Abstract:Objective To explore the clinical value of Japan narrow-band imaging expert team (JNET)classification based on magnifying endoscopy with narrow-band imaging (ME-NBI) combined with endoscopic ultrasonography (EUS) in judging the invasion depth of neoplastic lesions. Methods A retrospective analysis was performed on the clinical data of 86 patients with colorectal polypoid lesions diagnosed in Hengshui People's Hospital from October 2019 to October 2020. In all patients, the microvessels of lesions were typed according to JNET classification and evaluated by EUS for the depth of invasion before operation and were compared with pathological diagnosis(as the gold standard) after endoscopic mucosal resection, surgery and submucosal dissection. The reliabilities of EUS combined with JNET classification and JNET classification alone were analyzed in determining the invasion depth of colonic neoplastic lesions. Kappa test was used for consistency analysis between them. Results There were 94 lesions in 86 patients. The coincidence rate of JNET type 1 with pathological diagnosis of hyperplastic polyp/sessile serrated polyp (SSP) was 82.1% (23/28), that of type 2A with pathological diagnosis of low-grade intraepithelial neoplasia （LGIN） was 84.1% (32/38),that of type 2B and pathological diagnosis of high grade intraepithelial neoplasia(HGIN)/intramucosal carcinoma(M)/superficial submucosal invasive carcinoma(SM-s) was 65.2% (15/23), and that of type 3 and pathological diagnosis of deep submucosal invasive carcinoma(SM-d) was 61.1% (11/18). The accuracy of EUS combined with JNET was 97.06% (33/34) in determining the depth of lesion infiltration. The consistency test showed that the consistency between JNET classification alone and pathological diagnosis was medium (Kappa=0.580, P<0.01),while,there was a high consistency between EUS combined with JNET classification and pathological diagnosis (Kappa=0.940, P<0.01). Conclusions There is a certain coincidence rate between NET classification and pathological diagnosis, but that of some subtypes is still low. The accuracy of JNET classification combined with EUS is higher than that of JNET alone in judging the depth of invasion of colorectal neoplastic lesions.