【摘要】 目的 构建窄带成像（Narrow-Band Imaging,NBI）技术联合超声内镜检查（Endoscopic Ultrasonography,EUS）诊断结直肠癌（Colorectal Cancer,CRC）浸润深度的预测模型。方法 回顾性选取2017年3月至2022年5月于我院就诊的288例CRC患者为研究对象，使用随机数表法将患者以2∶1的比例分为训练集（n=192）和验证集（n=96）。根据训练集患者检查结果将其分为黏膜下癌组和黏膜内癌组，对比两组患者临床资料、NBI及EUS检查结果，采用LASSO和Logistic回归分析筛选黏膜下癌的独立影响因素，构建列线图预测模型并进行评价。结果 在训练集192例患者中，黏膜下癌48例，黏膜内癌144例，两组患者在肿瘤直径、淋巴结转移、易出血及扩张血管方面的差异具有统计学意义（P<0.05），回归分析结果显示，肿瘤直径≥3 cm、淋巴结转移、易出血及扩张血管为CRC患者黏膜下癌的独立危险因素（P<0.05）。列线图预测模型在训练集和测试集中受试者工作特征曲线下面积分别为0.758(95%CI:0.682～0.834,P<0.001)和0.732(95%CI:0.682～0.782,P<0.001)，模型预测概率可较好拟合实际概率，临床决策曲线中列线图模型远离极端曲线，净获益值高。结论 肿瘤直径≥3 cm、淋巴结转移、易出血及扩张血管为CRC患者黏膜下癌的独立危险因素，根据以上因素构建的预测模型可为临床诊断提供参考。更多还原

【Abstract】 Objective To construct a predictive model for diagnosing the depth of invasion of colorectal cancer(CRC) using narrowband imaging(NBI) technology combined with endoscopic ultrasonography(EUS).Methods A total of 288 patients with CRC who were admitted to our hospital from March 2017 to May 2022 were selected as the research subjects retrospectively.The patients were divided into training set(n=192) and verification set(n=96) with a ratio of 2∶1 by using random number table method.The patients were divided into submucosal cancer group and intramucosal cancer group according to the examination results of the training set.The clinical data,NBI and EUS examination results between the two groups of patients were compared.The independent influencing factors of submucosal cancer were screened by LASSO and Logistic regression analysis,and the nomogram prediction model was constructed and evaluated.Results Among the 192 patients in the training set,48 cases were submucous carcinoma and 144 cases were intramucosal carcinoma.The difference between the two groups in tumor diameter,lymph node metastasis,bleeding and vasodilation was statistically significant(P<0.05).The results of regression analysis showed that tumor diameter ≥3 cm,lymph node metastasis,bleeding and vasodilation were independent risk factors for submucous carcinoma in CRC patients(P<0.05).The area under the receiver operating characteristic curve of the nomogram prediction model in the training set and the test set was 0.758(95% CI:0.682-0.834,P<0.001) and 0.732(95% CI:0.682-0.782,P<0.001),respectively.The prediction probability of the model was well fitted to the actual probability.The nomogram model in the clinical decision curve was far from the extreme curve,and the net benefit rate was high.Conclusion Tumor diameter ≥3 cm,lymph node metastasis,bleeding and vasodilation are independent risk factors for submucosal carcinoma in CRC patients.The predictive model based on the above factors can provide reference for clinical diagnosis.更多还原