目的 探讨蓝激光成像放大内镜（BLI-ME）联合JNET分型对结直肠肿瘤性病变的诊断价值。 方法 回顾性分析2018年12月至2021年12月在华中科技大学同济医学院附属武汉中心医院接受BLI-ME检查的结直肠新生性病变患者的内镜图片，内镜医师根据JNET分型对上述病变图片进行分析并初步诊断病变的病理性质，后与内镜下处理或外科手术切除后标本的病理结果进行对比分析。 结果 入选93例患者共174处结直肠新生性病变，包括非肿瘤性病变（增生性病变及炎性病变）64处、低级别上皮内瘤变72处、高级别上皮内瘤变及黏膜下浅层浸润癌28处、黏膜下深层浸润癌10处。在BLI-ME下，JNET分型总体诊断结直肠肿瘤性病变的敏感度、特异度、阳性预测值、阴性预测值、准确率分别为93.6%、90.6%、94.5%、89.2%、92.5%；按病变大小<10 mm、10~20 mm、≥20 mm进行分类，三者分别对应的JNET分型总体诊断结直肠肿瘤性病变的准确率分别为91.0%、97.7%、96.8%，差异无统计学意义（P>0.05）。 结论 BLI-ME下应用JNET分型对结直肠肿瘤性与非肿瘤性病变具有较好的鉴别诊断能力，具有临床推广应用价值。 Objective To study the diagnostic value of Japan NBI Expert Team (JNET) classification combined with blue laser imaging magnifying endoscopy (BLI-ME) for colorectal neoplastic lesions. Methods A retrospective analysis was performed for endoscopic images taken during BLI-ME in patients with colorectal neoplastic lesions who underwent the procedure at the Central Hospital of Wuhan affiliated to Huazhong University of Science and Technology Tongji Medical College between December 2018 and December 2021. These images were reviewed and preliminarily decided on histological nature by endoscopists according to JNET classification, and then the results were compared with the findings of endoscopic or post-surgical pathology. Results A total of 174 lesions from 93 patients were studied, including 64 non-neoplastic (hyperplastic and inflammatory) lesions, 72 lesions with low grade intraepithelial neoplasia, 28 lesions with high grade intraepithelial neoplasia or superficial submucosal invasive cancers, and 10 with deep submucosal invasive cancers. The diagnostic sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of overall JNET classification with BLI-ME for colorectal neoplastic lesions was 93.6%, 90.6%, 94.5%, 89.2%, and 92.5%, respectively. Stratified by lesion size below 10 mm, 10 to 19 mm, and ≥ 20mm, the overall diagnostic accuracy of JNET classification was 91.0%, 97.7%, and 96.8%, respectively, showing no significant difference (P>0.05) . Conclusion JNET classification with BLI-ME may effectively distinguish between neoplastic and non-neoplastic colorectal lesions, and thus warrants widespread use in clinical practice.
Abstract
Objective To study the diagnostic value of Japan NBI Expert Team (JNET) classification combined with blue laser imaging magnifying endoscopy (BLI-ME) for colorectal neoplastic lesions. Methods A retrospective analysis was performed for endoscopic images taken during BLI-ME in patients with colorectal neoplastic lesions who underwent the procedure at the Central Hospital of Wuhan affiliated to Huazhong University of Science and Technology Tongji Medical College between December 2018 and December 2021. These images were reviewed and preliminarily decided on histological nature by endoscopists according to JNET classification, and then the results were compared with the findings of endoscopic or post-surgical pathology. Results A total of 174 lesions from 93 patients were studied, including 64 non-neoplastic (hyperplastic and inflammatory) lesions, 72 lesions with low grade intraepithelial neoplasia, 28 lesions with high grade intraepithelial neoplasia or superficial submucosal invasive cancers, and 10 with deep submucosal invasive cancers. The diagnostic sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of overall JNET classification with BLI-ME for colorectal neoplastic lesions was 93.6%, 90.6%, 94.5%, 89.2%, and 92.5%, respectively. Stratified by lesion size below 10 mm, 10 to 19 mm, and ≥ 20mm, the overall diagnostic accuracy of JNET classification was 91.0%, 97.7%, and 96.8%, respectively, showing no significant difference (P>0.05) . Conclusion JNET classification with BLI-ME may effectively distinguish between neoplastic and non-neoplastic colorectal lesions, and thus warrants widespread use in clinical practice.
万方数据