An analysis of the accuracy of the Kothari method with respect to shape and size of the intra-cerebral hemorrhagic collections in stroke patients of Makati Medical Center


Jennifer So Ng


College of Science

Document Type

Archival Material/Manuscript

Publication Date



Background and Purpose- CT scan remains the standard diagnostic technique for intracranial hemorrhage and approximating the volume of the hemorrhagic collection is the strongest predictor of 30-day mortality rate for all locations of bleed. As such, emergent and accurate volume determination is mandatory to provide immediate, necessary and appropriate therapeutic or surgical interventions. Current practice is to use the fast Kothari method (BC/2) of volume approximation in such cases. The aim of the study is to determine the efficiency of the Kothari method in measuring hemorrhagic collections with regards to its shape and size.

Methods- The study reviewed the computed tomography scans of the head of 55 patients with intraparenchymal ICH. Shapes of the ICH were then divided into (A) round-to-ellipsoid, (B) irregular with frayed margins and (C) multinodular to separated subsets. Sizes of the ICH were also divided into (volume A) 0.1-30 ml, (volume B) more than 30 ml to less than 60 ml and (volume C) more than 60 ml subsets. The volumes approximated by the Kothari method (ABC/2) were then compared with volumes analyzed by the computer-assisted free-hand volume analyzer with regards to hematoma size and shape.

Results- The most common shape analyzed in the study was shape C (multinodular to separated) comprising 41.82% of the population, followed by shape A (round-to-ellipsoid) at 34.55% and shape B ICH (irregular with frayed margins) at 23.64%. While the most common size distribution was in the range of volume A (0.1-30 ml) comprising 56.36% of the analyzed hemorrhagic collections, followed by the volume B (>30 - <= 60 ml) range at 25.45% and volume C (>60 ml range) at 18.18%.

The mean percent difference in the volumes A and B categories were 0.04% and 0.45% respectively, and seen in volume A at 22.98%, followed by volumes B (14.55%) and C (13.07%). The great percent deviation is due to the single outlier in this subset which is of Shape C category, with 87.5% deviation from the baseline. Due to the single outlier, the resulting volumes are more spread out in the volume A category when compared to volume C and is least varied in volume B.

With respect to hemorrhagic shapes, Kothari Method provided volume underestimation for Shape A (68.42%) and Shape B (61.54%) and volume overestimation on Shape C (60.87%). As for the size, it provided volume overestimation for volume A (61.29%) and volume C (60%). In the volume B range, there was equal percentage of over and underestimating the volume of the hemorrhagic collection.

The overall mean volume using Kothari method for shape A is 11.17 ml (versus 12.23 ml using volumetric method) with mean underestimation of 4.81 %; for shape B, it is 46.27 ml (versus 44.02 ml) with mean underestimation of 4.13%; for shape C, it is 51.40 ml (versus 50.47 ml) with 8.35% overestimation of the values. The percent deviation is more varied in the shape C category (22.48%) when compared to shapes A and B (11.71 % and 19.24%, respectively).

The mean volume calculated using the Kothari method in the shape A category was underestimated in 66.67% of the subset by 4.41 % and the standard deviation is 11.91 %. Only 1 of the hemorrhagic collection was in volume B category and was underestimated by 12.04%. In shape B category (irregular with frayed margins), there was underestimation of the volume by 18.49% and 19.36% in volume categories A (100%) and B (60%) respectively and overestimation in 100% of the subset by 16.14% in volume C range. The standard deviations are as follows: 11.36%, 19.36% and 8.25%, respectively.

In shape C category (multinodular to separated), there was overestimation of the hemorrhagic volume subset by 21.64% and 3.50% in volume categories A (75%) and B (62.5%) respectively and underestimation in 42.86% of the subset by 1.28% in volume C range. The standard deviations are as follows: 31.81% (volume A), 11.95% (volume B) and 11.29% (volume C).

Majority of the hemorrhagic collections analyzed are in the range of volume A (56.36%) category and most of which are of shape A (58.06%), followed by the volumes B (25.45%) and C (18.18%). In volume B range, most assumed the shape C (57.14%), followed by shape B (35.71%). Similar results were gathered for the analyzed volumes ,within the range of volume C, majority falls with shape C (70%), followed by shape B (30%) and none assumed shape A.

Test of lndependence: Fisher's Exact test. Since the p-value<0.0001 <0.05, Level of significance=0.05 Size and shape are related (p-value<0.0001).

One-Sample T-tests: As discussed, the Kothari Method is said to accurately determine the volume of hemorrhagic collections if the percentage deviation from value given by volumetric method = 0. Level of significance=0.05. Since the p-value=0.7446 (not less than 0.05). None of the percentage deviations for categories of size and of shape significantly differ from 0.

Conclusions- The categorization of hemorrhagic collections as round-ellipsoid, irregular with frayed margins, or multinodular to separated depends on the size of the collection. There is insufficient evidence to say that the average percentage deviation of volumes given by Kothari Method with respect to various shapes and sizes differs from 0 (p-value=0.7446).



Nervous System Diseases



Upload File


This document is currently not available here.