中老年患者结肠镜下息肉切除术后再检出 列线图预测模型的构建与验证

Construction and Validation of a Nomogram Prediction Model for Polyp Re-Detection after Endoscopic Resection in Middle-Aged and Elderly Patients

摘要: 目的：探讨中老年患者结肠镜下息肉切除术后再检出的危险因素，并建立列线图预测模型。方法：收集我院350例行结肠镜下息肉切除术的中老年患者的临床资料，按照7:3的比例随机分为训练集(244例)和验证集(106例)，随访至术后12个月，分别将两组数据集中息肉再检出患者纳入再检出组，其余纳入未再检出组。比较训练集中两组临床资料；经Logistic回归模型分析再检出组的影响因素，建立列线图预测模型并对其效能进行评价。结果：多因素Logistic回归分析结果表明，高血压病史、息肉病理类型、息肉位置、息肉数量及血清甘油三酯水平均为中老年患者结肠镜下息肉切除术后再检出的独立影响因素(P < 0.05)；列线图模型预测训练集息肉再检出的AUC为0.895，特异度为86.89%，敏感度为80.33%；预测验证集的AUC为0.805，特异度为70.67%，敏感度为80.65%；训练集、验证集的Calibration曲线经Hosmer-Lemeshow检验，差异均无统计学意义(训练集：χ² = 8.596，P = 0.378，验证集：χ² = 8.220，P = 0.412)；训练集、验证集分别在广泛风险阈值内获取临床净收益。结论：高血压病史、息肉病理类型、息肉位置、息肉数量及血清甘油三酯水平是中老年患者结肠镜下息肉切除术后再检出的独立危险因素，以此为根据建立的列线图模型对预测术后再检出具有良好的临床效能。

Abstract: Objective: To investigate risk factors for polyp re-detection after endoscopic resection of colorectal polyps in middle-aged and elderly patients, and to develop a nomogram-based prediction model. Methods: Clinical data of 350 middle-aged and elderly patients who underwent endoscopic resection of colorectal polyps in our hospital were collected. Patients were randomly assigned at a 7:3 ratio to a training set (n = 244) and a validation set (n = 106). All patients were followed up for 12 months postoperatively. In each dataset, patients with polyp re-detection were included in the re-detection group, and the remaining patients were included in the non-re-detection group. Clinical characteristics between the two groups in the training set were compared. Logistic regression analysis was performed to identify factors associated with postoperative polyp re-detection, and a nomogram prediction model was constructed and evaluated for performance. Results: Multivariable logistic regression analysis showed that a history of hypertension, polyp pathological type, polyp location, number of polyps, and serum triglyceride level were independent predictors of polyp re-detection after endoscopic resection in middle-aged and elderly patients (P < 0.05). The nomogram achieved an AUC of 0.895 in the training set, with a specificity of 86.89% and a sensitivity of 80.33%. In the validation set, the AUC was 0.805, with a specificity of 70.67% and a sensitivity of 80.65%. Calibration curves for both the training and validation sets showed no significant differences according to the Hosmer-Lemeshow test (training set: χ² = 8.596, P = 0.378; validation set: χ² = 8.220, P = 0.412). Decision curve analysis indicated that the model provided clinical net benefit across a wide range of risk thresholds in both datasets. Conclusion: A history of hypertension, polyp pathological type, polyp location, number of polyps, and serum triglyceride level are independent risk factors for polyp re-detection after endoscopic colorectal polypectomy in middle-aged and elderly patients. The nomogram developed based on these factors demonstrates good clinical utility for predicting postoperative polyp re-detection.