ADSeg: A flap-attention-based deep learning approach for aortic dissection segmentation

Accurate and rapid segmentation of the lumen in an aortic dissection (AD) is an important prerequisite for risk evaluation and medical planning for patients with this serious condition. Although some recent studies have pioneered technical advances for the challenging AD segmentation task, they generally neglect the intimal flap structure that separates the true and false lumens. Identification and segmentation of the intimal flap may simplify AD segmentation, and the incorporation of long-distance z axis information interaction along the curved aorta may improve segmentation accuracy. This study proposes a flap attention module that focuses on key flap voxels and performs operations with long-distance attention. In addition, a pragmatic cascaded network structure with feature reuse and a two-step training strategy are presented to fully exploit network representation power. The proposed ADSeg method was evaluated on a multicenter dataset of 108 cases, with or without thrombus; ADSeg outperformed previous state-of-the-art methods by a significant margin and was robust against center variation.

Volume Analysis to Predict the Long-Term Evolution of Residual Aortic Dissection after Type A Repair

Background: The aim of this study was to evaluate the aortic diameter and volume during the first year after a type A repair to predict the long-term prognosis of a residual aortic dissection (RAD). Methods: All patients treated in our center for an acute type A dissection with a RAD and follow-up > 3 years were included. We defined two groups: group 1 with dissection-related events (defined as an aneurysmal evolution, distal reintervention, or aortic-related death) and group 2 without dissection-related events. The aortic diameters and volume analysis were evaluated on three postoperative CT scans: pre-discharge (T1), 3−6 months (T2) and 1 year (T3). Results: Between 2009 and 2016, 54 patients were included. Following a mean follow-up of 75.4 months (SD 31.5), the rate of dissection-related events was 62.9% (34/54). The total aortic diameters of the descending thoracic aorta were greater in group 1 at T1, T2 and T3, with greater diameters in the FL (p < 0.01). The aortic diameter evolution at 3 months was not predictive of long-term dissection-related events. The total thoracic aortic volume was significantly greater in group 1 at T1 (p < 0.01), T2 (p < 0.01), and T3 (p < 0.01). At 3 months, the increase in the FL volume was significantly greater in group 1 (p < 0.01) and was predictive for long-term dissection-related events. Conclusion: This study shows that an initial CT scan volume analysis coupled with another at 3 months is predictive for the long-term evolution in a RAD. Based on this finding, more aggressive treatment could be given at an earlier stage.

Automated segmentation and quantification of the healthy and diseased aorta in CT angiographies using a dedicated deep learning approach

OBJECTIVES

To develop and validate a deep learning-based algorithm for segmenting and quantifying the physiological and diseased aorta in computed tomography angiographies.

METHODS

CTA exams of the aorta of 191 patients (68.1 ± 14 years, 128 male), performed between 2015 and 2018, were retrospectively identified from our imaging archive and manually segmented by two investigators. A 3D U-Net model was trained on the data, which was divided into a training, a validation, and a test group at a ratio of 7:1:2. Cases in the test group (n = 41) were evaluated to compare manual and automatic segmentations. Dice similarity coefficient (DSC), mean surface distance (MSD), and Hausdorff surface distance (HSD) were extracted. Maximum diameter, effective diameter, and area were quantified and compared between both segmentations at eight anatomical landmarks, and at the maximum area of an aneurysms if present (n = 14). Statistics included error calculation, intraclass correlation coefficient, and Bland-Altman analysis.

RESULTS

A DSC of 0.95 [0.94; 0.95] and an MSD of 0.76 [0.06; 0.99] indicated close agreement between segmentations. HSD was 8.00 [4.47; 10.00]. The largest absolute errors were found in the ascending aorta with 0.8 ± 1.5 mm for maximum diameter and at the coeliac trunk with - 30.0 ± 81.6 mm² for area. Results for absolute errors in aneurysms were - 0.5 ± 2.3 mm for maximum diameter, 0.3 ± 1.6 mm for effective diameter, and 64.9 ± 114.9 mm² for area. ICC showed excellent agreement (> 0.9; p < 0.05) between quantitative measurements.

CONCLUSIONS

Automated segmentation of the aorta on CTA data using a deep learning algorithm is feasible and allows for accurate quantification of the aortic lumen even if the vascular architecture is altered by disease.

KEY POINTS

• A deep learning-based algorithm can automatically segment the aorta, mostly within acceptable margins of error, even if the vascular architecture is altered by disease. • Quantifications performed in the segmentations were mostly within clinically acceptable limits, even in pathologically altered segments of the aorta.

Association between aortic arch angulation and bird-beak configuration after thoracic aortic stent graft repair of type B aortic dissection

OBJECTIVES

The goal of this study was to investigate factors favouring the bird-beak configuration after thoracic endovascular aortic repair (TEVAR) for type B aortic dissection.

METHODS

We retrospectively analysed 76 patients with type B aortic dissection who underwent landing zone 1 and 2 TEVAR from December 2015 to January 2018. Preoperative aortic arch geometry (aortic arch length, maximal diameter and angulation), stent graft details and operative details were evaluated. A bird-beak configuration was defined as a ≥5-mm gap between the proximal end of the stent and the aortic wall of the lesser curvature.

RESULTS

Patients were stratified into those with (n = 46) and without (n = 30) a bird-beak configuration. The baseline demographics, dissection chronicity, clinical features and implanted devices were largely similar between the 2 groups. No significant difference was observed in the arch length or maximal arch diameter. However, the mean aortic arch angulation was greater in patients with than without a bird-beak configuration (61.4° vs 51.3°; P < 0.001). No influence of either the stent graft brand or the proximal stent graft type was observed. The multivariable analysis showed that the aortic arch angulation was an independent risk factor for a bird-beak configuration (odds ratio 1.15, 95% confidence interval 1.07-1.24; P < 0.001). A cut-off angle of 59.15° was predictive of a bird-beak configuration (sensitivity 59%; specificity 77%).

CONCLUSIONS

The preoperative aortic arch angulation was an independent predictor of a postoperative bird-beak configuration in patients with type B aortic dissection who underwent TEVAR that involved the aortic arch. An angle of >59.15° may imply a relatively hostile anatomy with a higher risk of a bird-beak configuration.

Comparison of diametric and volumetric changes in Stanford type B aortic dissection patients in assessing aortic remodeling post-stent graft treatment

Background

The study aims to analyze the correlation between the maximal diameter (both axial and orthogonal) and volume changes in the true (TL) and false lumens (FL) after stent-grafting for Stanford type B aortic dissection.

Method

Computed tomography angiography was performed on 13 type B aortic dissection patients before and after procedure, and at 6 and 12 months follow-up. The lumens were divided into three regions: the stented area (Region 1), distal to the stent graft to the celiac artery (Region 2), and between the celiac artery and the iliac bifurcation (Region 3). Changes in aortic morphology were quantified by the increase or decrease of diametric and volumetric percentages from baseline measurements.

Results

At Region 1, the TL diameter and volume increased (pre-treatment: volume =51.4±41.9 mL, maximal axial diameter =22.4±6.8 mm, maximal orthogonal diameter =21.6±7.2 mm; follow-up: volume =130.7±69.2 mL, maximal axial diameter =40.1±8.1 mm, maximal orthogonal diameter =31.9+2.6 mm, P<0.05 for all comparisons), while FL decreased (pre-treatment: volume =129.6±150.5 mL; maximal axial diameter =43.0±15.8 mm; maximal orthogonal diameter =28.3±12.6 mm; follow-up: volume =66.6±95.0 mL, maximal axial diameter =24.5±19.9 mm, maximal orthogonal diameter =16.9±13.7, P<0.05 for all comparisons). Due to the uniformity in size throughout the vessel, high concordance was observed between diametric and volumetric measurements in the stented region with 93% and 92% between maximal axial diameter and volume for the true/false lumens, and 90% and 92% between maximal orthogonal diameter and volume for the true/false lumens. Large discrepancies were observed between the different measurement methods at regions distal to the stent graft, with up to 46% differences between maximal orthogonal diameter and volume.

Conclusions

Volume measurement was shown to be a much more sensitive indicator in identifying lumen expansion/shrinkage at the distal stented region.

Twelve-Month Computed Tomography Follow-Up after Thoracic Endovascular Repair for Acute Complicated Aortic Dissection

BACKGROUND

To explore the impact of thoracic endovascular aortic repair (TEVAR) on aortic remodeling (AR) and the relationship between AR and complications after TEVAR.

METHODS

A total of 56 patients (2 type IIIA aortic dissection [AD] and 54 type IIIB AD) with complicated acute type B aortic dissection suitable for TEVAR were prospectively enrolled. There were 44 men (78%) and 12 women (22%) with an average age of 54 ± 13.8 years. Aortic enhanced computed tomography (CT) was performed pre-TEVAR and 3, 6, and 12 months postoperatively. The morphological changes in AR, namely aortic volume and false lumen thrombosis, were obtained by analyzing the CT data. The effect of TEVAR on AR was determined by the morphological changes in the aorta. The relationship between AR index, false lumen thrombosis, and complications was analyzed.

RESULTS

The volume of the thoracic aortic true lumen gradually increased post-TEVAR, whereas the volume of the thoracic aortic false lumen gradually decreased. The volume of abdominal aortic total lumen and false lumen increased 6 months postoperatively. The AR index increased significantly 3 months postoperatively, which was negatively correlated with complications and mortality. The thoracic and abdominal aortic false lumen thrombosis developed gradually after TEVAR, and the degree of thoracic aortic false lumen thrombosis was negatively correlated with complications and mortality.

CONCLUSIONS

TEVAR promotes AR. AR index and the degree of thoracic aortic false lumen thrombosis can serve as predictors of complications and mortality.

Critical appraisal of multidimensional CT measurements following acute open repair of type A aortic dissection

INTRODUCTION

To identify patients with aneurysmal degeneration of the native aorta following type A aortic dissection (TAAD), reproducible serial measurements of aortic dimensions are critical. We used a systematic workflow for measuring aortic geometry following TAAD, using computed tomography angiography data, and test its reproducibility.

METHODS

The workflow for aortic measurements included centerline generation, luminal diameter, and area measurement at six anatomically defined locations along the aorta and luminal volumetric measurements in the descending aorta. Two independent observers measured the aortic geometry in 20 surgically repaired TAAD patients, preoperatively and at 3 months follow-up. To test reproducibility, intraobserver and interobserver agreement scores were analyzed using a concordance correlation coefficient (CCC).

RESULTS

The interobserver agreement scores of the diameter, area, and volumetric measurements in the descending aorta were acceptable. The agreement scores of the area measurements were highest, with CCCs ranging from 0.909 to 0.984. Luminal diameter measurements scored lower than luminal area measurements and were least reproducible at the mid aortic arch (CCC < 0.886). Overall, intraobserver agreement scores were better than interobserver agreement scores (SD of mean difference was 1.89 vs 1.94 for intraobserver vs interobserver diameter measurements, and 0.61 vs 0.66 for area measurements).

CONCLUSION

Although overall reproducibility was acceptable in descending aortic measurements, our results show that it remains challenging to reliably measure luminal diameters, compared with areas. To aid identification of early adverse remodeling following acute TAAD, novel two- and three-dimensional measurement techniques are needed that capture locoregional changes in the false lumen and true lumen morphology more accurately.

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

PURPOSE

This study sought to establish a robust and fully automated Type B aortic dissection (TBAD) segmentation method by leveraging the emerging deep learning techniques.

METHODS

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.

RESULTS

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 < 0.05). Each label volume from CNN3 showed excellent agreement with the ground truth, with mean volume differences of -31.05 (-82.76 to 20.65) ml, 4.79 (-11.04 to 20.63) ml, and 8.67(-11.40 to 28.74) ml for the whole aorta, TL, and FL, respectively. The segmentation speed of CNN3 was 0.038 ± 0.006 s/image.

CONCLUSION

Deep learning-based model provides a promising approach for accurate and efficient segmentation of TBAD and makes it possible for automated measurements of TBAD anatomical features.

Current evidence in predictors of aortic growth and events in acute type B aortic dissection

OBJECTIVES

Acute type B aortic dissection can have a stable course or evolve into aneurysm and subsequent adverse events. The aim of this systematic review was to analyze the morphologic predictors of an adverse course to establish their validity based on consistency of results.

METHODS

Fifty-one studies were included in this review, reporting on aortic size, false lumen (FL) size, primary entry tear (ET) size and location, status of FL thrombosis, number of ETs, branch vessels involvement, and FL longitudinal extent.

RESULTS

Some predictors showed good consistency, whereas others did not. Aortic size was the most investigated predictor. A larger diameter at presentation predicted worse outcomes, with few exceptions. Both FL size and size relative to true lumen size also predicted an adverse course, although a standardized measurement method was not used. Regarding primary ET size and location, evidence was sparse and somewhat conflicting. Although FL complete thrombosis was consistently associated with a more benign course, the role of partial thrombosis remained unclear and the concept of FL saccular formation might account for the inconsistency, but further evidence is needed. A higher number of re-entry tears was considered to be protective against false channel expansion, but results need to be confirmed. The predictive role of branch vessels involvement and FL longitudinal extent remain controversial.

CONCLUSIONS

Among several predictors of aortic growth and events in acute type B aortic dissection, controversial and even conflicting results have been described. Consistent evidence has been demonstrated only for two predictors: aortic size at presentation is associated with adverse events and total FL thrombosis has a protective role.