| 摘要: |
| 目的：基于临床与超声组学特征建立预后模型并构建诺莫图(nomogram)评估急性心肌梗死后左心室壁瘤(PI-LVA)患者远期预后。 方法：回顾性分析遂宁市中心医院心血管内科2019年7月-2022年5月住院治疗的PI-LVA患者,结合纳入、排查标准最终153例入选并进行随访,终点为全因死亡。将临床变量纳入多因素Cox回归分析筛选与远期预后相关的危险因素。分析所有患者的超声心动图,采集心尖两腔心切面收缩期且包含完整室壁瘤部分的图像,使用ITK-SNAP进行分割获取感兴趣区域(ROI),使用pyradiomics提取超声组学特征,并采用最小绝对收缩和选择算子(Lasso)回归进行筛选。结合临床与超声组学特征构建模型并绘制nomogram,并进行可行性验证。 结果：多因素Cox回归分析显示：Tp-Te间期、右冠重度狭窄与LVEF<35%是影响PI-LVA患者远期预后的独立危险因素；总共从ROI中提取出107个特征,经过Lasso回归分析筛选出3个显著性特征。结合临床与影像组学筛选的6个变量构建预测模型并绘制nomogram,其1000次bootstrap校正的一致性指数为0.849[95%CI(0.769-0.896)]。多时点ROC曲线计算24个月、36个月、60个月曲线下面积为0.850、0.831、0.898,校准曲线显示nomogram预测风险与实际风险良好的一致性,决策曲线分析显示nomogram具有较好的临床效用。 结论：基于临床与超声组学特征构建的nomogram可有效地预测PI-LVA患者远期预后,为临床诊疗提供潜在的指导意义。 |
| 关键词: 急性心肌梗死 左心室壁瘤 超声组学 诺莫图 预后 |
| DOI： |
| 投稿时间：2024-03-24修订日期：2024-04-21 |
| 基金项目: |
|
| Construction of nomograms based on clinical and ultrasonic omics features to predict the prognosis of patients with left ventricular aneurysm after myocardial infarction |
| LI Tian Jiao,LIU Lei,LUO Cheng Lin,HOU Qin,YE Mao |
| (Department of cardiology,Suining Central Hospital) |
| Abstract: |
| Objective：A prognostic model was established based on clinical and ultrasonic omics features, and a nomogram was constructed to evaluate the long-term prognosis of patients with post-infarction left ventricular aneurysm (PI-LVA). Methods：Retrospective analysis of patients with PI-LVA admitted to the Department of Cardiovascular Medicine of Suining Central Hospital from July 2019 to May 2022. Combined with inclusion and exclusion criteria, 153 patients were finally selected and followed up, with the endpoint of all-cause mortality. Clinical variables were included in a multivariable Cox regression analysis to screen for risk factors associated with long-term prognosis. Echocardiograms of all patients were analyzed, and images of the apical two-chamber heart section during systole, including the complete wall aneurysm portion, were acquired. ITK-SNAP was used for segmentation to obtain the region of interest (ROI), and pyradiomics was used to extract ultrasonic omics features. Lasso regression was used for screening. A model was constructed based on clinical and ultrasonic omics features, and a nomogram was drawn for feasibility verification. Results：Multivariate Cox regression analysis showed that the Tp-Te interval, severe stenosis of the right coronary artery, and LVEF <35% were independent risk factors affecting the long-term prognosis of patients with PI-LVA. A total of 107 features were extracted from the ROI, and 3 significant features were selected by Lasso regression analysis. A prediction model was constructed based on 6 variables screened from clinical and imaging omics, and a nomogram was drawn. The consistency index of 1000 bootstrap corrections was 0.849 [95% CI (0.769-0.896)]. The area under the ROC curve at multiple time points of 24 months, 36 months, and 60 months was calculated to be 0.850, 0.831, and 0.898, respectively. The calibration curve showed good consistency between the predicted risk and the actual risk of the nomogram. The decision curve analysis showed that the nomogram had good clinical utility. Conclusion：The nomogram constructed based on clinical and ultrasonic omics features can effectively predict the long-term prognosis of patients with PI-LVA, providing potential guidance for clinical diagnosis and treatment. |
| Key words: Acute myocardial infarction Left ventricular aneurysm Ultrasonic proteomics Nomogram Prognosis |
