Automated eigenvector analysis for quantification of color M-mode Doppler filling patterns of the left ventricle in an ischemic canine model

The feasibility of an automated set of computer algorithms for quantification of color M-mode Doppler patterns of the left ventricle (LV) was investigated at baseline and during ischemia, and these parameters were related to invasive measures. Six open-chest anesthetized dogs were instrumented by a multisensor high fidelity catheter; with sensors at base of LV, in apex, and in left atrium. Early diastolic filling waves were processed by eigenvector analysis, with the parameters: centroid (s,t), eigenvalues (ss, tt), and angle of rotation (theta). Maximum -dP/dt decreased from 1114/spl plusmn/174 to 562/spl plusmn/91 mmHg/s (p<0.05), and tau increased from 73/spl plusmn/4 to 122/spl plusmn/16 ms (p<0.05). Timing parameters of the eigenvector analysis increased significantly, while theta, which is similar to flow propagation, decreased from -0.71/spl plusmn/0.30 to -3.39/spl plusmn/0.66 degrees (p<0.01). Tau showed a high correlation with theta (r=-0.95, p<0.0001). Automated eigenvector analysis of color M-mode patterns is feasible, and quantifies significant changes during ischemia.