Assessment of fully automatic segmentation of pulmonary artery and aorta on noncontrast CT with optimal surface graph cuts

Abstract

Purpose Accurate segmentation of the pulmonary arteries and aorta is important due to the association of the diameter and the shape of these vessels with several cardiovascular diseases and with the risk of exacerbations and death in patients with chronic obstructive pulmonary disease. We propose a fully automatic method based on an optimal surface graph‐cut algorithm to quantify the full 3D shape and the diameters of the pulmonary arteries and aorta in noncontrast computed tomography (CT) scans. Methods The proposed algorithm first extracts seed points in the right and left pulmonary arteries, the pulmonary trunk, and the ascending and descending aorta by using multi‐atlas registration. Subsequently, the centerlines of the pulmonary arteries and aorta are extracted by a minimum cost path tracking between the extracted seed points, with a cost based on a combination of lumen intensity similarity and multiscale medialness in three planes. The centerlines are refined by applying the path tracking algorithm to curved multiplanar reformatted scans and are then smoothed and dilated nonuniformly according to the extracted local vessel radius from the medialness filter. The resulting coarse estimates of the vessels are used as initialization for a graph‐cut segmentation. Once the vessels are segmented, the diameters of the pulmonary artery (PA) and the ascending aorta (AA) and the PA:AA ratio are automatically calculated both in a single axial slice and in a 10 mm volume around the automatically extracted PA bifurcation level. The method is evaluated on noncontrast CT scans from the Danish Lung Cancer Screening Trial (DLCST). Segmentation accuracy is determined by comparing with manual annotations on 25 CT scans. Intraclass correlation (ICC) between manual and automatic diameters, both measured in axial slices at the PA bifurcation level, is computed on an additional 200 CT scans. Repeatability of the automated 3D volumetric diameter and PA:AA ratio calculations (perpendicular to the vessel axis) are evaluated on 118 scan–rescan pairs with an average in‐between time of 3 months. Results We obtained a Dice segmentation overlap of 0.94 ± 0.02 for pulmonary arteries and 0.96 ± 0.01 for the aorta, with a mean surface distance of 0.62 ± 0.33 mm and 0.43 ± 0.07 mm, respectively. ICC between manual and automatic in‐slice diameter measures was 0.92 for PA, 0.97 for AA, and 0.90 for the PA:AA ratio, and for automatic diameters in 3D volumes around the PA bifurcation level between scan and rescan was 0.89, 0.95, and 0.86, respectively. Conclusion The proposed automatic segmentation method can reliably extract diameters of the large arteries in non‐ECG‐gated noncontrast CT scans such as are acquired in lung cancer screening.

Finite element modeling with patient‐specific geometry to assess clinical risks of percutaneous pulmonary valve implantation

“…Improving the efficiency of image segmentation is the aim of many advanced computational studies that utilize machine learning and artificial intelligence tools. [33][34][35][36] Combining our FE workflow with automatic image segmentation would result in an improved clinical tool for assessing the risks of PPVI.…”

“…While our models have significantly faster run times than previous studies, image segmentation and model generation are still relatively slow and tedious processes. Improving the efficiency of image segmentation is the aim of many advanced computational studies that utilize machine learning and artificial intelligence tools 33–36 . Combining our FE workflow with automatic image segmentation would result in an improved clinical tool for assessing the risks of PPVI.…”

Pulmonary arteries segmentation from CT images using PA‐Net with attention module and contour loss

“…However, few studies on pulmonary artery segmentation in CT images. [19][20][21][22][23][24] Linguraru et al 19 proposed a semiautomatic tool for analyzing lung CTA, using level sets and geodesic active contours to segment pulmonary arteries. Zhang et al 20 proposed a pulmonary artery segmentation and arterial tree slice tracking algorithm based on region growth and slice marching (RGSM) to segment the pulmonary artery.…”

“…In traditional methods, vascular segmentation has been reported in extensive literature in medical image research, and a wide range of technologies have been used, such as threshold segmentation, 12,13 region growth, 14,15 active contour, 16–18 etc. However, few studies on pulmonary artery segmentation in CT images 19–24 . Linguraru et al 19 .…”

“…Liu et al 23 proposed a variable velocity geodesic active contour (GAC) model to drive the adaptive evolution of the curve inward or outward, and realize the automatic segmentation of the pulmonary artery. Sedghi et al 24 proposed an automatic method based on the optimal surface graph cutting algorithm, which makes segmentation of high-curvature surfaces possible.…”