European heart journal. Digital health · 2026
Machine-learning approach on echocardiography to improve the detection of transthyretin amyloid cardiomyopathy: GRAAL algorithm.
СтатьяНазвание 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.
Fraix Antoine, Huttin Olivier, Lacomblez Claire, Pace Nathalie, Marie Pierre-Yves, Mandry Damien, Claudin Marine, Sadoul Nicolas, Filippetti Laura, Donal Erwan, Lairez Olivier, Lointier Emmanuelle, Zaroui Amira, Damy Thibaud, Selton-Suty Christine, Girerd Nicolas
Аннотация
AIMS: Transthyretin amyloid cardiomyopathy (ATTR-CM) is an increasingly recognized cause of heart failure, yet detection remains challenging due to its echocardiographic similarities with age- and hypertension-related cardiac remodelling.
METHODS AND RESULTS: We retrospectively included 260 patients (76.5 ± 12.9 years old, 59.6% male) referred for suspected ATTR-CM. A supervised machine-learning diagnosis algorithm differentiating patients with (n = 111) and without (n = 149) ATTR-CM based on echocardiographic data, and subsequently validated in an external multicentre cohort of 454 patients (76.3 ± 12.6 years old, 69.1% male). Patients with ATTR-CM had a lower systolic function [left ventricular ejection fraction 47 ± 11 vs. 54 ± 12%, P < 0.00; global longitudinal strain (GLS) 11.0 ± 3.7 vs. 14.1 ± 4.5%, P < 0.001] and more significant relative apical longitudinal sparing (RALS) (1.5 ± 1.2 vs. 0.9 ± 0.4, P < 0.001) compared with controls. Machine learning identified right ventricular free wall thickness (RVFWT), RALS, GLS, and LV mass index as key variables for detecting ATTR-CM [AUC 0.90 (0.86-0.94); P < 0.001]. These variables enhanced diagnostic accuracy compared with the increased wall thickness guideline score [increase in C-index of 0.17 (0.11-0.23), P < 0.001]. Diagnostic performance was confirmed in the validation multicentre cohort [AUC of 0.83 (0.80-0.87), P < 0.001].
CONCLUSION: The simple GRAAL algorithm (GLS, RVFWT, Apical spAring, LV Mass) enhances detection accuracy for ATTR-CM and improves patient selection for bone scintigraphy.
METHODS AND RESULTS: We retrospectively included 260 patients (76.5 ± 12.9 years old, 59.6% male) referred for suspected ATTR-CM. A supervised machine-learning diagnosis algorithm differentiating patients with (n = 111) and without (n = 149) ATTR-CM based on echocardiographic data, and subsequently validated in an external multicentre cohort of 454 patients (76.3 ± 12.6 years old, 69.1% male). Patients with ATTR-CM had a lower systolic function [left ventricular ejection fraction 47 ± 11 vs. 54 ± 12%, P < 0.00; global longitudinal strain (GLS) 11.0 ± 3.7 vs. 14.1 ± 4.5%, P < 0.001] and more significant relative apical longitudinal sparing (RALS) (1.5 ± 1.2 vs. 0.9 ± 0.4, P < 0.001) compared with controls. Machine learning identified right ventricular free wall thickness (RVFWT), RALS, GLS, and LV mass index as key variables for detecting ATTR-CM [AUC 0.90 (0.86-0.94); P < 0.001]. These variables enhanced diagnostic accuracy compared with the increased wall thickness guideline score [increase in C-index of 0.17 (0.11-0.23), P < 0.001]. Diagnostic performance was confirmed in the validation multicentre cohort [AUC of 0.83 (0.80-0.87), P < 0.001].
CONCLUSION: The simple GRAAL algorithm (GLS, RVFWT, Apical spAring, LV Mass) enhances detection accuracy for ATTR-CM and improves patient selection for bone scintigraphy.