Precision Phenotyping of Dilated Cardiomyopathy Using Multidimensional Data

Abstract

Background: Dilated cardiomyopathy (DCM) is a final common manifestation of heterogenous etiologies. Adverse outcomes highlight the need for disease stratification beyond ejection fraction.

Objectives: The purpose of this study was to identify novel, reproducible subphenotypes of DCM using multiparametric data for improved patient stratification.

Methods: Longitudinal, observational UK-derivation (n = 426; median age 54 years; 67% men) and Dutch-validation (n = 239; median age 56 years; 64% men) cohorts of DCM patients (enrolled 2009-2016) with clinical, genetic, cardiovascular magnetic resonance, and proteomic assessments. Machine learning with profile regression identified novel disease subtypes. Penalized multinomial logistic regression was used for validation. Nested Cox models compared novel groupings to conventional risk measures. Primary composite outcome was cardiovascular death, heart failure, or arrhythmia events (median follow-up 4 years).

Results: In total, 3 novel DCM subtypes were identified: profibrotic metabolic, mild nonfibrotic, and biventricular impairment. Prognosis differed between subtypes in both the derivation (P < 0.0001) and validation cohorts. The novel profibrotic metabolic subtype had more diabetes, universal myocardial fibrosis, preserved right ventricular function, and elevated creatinine. For clinical application, 5 variables were sufficient for classification (left and right ventricular end-systolic volumes, left atrial volume, myocardial fibrosis, and creatinine). Adding the novel DCM subtype improved the C-statistic from 0.60 to 0.76. Interleukin-4 receptor-alpha was identified as a novel prognostic biomarker in derivation (HR: 3.6; 95% CI: 1.9-6.5; P = 0.00002) and validation cohorts (HR: 1.94; 95% CI: 1.3-2.8; P = 0.00005).

Conclusions: Three reproducible, mechanistically distinct DCM subtypes were identified using widely available clinical and biological data, adding prognostic value to traditional risk models. They may improve patient selection for novel interventions, thereby enabling precision medicine.

Keywords: heart; machine learning; proteomics.

Graphical abstract

Figure 1

An Overview of the Study Analysis Pipeline Machine learning approaches were applied to multiparametric data (clinical, imaging, genetics, biomarkers) from a prospectively recruited UK derivation cohort of patients with dilated cardiomyopathy (DCM) and identified 3 novel reproducible subtypes of disease: mild nonfibrotic, profibrotic metabolic, and biventricular impairment. Multinomial logistic regression was used to create a model to place patients in the independent Dutch validation cohort into corresponding subtypes. Composite survival differed between novel subtypes in both the derivation and validation cohorts. CMR = cardiovascular magnetic resonance; ECG = electrocardiogram; LAVi = indexed left atrial volume; LGE = late gadolinium enhancement; LVESVi = indexed left ventricular end-systolic volume; NL = the Netherlands; PCA = principal component analysis; RVESVi = indexed right ventricular end systolic volume; UK = United Kingdom.

Figure 2

UK Derivation Cohort: Outcome for the 3 Novel Disease Subtypes Phenotypic group 2 (PG2) is a novel distinct profibrotic metabolic subtype of DCM. PG2 patients had the highest rates of diabetes mellitus; all had midwall myocardial fibrosis, and they had experienced more ventricular tachycardia compared with the other groups. PG2 patients had intermediate values between PG1 and PG3 for several left ventricular measurements (left ventricular ejection fraction and end-diastolic and end-systolic volumes) but similar right ventricular structure and function to PG1. Composite survival consists of major arrhythmic events, major heart failure events, or cardiovascular mortality. Outcome varied by these novel DCM disease subtypes. P value is computed by the log-rank test.

Figure 3

Dutch Validation Cohort: Outcome for the 3 Disease Subtypes The novel disease subtypes in the validation cohort also vary by adverse event risk. Composite survival consists of major arrhythmic events, major heart failure events, or cardiovascular mortality. P value is computed by the log-rank test.

Figure 4

IL4RA Is a Novel Dilated Cardiomyopathy Prognostic Biomarker IL4RA was strongly associated with outcome in both derivation and validation cohorts. Unadjusted and adjusted HRs for IL4RA in the derivation cohort are shown. HRs are presented per 1-SD (ie, standardized to make a fair comparison across biomarkers). In adjusted analyses, IL4RA remained of prognostic utility in addition to clinical factors∗ that predicted outcome (indexed left atrial volume, left ventricular ejection fraction, midwall fibrosis on cardiac magnetic resonance, and a history of nonsustained ventricular tachycardia), as well as in addition to conventional prognostic biomarkers NT-proBNP and high sensitivity troponin I. This suggests that IL4RA is a novel prognostic marker for dilated cardiomyopathy. IL4RA = interleukin 4 receptor alpha; NT-proBNP = N-terminal pro-B-type natriuretic peptide.

Central Illustration

Machine Learning Approaches to Dilated Cardiomyopathy Identify 3 Novel Disease Subtypes Machine learning approaches applied to a prospectively recruited UK derivation cohort of patients with dilated cardiomyopathy identified 3 novel reproducible disease subtypes: mild nonfibrotic, profibrotic-metabolic, and biventricular impairment. Prognosis varied among groups and was reproduced in the independent Dutch validation cohort. The novel profibrotic-metabolic subtype had a high rate of diabetes, universal myocardial fibrosis, elevated creatinine, and preserved right ventricular function. For clinical application, 5 variables were sufficient for classification. CMR = cardiovascular magnetic resonance; LAVi = indexed left atrial volume; LGE = late gadolinium enhancement; LV = left ventricular; LVESVi = indexed left ventricular end-systolic volume; RV = right ventricular; RVESVi = indexed right ventricular end systolic volume.