Interobserver variability in interpreting contrast left ventriculograms (CASS)

Randomly selected left ventricular cineangiograms performed by 14 medical centers (clinical sites) participating in a large multicenter study of medical and surgical treatment of coronary artery disease (CASS) were systematically recalled and reread by one of four designated quality control centers. The end diastolic volume (EDV), end systolic volume (ESV), and ejection fraction (EF) calculated by the quality control center were compared with those calculated by the clinical site, and correlation coefficients of 0.71, 0.84, and 0.79, respectively, were obtained. Regional left ventricular (LV) wall motion abnormalities were assessed by dividing the LV wall as viewed in the right anterior oblique (RAO) and left anterior oblique (LAO) views into five segments each and grading the motion of each segment according to six categories of increasing abnormalities from normal to aneurysm. When discrepancies based on "nearest neighbor" differences in naming adjacent segments were eliminated and only marked differences of more than one degree of severity of wall motion abnormality were considered, interobserver differences were 7% and 10% for RAO and LAO segments, respectively. Agreement was less frequent in designating a segment "aneurysmal." When each LV segment was assigned a numerical score based on increasing severity of wall motion abnormality and a total LV score for each ventriculogram computed, the quality control and clinical site readings were in good agreement with a correlation coefficient of 0.83.