Department of Radiology, Beijing Anzhen Hospital, Capital Medical University, 2(nd) Anzhen Road, Chaoyang District, Beijing, China.
Cardiovascular Research Centre, Royal Brompton Hospital, SW3 6NP, London, UK; National Heart and Lung Institute, Imperial College London, London, SW7 2AZ, UK.
Eur J Radiol. 2021 Jan;134:109420. doi: 10.1016/j.ejrad.2020.109420. Epub 2020 Nov 24.
To develop a fully automatic multiview shape constraint framework for comprehensive coronary artery calcium scores (CACS) quantification via deep learning on nonenhanced cardiac CT images.
In this retrospective single-centre study, a multi-task deep learning framework was proposed to detect and quantify coronary artery calcification from CT images collected between October 2018 and March 2019. A total of 232 non-contrast cardiac-gated CT scans were retrieved and studied (80 % for model training and 20 % for testing). CACS results of testing datasets (n = 46), including Agatston score, calcium volume score, calcium mass score, were calculated fully automatically and manually at total and vessel-specific levels, respectively.
No significant differences were found in CACS quantification obtained using automatic or manual methods at total and vessel-specific levels (Agatston score: automatic 535.3 vs. manual 542.0, P = 0.993; calcium volume score: automatic 454.2 vs. manual 460.6, P = 0.990; calcium mass score: automatic 128.9 vs. manual 129.5, P = 0.992). Compared to the ground truth, the number of calcified vessels can be accurate recognized automatically (total: automatic 107 vs. manual 102, P = 0.125; left main artery: automatic 15 vs. manual 14, P = 1.000 ; left ascending artery: automatic 37 vs. manual 37, P = 1.000; left circumflex artery: automatic 22 vs. manual 20, P = 0.625; right coronary artery: automatic 33 vs. manual 31, P = 0.500). At the patient's level, there was no statistic difference existed in the classification of Agatston scoring (P = 0.317) and the number of calcified vessels (P = 0.102) between the automatic and manual results.
The proposed framework can achieve reliable and comprehensive quantification for the CACS, including the calcified extent and distribution indicators at both total and vessel-specific levels.
通过深度学习在非增强心脏 CT 图像上开发一种用于全面冠状动脉钙评分（CACS）定量的全自动多视图形状约束框架。
在这项回顾性单中心研究中，提出了一种多任务深度学习框架，用于从 2018 年 10 月至 2019 年 3 月采集的 CT 图像中检测和量化冠状动脉钙化。共检索并研究了 232 例非对比心脏门控 CT 扫描（80%用于模型训练，20%用于测试）。分别在总水平和血管特异性水平上，全自动和手动计算测试数据集（n=46）的 CACS 结果，包括 Agatston 评分、钙体积评分、钙质量评分。
在总水平和血管特异性水平上，使用自动或手动方法获得的 CACS 定量无显著差异（Agatston 评分：自动 535.3 与手动 542.0，P=0.993；钙体积评分：自动 454.2 与手动 460.6，P=0.990；钙质量评分：自动 128.9 与手动 129.5，P=0.992）。与真实值相比，自动方法可以准确识别钙化血管的数量（总：自动 107 与手动 102，P=0.125；左主干：自动 15 与手动 14，P=1.000；左升主动脉：自动 37 与手动 37，P=1.000；左回旋支：自动 22 与手动 20，P=0.625；右冠状动脉：自动 33 与手动 31，P=0.500）。在患者水平，自动和手动结果在 Agatston 评分的分类（P=0.317）和钙化血管的数量（P=0.102）方面无统计学差异。
该框架可以可靠、全面地定量 CACS，包括总水平和血管特异性水平的钙化程度和分布指标。
