Quantification of myocardial blood flow using dynamic 320-row multi-detector CT with Gauss curve fittingPET

Objectives: Measurements of myocardial blood flow (MBF) and coronary flow reserve (CFR) using multi-detector CT (MDCT) have been reported for evaluating the physiological significance of coronary lesions. However, there was a case in which the beginning of the curve was not stabilized due to deviation of imaging timing or motion. To obtain stable MBF even in such cases, Gaussian curve correction was useful to interpolate the unstable part of the curve. We developed a new algorithm using Gauss curve fitting in the density curve of the myocardial tissue. The aim of this study was to quantify the MBF using dynamic CT with 320-row MDCT with our new algorithm and compared to the conventional method.

Methods: Fifteen patients (64±11 years old) with suspected coronary artery disease were enrolled. All patients refrained from drinking caffeinated beverages for at least 24 hours, eating for more than 6 hours and smoking for at least 4 hours before the scan. Rest dynamic CT was followed by Stress dynamic CT with intravenous ATP infusion (0.016 mg/kg/min) using a second-generation 320-row MDCT system (Aquilion ONE, ViSION Edition, Toshiba Medical Systems). Time attenuation curves (TACs) of Hounsfield (HU) value in the left ventricle (LV) chamber and LV myocardium were derived from the dynamic CT data. HU value of descending aorta was subtracted to estimate the stable dynamic curve from HU value of LV pool. Gauss curves are added to smooth the dynamic curves. Using single-tissue compartment model and Renkin-Crone model analysis, regional MBFs were calculated in the Left anterior descending coronary artery (LAD), Right coronary artery (RCA) and Left circumflex coronary artery (LCX) territories from dynamic CT. These estimates were compared with MBF from 15O-H2O PET.

Results: MBF from our new method yielded significant better relationship with MBF from 15O-H2O PET compared to the conventional method. (R=0.851, 0.612, respectively; p < 0.001) (Figure a b). Bland -Altman analysis stated that Gauss curve fitting has less variation in correlation than conventional methods (mean=0.061, -0.323 SD=1.13, 0.960 respectively)(Figure c d).

Conclusion: We developed a new algorithm to estimate MBF with Gauss curve fitting in the TACs of the dynamic CT data. Gauss curve fitting was useful to interpolate MBF from dynamic CT scan.

• Download figure
• Open in new tab
• Download powerpoint