转化医学杂志

目的 分析急性缺血性卒中（AIS）患者隐匿性癌症（OCC）的影响因素并建立预测模型。方法 选取150例AIS合并OCC患者作为OCC组，按照1∶1比例选取同期未合并OCC的AIS患者作为非OCC组，收集所有患者临床资料。采用LASSO回归筛选影响AIS患者OCC的因素，多因素条件Logistic回归确定其独立影响因素；RStudio软件建立AIS患者OCC的预测模型，一致性指数评估模型区分能力，校准曲线评估模型准确性，决策曲线评估模型临床效益，霍斯默-莱梅肖检验模型拟合优度，受试者工作特征（ROC）曲线分析模型预测能效。结果 多因素条件Logistic回归分析显示，AIS患者OCC的独立危险因素为不明原因型、多血管区域梗死、D-二聚体（D-D）高、纤维蛋白原（FIB）高，独立保护因素为高血压、血红蛋白高（P<0.05）。基于AIS患者OCC的独立影响因素建立预测模型，模型的一致性指数为0.894，预测模型的校准曲线走形接近理想曲线，决策曲线高于两条极端曲线，霍斯默-莱梅肖检验P>0.05。ROC曲线显示，该预测模型预测AIS患者OCC的曲线下面积为0.894。结论 不明原因型、多血管区域梗死、高血压、D-D、FIB、血红蛋白是AIS患者OCC的独立影响因素，基于此建立的预测模型对AIS患者OCC具有较高的预测能效。

Objective To analyze the factors influencing occult cancer(OCC) in patients with acute ischemic stroke(AIS) and develop a predictive model. Methods A total of 150 AIS patients with OCC were selected as the OCC group, and AIS patients without OCC, selected in a 1:1 ratio, were selected as the non-OCC group. Clinical data of all patients were collected. LASSO regression was used to screen for factors influencing OCC in AIS patients, and multivariate conditional logistic regression was applied to identify independent influencing factors. The prediction model was developed using RStudio software,with the concordance index assessing the model's discriminative ability, the calibration curve evaluating model accuracy, the decision curve analyzing clinical benefit, the Hosmer-Lemeshow test assessing model fit, and the receiver operating characteristic(ROC) curve analyzing predictive performance. Results Multivariate conditional logistic regression analysis identified stroke of undetermined etiology, multiple vascular territory infarctions, elevated D-dimer(D-D), and high fibrinogen(FIB) as independent risk factors for OCC in AIS patients, while hypertension and elevated hemoglobin were independent protective factors(P<0.05). A predictive model was constructed based on these independent influencing factors, with a concordance index of 0.894. The calibration curve closely aligned with the ideal curve, and the decision curve was superior to the two extreme curves.The Hosmer-Lemeshow test yielded P>0.05. The ROC curve indicated that the area under the curve(AUC) for predicting OCC in AIS patients was 0.894. Conclusion Stroke of undetermined etiology, multiple vascular territory infarctions, hypertension,D-D,FIB,and hemoglobin are independent influencing factors for OCC in AIS patients.The developed prediction model demonstrates high predictive efficacy.