Fully automatic segmentation of type B aortic dissection from CTA images enabled by deep learning

Deep learning-based algorithm provides an automated segmentation solution for type B aortic dissection (TBAD) on original CTAs. The serial multi-task based CNN achieved the best Dice coefficient scores (0.93 ± 0.01, 0.93 ± 0.01, and 0.91 ± 0.02 for the whole aorta, true lumen, and false lumen, respectively) and obtained an aortic lumen volume close to that of the ground truth, with an acceptable segmentation speed of 0.038 ± 0.006 s per slice. These findings indicate that the proposed method promises to enable clinicians to extract a large amount of morphological information accurately and rapidly in clinical settings, and greatly facilitate the TBAD anatomical features measurement process. TBAD = type B aortic dissection; CNN = convolutional neural network; CTA = computed tomography angiography. [Display omitted] •Lumen segmentation is the first step for measurement of type B aortic dissection.•Deep learning model can segment each lumen automatically on CTA.•The mean Dice coefficients of each lumen are above 90%.•Lumen volume obtained by deep learning lies within currently recognized variances.•Our preliminary work lays the foundation for automated measurement in the future. This study sought to establish a robust and fully automated Type B aortic dissection (TBAD) segmentation method by leveraging the emerging deep learning techniques. Preoperative CTA images of 276 patients with TBAD were retrospectively collected from January 2011 to December 2018. Using a reproducible manual segmentation protocol of three labels (whole aorta, true lumen (TL), and false lumen (FL)), a ground truth database (n = 276) was established and randomly divided into training and testing sets in a rough 8:1 ratio. Three convolutional neural network (CNN) models were developed on the training set (n = 246): single one-task (CNN1), single multi-task (CNN2), and serial multi-task (CNN3) models. Performance was evaluated using the Dice coefficient score (DCS) and lumen volume accuracy on the testing set (n = 30). Pearson correlation, Intra-class correlation coefficients and Bland–Altman plots were used to evaluate the inter-observer measurement agreement. CNN3 performed the best, with mean DCSs of 0.93 ± 0.01, 0.93 ± 0.01 and 0.91 ± 0.02 for the whole aorta, TL, and FL, respectively (p