Implementation Feasibility and Clinical Utility of Phenomapping in Dilated and Hypokinetic Non-dilated Non-ischemic Cardiomyopathy
Year: 2026
Type: article
Study design: Observational Study
Abstract: AIMS: Despite phenotypic heterogeneity, the management of idiopathic non-ischemic cardiomyopathy (NICM) remains generalized and dominantly guided by left ventricular (LV) ejection fraction (EF) and NYHA class. We aimed to develop a point-of-service phenomapping pathway for dilated cardiomyopathy (DCM) and hypokinetic non-dilated cardiomyopathy (HNDC) at time of CMR imaging leveraging the integration of i) patient self-reported health status, ii) imaging phenotype, and the iii) electronic health record (EHR). METHODS AND RESULTS: 1,355 patients (752 DCM and 603 HNDC) were followed for the composite MACE outcome of mortality, heart failure hospitalization or ventricular arrhythmia. K-means clustering was independently applied to DCM and HNDC derivation cohorts using 55 routinely captured features (16 patient-reported health variables, 19 CMR-derived, 14 EHR-derived, and 6 combined features), identifying three phenogroups in each cohort. Over a median of 3.9 and 3.7 years, 145 DCM patients (22%) and 65 HNDC patients (13%), respectively, experienced MACE. Adjusting for age, sex, NYHA class, and LV EF, phenogroup membership was independently associated with future outcomes. In DCM, the adjusted HR (95% CI) of phenogroups 2 and 3 versus Phenogroup 1 were 2.8 (1.6-4.8) and 3.5 (1.9-6.5), respectively (p<0.001 for both). In HNDC, Phenogroup 3 demonstrated an adjusted HR of 4.2 (1.7-10.2; p=0.001) versus Phenogroup 1. Prospective phenogroup assignment was enabled by a decision tree-classifier and validated in a temporal hold-out cohort. CONCLUSION: Phenomapping delivers incremental value to LVEF and NYHA for identification of high-risk NICM patients. These findings support value from standardized, point-of-service data integration for the delivery of phenomics-driven personalized care.
Language: English
Subfield: Cardiology and Cardiovascular Medicine
Field: Medicine
Domain: Health Sciences
Keywords: phenomapping, phenomics, dilated cardiomyopathy, non-ischemic cardiomyopathy, patient-reported health status, cardiovascular magnetic resonance, electronic health record integration, phenogroups, major adverse cardiovascular events, risk stratification, decision tree classification, K-means clustering, individualized care
Open Access status: hybrid
Estimated APC paid: $5,203
OpenAlex ID: W7236620051
