Factors affecting and computation of myocardial perfusion reference images

Many quantitative analysis methods for myocardial perfusion studies require as a central step a comparison with a 'normal' or average density distribu tion map or reference image. It has been recognized, however, that the norm al distribution can be affected by patient attributes, including sex and we ight or body habitus, and by acquisition attributes, including the choice o f tracer and the position of the patient during imaging. Some authors have proposed separate reference images for the sexes and the tracer. This appro ach fails if a large number of binary attributes have to be considered, sin ce one would need 2(n) reference images for each attribute. The problem is compounded when continuous attributes (e.g. age and weight) are included, e specially if the approach is to average separate homogeneous groups for eac h attribute. We propose to create case-specific reference images for the in terpretation of myocardial perfusion studies by creating a model based on t he influence of each attribute. From a non-homogeneous population of normal cases, or cases presumed to be normal on the basis of the Diamond and Forrester stratification, the effect of patient and study attributes on the density distribution in the stress image and the density differences between rest and stress images were compu ted. The effects are computed by multi-linear regression, to account for cr oss-correlation. Significance is assigned on the basis of a partial Fisher test. The data are myocardial perfusion images matched in 3D to a template by an elastic transformation. Even though there was some cross-correlation in the data, we were able to s how independent effects of sex, position (prone or supine), age, weight, tr acer combination and stress method (exercise, persantine and adenosine). Ta ken as a whole, the multi-linear regression demonstrated a significant effe ct in 72% of the pixels within the myocardial volume. In addition, the dist ribution predicted by the model was equivalent to average images from homog eneous matched groups. In conclusion, our approach makes it possible to pro duce case-specific reference images without the need for multiple homogeneo us large groups to produce averages for each possible patient or study attr ibute. ((C) 1999 Lippincott Williams & Wilkins).