An objective feature extraction technique applied to the Doppler waveforms from the groin: a prospective study.

A group of 80 patients presenting with lower limb ischaemia were classified according to their arteriographic findings. The maximum frequency envelope from the Doppler waveforms obtained from the common femoral artery were retrospectively analysed using a mathematical feature extraction technique, principal component analysis (PCA). Two of the three coefficients uniquely describing each waveform were plotted and classified according to their arteriographic groupings. The best separating planes were then defined. Twenty-eight new patients were subsequently prospectively assessed using the principal components and separating planes defined in the previous study. The results were compared with the arteriographic findings and showed that PCA represents a significant improvement in diagnostic accuracy when compared with other techniques.