基于超声心动图的急性STEMI患者近期心力衰竭智能预测系统构建及临床应用OA
Construction and Clinical Application of an Intelligent Prediction System for Short-Term Heart Failure in Patients with Acute STEMI Based on Echocar-diography
目的 构建基于超声心动图的急性ST段抬高型心肌梗死(STEMI)近期心力衰竭智能预测系统,并探讨其临床应用价值.方法 回顾性分析 2022 年 1 月至 2024 年 4 月我院收治的 300 例急性STEMI患者的临床资料,以 2∶1 比例随机将急性STE-MI患者进一步分为训练集(n=200)和验证集(n=100).所有患者均行超声心动图检查,统计其 3 个月内心力衰竭发生情况.采用GoogLeNet Inception-v3 架构构建心力衰竭智能预测系统,分析医师人工评估与智能预测系统对急性STEMI患者近期心力衰竭的预测效能,并采用Kappa检验两种预测方式与实际心力衰竭发生的一致性.结果 验证集中,医师人工评估预测急性STEMI患者近期心力衰竭的灵敏度、特异度、准确度分别为 72.00%、80.00%、78.00%,智能预测系统预测急性STEMI患者近期心力衰竭的灵敏度、特异度、准确度分别为 92.00%、88.00%、89.00%,智能预测系统的预测准确度明显高于医师人工评估预测(P＜0.05).Kappa检验结果显示,医师人工评估、智能预测系统对急性STEMI患者近期心力衰竭的预测结果均与实际心力衰竭发生的一致性较好(Kappa值=0.470、0.732,P＜0.001),其中智能预测系统的一致性程度更高.结论 基于超声心动图构建的智能预测系统对急性STEMI患者近期心力衰竭具有良好预测效能,且优于单纯医师人工评估,可用于急性STEMI早期心力衰竭高危人群筛查,为个体化防治提供客观影像学参考.
Objective To construct an intelligent prediction system for short-term heart failure in patients with acute ST-segment elevation myocardial infarction(STEMI)based on echocardiography,and to explore its clinical application value.Methods The clinical data of 300 patients with acute STEMI admitted to our hospital from January 2022 to April 2024 were retrospectively analyzed.The patients were randomly divided into a training set(n=200)and a validation set(n=100)at a ratio of 2∶1.All patients underwent echocardiographic examination,and the incidence of heart failure within 3 months was recorded.An intelligent heart failure prediction system was constructed using the GoogLeNet Inception-v3 architecture.The predictive performance of physician manual assessment and the intelligent prediction system for short-term heart failure in acute STEMI patients was analyzed,and the consistency between the two prediction methods and actual heart failure occurrence was evaluated using the Kappa test.Results In the validation set,the sensitivity,specificity,and accuracy of physician manual assessment for predicting short-term heart failure were 72.00%,80.00%,and 78.00%,respectively,while those of the intelligent prediction system were 92.00%,88.00%,and 89.00%,respectively.The predictive accuracy of the intelligent system was significantly higher than that of physician manual assessment(P<0.05).The Kappa test showed that both physician manual assessment and the intelligent prediction system had good consistency with actual heart failure occurrence(Kappa values=0.470 and 0.732,respectively;P<0.001),with the intelligent system demonstrating a higher degree of consistency.Conclusions The intelligent prediction system constructed based on echocardiography exhibits good predictive performance for short-term heart failure in acute STEMI patients,outperforming physician manual assessment alone.It can be used for early screening of high-risk heart failure populations in acute STEMI and provides objective imaging references for individualized prevention and treatment.
李慧;刘春节;倪文璐;李志娟
河南大学第一附属医院 超声医学科,河南 洛阳 471000河南大学第一附属医院 超声医学科,河南 洛阳 471000河南大学第一附属医院 超声医学科,河南 洛阳 471000河南大学第一附属医院 心血管内科,河南 洛阳 471000
医药卫生
急性ST段抬高型心肌梗死心力衰竭超声心动图智能预测系统
Acute ST-segment elevation myocardial infarctionHeart failureEchocardiographyIntelligent prediction system
《临床医学工程》 2026 (8)
921-925,5
洛阳市2022年度医疗卫生指导性科技计划项目(项目编号:2022055Y)
评论