The Mina and Everard Goodman Faculty of Life Sciences, Bar-Ilan University, Ramat-Gan, Israel; Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel.
Sheba Medical Center affiliated with the Sackler School of Medicine, Tel-Aviv University, Ramat-Gan, Israel; The Mina and Everard Goodman Faculty of Life Sciences. Adult Bone Marrow Transplantation Service, Memorial Sloan Kettering Cancer Center, New York, NY, USA.
J Cardiol. 2021 Nov;78(5):439-446. doi: 10.1016/j.jjcc.2021.06.002. Epub 2021 Jun 19.
Various prognostic models for mortality prediction following ST-segment elevation myocardial infarction (STEMI) have been developed over the past two decades. Our group has previously demonstrated that machine learning (ML)-based models can outperform known risk scores for 30-day mortality post-STEMI. The study aimed to redevelop an ML-based random forest prediction model for 30-day mortality post-STEMI and externally validate it on a large cohort.
This was a retrospective, supervised learning, data mining study developed on the Acute Coronary Syndrome Israeli Survey (ACSIS) registry and the Myocardial Ischemia National Audit Project (MINAP) for external validation. Patients included received reperfusion therapy for STEMI between 2006 and 2016. Discrimination and calibration performances were assessed for two developed models and compared with the Global Registry of Acute Cardiac Events (GRACE) score.
The ACSIS cohort (2,782 included /15,212 total) and MINAP cohort (22,693 included/735,000 total) were significantly different in most variables, yet similar in 30-day mortality rate (4.3-4.4%). Random forest models were developed on the ACSIS cohort with a full model including all 32 variables and a simple model including the 10 most important ones. Features' importance was calculated using the varImp function measuring how much each feature contributes to the data's homogeneity. Applying the optimized models on the MINAP validation cohort showed high discrimination of area under the curve (AUC) = 0.804 (0.786-0.822) for the full model, and AUC = 0.787 (0.748-0.780) using the simple model, compared with the GRACE risk score discrimination of AUC = 0.764 (0.748-0.780). All models were not well calibrated for the MINAP data. Following Platt scaling on 20% of the MINAP data, the random forest models calibration improved while the GRACE calibration did not change.
The random forest predictive model for 30-day mortality post STEMI, developed on the ACSIS national registry, has been validated in the MINAP large external cohort and can be applied early at admission for risk stratification. The model performed better than the commonly used GRACE score. Furthermore, to the best of our knowledge, this is the first externally validated ML-based model for STEMI.
在过去的二十年中，已经开发出了各种用于预测 ST 段抬高型心肌梗死（STEMI）后死亡率的预后模型。我们的团队之前已经证明，基于机器学习（ML）的模型可以在 30 天后 STEMI 的死亡率方面优于已知的风险评分。本研究旨在重新开发一个基于 ML 的随机森林预测模型，用于预测 STEMI 后 30 天的死亡率，并在一个大型队列中进行外部验证。
这是一项回顾性、有监督的学习、数据挖掘研究，在急性冠状动脉综合征以色列调查（ACSIS）登记处和心肌缺血国家审计项目（MINAP）上进行开发，并在外部进行验证。纳入的患者在 2006 年至 2016 年间接受了 STEMI 的再灌注治疗。评估了两种开发模型的区分度和校准性能，并与全球急性心脏事件登记（GRACE）评分进行了比较。
ACSIS 队列（包括 2782 名患者/共 15212 名患者）和 MINAP 队列（包括 22693 名患者/共 735000 名患者）在大多数变量上存在显著差异，但 30 天死亡率相似（4.3-4.4%）。随机森林模型在 ACSIS 队列上进行了开发，全模型包括所有 32 个变量，简单模型包括 10 个最重要的变量。使用 varImp 函数计算特征的重要性，该函数测量每个特征对数据同质性的贡献程度。将优化后的模型应用于 MINAP 验证队列，全模型的曲线下面积（AUC）为 0.804（0.786-0.822），简单模型的 AUC 为 0.787（0.748-0.780），而 GRACE 风险评分的 AUC 为 0.764（0.748-0.780）。所有模型在 MINAP 数据上的校准效果均不理想。对 MINAP 数据的 20%进行 Platt 缩放后，随机森林模型的校准效果得到了改善，而 GRACE 模型的校准效果没有变化。
在 ACSIS 国家登记处开发的用于预测 STEMI 后 30 天死亡率的随机森林预测模型，已在 MINAP 大型外部队列中进行了验证，可在入院时早期用于风险分层。该模型的表现优于常用的 GRACE 评分。此外，据我们所知，这是第一个针对 STEMI 的经过外部验证的基于 ML 的模型。
