Multidetector computed tomography sizing of aortic annulus prior to transcatheter aortic valve replacement (TAVR): Variability and impact of observer experience

The Feasibility of Deep Learning-Based Reconstruction for Low-Tube-Voltage CT Angiography for Transcatheter Aortic Valve Implantation

OBJECTIVE

The purpose of this study is to evaluate the efficacy of deep learning reconstruction (DLR) on low-tube-voltage computed tomographic angiography (CTA) for transcatheter aortic valve implantation (TAVI).

METHODS

We enrolled 30 patients who underwent TAVI-CT on a 320-row CT scanner. Electrocardiogram-gated coronary CTA (CCTA) was performed at 100 kV, followed by nongated aortoiliac CTA at 80 kV using a single bolus of contrast material. We used hybrid-iterative reconstruction (HIR), model-based IR (MBIR), and DLR to reconstruct these images. The contrast-to-noise ratios (CNRs) were calculated. Five-point scales were used for the overall image quality analysis. The diameter of the aortic annulus was measured in each reconstructed image, and we compared the interobserver and intraobserver agreements.

RESULTS

In the CCTA, the CNR and image quality score for DLR were significantly higher than those for HIR and MBIR ( P < 0.01). In the aortoiliac CTA, the CNR for DLR was significantly higher than that for HIR ( P < 0.01) and significantly lower than that for MBIR ( P ≤ 0.02). The image quality score for DLR was significantly higher than that for HIR ( P < 0.01). No significant differences were observed between the image quality scores for DLR and MBIR. The measured aortic annulus diameter had high interobserver and intraobserver agreement regardless of the reconstruction method (all intraclass correlation coefficients, >0.89).

CONCLUSIONS

In low tube voltage TAVI-CT, DLR provides higher image quality than HIR, and DLR provides higher image quality than MBIR in CCTA and is visually comparable to MBIR in aortoiliac CTA.

Relevance of Motion Artifacts in Planning Computed Tomography on Outcomes After Transcatheter Aortic Valve Implantation

Background

Motion artifacts in planning computed tomography (CT) for transcatheter aortic valve implantation (TAVI) can potentially skew measurements required for procedural planning. Whether such artifacts may affect safety or efficacy has not been studied.

Methods

We conducted a retrospective analysis of 852 consecutive patients (mean age, 82 years; 47% women) undergoing TAVI-planning CT at a tertiary care center. Two independent observers divided CTs according to the presence of motion artifacts at the annulus level (Motion vs. Normal group). Endpoints included surrogate markers for inappropriate valve selection: annular rupture, valve embolization or misplacement, need for a new permanent pacemaker, paravalvular leak (PVL), postprocedural transvalvular gradient, all-cause death.

Results

Forty-six (5.4%) patients presented motion artifacts on TAVI-planning CT (Motion group). These patients had more preexisting heart failure, moderate-severe mitral regurgitation, and atrial fibrillation. Interobserver variability of annular measurement (Normal vs. Motion group) did not differ for mean annular diameter but was significantly different for perimeter and area. Presence of motion artifacts on planning CT did not affect the prevalence of PVL (≥moderate PVL 0% vs. 2.5% p = 0.5), mean transvalvular gradient (6±3 mmHg vs 7±5 mmHg, p = 0.1), or the need for additional valve implantation (0% vs. 2.8%, p = 0.6). One annular rupture occurred (Normal group). Pacemaker implantation, procedural duration, hospital stay, 30-day outcomes, and all-cause mortality did not differ between the groups.

Conclusions

Motion artifacts on planning CT were found in about 5% of patients. Measurements for valve selection were possible without the need for repeat CT, with mean diameter-derived annulus measurement being the most accurate. Motion artifacts were not associated with worse outcomes.

TAVI-PREP: A Deep Learning-Based Tool for Automated Measurements Extraction in TAVI Planning

BACKGROUND

Transcatheter aortic valve implantation (TAVI) is a less invasive alternative to open-heart surgery for treating severe aortic stenosis. Despite its benefits, the risk of procedural complications necessitates careful preoperative planning.

METHODS

This study proposes a fully automated deep learning-based method, TAVI-PREP, for pre-TAVI planning, focusing on measurements extracted from computed tomography (CT) scans. The algorithm was trained on the public MM-WHS dataset and a small subset of private data. It uses MeshDeformNet for 3D surface mesh generation and a 3D Residual U-Net for landmark detection. TAVI-PREP is designed to extract 22 different measurements from the aortic valvular complex. A total of 200 CT-scans were analyzed, and automatic measurements were compared to the ones made manually by an expert cardiologist. A second cardiologist analyzed 115 scans to evaluate inter-operator variability.

RESULTS

High Pearson correlation coefficients between the expert and the algorithm were obtained for most parameters (0.90-0.97), except for left and right coronary height (0.8 and 0.72, respectively). Similarly, the mean absolute relative error was within 5% for most measurements, except for left and right coronary height (11.6% and 16.5%, respectively). A greater consensus was observed among experts than when compared to the automatic approach, with TAVI-PREP showing no discernable bias towards either the lower or higher ends of the measurement spectrum.

CONCLUSIONS

TAVI-PREP provides reliable and time-efficient measurements of the aortic valvular complex that could aid clinicians in the preprocedural planning of TAVI procedures.

Pre-TAVI aortic annulus sizing: comparison between manual and semi-automated new generation software measurements in operators with different experience

OBJECTIVES

Aim of the study is to compare manual and semi-automatic measurements for aortic annulus assessment among different operators.

METHODS

Eighty patients who underwent TAVI were retrospectively enrolled. The measurements manually performed by an experienced reader for aortic annulus (minimum and maximum diameters, perimeter, area), annulus-to-coronary ostia distance and time needed for the whole evaluation, were collected. The same operator (observer 1) and two less experienced readers (observer 2 and 3, with >5 years and 1 year of experience, respectively) assessed the same measurements using a semi-automatic software. Differences between manual and semi-automatic measurements, reading time and suggested valves size derived by CT were compared.

RESULTS

Very good correlations were found between manual and software-aided measurements for aortic annulus area and perimeter in comparison with standard measurements for the three readers (ICC range 0.81-0.98). Good correlations were found for the distance with coronary ostia(0.75-0.79). The same area-derived prosthesis size for manual and semi-automatic measurements was selected in 96% of cases for observer 1; very good correlations were also found for observer 2 and 3 (ICC = 0.89 and 0.88, respectively). Using semi-automatic measurements, the mean time needed for CT images was significantly lower for observers 1 and 2 (1.50 and 1.72versus 3.14 min), respectively.

CONCLUSIONS

Pre-TAVI CT using semi-automatic software allows accurate and reproducible measurements, reducing reconstruction time up to 50% and is reliable even for operators with different experience.

ADVANCES IN KNOWLEDGE

The use of semi-automatic dedicated software for CT in TAVI planning is reliable even for operators without long time experience and allows accurate and reproducible measurements improving pre-TAVI workflow.

Impact of observer experience on multi-detector computed tomography aortic valve morphology assessment and valve size selection for transcatheter aortic valve replacement

Transcatheter aortic valve replacement (TAVR) has become the standard treatment for aortic stenosis in older patients. It increasingly relies on accurate pre-procedural planning using multidetector computed tomography (MDCT). Since little is known about the required competence levels for MDCT analyses, we comprehensively assessed MDCT TAVR planning reproducibility and accuracy with regard to valve selection in various healthcare workers. 20 randomly selected MDCT of TAVR patients were analyzed using dedicated software by healthcare professionals with varying backgrounds and experience (two structural interventionalists, one imaging specialist, one cardiac surgeon, one general physician, and one medical student). Following the analysis, the most appropriate Edwards SAPIEN 3™ and Medtronic CoreValve valve size was selected. Intra- and inter-observer variability were assessed. The first structural interventionalist was considered as reference standard for inter-observer comparison. Excellent intra- and inter-observer variability was found for the entire group in regard to the MDCT measurements. The best intra-observer agreement and reproducibility were found for the structural interventionalist, while the medical student had the lowest reproducibility. The highest inter-observer agreement was between both structural interventionalists, followed by the imaging specialist. As to valve size selection, the structural interventionalist showed the highest intra-observer reproducibility, independent of the brand of valve used. Compared to the reference structural interventionalist, the second structural interventionalist showed the highest inter-observer agreement for valve size selection [ICC 0.984, 95% CI 0.969-0.991] followed by the cardiac surgeon [ICC 0.947, 95%CI 0.900-0.972]. The lowest inter-observer agreement was found for the medical student [ICC 0.507, 95%CI 0.067-0.739]. While current state-of-the-art MDCT analysis software provides excellent reproducibility for anatomical measurements, the highest levels of confidence in terms of valve size selection were achieved by the performing interventional physicians. This was most likely attributable to observer experience.

Preoperative TAVR Planning: How to Do It

Transcatheter aortic valve replacement (TAVR) is a well-established treatment option for patients with severe symptomatic aortic stenosis (AS) whose procedural efficacy and safety have been continuously improving. Appropriate preprocedural planning, including aortic valve annulus measurements, transcatheter heart valve choice, and possible procedural complication anticipation is mandatory to a successful procedure. The gold standard for preoperative planning is still to perform a multi-detector computed angiotomography (MDCT), which provides all the information required. Nonetheless, 3D echocardiography and magnet resonance imaging (MRI) are great alternatives for some patients. In this article, we provide an updated comprehensive review, focusing on preoperative TAVR planning and the standard steps required to do it properly.

Automatic Aortic Valve Cusps Segmentation from CT Images Based on the Cascading Multiple Deep Neural Networks

Accurate morphological information on aortic valve cusps is critical in treatment planning. Image segmentation is necessary to acquire this information, but manual segmentation is tedious and time consuming. In this paper, we propose a fully automatic aortic valve cusps segmentation method from CT images by combining two deep neural networks, spatial configuration-Net for detecting anatomical landmarks and U-Net for segmentation of aortic valve components. A total of 258 CT volumes of end systolic and end diastolic phases, which include cases with and without severe calcifications, were collected and manually annotated for each aortic valve component. The collected CT volumes were split 6:2:2 for the training, validation and test steps, and our method was evaluated by five-fold cross validation. The segmentation was successful for all CT volumes with 69.26 s as mean processing time. For the segmentation results of the aortic root, the right-coronary cusp, the left-coronary cusp and the non-coronary cusp, mean Dice Coefficient were 0.95, 0.70, 0.69, and 0.67, respectively. There were strong correlations between measurement values automatically calculated based on the annotations and those based on the segmentation results. The results suggest that our method can be used to automatically obtain measurement values for aortic valve morphology.