Abstract 21259: Pericardial Fat Volume Measured From Non-contrast CT Significantly Improves Prediction Of Myocardial Ischemia
Background: Pericardial fat volume (PFV) and thoracic fat volume (TFV) measured from noncontrast CT (NCT) has been shown to be associated with an increase in Coronary Calcium Score (CCS) and major adverse cardiovascular event risk. We examined the association between these measures and myocardial ischemia by single photon emission computed tomography (SPECT).
Methods: From a cohort of 1777 consecutive patients without known coronary artery disease (CAD) in whom NCT was performed within 6 months of myocardial perfusion SPECT (MPS), we compared 73 patients with ischemia by MPS to 146 patients with normal MPS matched by age, gender, CCS categories, symptoms and risk factors for CAD. TFV was measured automatically and PFV was measured semi-automatically by an experienced reader who defined the pericardial contour, within which fat voxels were automatically identified. MPS defect assessment was performed using computer-assisted visual interpretation by an expert reader using the standard 17-segment, 5 point-scoring model (≥5% perfusion stress-rest difference = ischemia). Prediction of ischemia by CCS, PFV and TFV were examined using conditional multivariable logistic regression, Receiver Operating Characteristic (ROC) and Net Reclassification Improvement (NRI) analysis.
Results: Cases had higher mean PFV (99.1±42.9 vs. 80.1±31.8 cm3, p=0.0003) and TFV (196.1±82.7 vs. 160.8±72.1 cm3, p=0.001) and higher frequencies of PFV>125 cm3 (22% vs. 8%, p=0.004). After adjusting for CCS, PFV and TFV remained the strongest predictor of ischemia [odds ratio (OR) 2.97, 95% Confidence interval (CI) 1.53–5.76 for each doubling of PFV; OR 2.79, 95%CI 1.5–5.2 for TFV]. ROC area-under-the curve improved significantly (0.75 vs. 0.69, p=0.04) when PFV (or TFV) was added to clinical risk factors and CCS. NRI for ischemia prediction increased significantly when PFV was added to clinical risk factors and CCS (31.5%, 95% CI: 15.1–47.8%, p<0.001); an additional 16% of patients with ischemia were reclassified as high risk, and an additional 15% without ischemia were reclassified as low risk.
Conclusion: Addition of pericardial fat to a prediction model based on clinical risk factors and CCS significantly improves prediction of ischemia, in patients without history of CAD.
- © 2010 by American Heart Association, Inc.