Effects of cardiac phase on diameter measurements from coronary cineangiograms

The end-diastole is the optimum portion of the cardiac cycle for computerized analysis of coronary angiograms, since vessel movement is minimal and the vessels are relatively straight. Since it is frequently not possible to find suitable end-diastolic frames, the authors initiated a study to determine the penalty in precision that is incurred when other than end-diastolic frames are processed. Computer measurements of average and minimum diameter were made for every frame of two complete cardiac cycles in angiograms of 20 subjects who had received bypass grafts. To estimate the effects of phase on the measurement, six sampling schemes were defined: beginning systole, end-diastole, mid-diastole, anywhere in diastole, anywhere in systole, and anywhere in the cardiac cycle. Two frames were selected for each scheme and the mean and standard deviation (SD) computed. Frames for the last three schemes above were randomly selected 50 times and the results averaged. The SD of average diameter is half that of the minimum diameter and relatively independent of cardiac phase. The SD of the minimum diameter is lowest for end-diastole, but an approximately 25% penalty in variability is paid when sampling elsewhere