Prognostic Value of Echocardiographic Parameters Based on Machine Learning Approach

Valeur Pronostique Dans Une Population d'Amylose Cardiaque Des paramètres échocardiographiques basée Sur Une Approche d'Apprentissage Automatique

Transthyretin cardiac amyloidosis is an increasingly recognized cause of heart failure with preserved ejection fraction. Its diagnosis is currently based on a non-invasive method including biology and imaging. Still currently incurable, the evolution of this pathology is burdened by numerous comorbidities, including iterative hospitalizations for heart failure leading to death. The Machine Learning approach has already shown its efficiency in terms of diagnosis but its prognostic approach has not yet been studied.

Inclusion Criteria:

• Patients with suspicion of transthyretin cardiac amyloidosis
• Age ≥18 years

Exclusion Criteria:

• Lack of data to confirm or overturn the transthyretin amyloidosis diagnostic
• Echocardiographic data not allowing deep analysis (technical default, bad echogenicity of the patient)
• Final diagnostic of AL or AA amyloidosis