Quantitative sorting of normal and abnormal coronary flow wave form shapes

The normal phasic flow wave form in an epicardial coronary artery has a distinct characteristic shape, which reflects the interaction between the coronary tree, myocardial function and hemodynamic conditions. Since clinical measurements of phasic coronary wave forms are becoming available, determination of abnormal coronary flow wave forms is important. The authors suggest here an objective and automatic method to discriminate between normal and abnormal flow wave forms based on the Karhunen-Loeve transform (KLT), and experimentally test it. The normal flow domain was represented by the resting flow waves measured in the left anterior descending arteries in 31 anesthetized dogs. The abnormal flow conditions, imposed and tested experimentally, were varying stenosis severity and severely reduced left ventricular pressure. In addition, the effects of reactive hyperemia on the shape of the flow were examined. The sorting index was based on the mean-square error (MSE) calculated for each flow signal based on a truncated KLT expansion. The results show excellent discrimination between the normal and the abnormal groups. During reactive hyperemia, however, MSE did not change significantly. These results indicate that the shape of abnormal coronary flow wave forms can be identified and discriminated from normal wave forms.