LAUSR

To the Editor: We read with interest the recent article by Zhao et al that investigated the accuracy, reliability, and efficiency of several formula-based methods for estimating intracerebral hematoma volume compared with a gold standard computer-based volumetric analysis—3D Slicer. Three new formula methods were investigated in this study that were derived from existing formulas, among them was 2.5/6ABC. Formulas with ABC factors attempt to estimate hematoma volume based on the length of perpendicular lines in 3 orthogonal planes that are applied to a standardized formula. However, geometrically intracerebral hemorrhages do not abide by the principles of standard geometric structures. In engineering, advanced computational methods are used to calculate the volumes and surface areas of complex structures, as is the case with the 3D Slicer when calculating the hematoma volume. Formula-based methods attempt to define the shape of a hematoma with standardized geometric formulas. However, a one-size-fits-all approach may significantly overor underestimate intracerebral hematoma volume. Zhao et al found that 2.5/6ABC was more accurate than 1/2ABC, 1/3ABC, and π/6ABC. The theoretical basis for the 2.5/6ABC method is to average the commonly overestimating 1/2ABC with the commonly underestimating 1/3ABC methods. The authors conclude that 2.5/6ABC is preferred when maximal slice area S cannot be obtained. Given that a hematoma can have varying shapes, one alternative approach would be to report hematoma volume with a range of upper and lower boundaries. Using applied mathematics in other industries to solve similar volume estimation problems, it has been shown that averaging estimated volume boundaries provides an accurate estimate of the true volume of an irregular shaped object without computer-based 3-dimensional reconstruction of an object. While the range for small hematomas would be minuscule, the range of large hematomas could have a profound clinical impact. Mathematically, 1/2ABC formula likely represents an upper boundary of hematoma volume, while 1/3ABC formula likely represents the lower boundary of hematoma volume, with an average of 2.5/6ABC. If we were to consider π/6ABC and 1/3ABC, upper and lower boundaries, respectively, our potential range of volumes would be larger, and the average would yield (π+2/12) ABC. The same principle can be applied to formula-based methods that include maximal slice area S with 2/3SH and 1/2SH representing the upper and lower boundaries, respectively. The average of these 2 formulas would yield 3.5/6SH. As we have seen from a prior study, 1/3ABC is better suited for irregular hematomas and 1/2ABC works better for small regularly shaped hematomas. It is reasonable to consider reporting hematoma volume as a likely range with the average of the boundaries. Based on inspection of the hematoma shape on computed tomography, one can make a necessary assumption as to which boundary the true volume is likely closer to. This has an important clinical implication as a clinician may prefer to rely on the upper versus the lower boundaries depending on the unique circumstances and weighing the relevant risks and benefits when making a critical clinical decision with stroke patient’s family. In summary, Zhao et al demonstrated that 2.5/6ABC—the average of 1/2ABC and 1/3ABC boundaries—provided improved accuracy of hematoma volume estimation to 1/2ABC and 1/3ABC on their own. Questions still arise as to whether there are other modified formulas to better define hematoma volume boundaries and whether there is a clinical benefit in reporting hematoma volume with a range.