Automated MSCT Analysis for Planning Left Atrial Appendage Occlusion Using Artificial Intelligence

A Systematic Literature Review of 3D Deep Learning Techniques in Computed Tomography Reconstruction

Computed tomography (CT) is used in a wide range of medical imaging diagnoses. However, the reconstruction of CT images from raw projection data is inherently complex and is subject to artifacts and noise, which compromises image quality and accuracy. In order to address these challenges, deep learning developments have the potential to improve the reconstruction of computed tomography images. In this regard, our research aim is to determine the techniques that are used for 3D deep learning in CT reconstruction and to identify the training and validation datasets that are accessible. This research was performed on five databases. After a careful assessment of each record based on the objective and scope of the study, we selected 60 research articles for this review. This systematic literature review revealed that convolutional neural networks (CNNs), 3D convolutional neural networks (3D CNNs), and deep learning reconstruction (DLR) were the most suitable deep learning algorithms for CT reconstruction. Additionally, two major datasets appropriate for training and developing deep learning systems were identified: 2016 NIH-AAPM-Mayo and MSCT. These datasets are important resources for the creation and assessment of CT reconstruction models. According to the results, 3D deep learning may increase the effectiveness of CT image reconstruction, boost image quality, and lower radiation exposure. By using these deep learning approaches, CT image reconstruction may be made more precise and effective, improving patient outcomes, diagnostic accuracy, and healthcare system productivity.

Intra-Cardiac versus Transesophageal Echocardiographic Guidance for Left Atrial Appendage Occlusion with a Watchman FLX Device

This study aimed to compare the peri-procedural success and complication rate within a large registry of intra-cardiac echocardiography (ICE)- vs. transesophageal echocardiography (TEE)-guided left atrial appendage occlusion (LAAO) procedures with a Watchmann FLX device. Data from 772 LAAO procedures, performed at 26 Italian centers, were reviewed. Technical success was considered as the final implant of a Watchmann FLX device in LAA; the absence of pericardial tamponade, peri-procedural stroke and/or systemic embolism, major bleeding and device embolization during the procedure was defined as a procedural success. One-year stroke and major bleeding rates were evaluated as outcome. ICE-guided LAA occlusion was performed in 149 patients, while TEE was used in 623 patients. Baseline characteristics were similar between the ICE and TEE groups. The technical success was 100% in both groups. Procedural success was also extremely high (98.5%), and was comparable between ICE (98.7%) and TEE (98.5%). ICE was associated with a slightly longer procedural time (73 ± 31 vs. 61.9 ± 36 min, p = 0.042) and shorter hospital stay (5.3 ± 4 vs. 5.8 ± 6 days, p = 0.028) compared to the TEE group. At one year, stroke and major bleeding rates did not differ between the ICE and TEE groups. A Watchmann FLX device showed high technical and procedural success rate, and ICE guidance does not appear inferior to TEE.

Adversarial-based latent space alignment network for left atrial appendage segmentation in transesophageal echocardiography images

Left atrial appendage (LAA) is a leading cause of atrial fibrillation and thrombosis in cardiovascular disease. Clinicians can rely on LAA occlusion (LAAO) to effectively prevent and treat ischaemic strokes attributed to the LAA. The correct selection of the LAAO is one of the most critical stages in the successful surgical process, which relies on the quantification of the anatomical structure of the LAA for successful intervention in LAAO. In this paper, we propose an adversarial-based latent space alignment framework for LAA segmentation in transesophageal echocardiography (TEE) images by introducing prior knowledge from the label. The proposed method consists of an LAA segmentation network, a label reconstruction network, and a latent space alignment loss. To be specific, we first employ ConvNeXt as the backbone of the segmentation and reconstruction network to enhance the feature extraction capability of the encoder. The label reconstruction network then encodes the prior shape features from the LAA labels to the latent space. The latent space alignment loss consists of the adversarial-based alignment and the contrast learning losses. It can motivate the segmentation network to learn the prior shape features of the labels, thus improving the accuracy of LAA edge segmentation. The proposed method was evaluated on a TEE dataset including 1,783 images and the experimental results showed that the proposed method outperformed other state-of-the-art LAA segmentation methods with Dice coefficient, AUC, ACC, G-mean, and Kappa of 0.831, 0.917, 0.989, 0.911, and 0.825, respectively.

FEops HEARTguide Patient-Specific Computational Simulations for WATCHMAN FLX Left Atrial Appendage Closure: A Retrospective Study

Background

Three-dimensional transesophageal echocardiography (3D-TEE) is the primary imaging tool for left atrial appendage closure planning. The utility of cardiac computed tomography angiography (CCTA) and patient-specific computational models is unknown.

Objectives

The purpose of this study was to evaluate the accuracy of the FEops HEARTguide patient-specific computational modeling in predicting appropriate device size, location, and compression of the WATCHMAN FLX compared to intraprocedural 3D-TEE.

Methods

Patients with both preprocedural and postprocedural CCTA and 3D-TEE imaging of the LAA who received a WATCHMAN FLX left atrial appendage closure device were studied (n = 22). The FEops HEARTguide platform used baseline CCTA imaging to generate a prediction of device size(s), device position(s), and device dimensions. Blinded (without knowledge of implanted device size/position) and unblinded (implant device size/position disclosed) simulations were evaluated.

Results

In 16 (72.7%) patients, the blind simulation predicted the final implanted device size. In these patients, the 3D-TEE measurements were not significantly different and had excellent correlation (Pearson correlation coefficient (r) ≥ 0.90). No patients had peridevice leak after device implant. In the 6 patients for whom the model did not predict the implanted device size, a larger device size was ultimately implanted as per operator preference. The model measurements of the unblinded patients demonstrated excellent correlation with 3D-TEE.

Conclusions

This is the first study to demonstrate that the FEops HEARTguide model accurately predicts WATCHMAN FLX device implantation characteristics. Future studies are needed to evaluate if computational modeling can improve confidence in sizing, positioning, and compression of the device without compromising technical success.

Left Atrial Appendage Occlusion: Current Advances and Remaining Challenges

The field of left atrial appendage occlusion is rapidly evolving. However, several issues remain including the limited randomized efficacy data, peri-device leak, device-related thrombus, and the ongoing refinement of procedural techniques. In this article, we provide a contemporary overview of left atrial appendage occlusion focusing on 4 key remaining challenges: efficacy data, peri-device leak, device-related thrombus, and procedural optimization.