Reproducibility and accuracy of ABC/2 and 3D techniques in ICH volume measurement

Abstract

Purpose The aims of this study were to validate 3D volume measurement and to compare ICH volume determined by conventional measurements. Materials and methods Data of all 50 patients were retrospectively collected. The ICH volume was compared using ABC/2 technique and 3D volumetric technique by two observers. Interobserver and intraobserver variations were calculated. Variability of measurements was analysed. Results The mean volume was 41.43 ± 3.0 mm 3 using 3D technique versus 45.37 ± 3.9 mm 3 using ABC/2 technique. The coefficients of variation for interobserver and intraobserver variability of 3D and ABC/2 technique were 5.03% versus 6.28% and 2.78% versus 8.77%. ABC/2 technique generally overestimated ICH volume. The 95% confidence interval (CI) of 3D and ABC/2 technique was 29.4–53.5 mm 3 and 32–58.6 mm 3. There was no significance difference in volume measurement by variance ratio test (F-test), where p = 0.527. The correlation coefficient of ABC/2 and 3D method was 0.99 ( p < 0.001, 95% CI 0.98–0.99). Box-and-whisker graph showed that ABC/2 technique generally overestimated the ICH volume with irregular margin. The results showed that (a) ICH volumes estimated by the ABC/2 technique do not significantly differ from 3D technique. (b) ABC/2 technique was accurate when estimating volume with regular margin but overestimated volume in irregular margin. (c) Reproducibility of 3D technique in ICH volume with irregular margin was higher. (d) Reproducibility of ABC/2 technique in ICH volume with regular margin was higher. Conclusions 3D technique is the most reproducible technique. Although the ABC/2 technique slightly overestimates ICH volume with irregular margin, the difference was not significant. In view of the software requirement for 3D measurement, ABC/2 technique is considered a rapid and reliable volume measurement technique.