Machine Learning (ML) allows integration of the numerous variables delivered by cardiac PET/CT, while traditional survival analysis can provide explainable prognostic estimates from a restricted number of input variables. We implemented a ML-and-survival analysis of multimodal PET/CT data to identify patients who developed myocardial infarction (MI) or death in long-term follow up. Data from 739 intermediate risk patients who underwent coronary CT and selectively stress O-water-PET perfusion were analyzed for the occurrence of MI and all-cause mortality. Images were evaluated segmentally for atherosclerosis and absolute myocardial perfusion through 75 variables that were integrated through ML into an ML-CCTA and an ML-PET score. These scores were then modeled along with clinical variables through Cox regression. This model was compared against an expert interpretation-based and a calcium score-based model. Compared with expert- and calcium score-based models, the hybridized ML-survival model showed the highest performance (CI .81 vs .71 and .64). The strongest predictor for outcomes was the ML-CCTA score. Prognostic modeling of PET/CT data for the long-term occurrence of adverse events may be improved through ML imaging score integration and subsequent traditional survival analysis with clinical variables. This of methods offers an alternative to traditional survival modeling of conventional expert image scoring and interpretation.

中文翻译：

---

混合机器学习在心脏 PET/CT 成像生存分析中的应用

机器学习 (ML) 可以整合心脏 PET/CT 提供的众多变量，而传统的生存分析可以根据有限数量的输入变量提供可解释的预后估计。我们对多模态 PET/CT 数据进行了 ML 和生存分析，以确定在长期随访中发生心肌梗死 (MI) 或死亡的患者。对 739 名接受冠状动脉 CT 和选择性应激 O-水-PET 灌注的中危患者的数据进行了 MI 发生和全因死亡率的分析。通过 75 个变量对图像进行分段评估，以了解动脉粥样硬化和绝对心肌灌注，这些变量通过 ML 整合到 ML-CCTA 和 ML-PET 评分中。然后通过 Cox 回归对这些评分与临床变量进行建模。该模型与基于专家解释和基于钙评分的模型进行了比较。与基于专家和钙评分的模型相比，混合 ML 生存模型表现出最高的性能（CI 0.81 vs 0.71 和 0.64）。结果最强的预测指标是 ML-CCTA 评分。通过 ML 成像评分整合以及随后与临床变量的传统生存分析，可以改善 PET/CT 数据对长期发生不良事件的预后建模。这种方法提供了传统专家图像评分和解释的传统生存建模的替代方案。