Abstract
Background: Mitral regurgitation (MR) is one of the most common valvular heart diseases, though many patients with severe MR are unsuitable for surgical interventions. Transcatheter edge-to-edge repair (TEER) is an established non-surgical procedure that allows safe and effective reduction of MR in highrisk patients. Appropriate patient selection is critical and achieved through clinical assessment and imaging derived measurement of the mitral valve (MV) apparatus to optimize procedural success.
Objective: We have developed a technology to automatically measure the MV structure based on computed tomography (CT) imaging data. The primary objective of this study is to compare automatically derived measurements to manually derived measurements performed by expert readers, in patients with MV disease.
Methods: The MV and left-sided cardiac chambers were assessed by 4D CT in 28 patients with MV disease of varying type and severity. MV measurements (MVM) were performed with a fully automatic MV shape extraction algorithm, which provides a means to automatically measure the annulus area, annulus perimeter, anteroposterior (AP) diameters, the inter-commissural (IC) diameters, D-shaped perimeter, and aorto-mitral (AM) angle. These automatically derived measurements were compared to manual measurements performed by experienced cardiologist designated as expert readers.
Results: Our results suggest that for annulus area, annulus circumference, D-shaped annulus circumference, and intercommissural diameter measurements there is a strong correlation between MVM and expert readers (all correlation coefficients( CC), r > 0.8). There was slightly less observer (Obs) variability between MVM and the expert readers than between individual expert readers.
Conclusion: The measurement of structural dimensions of the MV using this novel technology demonstrates a significant and strong correlation to manually derived measurements performed by expert readers, and its variability is consistent with inter-Obs variability. We found that key structures of the MV could be measured within 10 seconds and with a 100% success rate. We limited the number of structures measured for this evaluation; however, it is possible to perform several additional measurements that are currently being evaluated. We believe that a streamlined workflow with a more efficient resource utilization can be achieved using this automatic, accurate, and non-Obs dependent technology.
Published on: March 21, 2025
doi: 10.17756/micr.2025-114
Citation: Pham MHC, Kofoed KF, Chapman JV, de Backer O, Linde JJ1, et al. 2025. Preclinical Validation of Automatic Mitral Valve Segmentation and Measurement Technology for Planning Transcatheter Edge-to-edge Valve Repair Using Computed Tomography. J Med Imaging Case Rep 9(1): 11-16.
Downloads