Abstract
435
Objectives: To estimate coronary flow reserve (CFR = stress myocardial blood flow (MBF) / rest MBF) in the ischemic myocardial lesion with 15O-H2O PET, we developed a new algorithm to improve the detectability of CFR using electro-cardiogram (ECG)-gated dynamic myocardial PET with 15O-H2O, which would reduce the left ventricular (LV) wall motion artifact. Estimated CFR from the new method were compared with those from the conventional non-gated dynamic PET.
Methods: Twenty-seven patients with ischemic heart disease (67±11 years old, male 17, female 10) underwent dynamic 15O-H2O PET during rest and pharmacological (ATP) stress. The PET study was performed using Philips Gemini TF64, list mode 3D PET data were acquired with ECG signals. For each scan, 500MBq of 15O-H2O was infused slowly for 2 min, dynamic data were acquired for 6 min. Using list mode PET data and ECG signals, both non-gated dynamic images and ECG-gated end-diastolic dynamic images were reconstructed. Coronary arteriograms were also carried out, totally 50 ischemic segments with over 90% stenosis and 128 normal segments were evaluated in these patients. Rest and stress myocardial blood flow (MBF) in 17 segments in each LV wall were estimated. A whole myocardial ROI was positioned and divided into 17 segments. The regional myocardial ROI curve R(t) and left ventricular ROI curve LV(t) were calculated, regional MBF, CFR, perfusable tissue fraction (PTF) and spillover fraction Va in each myocardial ROI were estimated by these curves with single-tissue compartment model analysis with non-linear least squared method as following equations; R(t) = PTF Ct(t) + Va Ca(t), LV(t) = b Ca(t) + (1-b) Ct(t), dCt(t) / dt = MBF Ca(t) - (MBF/p) Ct(t), where Ca(t) and Ct(t) were true arterial and myocardial tissue curves, parameters p and b were partition coefficient in the myocardium and recovery coefficient of the LV ROI counts, respectively. Both non-gated dynamic images and ECG-gated end-diastolic dynamic images were analyzed. Development environment for these analyzing tools were Visual C++ and C#.
Results: In the ischemic segments, stress MBF values with ECG-gated PET were estimated significantly lower (0.999±0.264 (ml/min/g)) than those with non-gated PET (1.241±0.443 (ml/min/g), p<0.001). In the ischemic segments, CFR values with ECG-gated PET were estimated significantly lower (1.226±0.291) than those with non-gated PET (1.694±0.714, p<0.001). However, in the normal segments, ECG-gated PET dedicated no significantly different CFR (2.154±0.755) compared with those from non-gated PET (2.238±0.790).
Conclusions: We developed a technique to estimate the coronary flow reserve in the ischemic myocardial lesion using ECG-gated dynamic myocardial PET with 15O-H2O. ECG-gated PET enabled to suppress the LV wall motion artifact, and could estimate the decrease of the coronary flow reserve in the ischemic myocardial lesion. ECG-gated dynamic myocardial PET improved the detectability of the ischemic myocardial lesion than the conventional non-gated myocardial dynamic PET.