Applications of Artificial Intelligence in Amyloidosis

Cardiac amyloidosis (CA) has been traditionally considered a rare disease even if its real prevalence is unknown due to frequent misdiagnoses. Patients with rare conditions, such as CA, are difficult to identify, given the similarity of disease manifestations to more prevalent disorders. Artificial intelligence (AI) could enable the detection of this rare disease. In this chapter, we summarize evidence regarding the use of AI models with clinical variables, biomarkers, electrocardiograms, and imaging techniques (e.g., echocardiography, cardiac magnetic resonance) as inputs for the diagnosis of CA. These models could be particularly useful in clinical practice for (1) screening of asymptomatic patients and (2) early diagnosis in patients presenting with initial clinical manifestations (e.g., signs and symptoms of heart failure). Furthermore, AI models using cardiac magnetic resonance and myocardial texture assessed by native T1 or T2 can help differentiate CA from other forms of left ventricular hypertrophy. Finally, some AI models (cluster analyses) have been associated with the outcome.