BACKGROUND AND PURPOSE: The 1/2ABC formula is the method most commonly used in clinical practice to rapidly estimate intracerebral hemorrhage (ICH) volume. We aimed to compare this method with the alternative '2/3Sh' formula for both regularly and irregularly-shaped hematomas. METHODS: Computed tomography (CT) images from 344 ICH patients (median volume: 16·66 ml) were retrospectively reviewed. According to the maximum slice, the shape was classified into regular or irregular (multilobular, conical, and other). Volumes as determined by the 1/2ABC and 2/3Sh formulas were compared against the gold standard, computer-assisted planimetry, for various hematoma shapes. RESULTS: With the 1/2ABC method, errors were seen non-significantly more frequently for irregularly-shaped hematomas [OR: 2.85 (95% CI: 0.65-12.50)]. The 1/2ABC method misclassified a greater proportion of hematomas as greater or less than 30 ml in volume: 7.0% (95% CI: 6.0-9.9%). Both the 1/2ABC and 2/3Sh formulas correlated well with gold standard (correlation coefficients >0.9 for each shape). While there was no statistically significant measurement error bias for either method, the 95% confidence intervals of the limit of agreement for 2/3Sh were tighter: -0.22 ml (-4.7-4.25 ml) versus 2·50 ml (-10.35-15.34 ml). Measurement errors were significantly greater with the 1/2ABC method, for both regular and irregular hematomas [1.17 ml (0.48-2.83 ml) versus 0.88 ml (0.42-1.68 ml) and 2.65 ml (1.07-5.88 ml) versus 0·99 ml (0.47-2.28 ml); P<0.05, respectively], although the magnitude of error would only rarely be clinically relevant for regular hematomas. Errors were most evident in assessing multilobular-shaped hematomas [6.49 ml (3.35-13.98 ml) versus 1.86 ml (0.96-9.94 ml); P<0·001]. CONCLUSIONS: The 2/3Sh formula leads to fewer clinically-relevant hematoma volume misclassifications than the 1/2ABC formula, and is particularly superior in estimating volumes of irregularly-shaped hematomas.
BACKGROUND AND PURPOSE: The 1/2ABC formula is the method most commonly used in clinical practice to rapidly estimate intracerebral hemorrhage (ICH) volume. We aimed to compare this method with the alternative '2/3Sh' formula for both regularly and irregularly-shaped hematomas. METHODS: Computed tomography (CT) images from 344 ICHpatients (median volume: 16·66 ml) were retrospectively reviewed. According to the maximum slice, the shape was classified into regular or irregular (multilobular, conical, and other). Volumes as determined by the 1/2ABC and 2/3Sh formulas were compared against the gold standard, computer-assisted planimetry, for various hematoma shapes. RESULTS: With the 1/2ABC method, errors were seen non-significantly more frequently for irregularly-shaped hematomas [OR: 2.85 (95% CI: 0.65-12.50)]. The 1/2ABC method misclassified a greater proportion of hematomas as greater or less than 30 ml in volume: 7.0% (95% CI: 6.0-9.9%). Both the 1/2ABC and 2/3Sh formulas correlated well with gold standard (correlation coefficients >0.9 for each shape). While there was no statistically significant measurement error bias for either method, the 95% confidence intervals of the limit of agreement for 2/3Sh were tighter: -0.22 ml (-4.7-4.25 ml) versus 2·50 ml (-10.35-15.34 ml). Measurement errors were significantly greater with the 1/2ABC method, for both regular and irregular hematomas [1.17 ml (0.48-2.83 ml) versus 0.88 ml (0.42-1.68 ml) and 2.65 ml (1.07-5.88 ml) versus 0·99 ml (0.47-2.28 ml); P<0.05, respectively], although the magnitude of error would only rarely be clinically relevant for regular hematomas. Errors were most evident in assessing multilobular-shaped hematomas [6.49 ml (3.35-13.98 ml) versus 1.86 ml (0.96-9.94 ml); P<0·001]. CONCLUSIONS: The 2/3Sh formula leads to fewer clinically-relevant hematoma volume misclassifications than the 1/2ABC formula, and is particularly superior in estimating volumes of irregularly-shaped hematomas.
Authors: Jay F Yu; Andrew D Nicholson; Jeffrey Nelson; Matthew D Alexander; Stephanie H Tse; Steven W Hetts; J Claude Hemphill; Helen Kim; Daniel L Cooke Journal: Interv Neuroradiol Date: 2018-01-17 Impact factor: 1.610