Evaluating the adverse outcome of subtypes of heart failure with preserved ejection fraction defined by machine learning: a systematic review focused on defining high risk phenogroups

Abstract

The ability to distinguish clinically meaningful subtypes of heart failure with preserved ejection fraction (HFpEF) has recently been examined by machine learning techniques but studies appear to have produced discordant results. The objective of this study is to synthesize the types of HFpEF by examining their features and relating them to phenotypes with adverse prognosis. A systematic search was conducted using the search terms “Diastolic Heart Failure” OR “heart failure with preserved ejection fraction” OR “heart failure with normal ejection fraction” OR “HFpEF” AND “machine learning” OR “artificial intelligence” OR 'computational biology'. Ten studies were identified and they varied in their prevalence of ten clinical variables: age, sex, body mass index (BMI) or obesity, hypertension, diabetes mellitus, coronary artery disease, atrial fibrillation, chronic kidney disease, chronic obstructive pulmonary disease or symptom severity (NYHA class or BNP). The clinical findings associated with the different phenotypes in > 85 % of studies were age, hypertension, atrial fibrillation, chronic kidney disease and worse symptoms severity; an adverse outcome was in 65 % to 85 % of studies identified diabetes mellitus and female sex and in less than 65 % of studies was body mass index or obesity, and coronary artery disease. COPD was a relevant factor in only 33 % of studies. Adverse clinical outcome - death or admission to hospital (for heart failure) defined phenogroups with the worst outcome. Combining the 4 studies that calculated the MAGGIC score showed a significant (p<0.05) linear relationship between MAGGIC score and outcome, using the one-year event rate. A new score based on strength of the evidence of the HFpEF studies analyzed here, using 9 variables (eliminating COPD), showed a significant (p<0.009) linear relationship with one-year event rate. Three studies examined biomarkers in detail and the ones most prominently related to outcome or consistently found in the studies were GDF15, FABP4, FGF23, sST2, renin and TNF. The dominant factors that identified phenotypes of HFpEF with adverse outcome were hypertension, atrial fibrillation, chronic kidney disease and worse symptoms severity. A new simplified score, based on clinical factors, was proposed to assess prognosis in HFpEF. Several biomarkers were consistently elevated in phenogroups with adverse outcomes and may indicate the underlying mechanism or pathophysiology specific for phenotypes with an adverse prognosis.

Machine learning for atrial fibrillation risk prediction in patients with sleep apnea and coronary artery disease

“…The results of our cluster analyses identified seven phenotypes for each group of study (AF, CAD, OSA) and showed that HBP was the most prevalent comorbidity associated to CAD, OSA, and AF. Cluster analytic techniques used in several studies have proposed phenotypic groups for CVD risk and include HBP, AF, CAD and renal disfunction as comorbidities with the highest risk for heart failure and death ( 46 ). The role of OSA in the phenotypes groups of CVD risk has not been extensively studied.…”

Artificial intelligence streamlines diagnosis and assessment of prognosis in Brugada syndrome: a systematic review and meta-analysis

“…Effective space manipulation leads to improved model performance by ensuring representations of both text and image modalities are compatible and can be efficiently combined to make diagnostic predictions. In contrast, MED-UniC models involve medical data streams from multiple sources of data, such as imaging (e.g., radiography) and text data (e.g., consult notes) [31] . Medical vision and language pre-training (MED-VLP) hopes to integrate and jointly process these data to generalize representations from large-scale medical image-text data.…”