| 王昆,陆小宁,臧其威.基于血清微核糖核酸联合检测构建非小细胞肺癌根治术后早期复发预测模型[J].安徽医药,2026,30(11):2205-2211. |
| 基于血清微核糖核酸联合检测构建非小细胞肺癌根治术后早期复发预测模型 |
| Construction of early recurrence prediction model for non-small cell lung cancer patients after radical surgery based on combined detection of serum miRNA |
| |
| DOI：10.3969/j.issn.1009-6469.2026.11.016 |
| 中文关键词: 癌，非小细胞肺 肺切除术 复发 微小核糖核酸 列线图 |
| 英文关键词: Carcinoma,non-small-cell lung Pneumonectomy Recurrence MicroRNA Nomogram |
| 基金项目:宿迁市第一人民医院科研项目( SY202207) |
|
| 摘要点击次数: 0 |
| 全文下载次数: 0 |
| 中文摘要: |
| 目的探讨基于血清微核糖核酸( miRNA)构建的预测模型对非小细胞肺癌( NSCLC)病人行肺癌根治术后早期复发预测性能。方法前瞻性研究 2023年 1月至 2024年 6月宿迁市第一人民医院收治的 217例 NSCLC病人作为研究对象，所有病人均行肺叶切除术，术后连续随访 12个月，收集病人临床资料和 miRNA，通过 LASSO回归筛选变量，并经多因素 logistic回归分析确定独立危险因素，基于相关危险因素分别构建临床模型和 miRNA模型，采用受试者操作特征曲线( ROC曲线)和决策曲线分析( DCA)验证模型的预测性能。结果纳入的 NSCLC病人术后复发率为 19.35%(42/217)。根据临床资料和血清 miRNA分别筛选出 8个和 11个关键变量纳入多因素 Cox回归分析， TNM分期、分化程度、纵隔淋巴结转移、肿瘤长径、癌胚抗原、糖类抗原125、细胞角蛋白 19可溶性片段、神经元特异性烯醇化酶为肺癌根治术后早期复发的临床危险因素(P<0.05)； miR-34a、miR-129、 miR-184、miR-200b、miR-638为肺癌根治术后早期复发的血清 miRNA危险因素( P<0.05)。 ROC曲线分析结果显示， miRNA模型的训练集 ROC曲线下面积( AUC)=0.98，95%CI:(0.96，0.99)测试集 AUC=0.96，95%CI:(0.93，1.00)，临床模型的 AUC=0.90， 95%CI:(0.86，0.95)，Delong检验结果显示， miRNA模型的训练集，和测试集 AUC均显著大于临床模型(Z=3.44、3.18，P<0.001)。校准曲线分析结果显示， miRNA模型的训练集校准斜率和 Brier评分分别为 0.92和 0.06，miRNA模型的测试集校准斜率和 Brier评分分别为 0.88和 0.08，临床模型校准斜率和 Brier评分分别为 0.85和 0.10。DCA曲线分析结果显示，在高风险阈值 0.05~0.95，采用 miRNA模型对病人进行预防性干预，相比均干预和未干预，可获得更高的标准化净效益；在高风险阈值 0.10~0.75，采用临床模型对病人进行预防性干预，相比均干预和未干预，可获得更高的标准化净效益，且 miRNA模型的标准化净效益高于临床模型。结论基于 miRNA构建的预测模型可准确预测 NSCLC病人行根治术后早期复发风险，且预测性能优于临床模型，可为病人治疗方案的优化与调整提供更可靠的参考。 |
| 英文摘要: |
| Objective To evaluate the predictive performance of a prediction model based on serum micro ribonucleic acid (miRNA)for early recurrence in patients withnon-small cell lung cancer (NSCLC)after radical resection.Methods A total of 217 patients withNSCLC admitted to Suqian First People's Hospital from January 2023 to June 2024 were enrolled as study subjects. All patients under-went lobectomy and were followed up continuously for 12 months after surgery. Clinical data and serum miRNA profiles were collected,and LASSO regression was used to screen variables. Multivariate logistic regression analysis was conducted to identify independent riskfactors. Based on the relevant risk factors, a clinical model and an miRNA-based model were constructed, and the predictive perfor-mance of the models was validated using receiver operating characteristic (ROC) curves and decision curve analysis (DCA). Results The postoperative recurrence rate of the NSCLC patients included in this study was 19.35% (42/217). Eight clinical variables and 11key serum miRNA variables were selected respectively and included in multivariate Cox regression analysis. TNM staging, degree ofdifferentiation , mediastinal lymph node metastasis, tumor diameter, carcinoembryonic antigen, carbohydrate antigen 125, cytokeratin19 fragment antigen 21-1, and neuron-specific enolase were identified as clinical risk factors for early recurrence of lung cancer after radical surgery (P<0.05); MiR-34a, miR-129, miR-184, miR-200b, and miR-638 were identified as serum miRNA risk factors for early recurrence of lung cancer after radical surgery (P<0.05). The ROC curve analysis results showed that the area under the curve (AUC) of the miRNA model was 0.98[95%CI:(0.96,0.99)]in the training set,and 0.96[95%CI:(0.93,1.00)] in the test set, while the AUC of the clin-ical model was 0.90[95%CI:(0.86,0.95)]. Delong test results showed that the AUCs of both the training and test sets of the miRNA mod-el were significantly greater than those of the clinical model (Z=3.44, 3.18, respectively; P<0.001). Calibration curve analysis resultsshowed that the calibration slope and Brier scoreswere 0.92 and 0.06 for the training set of the miRNA model, 0.88 and 0.08 for the testset of the miRNA model, and 0.85 and 0.10 for the clinical model, respectively. Decision curve analysis (DCA) results showed that with-in the high-risk threshold range of 0.05 to 0.95, preventive intervention guided by the miRNA model yielded a higher standardized netbenefit compared to both intervention and nonintervention; Within the high-risk threshold range of 0.10 to 0.75, the clinical modelalsoyielded a higher standardized net benefit, compared to both intervention and nonintervention, and the standardized net benefit of themiRNA models was greater than that of the clinical model.Conclusions The prediction model based on miRNA can accurately pre-dict the risk of early recurrence in patients with NSCLC after radical surgery, and its predictive performance is superior to that of theclinical model, providinga more reliable reference for optimizing and adjusting therapeutic regimens for patients. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
