3D Automatic Segmentation of Aortic Computed Tomography Angiography Combining Multi-View 2D Convolutional Neural Network
- PDF / 1,314,342 Bytes
- 11 Pages / 593.972 x 792 pts Page_size
- 71 Downloads / 195 Views
Cardiovascular Engineering and Technology (Ó 2020) https://doi.org/10.1007/s13239-020-00481-z
Original Article
3D Automatic Segmentation of Aortic Computed Tomography Angiography Combining Multi-View 2D Convolutional Neural Networks ALICE FANTAZZINI
,1,2 MARIO ESPOSITO,2 ALICE FINOTELLO,3 FERDINANDO AURICCHIO,4 BIANCA PANE,5 CURZIO BASSO,2 GIOVANNI SPINELLA,5 and MICHELE CONTI4
1
Department of Experimental Medicine, University of Genoa, Via Leon Battista Alberti, 2, 16132 Genoa, Italy; 2Camelot Biomedical Systems S.r.l, Via Al Ponte Reale, 2, 16124 Genoa, Italy; 3Department of Integrated Surgical and Diagnostic Sciences, University of Genoa, Genoa, Italy; 4Department of Civil Engineering and Architecture, University of Pavia, Pavia, Italy; and 5Vascular and Endovascular Surgery Unit, IRCCS Ospedale Policlinico San Martino, University of Genoa, Genoa, Italy (Received 8 April 2020; accepted 22 July 2020) Associate Editor Patrick Segers oversaw the review of this article.
Abstract Purpose—The quantitative analysis of contrast-enhanced Computed Tomography Angiography (CTA) is essential to assess aortic anatomy, identify pathologies, and perform preoperative planning in vascular surgery. To overcome the limitations given by manual and semi-automatic segmentation tools, we apply a deep learning-based pipeline to automatically segment the CTA scans of the aortic lumen, from the ascending aorta to the iliac arteries, accounting for 3D spatial coherence. Methods—A first convolutional neural network (CNN) is used to coarsely segment and locate the aorta in the whole sub-sampled CTA volume, then three single-view CNNs are used to effectively segment the aortic lumen from axial, sagittal, and coronal planes under higher resolution. Finally, the predictions of the three orthogonal networks are integrated to obtain a segmentation with spatial coherence. Results—The coarse segmentation performed to identify the aortic lumen achieved a Dice coefficient (DSC) of 0.92 ± 0.01. Single-view axial, sagittal, and coronal CNNs provided a DSC of 0.92 ± 0.02, 0.92 ± 0.04, and 0.91 ± 0.02, respectively. Multi-view integration provided a DSC of 0.93 ± 0.02 and an average surface distance of 0.80 ± 0.26 mm on a test set of 10 CTA scans. The generation of the ground truth dataset took about 150 h and the overall training process took 18 h. In prediction
Address correspondence to Alice Fantazzini, Department of Experimental Medicine, University of Genoa, Via Leon Battista Alberti, 2, 16132 Genoa, Italy. Electronic mail: [email protected], [email protected]
phase, the adopted pipeline takes around 25 ± 1 s to get the final segmentation. Conclusion—The achieved results show that the proposed pipeline can effectively localize and segment the aortic lumen in subjects with aneurysm. Keywords—Aorta segmentation, Deep learning, Convolutional neural network, Multi-view integration.
INTRODUCTION Abdominal aortic aneurysm (AAA) is a vascular disease involving pathologic dilatations of the abdominal aorta up to more than 3 c
Data Loading...
