Doppler ultrasound diagnoses are carried out mainly in cardiovascular clinics,where they have become routine tests. The diagnostic procedure is complex and takes time, imposing considerable burdens on both patient and technician. To solve this problem, methods of automating ultrasonic Doppler blood flow measurements have been under development for quite some time. The present study considers the possibility of an automated diagnostic support technology using the Mahalanobis-Taguchi system, based on automated Doppler measurement. An analysis using automatically traced Doppler waveforms and measured diagnostic indicators (values) related to the left ventricle distinguished between healthy and diseased subjects with sufficient accuracy. An analysis using only the measured diagnostic indicators to determine the presence of heart conditions had difficulty in distinguishing between healthy and diseased subjects, but favorable results were obtained from analyses targeted at some specific diseases (DCM, MS, MR). Further study will be carried out to improve the performance of the automatic diagnostic support technology by combining the diagnostic indicators found to be effective with automatically traced Doppler waveforms.
