Effect of filming projection and interobserver variability on angiographic biplane left ventricular volume determination.

Abstract

Although biplane right anterior oblique-left anterior oblique (RAO/LAO) quantitative left ventricular (LV) angiography is commonly performed, justification of LV volume calculation using the area length method (originally formulated from anteroposterior-lateral (AP/LAT) angiograms) has been limited. To assess whether RAO/LAO and AP/LAT LV volumes are similar when computed by the area length method formula, we performed biplane cine LV angiography in both RAO/LAO and AP/LAT projections in random sequence in 21 patients and four LV models of known volume. LV silhouettes were drawn independently by two trained observers. Calculated angiographic volume of the models correlated almost exactly with their true volume (r = 0.999), establishing the absolute accuracy of this system. Rotation of the LV models through 90 degrees of obliquity at 10 degree increments demonstrated a mean change from true volume of only -5.4 +/- 0.7% (p less than 0.001). In the patient studies, rotation to the 30 degree RAO/60 degree LAO position was associated with significant changes in magnitude of biplane areas and long axes, but area length volume estimates were unchanged. Excellent correlation was found between area length calculated AP/LAT and RAO/LAO volumes with r = 0.90, 0.97, and 0.91 for end-diastolic volume (EDV), end-systolic volume (ESV) and ejection fraction (EF), respectively. Furthermore, interobserver agreement in volume assessment was excellent, with r = 0.98, 0.99, and 0.94 between observers for EDV, ESV, and EF, respectively. Interobserver and inter-method variability for estimates of LV volume and EF ranged from 5--10%. We conclude that when using RAO/LAO LV angiography, volume calculation by the area length method is justified.