Department of Medicine, Division of Cardiology, University Medical Center Utrecht, Utrecht University, Heidelberglaan 100, 3584 CX, Utrecht, The Netherlands.
Department of Biomedical Engineering and Physics, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, The Netherlands.
Clin Res Cardiol. 2023 Mar;112(3):363-378. doi: 10.1007/s00392-022-02088-x. Epub 2022 Sep 6.
Arrhythmogenic right ventricular cardiomyopathy (ARVC) is diagnosed according to the Task Force Criteria (TFC) in which cardiovascular magnetic resonance (CMR) imaging plays an important role. Our study aims to apply an automatic deep learning-based segmentation for right and left ventricular CMR assessment and evaluate this approach for classification of the CMR TFC.
We included 227 subjects suspected of ARVC who underwent CMR. Subjects were classified into (1) ARVC patients fulfilling TFC; (2) at-risk family members; and (3) controls. To perform automatic segmentation, a Bayesian Dilated Residual Neural Network was trained and tested. Performance of automatic versus manual segmentation was assessed using Dice-coefficient and Hausdorff distance. Since automatic segmentation is most challenging in basal slices, manual correction of the automatic segmentation in the most basal slice was simulated (automatic). CMR TFC calculated using manual and automatic segmentation were compared using Cohen's Kappa (κ).
Automatic segmentation was trained on CMRs of 70 subjects (39.6 ± 18.1 years, 47% female) and tested on 157 subjects (36.9 ± 17.6 years, 59% female). Dice-coefficient and Hausdorff distance showed good agreement between manual and automatic segmentations (≥ 0.89 and ≤ 10.6 mm, respectively) which further improved after simulated correction of the most basal slice (≥ 0.92 and ≤ 9.2 mm, p < 0.001). Pearson correlation of volumetric and functional CMR measurements was good to excellent (automatic (r = 0.78-0.99, p < 0.001) and automatic (r = 0.88-0.99, p < 0.001) measurements). CMR TFC classification using automatic segmentations was comparable to manual segmentations (κ 0.98 ± 0.02) with comparable diagnostic performance.
Combining automatic segmentation of CMRs with correction of the most basal slice results in accurate CMR TFC classification of subjects suspected of ARVC.
心律失常性右室心肌病（ARVC）的诊断依据是工作组标准（TFC），心血管磁共振（CMR）成像在此标准中发挥着重要作用。我们的研究旨在应用一种基于深度学习的自动分割方法对左右心室 CMR 进行评估，并评估该方法对 CMR TFC 的分类能力。
我们纳入了 227 名疑似 ARVC 的患者，这些患者均接受了 CMR 检查。患者被分为（1）符合 TFC 的 ARVC 患者；（2）有患病风险的家族成员；（3）对照组。为了进行自动分割，我们训练和测试了一种贝叶斯扩张残差神经网络。使用 Dice 系数和 Hausdorff 距离评估自动分割与手动分割的性能。由于在基底切片中进行自动分割最具挑战性，因此模拟了对最基底切片中的自动分割进行手动校正（自动校正）。使用 Cohen's Kappa（κ）比较使用手动和自动分割计算的 CMR TFC。
我们在 70 名患者（39.6±18.1 岁，47%为女性）的 CMR 上进行了自动分割的训练，并在 157 名患者（36.9±17.6 岁，59%为女性）上进行了自动分割的测试。手动分割与自动分割之间的 Dice 系数和 Hausdorff 距离具有很好的一致性（分别≥0.89 和≤10.6mm），而在模拟校正最基底切片后，这种一致性进一步提高（分别≥0.92 和≤9.2mm，p<0.001）。容积和功能 CMR 测量的 Pearson 相关性很好到极好（自动（r=0.78-0.99，p<0.001）和自动（r=0.88-0.99，p<0.001）测量）。使用自动分割进行 CMR TFC 分类与手动分割相当（κ为 0.98±0.02），且具有相当的诊断性能。
将 CMR 的自动分割与最基底切片的校正相结合，可准确对疑似 ARVC 的患者进行 CMR TFC 分类。
