Comparison of Intracerebral Hemorrhage Volume Calculation Methods and Their Impact on Scoring Tools.

Khan, Muhib, Grayson L Baird, Roderick Elias, Joshua Rodriguez-Srednicki, Shadi Yaghi, Sandra Yan, Scott Collins, et al. 2017. “Comparison of Intracerebral Hemorrhage Volume Calculation Methods and Their Impact on Scoring Tools”. Journal of Neuroimaging : Official Journal of the American Society of Neuroimaging 27 (1): 144-48.


BACKGROUND: Intracerebral hemorrhage (ICH) volumes are frequently used for prognostication and inclusion of patients in clinical trials. We sought to compare the original ABC/2 method and sABC/2, a simplified version with the planimetric method.

METHODS: We retrospectively reviewed admission head CT scans of consecutive ICH patients admitted to a single academic center from July 2012 to April 2013. We assessed ICH volume on the admission. In ABC/2 method, A = greatest hemorrhage diameter by CT, B = diameter perpendicular to A, C = the approximate number of CT slices with hemorrhage multiplied by the slice thickness. C is weighted by area as < 25%, 25-50%, or > 75%. However, in the sABC/2 method, C is the total number of cuts with ICH without any weighting. Bland-Altman plots were generated for both the ABC/2 and sABC/2 methods in comparison to the planimetric method.

RESULTS: One hundred thirty-five patients with spontaneous ICH were included in the final analysis. Bland-Altman analysis illustrated that both ABC/2 and sABC/2 were concordant with the planimetric method. ABC/2 had more bias than sABC/2 (47% vs. 5%, respectively) with no evidence of a linear trend. For differentiating a volume threshold of 30 mL, ABC/2 was less sensitive but more specific than sABC/2 (P < .0001). Concordance between planimetry, ABC/2, and sABC/2 was high, evidenced by most coefficients exceeding .90.

CONCLUSION: Simplified ABC/2 (sABC/2) method performs better than ABC/2 in calculating ICH volumes. Moreover, it is better in differentiating a volume threshold of 30 mL. These findings may have implications for outcomes prediction and clinical trials inclusion.

Last updated on 07/10/2023