procedures were performed by BE expert endoscopists, after standard instructions and training, samples were deemed inadequate in a substantial number of cases.  Sample adequacy improved to >90% after additional training instructions, video tutorial and individual feedback on sample adequacy. These findings underscore the critical role of a proper brushing technique in WATS3D sampling and the necessity of reporting inadequacy rates in WATS3D studies.