Automated quantitation of pericardiac fat from noncontrast CT

Increased abdominal visceral fat has been shown to be a cardiovascular risk factor. Preliminary studies indicate that pericardiac fat (PF) may provide similar information. We aimed to develop new software (QFAT) for automatic quantitation of PF from noncontrast cardiac CT and compare PF measures to other cardiovascular risk factors.

QFAT accepts user-defined range of noncontrast transverse cardiac CT slices, automatically segments the heart, and determines PF volume (PFV) as contiguous pericardial fat voxels. PFV normalized to cardiac volume defines PF ratio (PFR). QFAT and manual processing (MAN) was performed in 105 patients (mean BMI, 27; range, 17-41) by 2 observers.

Mean processing time was 20 +/- 4 seconds for QFAT, and 9 +/- 6 minutes for MAN. There was excellent agreement between QFAT and MAN for PFV (R = 0.98) and PFR (R = 0.98). MAN and QFAT interobserver variability were comparable. Interscan and interscanner variability for PFV and PFR were comparable to corresponding interobserver variability. PFV (R = 0.88, P < 0.0001) and PFR (R = 0.81, P < 0.0001) correlated strongly with abdominal visceral fat area, moderately with BMI (R = 0.58, P < 0.0001 and R = 0.48, P < 0.0001), and weakly with abdominal subcutaneous fat area (R = 0.33, P < 0.0001 and R = 0.32, P = 0.001).

PFV and PFR can be accurately and automatically quantified from noncontrast CT acquired for coronary calcium screening and may provide complementary information regarding cardiovascular risk.