Thirty-three cases of anthracycline related cardiotoxicity occurred in our institution in patients with a previously negative cardiovascular history, physical examination, and normal ECG. A total of 95 RNCA studies were performed in this group (73 studies included both rest and exercise RNCA). Twenty-one patients had two or more serial studies. Seventeen had a normal, baseline prechemotherapy study and 16 had studies done following the initiation of therapy. Fourier analysis, consisting of amplitude and phase images, were created for each study. In the subset with a baseline study, the rest LVEF became abnormal first in two of 17 patients (12%), the rest or stress LVEF in ten of 17 (59%), the rest Fourier image analysis in ten of 17 (59%), and the rest or stress Fourier image analysis in 16 of 17 (94%). In the subset without a baseline study, the rest LVEF was abnormal in ten of 16 (63%), the rest or stress LVEF in 15 of 16, (94%), the rest Fourier image analysis in 16 of 16 (100%), and the rest or stress Fourier image analysis in 16 of 16 (100%). The authors conclude that: 1) the exercise RNCA is superior to the rest RNCA alone in the early detection of anthracycline related cardiotoxicity, 2) the single most sensitive indicator of cardiotoxicity is Fourier image analysis; and 3) sequential rest and stress RNCA studies with Fourier amplitude and phase analysis is the most sensitive, noninvasive method of evaluating patients who receive potentially cardiotoxic agents.