Liu Mengfei, Qi Zifan, Zhou Ren, Guo Chuanhai, Liu Anxiang, Yang Haijun, Li Fenglei, Duan Liping, Shen Lin, Wu Qi, Liu Zhen, Pan Yaqi, Liu Fangfang, Liu Ying, Chen Huanyu, Hu Zhe, Cai Hong, He Zhonghu, Ke Yang
Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department of Genetics, Peking University Cancer Hospital & Institute, Beijing, China.
Endoscopy Center, Anyang Cancer Hospital, Anyang, Henan Province, China.
Clin Transl Gastroenterol. 2025 May 1;16(5):e00835. doi: 10.14309/ctg.0000000000000835.
Image-based diagnostic tools that aid endoscopists to biopsy putative esophageal malignant lesions are essential for ensuring the standardization and quality of Lugol chromoendoscopy. But there is no such model available yet.
We developed a diagnostic model using endoscopic Lugol-unstained lesions (LULs) features and baseline data from 1,099 individuals enrolled from a large-scale population-based ESCC screening cohort. Six hundred three participants from a clinical outpatient cohort were included as the external validation set. High-grade intraepithelial neoplasia and above lesions identified at baseline or within 1 year after screening were defined as outcome. The final model was determined using logistic regression analysis by the Akaike information criterion.
The optimal diagnostic model contained the size, irregularity, sharp border of LUL, age, and body mass index of the participant, with the area under the curve of 0.83 (95% confidence interval [CI]: 0.78-0.87) in the development set, 0.81 (95% CI: 0.77-0.86) in the internal validation set, and 0.87 (95% CI: 0.84-0.90) in the external set. This model stratified individuals with LULs into low-risk, moderate-risk, and high-risk groups based on tertiles of predicted probabilities. The high-risk group accounted for <40% participants but enriched 80.8% and 82.7% of high-grade intraepithelial neoplasia and above cases in the development and external validation sets, respectively, achieving detection ratios 16.2 and 11.0 times higher than the low-risk group.
Our model can help maintain consistency and accuracy in detecting esophageal malignancy through Lugol chromoendoscopy, particularly in primary healthcare units in high-risk rural areas.
基于图像的诊断工具可辅助内镜医师对疑似食管恶性病变进行活检，这对于确保卢戈氏染色内镜检查的标准化和质量至关重要。但目前尚无此类模型。
我们利用来自大规模基于人群的食管癌筛查队列中1099名个体的内镜卢戈氏未染色病变（LULs）特征和基线数据开发了一种诊断模型。将来自临床门诊队列的603名参与者作为外部验证集。将基线时或筛查后1年内确定的高级别上皮内瘤变及以上病变定义为结局。通过赤池信息准则使用逻辑回归分析确定最终模型。
最佳诊断模型包含LULs的大小、不规则性、边界清晰度、参与者的年龄和体重指数，在开发集中曲线下面积为0.83（95%置信区间[CI]：0.78 - 0.87），内部验证集中为0.81（95%CI：0.77 - 0.86），外部集中为0.87（95%CI：0.84 - 0.90）。该模型根据预测概率的三分位数将患有LULs的个体分为低风险、中风险和高风险组。高风险组参与者占比<40%，但在开发集和外部验证集中分别富集了80.8%和82.7%的高级别上皮内瘤变及以上病例，检测率分别比低风险组高16.2倍和11.0倍。
我们的模型可通过卢戈氏染色内镜检查在检测食管恶性肿瘤方面帮助保持一致性和准确性，特别是在高风险农村地区的基层医疗单位。
