Diagnostic Value Of Endoscopic Techniques In Evaluating Invasion Depth Of Laterally Spreading Tumors Of The Colon And Rectum

Objective:A retrospective study was conducted to compare the accuracy of magnifying endoscopic narrow band imaging(ME-NBI)and endoscopic ultrasonography(EUS)in predicting the invasion depth of laterally spreading tumors(LST)of the colon and rectum.Our study was designed to provide a theoretical basis for optimizing the clinical treatments.Methods:We conducted a retrospective review of medical records of colorectal LST patients with well-documented data who were hospitalized in the third Xiangya Hospital of Central South University from January2016 to March 2022.Our study enrolled 75 patients who both performed ME-NBI and EUS examinations and finally obtained pathological examination results through endoscopic therapy or surgical treatment.The patients’LST lesions were classified by the Japan Narrow Band Imaging Expert Team classification in order to predict invasion depth.We also collected the predicted invasion depth by the endoscopic ultrasonography(EUS).The final pathological results serve as the gold standard to compare the accuracy rate of ME-NBI or EUS.The accuracy rate was calculated by chi-square test.Results:We collected the clinical data of 75 LST patients,including34 males and 41 females,with an average age of 59.5±9.8 years and an average LST lesion size of 22.8±10.5 mm.The postoperative pathological results showed that there were 5 cases of hyperplastic polyps(HP)/sessile serrated polyps(SSP),36 cases of low-grade intraepithelial neoplasia(LGIN),18 cases of high-grade intraepithelial neoplasia(HGIN),5 cases of superficial submucosal invasive cancer(SM_S)and 11 cases of deep submucosal invasive cancer(SM_D).ME-NBI prediction results:There were 6 cases of J-NET type 1 and the invasion depth was predicted to be mucosal layer;There were 44 cases of J-NET type 2a:the invasion depth of 43 cases was predicted to be mucosal layer and 1 case was predicted to be deep or deeper submucosal layer;There were 17 cases of J-NET type2b:the invasion depth of 10 cases was predicted to be mucosal layer and 5cases was predicted to be superficial submucosal layer,and 2 cases was predicted to be deep or deeper submucosal layer;There were 8 cases of J-NET type 3 and the invasion depth was predicted to be deep or deeper submucosal layer.EUS prediction results:54 cases were predicted to be mucosal layer;13 cases were predicted to be superficial submucosal layer;8 cases were predicted to be deep or deeper submucosal layer.The overall accuracy rates of NBI-ME and EUS in evaluating the invasion depth of LST were 92.0%,80.0%(X~2=4.485,P<0.05).The diagnostic accuracy of ME-NBI to mucosal lesions,superficial submucosal lesions and deep or deeper submucosal lesions was 98.3%,60.0%,and 72.7%,respectively;The sensitivity was 98.3%,60.0%,and 72.7%,respectively;The specificity was 75.0%,97.1%,and 100%,respectively;The positive predictive values was 93.5%,60.0%,and 100%,respectively;The negative predictive values was 92.3%,97.1%,and 95.5%,respectively.The diagnostic accuracy of EUS to mucosal lesions,superficial submucosal lesions and deep or deeper submucosal lesions was 84.7%,60.0%,and63.6%,respectively;The sensitivity was 84.7%,60.0%,and 63.6%,respectively;The specificity was 75.0%,81.4%,and 98.4%,respectively;The positive predictive values was 92.6%,18.8%,and 87.5%,respectively;The negative predictive values was 57.1%,96.6%,and 94.0%,respectively.Conclusion:Both NBI-ME and EUS are effective methods to evaluate the invasion depth of LST,NBI-ME may be more accurate in evaluating the invasion depth of invasion of LST.