Department of Imaging, Medicine, and Biomedical Sciences, Cedars-Sinai Medical Center, 8700 Beverly Blvd, Los Angeles, CA 90048, USA.
Department of Nuclear Medicine, Taipei Veterans General Hospital, No. 201, Section 2, Shipai Rd, Taipei, Taiwan.
Eur Heart J Cardiovasc Imaging. 2020 May 1;21(5):549-559. doi: 10.1093/ehjci/jez177.
To optimize per-vessel prediction of early coronary revascularization (ECR) within 90 days after fast single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) using machine learning (ML) and introduce a method for a patient-specific explanation of ML results in a clinical setting.
A total of 1980 patients with suspected coronary artery disease (CAD) underwent stress/rest 99mTc-sestamibi/tetrofosmin MPI with new-generation SPECT scanners were included. All patients had invasive coronary angiography within 6 months after SPECT MPI. ML utilized 18 clinical, 9 stress test, and 28 imaging variables to predict per-vessel and per-patient ECR with 10-fold cross-validation. Area under the receiver operator characteristics curve (AUC) of ML was compared with standard quantitative analysis [total perfusion deficit (TPD)] and expert interpretation. ECR was performed in 958 patients (48%). Per-vessel, the AUC of ECR prediction by ML (AUC 0.79, 95% confidence interval (CI) [0.77, 0.80]) was higher than by regional stress TPD (0.71, [0.70, 0.73]), combined-view stress TPD (AUC 0.71, 95% CI [0.69, 0.72]), or ischaemic TPD (AUC 0.72, 95% CI [0.71, 0.74]), all P < 0.001. Per-patient, the AUC of ECR prediction by ML (AUC 0.81, 95% CI [0.79, 0.83]) was higher than that of stress TPD, combined-view TPD, and ischaemic TPD, all P < 0.001. ML also outperformed nuclear cardiologists' expert interpretation of MPI for the prediction of early revascularization performance. A method to explain ML prediction for an individual patient was also developed.
In patients with suspected CAD, the prediction of ECR by ML outperformed automatic MPI quantitation by TPDs (per-vessel and per-patient) or nuclear cardiologists' expert interpretation (per-patient).
利用机器学习（ML）优化快速单光子发射计算机断层扫描（SPECT）心肌灌注成像（MPI）后 90 天内早期冠状动脉血运重建（ECR）的血管预测，并引入一种在临床环境下对 ML 结果进行个体化解释的方法。
共纳入 1980 例疑似冠心病（CAD）患者，行 99mTc- sestamibi/tetrofosmin SPECT 负荷/静息 MPI 检查，均采用新一代 SPECT 扫描仪。所有患者在 SPECT MPI 后 6 个月内行有创冠状动脉造影。ML 利用 18 项临床、9 项负荷试验和 28 项影像学变量进行血管和患者水平的 ECR 预测，采用 10 倍交叉验证。比较了 ML 的接受者操作特征曲线下面积（AUC）与标准定量分析[总灌注缺损（TPD）]和专家解读的结果。共 958 例患者（48%）行 ECR。ML 预测 ECR 的血管 AUC（AUC 0.79，95%置信区间（CI）[0.77，0.80]）高于区域应激 TPD（0.71，[0.70，0.73]）、综合视图应激 TPD（AUC 0.71，95% CI [0.69，0.72]）或缺血性 TPD（AUC 0.72，95% CI [0.71，0.74]），均 P<0.001。ML 预测 ECR 的患者 AUC（AUC 0.81，95% CI [0.79，0.83]）也高于 TPD、综合视图 TPD 和缺血性 TPD 的应激，均 P<0.001。ML 对 ECR 预测的性能也优于核医学专家对 MPI 的解读。还开发了一种解释个体患者 ML 预测的方法。
在疑似 CAD 患者中，ML 预测 ECR 的性能优于 TPD（血管和患者水平）或核医学专家解读（患者水平）的自动 MPI 定量分析。
