目的:探讨影响急性脑梗死(ACI)预后的多模态CT影像学参数和临床危险因素,并基于此构建和验证Nomogram模型.方法:回顾性选取2020年6月至2022年8月在本院治疗的ACI患者155例,患者均随访1年统计预后情况.将155例患者按7∶3比例分为建模集(n=109)和验证集(n=46),根据建模集中患者预后情况分为预后良好组(n=74)和预后不良组(n=35).收集患者的CT血管造影(CTA)、CT灌注成像(CTP)结果和临床资料,通过单因素分析及多因素Logistic回归分析得出对应的CT影像学危险参数和临床危险因素,并基于回归分析法构建对应的预测模型,根据CT影像学危险参数和临床危险因素构建联合预测模型,使用R语言软件绘制对应的Nomogram图.采用受试者操作特征(ROC)曲线、校准曲线检验预测效能,临床决策曲线评价模型预测收益.使用验证集数据构建3种模型,绘制ROC曲线、校准曲线及决策曲线,对模型的预测效能进行外部验证.结果:联合模型的多因素Logistic回归分析结果显示,吸烟史、发病到开始治疗时间、白细胞介素(IL)-6、CTA结果、脑血容量(CBV)、脑血流量(CBF)、平均通过时间(MTT)、对比剂峰值时间(TTP)均与急性脑梗死不良预后存在明显关联,差异有统计学意义(P＜0.05).联合预测模型在建模集的ROC曲线下面积(AUC)为0.970,最佳截断值(0.480)对应的灵敏度、特异度分别为0.875、0.946,明显高于CT影像学参数模型和临床因素模型,校准曲线结果显示平均绝对误差(MAE)为0.049,联合模型预测概率与实际发生的概率之间差异不大,具有较好的临床实用性.临床决策曲线显示联合模型的临床收益较好.结论:基于多模态CT影像学参数和临床资料构建的ACI预后评估的Nomogram模型预测效能良好,可以准确预测急性脑梗死患者预后情况,为康复预后计划制定与医疗资源分配提供参考.
Objective:To investigate the multimodal CT imaging parameters and clinical risk factors affecting the prognosis of acute cerebral infarction(ACI),and establish a Nomogram model based on this analysis.Methods:A total of 155 ACI patients treated in our hospital from June 2020 to August 2022 were retrospectively selected.All patients were followed up for 1 year to evaluate their prognosis.155 patients were divided into modeling set(n=109)and validation set(n=46)according to a ratio of 7∶3.Patients in the modeling set were divided into good prognosis group(n=74)and bad prognosis group(n=35).The results of computer tomography angiography(CTA),computer tomography perfusion(CTP)and clinical data of the patient were collected.The corresponding CT imaging risk parameters and clinical risk factors were obtained through univariate analysis and multivariate Logistic regression analysis,and the corresponding prediction model was constructed based on the regression analysis.The combined prediction model was constructed according to the CT imaging risk parameters and clinical risk factors,and the corresponding Nomogram map was drawn using R language software.Receiver operating characteristic(ROC)curve and calibration curve were used to test the prediction efficiency.Clinical decision curve was used to evaluate the model's prediction benefits.Three models were constructed using validation set data,and ROC,calibration curve and decision curve were drawn.The prediction efficiency of the model is verified externally.Results:Multivariate Logistic regression analysis of the combined model showed that smoking history,time to onset of treatment,interleukin(IL)-6,CTA results,cerebral blood volume(CBV),cerebral blood flow(CBF),mean time to passage(mean)transit time(MTT)and time to peak(TTP)of contrast agent were significantly associated with adverse prognosis of acute cerebral infarction,and the difference was statistically significant(P＜0.05).The area under curve(AUC)of the combined prediction model in the modeling set was 0.970,and the sensitivity and specificity corresponding to the optimal cutoff value of 0.480 were 0.875 and 0.946,respectively,which were significantly higher than the CT imaging parameter model and clinical factor model.The calibration curve results showed that mean absolute error(MAE)was 0.049,and there was little difference between the predicted probability of the combined model and the actual probability,which had good clinical practicability.The clinical decision curve shows that the combined model has better clinical benefits.Conclusion:This Nomogram model based on multi-modal CT imaging parameters and clinical data can accurately predict the prognosis of patients with acute cerebral infarction,and provide references for the formulation of rehabilitation prognosis plan and allocation of medical resources.
田臻;杨扬;李明超
淮安市第五人民医院神经内科,江苏淮安 223300淮安市第一人民医院神经内科,江苏淮安 223300
灌注成像;血管造影;急性脑梗死;预后评估;列线图
perfusion imaging;angiography;acute cerebral infarction;prognosis assessment;Nomogram
《影像科学与光化学》 2024 (002)
104-112 / 9
淮安市自然科学研究计划项目(42)
10.7517/issn.1674-0475.231003
评论