Morphologie : bulletin de l'Association des anatomistes · 2019
Towards safe and efficient preoperative planning of transcatheter mitral valve interventions.
СтатьяВалидационное исследованиеНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Astudillo P, De Beule M, Dambre J, Mortier P
Аннотация
OBJECTIVE OF THE STUDY: Transcatheter mitral valve interventions are emerging as a viable alternative for patients at high risk. Two key aspects are crucial during the preoperative planning: left ventricular outflow tract assessment and anatomical analysis. Given that the manual anatomical analysis is time-consuming, an automated approach may introduce efficiency during preoperative planning. In this study, we present an automatic method to detect the mitral valve annulus and discuss possible implementation of this method in clinical practice.
PATIENTS: This retrospective study used the data of 71 patients collected from multiple centra. The mean age of this cohort was 74.2±13.1 years, and 56.1% of the patients were female and 43.9% male.
MATERIALS AND METHODS: We trained three deep learning models to segment the area around the mitral valve annulus. In a post-processing step, we extracted the mitral valve annulus from this segmentation. As a final step, clinically relevant measurements such as 2D perimeter, trigone-to-trigone (TT) distance, septal-to-lateral (SL) distance and commissure-to-commissure (IC) distance were derived from the predicted mitral valve annulus. The method was cross-validated with k-folding.
RESULTS: The predicted measurements showed excellent correlation with the manually obtained clinical measurements: 2D perimeter: R2=0.93, TT-distance: R2=0.86, SL-distance: R2=0.86 and IC-distance: R2=0.90. The total analysis time per patient of the automatic method was less than 1 second, which is an enormous speed-up as compared to the manual process (25minutes).
CONCLUSION: The efficiency and accuracy of the proposed method give the confidence to move towards implementation of this technology in clinical practice. We propose a possible implementation of this method in clinical practice, which, in our opinion, will facilitate safe and efficient preoperative planning of transcatheter mitral valve interventions.
PATIENTS: This retrospective study used the data of 71 patients collected from multiple centra. The mean age of this cohort was 74.2±13.1 years, and 56.1% of the patients were female and 43.9% male.
MATERIALS AND METHODS: We trained three deep learning models to segment the area around the mitral valve annulus. In a post-processing step, we extracted the mitral valve annulus from this segmentation. As a final step, clinically relevant measurements such as 2D perimeter, trigone-to-trigone (TT) distance, septal-to-lateral (SL) distance and commissure-to-commissure (IC) distance were derived from the predicted mitral valve annulus. The method was cross-validated with k-folding.
RESULTS: The predicted measurements showed excellent correlation with the manually obtained clinical measurements: 2D perimeter: R2=0.93, TT-distance: R2=0.86, SL-distance: R2=0.86 and IC-distance: R2=0.90. The total analysis time per patient of the automatic method was less than 1 second, which is an enormous speed-up as compared to the manual process (25minutes).
CONCLUSION: The efficiency and accuracy of the proposed method give the confidence to move towards implementation of this technology in clinical practice. We propose a possible implementation of this method in clinical practice, which, in our opinion, will facilitate safe and efficient preoperative planning of transcatheter mitral valve interventions.