DOI: 10.1161/circ.148.suppl_1.16201 ISSN: 0009-7322

Abstract 16201: Use of Cardiac Computed Tomography (CT) Imaging Biomarker Variables for the Prediction of Incident Heart Failure: Multi-Ethnic Study of Atherosclerosis

Aditya Mehta, Angelo De La Rosa, Spencer Hansen, Robyn McClelland, Alain G Bertoni, Matthew J Budoff
  • Physiology (medical)
  • Cardiology and Cardiovascular Medicine

Introduction: Several predictive heart failure (HF) models exist to determine incident HF. We aimed to create a model that uses cardiac CT imaging biomarkers to improve discrimination scores of incident HF.

Hypothesis: Cardiac CT variables increase predictive abilities of the Pooled Cohort Equations to Prevent HF (PCP-HF) score (a validated 10-year risk of new-onset HF prediction model) in all HF, HFrEF, and HFpEF.

Methods: MESA participants aged 45-84 years old and free of clinical CVD who completed a cardiac CT were included for study analysis. The outcome of interest was new-onset HF. Clinical risk factors were obtained. Cardiac CT variables analyzed included left ventricular size index (LVSi) and calcifications of coronary arteries (CAC), aortic valve (AVC), mitral valve (MVC), and thoracic aorta (TAC).

Results: Among 6,667 MESA study participants who underwent cardiac CT, 426 events of new-onset HF occurred during the follow-up period. Among the 426 events, 173 (40.6%) were categorized as HFrEF, 193 (45.3%) were categorized as HFpEF, and 60 (14.1%) had missing ejection fraction data. The reported data is based on the Cox model adjusted for all CT variables in one model with log(PCP-HF). For all incident HF (preserved and reduced), CAC (HR 1.10 95% CI 1.05-1.15, p<0.001) and LVSI (HR 1.14 95% CI 1.10-1.18, p<0.001) were statistically significant with an improvement in c-stat from .773 to .805. For HFpEF, CAC (HR 1.07 95% CI 1.01-1.14, p<0.025), AVC (HR 1.08 95% CI 1.00-1.16, p<0.037) and MVC (HR 1.12 95% CI 1.05-1.19, p<0.001) were statistically significant with an improvement in c-stat from .782 to .800. For HFrEF, CAC (HR 1.15 95% CI 1.08-1.23, p<0.001) and LVSI (HR 1.28 95% CI 1.23-1.34, p<0.001) were statistically significant with an improvement in c-stat from .751 to .815.

Conclusions: Current HF prediction models, including PCP-HF model, are based on clinical risk factors. As evidenced by this model CAC, AVC, MVC and LVSi have statistical significance with improvement in c-statistics of the PCP-HF model when adjusted for CT variables. HFpEF and HFrEF have different CT variables with enhanced predictive ability.This study emphasizes the need to study the utility of cardiac CT imaging biomarkers in the development of an incident HF predictive model.

More from our Archive