Valvular & structural heart diseaseCohort study
Machine Learning Identification of Patient Phenoclusters in Aortic Regurgitation
- Multicentre cohort study with unsupervised machine learning clustering
- 4 US centres
- derivation (2 centres) and validation (2 centres) cohorts
Authors' conclusion
Unique patient phenoclusters of AR are described using a machine learning approach leveraging comprehensive CMR and clinical characterization.
Study context
- Population
- Patients with moderate or severe aortic regurgitation with clinical and CMR characterization
- Participants
- 972 patients
- Exposure
- Four machine-learning-derived phenoclusters
- Outcome
- All-cause death
- Follow-up
- Median 2.58 years
Main finding
As reported in the abstract
All-cause mortality by phenocluster
- Overall mortality rate
- 12%
- Clustering independently associated with survival after adjustment for time-dependent AVR and traditional risk markers (C statistic, validation cohort)
- 0.77 vs 0.75; P = 0.009
Observational data: associations, not effects.
Original publication
Machine Learning Identification of Patient Phenoclusters in Aortic Regurgitation. JACC Cardiovasc Imaging. Published 26 March 2025. doi:10.1016/j.jcmg.2025.01.006
Summarised from the published abstract; every number is checked against it