Key result
Phenomapping phenogroup 3 is linked to ~3.5-fold higher major cardiovascular events risk in non-ischemic cardiomyopathy.
Why the study?
Despite phenotypic heterogeneity, management of idiopathic non-ischemic cardiomyopathy remains generalized and largely guided by LVEF and NYHA class, motivating development of a point-of-service phenomapping pathway at the time of CMR imaging.
Does point-of-service phenomapping integrating patient-reported health status, CMR imaging, and EHR data improve risk stratification for MACE in patients with dilated and hypokinetic non-dilated non-ischemic cardiomyopathy?
Cohort (n=1,355)
Does point-of-service phenomapping integrating patient-reported health status, CMR imaging, and EHR data improve risk stratification for MACE in patients with dilated and hypokinetic non-dilated non-ischemic cardiomyopathy?
Answers are grounded in this paper and cite their sources.
Point-of-service phenomapping integrating CMR, EHR, and patient-reported data provides incremental prognostic value beyond LVEF and NYHA class for identifying high-risk patients with non-ischemic cardiomyopathy.
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.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) conducted a cohort in Idiopathic non-ischemic cardiomyopathy (dilated cardiomyopathy and hypokinetic non-dilated cardiomyopathy) (n=1,355). Phenogroup 3 (via point-of-service K-means clustering phenomapping) vs. Phenogroup 1 was evaluated on Composite MACE of mortality, heart failure hospitalization, or ventricular arrhythmia (HR 3.5, 95% CI 1.9-6.5, p=<0.001). Phenomapping independently predicted major cardiovascular events in non-ischemic cardiomyopathy, with Phenogroup 3 vs 1 showing aHR 3.5 in DCM and 4.2 in HNDC.
Synapse has enriched 10 related papers. Consider them for comparative context: