Improvement of estimation of regional myocardial blood flow in O-15 water PET heart study

Purpose: PET is the only technique that can provide noninvasive absolute quantification of regional myocardial blood flow (MBF) in humas. O-15 labeled water is an ideal tracer in measuring MBF, and dynamic water study has been done widely in the world. However, low-quality and noisy images from heart study cause the difficulties in analyzing these data, especially in high flow situations, such as induced by exercise or by medicine. The present study is designed to improve the analysis result for water paired rest/stress studies. Methods: Since the myocardial region is very difficult to identify in dynamic O-15 labeled water study image, traditional way is to use C-15 CO to locate myocardial regions. This is a cumbersome procedure for patients. In present study, MBF is estimated using O-15 water without additional C-15 CO measurement. Myocardial regions are enhanced in following ways: 1). Difference image: the summary of the second part minus the summary of first part in dynamic image. 2). ICA (independent component analysis) image. 3). Combination of difference image and ICA image. In rest/stress paired studies, ROI (region of intereset) is obtained from rest study (manually or automatically), then the rest ROI is copied to the corresponding stress study. Furthermore, copied ROI is corrected in three dimensions manually. Finally, one-compartment model is employed to calculate MBF. These features have been implemented in a heart PET analysis package, SpringThunder (http://users.utu.fi/chuhan/springthunder. html, developed by author using IDL language). This package can be run on IDL virtual machine (free download http://www.rsinc.com) environment Results: Difference image and/or ICA image or combination of both can provide statisfied myocardial regions in O-15 water heart dynamic study. These image processing techniques avoid the aid of additional C-.15 CO scanning. In a typical study (stress induced by drug), MBF values are as following (rest/stress(no corrected)/stress(corrected)): apex 0.84/1.28/2.87, anterior 0.65/2.83/3.82, septum 0.62/0.58/1.37, inferior 0.95/1.20/3.26, lateral 0.73/1.56/2,73. Conclusions: MBF can be estimated in O-15 water study without additional aid scan. Three-dimensional correction of ROIs in O-15 labeled water paired studies (rest/stress) is necessary and can improve MBF significantly. (authors)