Evaluation of an Echocardiography-Based Artificial Intelligence Algorithm for Early HFpEF Detection: A HeartShare/AMP-HF Cohort Study

Ahmad, FS; Usman, MS; Karnik, A; Segar, MW; Subramanian, V; Upton, R; Akerman, AP; Keshvani, N; Freed, BH; Shah, AM; Patel, RB; Chandra, A; Willett, D; Bertoni, A; Butler, J; Shah, S; Luo, Y; Shah, S; Pandey, A

doi:10.1016/j.jchf.2026.103420

Heart failure & cardiomyopathies

Evaluation of an Echocardiography-Based Artificial Intelligence Algorithm for Early Heart Failure with Preserved Ejection Fraction (HFpEF) Detection: A HeartShare/AMP-HF Cohort Study

JACC Heart FailPublished 29 Oct 2026Evaluation of an echocardiography-based artificial intelligence algorithm

Read the original article ↗PubMed

Cite

Authors' conclusion

AI-HFpEF identified the early HFpEF phenotype in community-dwelling adults and independently predicted incident clinical HFpEF among patients in a health system.

Study context

Population: Community-dwelling adults without a clinical heart failure (HF) diagnosis (MESA Early HF study and DHMS), and adults in the UTSW Health System with structurally normal echocardiograms and no prior HF (2018–2024)three cohorts · MESA 2,544 · DHMS 624 · UTSW cohort 6,876

Index test or model: AI-HFpEF vs H2FPEF for detecting early HFpEF
Target condition: Early HFpEF (impaired exercise capacity or health-related quality of life with objective evidence of elevated cardiac filling pressures in patients without clinically diagnosed HF)

Methodological details2 community cohorts, health system cohort, multivariable models, Cox regression

2 community cohorts
Health system cohort
Multivariable models
Cox regression

Performance

As reported in the abstract

Test or model	AUROC, MESA	AUROC, DHMS
AI-HFpEF	0.77	0.83
H2FPEF	0.70	0.77

AUROC, MESA

AI-HFpEF0.77

H2FPEF0.70

AUROC, DHMS

AI-HFpEF0.83

H2FPEF0.77

Further findings

Outcome	Estimate	P value
Early HFpEF (MESA)	OR 3.26	P < 0.001
Early HFpEF (DHMS)	OR 6.76	P < 0.001
Incident HFpEF (UTSW cohort) · median follow-up = 33 months>2-fold higher incident HFpEF risk	aHR 2.08	P < 0.001

Incident HFpEF (UTSW cohort) — Incident HFpEF: 3.4%

Early HFpEF (MESA)

AI-HFpEF diagnostic positivity associated with lower 6-minute walk distance and lower Kansas City Cardiomyopathy Questionnaire-12-Clinical Summary Score

Early HFpEF (DHMS)

Diagnostic positivity associated with shorter exercise duration and lower peak oxygen uptake

Incident HFpEF (UTSW cohort) · median follow-up = 33 months

Association persisted across all conventional diastolic dysfunction grades

Original publication

Evaluation of an Echocardiography-Based Artificial Intelligence Algorithm for Early HFpEF Detection: A HeartShare/AMP-HF Cohort Study. JACC Heart Fail. Published 29 October 2026. doi:10.1016/j.jchf.2026.103420

Read the original article ↗

Composed by AI from the published abstract and checked against it. How Briefs are made →