Multi-network approach for image segmentation in non-contrast enhanced cardiac 3D MRI of arrhythmic patients.

Left atrial appendage (LAA) is the source of thrombi formation in more than 90% of strokes in patients with nonvalvular atrial fibrillation. Catheter-based LAA occlusion is being increasingly applied as a treatment strategy to prevent stroke. Anatomical complexity of LAA makes percutaneous occlusion commonly performed under transesophageal echocardiography (TEE) and X-ray (XR) guidance especially challenging. Image fusion techniques integrating 3D anatomical models derived from pre-procedural imaging into the live XR fluoroscopy can be applied to guide each step of the LAA closure. Cardiac magnetic resonance (CMR) imaging gains in importance for radiation-free evaluation of cardiac morphology as alternative to gold-standard TEE or computed tomography angiography (CTA). Manual delineation of cardiac structures from non-contrast enhanced CMR is, however, labor-intensive, tedious, and challenging due to the rather low contrast. Additionally, arrhythmia often impairs the image quality in ECG synchronized acquisitions causing blurring and motion artifacts. Thus, for cardiac segmentation in arrhythmic patients, there is a strong need for an automated image segmentation method. Deep learning-based methods have shown great promise in medical image analysis achieving superior performance in various imaging modalities and different clinical applications. Fully-convolutional neural networks (CNNs), especially U-Net, have become the method of choice for cardiac segmentation. In this paper, we propose an approach for automatic segmentation of cardiac structures from non-contrast enhanced CMR images of arrhythmic patients based on CNNs implemented in a multi-stage pipeline. Two-stage implementation allows subdividing the task into localization of the relevant cardiac structures and segmentation of these structures from the cropped sub-regions obtained from previous step leading to efficient and effective way of automated cardiac segmentation.

Cardiac magnetic resonance imaging for preprocedural planning of percutaneous left atrial appendage closure.

Introduction: Percutaneous closure of the left atrial appendage (LAA) facilitates stroke prevention in patients with atrial fibrillation. Optimal device selection and positioning are often challenging due to highly variable LAA shape and dimension and thus require accurate assessment of the respective anatomy. Transesophageal echocardiography (TEE) and x-ray fluoroscopy (XR) represent the gold standard imaging techniques. However, device underestimation has frequently been observed. Assessment based on 3-dimensional computer tomography (CTA) has been reported as more accurate but increases radiation and contrast agent burden. In this study, the use of non-contrast-enhanced cardiac magnetic resonance imaging (CMR) to support preprocedural planning for LAA closure (LAAc) was investigated. Methods: CMR was performed in thirteen patients prior to LAAc. Based on the 3-dimensional CMR image data, the dimensions of the LAA were quantified and optimal C-arm angulations were determined and compared to periprocedural data. Quantitative figures used for evaluation of the technique comprised the maximum diameter, the diameter derived from perimeter and the area of the landing zone of the LAA. Results: Perimeter- and area-based diameters derived from preprocedural CMR showed excellent congruency compared to those measured periprocedurally by XR, whereas the respective maximum diameter resulted in significant overestimation (p < 0.05). Compared to TEE assessment, CMR-derived diameters resulted in significantly larger dimensions (p < 0.05). The deviation of the maximum diameter to the diameters measured by XR and TEE correlated well with the ovality of the LAA. C-arm angulations used during the procedures were in agreement with those determined by CMR in case of circular LAA. Discussion: This small pilot study demonstrates the potential of non-contrast-enhanced CMR to support preprocedural planning of LAAc. Diameter measurements based on LAA area and perimeter correlated well with the actual device selection parameters. CMR-derived determination of landing zones facilitated accurate C-arm angulation for optimal device positioning.

Impact of concomitant tricuspid regurgitation on outcome after edge-to-edge mitral valve repair.

AIMS: To evaluate the impact of tricuspid regurgitation (TR) on echocardiographic and functional outcome after mitral valve transcatheter edge-to-edge-repair (M-TEER). METHODS AND RESULTS: A total of 740 patients underwent M-TEER at our center from 2010 to 2021. Patients were analyzed according to severity of concomitant TR at the time of M-TEER procedure: low-grade TR (grade ≤I [trace-mild], 279 patients [37.7%]), moderate TR (grade II, 170 patients [23.0%]) and high-grade TR (grade III-V [severe-torrential], 291 patients [39.3%]). Patients with moderate to high-grade TR had higher morbidity. Procedural success of M-TEER was achieved similarly in all groups (98.2% vs. 97.6% vs. 95.9%, p = 0.22). TR severity decreased rapidly and consistently after M-TEER to only 48.0% of high-grade TR patients after 3 months (p < 0.001) and to 46.8% after 12 months (p = 0.99). High-grade TR patients had significantly higher mortality (21.5% vs. 18.2% vs. 11.1%, p = 0.003) up to 12 months after M-TEER. However, high-grade TR did not independently predict mortality (HR 1.302, 95% CI 0.937-1.810; p = 0.116). Echocardiographic and functional outcome was similar in both secondary and primary MR patients. CONCLUSIONS: High-grade concomitant TR did not independently predict adverse outcome following M-TEER. A wait-and-observe approach for these patients is reasonable.

TRI-SCORE is superior to EuroSCORE II and STS-Score in mortality prediction following transcatheter edge-to-edge tricuspid valve repair.

BACKGROUND: The development of transcatheter tricuspid edge-to-edge repair for tricuspid regurgitation is a therapeutic milestone but a specific periprocedural risk assessment tool is lacking. TRI-SCORE has recently been introduced as a dedicated risk score for tricuspid valve surgery. AIMS: This study analyzes the predictive performance of TRI-SCORE following transcatheter edge-to-edge tricuspid valve repair. METHODS: 180 patients who underwent transcatheter tricuspid valve repair at Ulm University Hospital were consecutively included and stratified into three TRI-SCORE risk groups. The predictive performance of TRI-SCORE was assessed throughout a follow-up period of 30 days and up to 1 year. RESULTS: All patients had severe tricuspid regurgitation. Median EuroSCORE II was 6.4% (IQR 3.8-10.1%), median STS-Score 8.1% (IQR 4.6-13.4%) and median TRI-SCORE 6.0 (IQR 4.0-7.0). 64 patients (35.6%) were in the low TRI-SCORE group, 91 (50.6%) in the intermediate and 25 (13.9%) in the high-risk groups. The procedural success rate was 97.8%. 30-day mortality was 0% in the low-risk group, 1.3% in the intermediate-risk and 17.4% in the high-risk groups (p < 0.001). During a median follow-up of 168 days mortality was 0%, 3.8% and 52.2%, respectively (p < 0.001). The predictive performance of TRI-SCORE was excellent (AUC for 30-day mortality: 90.3%, for one-year mortality: 93.1%) and superior to EuroSCORE II (AUC 56.6% and 64.4%, respectively) and STS-Score (AUC 61.0% and 59.0%, respectively). CONCLUSION: TRI-SCORE is a valuable tool for prediction of mortality after transcatheter edge-to-edge tricuspid valve repair and its performance is superior to EuroSCORE II and STS-Score. In a monocentric cohort of 180 patients undergoing edge-to-edge tricuspid valve repair TRI-SCORE predicted 30-day and up to one-year mortality more reliably than EuroSCORE II and STS-Score. AUC area under the curve, 95% CI 95% confidence interval.

Computed tomography angiography/magnetic resonance imaging-based preprocedural planning and guidance in the interventional treatment of structural heart disease.

Preprocedural planning and periprocedural guidance based on image fusion are widely established techniques supporting the interventional treatment of structural heart disease. However, these two techniques are typically used independently. Previous works have already demonstrated the benefits of integrating planning details into image fusion but are limited to a few applications and the availability of the proprietary tools used. We propose a vendor-independent approach to integrate planning details into periprocedural image fusion facilitating guidance during interventional treatment. In this work, we demonstrate the feasibility of integrating planning details derived from computer tomography and magnetic resonance imaging into periprocedural image fusion with open-source and commercially established tools. The integration of preprocedural planning details into periprocedural image fusion has the potential to support safe and efficient interventional treatment of structural heart disease.