Multimodal artificial intelligence for early detection and precision management of inflammatory and infiltrative cardiomyopathies.

BACKGROUND

Inflammatory and infiltrative cardiomyopathies, including cardiac sarcoidosis, transthyretin amyloidosis, and autoimmune myocarditis, are frequently underrecognized until advanced myocardial dysfunction develops. Conventional diagnosis of inflammatory and infiltrative cardiomyopathies depends on multimodality imaging and clinical integration, yet interpretation remains complex and delayed. AI (Artificial Intelligence) can improve diagnostic suspicion, tissue-based phenotyping, and risk-directed triage by integrating ECG, echocardiography, cardiac magnetic (CMR), positron emission tomography (PET)/scintigraphy, biomarkers, and Electronic Health Record (EHR) features.

METHODS

This narrative review synthesizes contemporary evidence (2018-2026) regarding multimodality imaging and emerging AI approaches for early detection and risk-directed management of inflammatory and infiltrative cardiomyopathies. We emphasize mechanistic integration across electrocardiography, echocardiography, CMR, PET, nuclear scintigraphy, and EHR-derived phenotypes.

RESULTS

Traditional functional assessment incompletely captures subclinical myocardial inflammation, interstitial expansion, and arrhythmogenic substrate. Multimodal AI platforms enable weak-signal fusion across imaging, electrophysiologic, and laboratory domains, identifying subvisual patterns not discernible through conventional interpretation. Disease-specific AI applications demonstrate promise in detecting occult cardiac sarcoidosis prior to overt clinical manifestation, differentiating amyloid phenotypes and expediting nonbiopsy diagnosis, and stratifying myocarditis trajectories. Importantly, AI-derived phenotypes may refine biopsy targeting, optimize immunosuppression timing, enhance arrhythmic risk stratification, and guide heart failure therapy escalation.

CONCLUSIONS

Transitioning from conventional functional assessment to multimodal, AI-supported phenotyping represents a paradigm shift in the evaluation of inflammatory and infiltrative cardiomyopathies. Carefully validated and clinically integrated AI tools have the potential to enable earlier diagnosis, individualized therapy, and improved cardiovascular outcomes.