Unimodal to multimodal: a systematic review of predictive machine learning models for valvular heart diseases - Report - MDSpire

Unimodal to multimodal: a systematic review of predictive machine learning models for valvular heart diseases

  • By

  • Valentine Ojonugwa Idakwo

  • Caren Strote

  • Christian Goelz

  • Qasrina Shafei

  • Thomas J. Stocker

  • Jörg Hausleiter

  • Solveig Vieluf

  • July 1, 2026

  • 0 min

Share

Clinical Report: Comprehensive Review of Predictive Machine Learning Models in VHD

Overview

This systematic review synthesizes evidence on predictive machine learning models for valvular heart disease (VHD), highlighting the performance of unimodal versus multimodal approaches.

Background

Valvular heart disease (VHD) significantly contributes to cardiovascular morbidity and mortality, especially in older populations. The complexity of VHD, characterized by various valve types and lesion characteristics, complicates clinical prediction and management. Machine learning (ML) has emerged as a promising tool to enhance diagnostic precision and treatment decision-making in VHD.

Data Highlights

Study TypePercentage
Single-lesion models for aortic stenosis38.5%
Retrospective datasets used86%
Internal validation reliance79%
Multimodal models developed8.2%
Performance increase of multimodal models6.3 percentage points

Key Findings

  • 195 studies were identified that met the inclusion criteria for predictive ML models in VHD.
  • 38.5% of studies developed single-lesion models specifically for aortic stenosis.
  • 86% of studies utilized retrospective datasets for their analyses.
  • 79% of studies relied on internal validation of their models.
  • Multimodal models showed a 6.3 percentage point increase in average performance compared to unimodal models.

Clinical Implications

The findings indicate that unimodal ML models are effective for various clinical tasks in VHD.

Conclusion

This review highlights the evolving landscape of predictive ML in VHD.

Related Resources & Content

  1. Basic Research in Cardiology, 2023 -- A Cardiologist's Perspective on Utilizing Machine Learning for Predicting Outcomes in Cardiovascular Disease
  2. Frontiers in Cardiovascular Medicine, 2026 -- Multicenter development and validation of machine-learning risk models to predict procedural complete revascularization and in-hospital heart failure in STEMI patients treated with primary PCI
  3. Frontiers in Medicine, 2026 -- Heart failure risk prediction based on machine learning and interpretability analysis
  4. JMIR Medical Informatics, 2026 -- Multimodal Fusion of Echocardiogram Images and Electronic Medical Records for Heart Disease Screening: Retrospective Algorithm Development and Validation Study
  5. ACC/AHA vs. ESC/EACTS guidelines on valvular heart disease: EJPC guideline comparison review | European Journal of Preventive Cardiology
  6. PARTNER 3: TAVR vs. Surgery in Low-Risk Patients at 7 Years - American College of Cardiology
  7. ACC/AHA vs. ESC/EACTS guidelines on valvular heart disease: EJPC guideline comparison review | European Journal of Preventive Cardiology | Oxford Academic
  8. PARTNER 3: TAVR vs. Surgery in Low-Risk Patients at 7 Years - American College of Cardiology
  9. Two-Year Outcomes of Transcatheter Edge-to-Edge Repair for Severe Tricuspid Regurgitation: The TRILUMINATE Pivotal Randomized Controlled Trial - PubMed

Original Source(s)

Unimodal to multimodal: a systematic review of predictive machine learning models for valvular heart diseases

  • by Valentine Ojonugwa Idakwo, Caren Strote, Christian Goelz, Qasrina Shafei, Thomas J. Stocker, Jörg Hausleiter, Solveig Vieluf

  • July 1, 2026

Related Content

Beyond cardiac fibroblasts: research advances on understanding and targeting intercellular communication networks in cardiac fibrosis

Sex-specific echocardiographic remodeling and prognostic value of left ventricular sphericity, pressure–dimension index, and myocardial work after durable LVAD implantation

Subclinical primary aldosteronism and major adverse cardiovascular events: evidence for a continuum of renin-independent aldosterone excess and a proposal for early detection