Endoscopic submucosal dissection and JNET classification for colorectal neoplasia: A North American academic center experience

OBJECTIVES: Endoscopic submucosal dissection (ESD) enables minimally invasive resection of superficial gastrointestinal neoplasms en bloc regardless of size. The Japan narrow band imaging expert team (JNET) classification utilizes optical magnification and narrow band imaging (NBI) to predict pathology. In North America, ESD is far from ubiquitous, and regional outcomes are not widely described. To date there are no North American studies describing the application and yield of the JNET classification as applied in the practice of ESD. METHODS: A retrospective, single‐center, cohort analysis was performed on a prospectively maintained database of ESD procedures. Between July 2016 and February 2023, all consecutive patients treated with ESD for colorectal lesions were identified and stratified by lesion location, JNET, NBI International Colorectal Endoscopic, lateral spreading tumors, and Paris classifications. Univariate analysis was used for clinicopathological data. p < 0.05 was considered statistically significant. RESULTS: A total of 112 patients were identified. One lesion, a lipoma, was excluded. Overall, 49.5% (55/111) of lesions were colonic and 50.5% (56/111) rectal. Most lesions were lateral spreading tumors (60.4%, 67/111). Overall, 96.4% (107/111) ESDs were successfully completed, 98.1% (105/107) en bloc, and 87.9% (94/107) R0. Adverse events occurred in 1.8% (2/111) of procedures. The median diameter was 4.0 cm and resected in a median time of 62.0 min. Overall, 70.1% (47/67) lesions were upstaged from pre‐ESD biopsy. JNET 2B showed 80.2% (95% CI 71.5–87.1) accuracy for high‐grade dysplasia or sm1. All JNET type 3 were ≥sm2 (p < 0.001). CONCLUSIONS: ESD permits safe and effective resection of superficial colorectal neoplasms. JNET classification was more accurate than pre‐resection biopsy at predicting histology in this series.