Graph neural networks for automatic extraction and labeling of the coronary artery tree in CT angiography.

Purpose: Automatic comprehensive reporting of coronary artery disease (CAD) requires anatomical localization of the coronary artery pathologies. To address this, we propose a fully automatic method for extraction and anatomical labeling of the coronary artery tree using deep learning. Approach: We include coronary CT angiography (CCTA) scans of 104 patients from two hospitals. Reference annotations of coronary artery tree centerlines and labels of coronary artery segments were assigned to 10 segment classes following the American Heart Association guidelines. Our automatic method first extracts the coronary artery tree from CCTA, automatically placing a large number of seed points and simultaneous tracking of vessel-like structures from these points. Thereafter, the extracted tree is refined to retain coronary arteries only, which are subsequently labeled with a multi-resolution ensemble of graph convolutional neural networks that combine geometrical and image intensity information from adjacent segments. Results: The method is evaluated on its ability to extract the coronary tree and to label its segments, by comparing the automatically derived and the reference labels. A separate assessment of tree extraction yielded an F1 score of 0.85. Evaluation of our combined method leads to an average F1 score of 0.74. Conclusions: The results demonstrate that our method enables fully automatic extraction and anatomical labeling of coronary artery trees from CCTA scans. Therefore, it has the potential to facilitate detailed automatic reporting of CAD.

Reconstruction and completion of high-resolution 3D cardiac shapes using anisotropic CMRI segmentations and continuous implicit neural representations.

Since the onset of computer-aided diagnosis in medical imaging, voxel-based segmentation has emerged as the primary methodology for automatic analysis of left ventricle (LV) function and morphology in cardiac magnetic resonance images (CMRI). In standard clinical practice, simultaneous multi-slice 2D cine short-axis MR imaging is performed under multiple breath-holds resulting in highly anisotropic 3D images. Furthermore, sparse-view CMRI often lacks whole heart coverage caused by large slice thickness and often suffers from inter-slice misalignment induced by respiratory motion. Therefore, these volumes only provide limited information about the true 3D cardiac anatomy which may hamper highly accurate assessment of functional and anatomical abnormalities. To address this, we propose a method that learns a continuous implicit function representing 3D LV shapes by training an auto-decoder. For training, high-resolution segmentations from cardiac CT angiography are used. The ability of our approach to reconstruct and complete high-resolution shapes from manually or automatically obtained sparse-view cardiac shape information is evaluated by using paired high- and low-resolution CMRI LV segmentations. The results show that the reconstructed LV shapes have an unconstrained subvoxel resolution and appear smooth and plausible in through-plane direction. Furthermore, Bland-Altman analysis reveals that reconstructed high-resolution ventricle volumes are closer to the corresponding reference volumes than reference low-resolution volumes with bias of [limits of agreement] -3.51 [-18.87, 11.85] mL, and 12.96 [-10.01, 35.92] mL respectively. Finally, the results demonstrate that the proposed approach allows recovering missing shape information and can indirectly correct for limited motion-induced artifacts.

Learning coronary artery calcium scoring in coronary CTA from non-contrast CT using unsupervised domain adaptation.

Deep learning methods have demonstrated the ability to perform accurate coronary artery calcium (CAC) scoring. However, these methods require large and representative training data hampering applicability to diverse CT scans showing the heart and the coronary arteries. Training methods that accurately score CAC in cross-domain settings remains challenging. To address this, we present an unsupervised domain adaptation method that learns to perform CAC scoring in coronary CT angiography (CCTA) from non-contrast CT (NCCT). To address the domain shift between NCCT (source) domain and CCTA (target) domain, feature distributions are aligned between two domains using adversarial learning. A CAC scoring convolutional neural network is divided into a feature generator that maps input images to features in the latent space and a classifier that estimates predictions from the extracted features. For adversarial learning, a discriminator is used to distinguish the features between source and target domains. Hence, the feature generator aims to extract features with aligned distributions to fool the discriminator. The network is trained with adversarial loss as the objective function and a classification loss on the source domain as a constraint for adversarial learning. In the experiments, three data sets were used. The network is trained with 1,687 labeled chest NCCT scans from the National Lung Screening Trial. Furthermore, 200 labeled cardiac NCCT scans and 200 unlabeled CCTA scans were used to train the generator and the discriminator for unsupervised domain adaptation. Finally, a data set containing 313 manually labeled CCTA scans was used for testing. Directly applying the CAC scoring network trained on NCCT to CCTA led to a sensitivity of 0.41 and an average false positive volume 140 mm3/scan. The proposed method improved the sensitivity to 0.80 and reduced average false positive volume of 20 mm3/scan. The results indicate that the unsupervised domain adaptation approach enables automatic CAC scoring in contrast enhanced CT while learning from a large and diverse set of CT scans without contrast. This may allow for better utilization of existing annotated data sets and extend the applicability of automatic CAC scoring to contrast-enhanced CT scans without the need for additional manual annotations. The code is publicly available at https://github.com/qurAI-amsterdam/CACscoringUsingDomainAdaptation.

Clinical impact of assessing thrombus age using magnetic resonance venography prior to catheter-directed thrombolysis.

OBJECTIVES: Magnetic resonance venography (MRV) is underutilized in the evaluation of thrombus properties prior to endovascular treatment but may improve procedural outcomes. We therefore investigated the clinical impact of using a dedicated MRV scoring system to assess thrombus characteristics prior to endovascular intervention for iliofemoral deep vein thrombosis (DVT). METHODS: This is a post hoc analysis of data from the CAVA trial ( Clinicaltrials.gov :NCT00970619). MRV studies of patients receiving ultrasound-accelerated catheter-directed thrombolysis (CDT) for iliofemoral DVT were reviewed. Thrombus age-related imaging characteristics were scored and translated into an overall score (acute, subacute, or old). MRV scores were compared to patient-reported complaints. MRV-scored groups were compared for CDT duration and success rate. RESULTS: Fifty-six patients (29 men; age 50.8 ± 16.4 years) were included. Using MRV, 27 thrombi were classified acute, 17 subacute, and 12 old. Based on patient-reported complaints, 11 (91.7%) of these old thrombi would have been categorized acute or subacute, and one (3.7%) of the acute thrombi as old. Average duration of CDT to > 90% restored patency differed significantly between groups (p < 0.0001): average duration was 23 h for acute thromboses (range: 19-25), 43 h for subacute (range: 41-62), and 85 h for old thromboses (range: 74-96). CDT was almost eleven times more successful in thromboses characterized as acute and subacute compared to old thromboses (OR: 10.7; 95% CI 2.1-55.5). CONCLUSION: A dedicated MRV scoring system can safely discriminate between acute, subacute, and old thromboses. MRV-based selection is predictive of procedural duration and success rate and can help avoid unnecessary complications. KEY POINTS: • Thrombus age, characterized by MRV as acute, subacute, and old, can predict CDT duration and probability of success. • Accurate pre-interventional MRV-based thrombus aging has the potential to facilitate identification of eligible patients and may thus prevent CDT-related complications.

The Impact of Percutaneous Coronary Intervention on Mortality in Patients With Coronary Lesions Who Underwent Transcatheter Aortic Valve Replacement.

OBJECTIVES: To analyze the effect of percutaneous coronary intervention (PCI) before transcatheter aortic valve replacement (TAVR) on all-cause and cardiovascular mortality after TAVR, differentiating between significant proximal lesions and the non-proximal (residual) lesions. METHODS: An institutional TAVR database was complemented with data on the extent of coronary artery disease (CAD), lesion location, lesion severity, and the location of PCI. Survival analysis was performed to investigate the impact on 6-month and 3-year mortality after TAVR in all patients and in subgroups of patients with significant proximal lesions (>70% diameter stenosis [DS], >50% DS in left main), the non-proximal residual lesions, and in a propensity score matched cohort. RESULTS: Among the 577 included patients, mean age was 83 years, 50% were female, and 31% had diabetes mellitus. Preprocedural PCI of unselected lesions was independently associated with increased 6-month mortality (hazard ratio, 2.2; 95% confidence interval, 1.0-4.6; P=.04), but selective PCI of significant proximal lesions did not have an association with higher mortality, nor did we find a significant effect of PCI on mortality in the propensity-matched cohort. CONCLUSION: Routine pre-TAVR PCI is not associated with mortality reduction in TAVR patients with coronary lesions in any segment or in patients with proximal coronary lesions. Despite the lack of a beneficial effect of routine pre-TAVR PCI, we cannot exclude a beneficial effect in a selection of patients with proximal lesions. Therefore, we strongly support the current clinical guidelines to only consider pre-TAVR PCI in proximal coronary lesions, while advocating a restrictive pre-TAVR PCI strategy.

Corrigendum: CT Angiography or Cardiac MRI for Detection of Coronary Artery Aneurysms in Kawasaki Disease.

[This corrects the article DOI: 10.3389/fped.2021.630462.].

CT Angiography or Cardiac MRI for Detection of Coronary Artery Aneurysms in Kawasaki Disease.

Background: Kawasaki disease (KD) is an acute vasculitis that mainly affects the coronary arteries. This inflammation can cause coronary artery aneurysms (CAAs). Patients with KD need cardiac assessment for risk stratification for the development of myocardial ischemia, based on Z-score (luminal diameter of the coronary artery corrected for body surface area). Echocardiography is the primary imaging modality in KD but has several important limitations. Coronary computed tomographic angiography (cCTA) and Cardiac MRI (CMR) are non-invasive imaging modalities and of additional value for assessment of CAAs with a high diagnostic yield. The objective of this single center, retrospective study is to explore the diagnostic potential of coronary artery assessment of cCTA vs. CMR in children with KD. Methods and Results: Out of 965 KD patients from our database, a total of 111 cCTAs (104 patients) and 311 CMR (225 patients) have been performed since 2010. For comparison, we identified 54 KD patients who had undergone both cCTA and CMR. CMR only identified eight patients with CAAs compared to 14 patients by cCTA. CMR missed 50% of the CAAs identified by cCTA. Conclusions: Our single center study demonstrates that cCTA may be a more sensitive diagnostic tool to detect CAAs in KD patients, compared to CMR.

Association of Successful Ultrasound-Accelerated Catheter-Directed Thrombolysis with Postthrombotic Syndrome: A Post Hoc Analysis of the CAVA Trial.

BACKGROUND: The CAVA trial did not show the anticipated risk reduction for postthrombotic syndrome (PTS) after thrombus removal via additional ultrasound-accelerated catheter-directed thrombolysis (UACDT) in patients with acute iliofemoral deep vein thrombosis (IFDVT). Difficulties in achieving an effective degree of recanalization through thrombolysis may have influenced outcomes. We therefore assessed whether successful UACDT (restored patency ≥ 90%) did reduce the development of PTS. METHODS: This CAVA trial post hoc analysis compared the proportion of PTS at 1-year follow-up between patients with successful UACDT and patients that received standard treatment only. In addition, clinical impact as well as determinants of successful thrombolysis were explored. RESULTS: UACDT was initiated in 77 (50.7%) patients and considered successful in 41 (53.2%, interrater agreement κ = 0.7, 95% confidence interval 0.47-0.83). PTS developed in 15/41 (36.6%) patients in the successful UACDT group versus 33/75 (44.0%) controls (p = 0.44). In this comparison, successful UACDT was associated with lower Venous Clinical Severity Score (3.50 ± 2.57 vs. 4.82 ± 2.74, p = 0.02) and higher EuroQOL-5D (EQ-5D) scores (40.2 ± 36.4 vs. 23.4 ± 34.4, p = 0.01). Compared with unsuccessful UACDT, successful UACDT was associated with a shorter symptom duration at inclusion (p = 0.05), and higher rates of performed adjunctive procedures (p < 0.001) and stent placement (p < 0.001). CONCLUSION: Successful UACDT was not associated with a reduced proportion of PTS 1 year after acute IFDVT compared with patients receiving standard treatment alone. There was, however, a significant reduction in symptom severity and improvement of generic quality of life according to the EQ-5D. Better patient selection and optimization of treatment protocols are needed to assess the full potential of UACDT for the prevention of PTS. TRIAL REGISTRATION NUMBER: ClinicalTrials.gov number, NCT00970619.

Minimal invasive aortic valve replacement: associations of radiological assessments with procedure complexity.

OBJECTIVES: Limited aortic annulus exposure during minimal invasive aortic valve replacement (mini-AVR) proves to be challenging and contributes to procedure complexity, resulting in longer procedure times. New innovations like sutureless valves have been introduced to reduce procedure complexity. Additionally, preoperative imaging could also contribute to reducing procedure times. Therefore, we hypothesize that Computed Tomography (CT)-image based measurements are associated with mini-AVR complexity. METHODS: One hundred patients who underwent a mini-sternotomy and had a preoperative CT scan were included. With a CT-based mini-AVR planning tool, we measured access distance, access angle, annulus dimensions, and calcium volume. The associations of these measurements with cardiopulmonary bypass (CPB) time and aortic cross-clamp (AoX) time were assessed using univariable and multivariable regression models. In the multivariable models, these measurements were adjusted for age and suture technique. RESULTS: In the univariable regression models, calcium volume and annulus dimensions were associated with longer CPB and AoX time. After adjusting for age and suture technique, increasing calcium volume was still associated with longer CPB (adjusted ß-coefficient 0.002, 95%-CI (0.005, 0.019), p-value = 0.002) and AoX time (adjusted ß-coefficient 0.010, 95%-CI (0.004, 0.016), p-value = 0.002). However, after adjusting for these confounders, the association between annulus dimensions and procedure times lost statistical significance. CONCLUSION: Increase in calcium volume are associated with longer CPB and AoX times, with age and sutureless valve implantation as independent confounders. In contrast to previous studies, access angle was not associated with procedure complexity.

Complementary role of cardiac computed tomography angiography in the diagnosis of prosthetic aortic valve endocarditis and septic coronary embolism - a case report.

A 73-year old man presented with a posterolateral ST-elevated myocardial infarction 9 months after biological aortic valve replacement for aortic valve stenosis. Invasive coronary angiography showed a filling defect across the left main coronary artery bifurcation extending into the left anterior descending artery and the ramus circumflex. Transthoracic echocardiography revealed a thickened prosthesis leaflet with signs of slight stenosis. Cardiac computed tomography angiography showed a mass on the left coronary cusp of the valve prosthesis, suggestive for vegetation or thrombus. The scan also revealed central luminal filling defects, indicative for thrombus or septic emboli. Blood cultures proved positive for Propionibacterium acnes, therefore the patient was treated for prosthetic valve endocarditis. Computed tomography angiography offers high diagnostic accuracy for detecting infective endocarditis and renders complementary information about valvular anatomy, coronary artery disease and the extension of infections.

A computed tomography-based planning tool for predicting difficulty of minimally invasive aortic valve replacement.

OBJECTIVES: Minimally invasive aortic valve replacement has proven its value over the last decade by its significant advancement and reduction in mortality, morbidity and admission time. However, minimally invasive aortic valve replacement is associated with some on-site difficulties such as limited aortic annulus exposure. Currently, computed tomography scans are used to evaluate the anatomical relationship among the intercostal spaces, ascending aorta and aortic valve prior to surgery. We hypothesized that quantitative measurements of access distance and access angle are associated with outcome and access difficulty. METHODS: We introduce a novel minimally invasive aortic valve replacement planning prototype that allows automatic measurements of access angle, access distance and aortic annulus dimensions. The prototype visualizes these measurements on the chest cage as ISO contours. The association of these measures with outcome parameters such as extracorporeal circulation time, aortic cross-clamping time and access difficulty score was assessed. We included 14 patients who received a new valve by ministernotomy. RESULTS: The mean access angle was 40.3 ± 5.1°. It was strongly associated with aortic cross-clamping time (Pearson correlation coefficient = 0.60, P = 0.02) and access difficulty score (Spearman's rank correlation coefficient = 0.57, P = 0.03). Access angles were significantly different between easy and difficult access groups (P = 0.03). There was no significant association between access distance and outcome parameters. CONCLUSIONS: Access angle is strongly associated with procedure complexity. The automated presentation of this measure suggests added value of the prototype in clinical practice.

Automated CTA based measurements for planning support of minimally invasive aortic valve replacement surgery.

Minimally invasive aortic valve replacement (mini-AVR) procedures are a valuable alternative to conventional open heart surgery. Currently, planning of mini-AVR consists of selection of the intercostal space closest to the sinotubular junction on preoperative computer tomography images. We developed an automated algorithm detecting the sinotubular junction (STJ) and intercostal spaces for finding the optimal incision location. The accuracy of the STJ detection was assessed by comparison with manual delineation by measuring the Euclidean distance between the manually and automatically detected points. In all 20 patients, the intercostal spaces were accurately detected. The median distance between automated and manually detected STJ locations was 1.4 [IQR= 0.91-4.7] mm compared to the interobserver variation of 1.0 [IQR= 0.54-1.3] mm. For 60% of patients, the fourth intercostal space was the closest to the STJ. The proposed algorithm is the first automated approach for detecting optimal incision location and has the potential to be implemented in clinical practice for planning of various mini-AVR procedures.

Automatic aortic root landmark detection in CTA images for preprocedural planning of transcatheter aortic valve implantation.

Transcatheter aortic valve implantation is currently a well-established minimal invasive treatment option for patients with severe aortic valve stenosis. CT Angiography is used for the pre-operative planning and sizing of the prosthesis. To reduce the inconsistency in sizing due to interobserver variability, we introduce and evaluate an automatic aortic root landmarks detection method to determine the sizing parameters. The proposed algorithm detects the sinotubular junction, two coronary ostia, and three valvular hinge points on a segmented aortic root surface. Using these aortic root landmarks, the automated method determines annulus radius, annulus orientation, and distance from annulus plane to right and left coronary ostia. Validation is performed by the comparison with manual measurements of two observers for 40 CTA image datasets. Detection of landmarks showed high accuracy where the mean distance between the automatically detected and reference landmarks was 2.81 ± 2.08 mm, comparable to the interobserver variation of 2.67 ± 2.52 mm. The mean annulus to coronary ostium distance was 16.9 ± 3.3 and 17.1 ± 3.3 mm for the automated and the reference manual measurements, respectively, with a mean paired difference of 1.89 ± 1.71 mm and interobserver mean paired difference of 1.38 ± 1.52 mm. Automated detection of aortic root landmarks enables automated sizing with good agreement with manual measurements, which suggests applicability of the presented method in current clinical practice.

Increased aortic tortuosity indicates a more severe aortic phenotype in adults with Marfan syndrome.

BACKGROUND: Patients with Marfan syndrome (MFS) have a highly variable occurrence of aortic complications. Aortic tortuosity is often present in MFS and may help to identify patients at risk for aortic complications. METHODS: 3D-visualization of the total aorta by MR imaging was performed in 211 adult MFS patients (28% with prior aortic root replacement) and 20 controls. A method to assess aortic tortuosity (aortic tortuosity index: ATI) was developed and reproducibility was tested. The relation between ATI and age, and body size and aortic dimensions at baseline was investigated. Relations between ATI at baseline and the occurrence of a clinical endpoint (aortic dissection, and/or aortic surgery) and aortic dilatation rate during 3 years of follow-up were investigated. RESULTS: ATI intra- and interobserver agreements were excellent (ICC: 0.968 and 0.955, respectively). Mean ATI was higher in 28 age-matched MFS patients than in the controls (1.92 ± 0.2 vs. 1.82 ± 0.1, p=0.048). In the total MFS cohort, mean ATI was 1.87 ± 0.20, and correlated with age (r=0.281, p<0.001), aortic root diameter (r=0.223, p=0.006), and aortic volume expansion rate (r=0.177, p=0.026). After 49.3 ± 8.8 months follow-up, 33 patients met the combined clinical endpoint (7 dissections) with a significantly higher ATI at baseline than patients without endpoint (1.98 ± 0.2 vs. 1.86 ± 0.2, p=0.002). Patients with an ATI>1.95 had a 12.8 times higher probability of meeting the combined endpoint (log rank-test, p<0.001) and a 12.1 times higher probability of developing an aortic dissection (log rank-test, p=0.003) compared to patients with an ATI<1.95. CONCLUSIONS: Increased ATI is associated with a more severe aortic phenotype in MFS patients.

Clinical study protocol for the ARCH project - computational modeling for improvement of outcome after vascular access creation.

Despite clinical guidelines and the possibility of diagnostic vascular imaging, creation and maintenance of a vascular access (VA) remains problematic: avoiding short- and long-term VA dysfunction is challenging. Although prognostic factors for VA dysfunction have been identified in previous studies, their potential interplay at a systemic level is disregarded. Consideration of multiple prognostic patient specific factors and their complex interaction using dedicated computational modeling tools might improve outcome after VA creation by enabling a better selection of VA configuration. These computational modeling tools are developed and validated in the ARCH project: a joint initiative of four medical centers and three industrial partners (FP7-ICT-224390). This paper reports the rationale behind computational modeling and presents the clinical study protocol designed for calibrating and validating these modeling tools. The clinical study is based on the pre-operative collection of structural and functional data at a vascular level, as well as a VA functional evaluation during the follow-up period. The strategy adopted to perform the study and for data collection is also described here.

Magnetic resonance angiographic assessment of upper extremity vessels prior to vascular access surgery: feasibility and accuracy.

A contrast-enhanced magnetic resonance angiography (CE-MRA) protocol for selective imaging of the entire upper extremity arterial and venous tree in a single exam has been developed. Twenty-five end-stage renal disease (ESRD) patients underwent CE-MRA and duplex ultrasonography (DUS) of the upper extremity prior to hemodialysis vascular access creation. Accuracy of CE-MRA arterial and venous diameter measurements were compared with DUS and intraoperative (IO) diameter measurements, the standard of reference. Upper extremity vasculature depiction was feasible with CE-MRA. CE-MRA forearm and upper arm arterial diameters were 2.94 +/- 0.67 mm and 4.05 +/- 0.84 mm, respectively. DUS arterial diameters were 2.80 +/- 0.48 mm and 4.38 +/- 1.24 mm; IO diameters were 3.00 +/- 0.35 mm and 3.55 +/- 0.51 mm. Forearm arterial diameters were accurately determined with both techniques. Both techniques overestimated upper arm arterial diameters significantly. Venous diameters were accurately determined with CE-MRA but not with DUS (forearm: CE-MRA: 2.64 +/- 0.61 mm; DUS: 2.50 +/- 0.44 mm, and IO: 3.40 +/- 0.22 mm; upper arm: CE-MRA: 4.09 +/- 0.71 mm; DUS: 3.02 +/- 1.65 mm, and IO: 4.30 +/- 0.78 mm). CE-MRA enables selective imaging of upper extremity vasculature in patients requiring hemodialysis access. Forearm arterial diameters can be assessed accurately by CE-MRA. Both CE-MRA and DUS slightly overestimate upper arm arterial diameters. In comparison to DUS, CE-MRA enables a more accurate determination of upper extremity venous diameters.

Impact of a quality improvement programme based on vascular access flow monitoring on costs, access occlusion and access failure.

BACKGROUND: Vascular access thrombosis is a substantial source of morbidity in chronic haemodialysis patients. Periodical access flow measurements can predict the presence of vascular access stenosis and provide an opportunity for early intervention to prevent subsequent vascular access thrombosis. By this system of quality improvement, vascular access-related costs might be reduced. The aim of this study was to analyse the cost impact of a quality improvement programme based on periodic access flow measurements. METHODS: The number and costs of vascular access interventions (summary of angiography, percutaneous transluminal angioplasty, catheter placement, hospitalization days and costs for surgery) in the period 2001-2003 (quality improvement period; QIP, 218.6 patient-years observed) were retrospectively compared with a reference period (RP, 1996-1998, 214.4 patient-years observed) during which no access flow was measured. All access flow measurements were done on a regular base and interventions were performed according to the Kidney Disease Outcome Quality Initiative. RESULTS: Surgical thrombectomy procedures were significantly less during the QIP (0.25 +/- 0.57 events/patient-year) compared with RP (0.63 +/- 1.06 events/patient-year; P = 0.000), whereas access loss was not significantly different. During the QIP, 205 radiological interventions were performed (0.88 +/- 1.16 events/patient-year), and in the RP around 48 (0.33 +/- 0.65 events/patient-year; P = 0.000). Access-related costs tended to be lower during the QIP compared with the RP. The cost reduction appeared to be limited to patients with arteriovenous graft (AVG), in which access-related costs were significantly lower during the QIP (2360.95 euro +/- 2838.17 euro patient-year) compared with the RP (4003.96 euro +/- 3810.92 euro patient-year; P = 0.012), but not in patients with arteriovenous fistula (AVF). CONCLUSION: A quality improvement programme based on periodical access flow measurement reduced the number of acute vascular access failures due to thrombotic events and also significantly reduced health care costs in patients with AVG, but not in patients with AVF. The quality improvement programme had no effect on access survival.