Quantification of myocardial perfusion for CAD diagnosis.

We introduce a computer based algorithm for objective quantification of myocardial perfusion to support the diagnosis of cad patients. This new method is based on conventional cine angiographic films. In order to achieve maximal quality of the digital subtraction angiography images, the sequence is synchronized with the ECG. Optionally, the digital images can be motion compensated by a two step matching method. The spatio-temporal spread of blood, or the so-called blush, through the microvasculature to the myocardium--indicated by dye injection--represents a characteristic pattern for the myocardial perfusion. This dynamic temporal pattern is characterized by typical features as the maximal value of blush intensity, of increase and of decrease velocity which correspond with the different phases of flooding in and washout. On the basis of 100 different temporal blush profiles, an algorithm is established which classifies the acquired blush patterns into 4 different grades.