This study aimed to evaluate the interobserver variability of right-to-left ventricular (RV/LV) diameter ratios on computed tomographic pulmonary angiographies at reduced tube potential.
Consecutive computed tomographic pulmonary angiographies performed at 80 kVp for subjects weighing less than 80 kg (n = 40) and 100 kVp for those weighing 80 kg or greater (n = 40) were selected as the low peak kilovoltage cohort, with age- and sex-matched studies performed at 120 kVp forming the controls (n = 40 each). The interobserver variability in RV/LV diameter ratio on axial and 4-chamber images was evaluated.
Interobserver correlation was excellent at both reduced tube voltages (correlation coefficients, 0.81–0.91; P < 0.001). Bland-Altman plots indicated higher variability at lower peak kilovoltage, confirmed only for axial measurements by multivariate regression adjusting for body weight and ventricular ratio (β coefficient = 0.04 and 0.03, P = 0.049 and 0.022, for 80 kVp and 100 kVp, respectively).
Lower tube voltage did not affect the reproducibility of RV/LV diameter ratio on 4-chamber images; for axial measurements, the increase in variability was 3% to 4% of the average ventricular ratio.