Accuracy of Myocardial Blood Flow Estimation From Dynamic Contrast-Enhanced Cardiac CT Compared With PET.

Background: The accuracy of absolute myocardial blood flow (MBF) from dynamic contrast-enhanced cardiac computed tomography acquisitions has not been fully characterized. We evaluate computed tomography (CT) compared with rubidium-82 positron emission tomography (PET) MBF estimates in a high-risk population. Methods: In a prospective trial, patients receiving clinically indicated rubidium-82 PET exams were recruited to receive a dynamic contrast-enhanced cardiac computed tomography exam. The CT protocol included a rest and stress dynamic portion each acquiring 12 to 18 cardiac-gated frames. The global MBF was estimated from the PET and CT exam. Results: Thirty-four patients referred for cardiac rest-stress PET were recruited. Of the 68 dynamic contrast-enhanced cardiac computed tomography scans, 5 were excluded because of injection errors or mismatched hemodynamics. The CT-derived global MBF was highly correlated with the PET MBF (r=0.92; P<0.001) with a mean difference of 0.7 +/- 26.4%. The CT MBF estimates were within 20% of PET estimates (P<0.02) with a mean of (1) MBF for resting flow of PET versus CT of 0.9 +/- 0.3 versus 1.0 +/- 0.2 mL/min per gram and (2) MBF for stress flow of 2.1 +/- 0.7 versus 2.0 +/- 0.8 mL/min per gram. Myocardial flow reserve was -14 +/- 28% underestimated with CT (PET versus CT myocardial flow reserve, 2.5 +/- 0.6 versus 2.2 +/- 0.6). The proposed rest+stress+computed tomography angiography protocol had a dose length product of 598 +/- 76 mGyxcm resulting in an approximate effective dose of 8.4 +/- 1.1 mSv. Conclusions: In a high-risk clinical population, a clinically practical dynamic contrast-enhanced cardiac computed tomography provided unbiased MBF estimates within 20% of rubidium-82 PET. Although unbiased, the CT estimates contain substantial variance with an standard error of the estimate of 0.44 mL/min per gram. Myocardial flow reserve estimation was not as accurate as individual MBF estimates.