We present an open-source software pipeline to create patient-specific left atrial (LA) models with fibre orientations and a fibrosis map, suitable for electrophysiology simulations. The semi-automatic pipeline takes as input a contrast enhanced magnetic resonance angiogram, and a late gadolinium enhanced (LGE) contrast magnetic resonance (CMR). Five operators were allocated 20 cases each from a set of 50 CMR datasets to create a total of 100 models to evaluate inter/intra-operator variability. Each output model consisted of (1) a labelled surface mesh open at the pulmonary veins (PV) and mitral valve (MV), (2) fibre orientations mapped from a diffusion tensor MRI human atlas, (3) fibrosis map from the LGE-CMR scan, and (4) simulation of local activation time (LAT) and phase singularity (PS) mapping. We evaluated reproducibility in our pipeline by comparing agreement in shape of the output meshes, fibrosis distribution in the LA body, and fibre orientations; simulations outputs were evaluated comparing total activation times of LAT maps, mean conduction velocity (CV), and structural similarity index measure (SSIM) of PS maps. Our workflow allows a single model to be created in 16.72 +/- 12.25 minutes. Results in this abstract are reported as inter/intra. Shape only differed noticeably with users' selection of the MV and the length of the PV from the ostia to the distal end; fibrosis agreement (0.91/0.99 ICC) and fibre orientation agreement (60.63/71.77 %) were high. LAT maps showed good agreement, the median of the absolute difference of the total activation times was 2.02ms/1.37ms. The average of the mean CV difference was -4.04mm/s / 2.1mm/s. PS maps showed a moderately good agreement with SSIM of 0.648/0.608. Although we found notable differences in the models due to user input, our tests show that operator variability was comparable to that of image resolution or fibre estimation.

翻译：我们提出了一种开源软件流程，用于创建适用于电生理仿真的患者特异性左心房（LA）模型，包含纤维取向和纤维化图谱。该半自动流程以对比增强磁共振血管造影和延迟钆增强（LGE）磁共振成像（CMR）作为输入。五名操作员从50组CMR数据集中各分配20个病例，共创建100个模型以评估操作员间/操作员内变异性。每个输出模型包括：（1）肺静脉（PV）和二尖瓣（MV）处开口的标记表面网格；（2）基于弥散张量MRI人体图谱映射的纤维取向；（3）LGE-CMR扫描的纤维化图谱；（4）局部激活时间（LAT）和相位奇点（PS）映射的仿真。我们通过比较输出网格形状、左心房体纤维化分布及纤维取向的一致性来评估流程的可重复性；仿真输出通过比较LAT映射的总激活时间、平均传导速度（CV）及PS映射的结构相似性指数（SSIM）进行评估。该工作流可在16.72±12.25分钟内完成单个模型创建。本文结果以操作员间/操作员内形式报告。形状仅在用户选择二尖瓣及肺静脉从开口到远端的长度时出现显著差异；纤维化一致性（0.91/0.99 ICC）和纤维取向一致性（60.63/71.77%）均较高。LAT映射显示出良好一致性，总激活时间绝对值中位数差异为2.02毫秒/1.37毫秒。平均传导速度差异均值为-4.04毫米/秒/2.1毫米/秒。PS映射显示中等水平的一致性，SSIM为0.648/0.608。尽管我们发现模型因用户输入而存在显著差异，但测试表明操作员变异性与图像分辨率或纤维估计的变异性相当。