Current European guidelines favor the use of endoscopic submucosal dissection (ESD) for the eradication of gastric dysplasia exceeding 1cm in size.  Despite its efficacy in “uprooting” superficial gastrointestinal lesions, ESD is often time-consuming, requiring additional expertise and resources compared to endoscopic mucosal resection (EMR). The aim of this study was to investigate the efficacy, safety and speed of an EMR variation capable of providing a deeper resection plain, cap-assisted EMR (capEMR), for the endoscopic management of gastric dysplastic lesions measuring up to 2cm. 

Consenting adults with documented by 2 expert histopathologists gastric dysplasia (low and high-grade) were eligible for endoscopic treatment. Lesions >2cm were treated with ESD, those <1cm with capEMR whereas patients with lesions in between 1 and 2cm could choose either treatment. In the latter group, if treatment failure using capEMR occurred, patients could receive subsequent treatment with ESD. For the present study, data from cases with gastric dysplasia in between 1 and 2cm in size, diagnosed and treated during the period 2019-2025 were retrospectively collected and analysed. Dysplastic lesions with a diameter up to 2cm were treated using capEMR as follows: using an endoscope with a near focus of 1.5mm and the tip of a crescent-shaped snare, pre-resection marking was performed followed by submucosal injection of succinylated gelatin and indigo carmine solution. Then the crescent-shaped snare was prelooped and fixed onto the ridge of an oblique cap mounted onto the tip of the scope. The lesion was then carefully suctioned into the cap (pseudopolyp formation), tightly enclosed within the snare and resected using blended/mixed current (PulseCut Slow effect 2, 120 watts). Coagulation of visible or bleeding vessels and capEMR defect closure were applied in all cases using coagulation forceps and through-the-scope clips. Midazolam and fentanyl were used for patient sedation.

Apart from demographics, en bloc, R0 resection (dysplasia-free margins), recurrence and adverse event rates were calculated and recorded. Additional parameters such as total operation time (from marking until sample acquisition), resection time per lesion, number of resected lesions per endoscopy and depth of resection (deepest layer within the specimen eg muscularis mucosa or submucosa) were also determined. 

Sixty-four patients (48% females), mean age ± SD 62 ± 8.4years underwent capEMR for 103 lesions of which the vast majority (93/103) harbored low-grade dysplasia (10/103 high-grade dysplasia, P<0.001) and were located in the antrum (77/103), incisura (12/103), corpus (12/103) and cardia (2/103)(P<0.001). Median (interquartile range) size of the lesions was 1.7cm (1.3-2.1). Median (interquartile range) total operation time was 13min (6-21) for a median of 2(1-5) lesions per endoscopic session. Median resection time per lesion was 3min (2.8-4.6). En bloc resection rates were 91% (94/103) whereas R0 resection rates were 99%. The submucosa was present in all but three (100/103, 97%) resection specimens. No recurrences were recorded during follow-up (9-64months). Adverse events included mild post-procedural pain in 59/64 cases, with only 23 cases requiring paracetamol analgesia, fever up to 38.1C in 7/64 cases and late onset bleeding in a single case upon resumption of rivaroxaban. The bleeding was successfully treated with additional hemoclip placement.

Based on this study capEMR is a fast, effective and safe endoscopic treatment modality for the eradication of gastric dysplasia of moderate size. The endoscopist’s expertise in capEMR, the location and histopathology of lesions (predominantly low-grade dysplasia of the antrum) and the routine application of endoscopic bleeding prophylaxis eg coagulation of vessels and clipping may have optimized outcomes. If these findings are confirmed in randomized controlled trials comparing capEMR with ESD then perhaps capEMR can be established as a less time-consuming alternative to ESD for treating gastric dysplasia up to 2cm.