Benjamin Wildman-Tobriner1, Michael S Enslow2, Rendon C Nelson2. 1. Department of Radiology, Duke University Medical Center, Durham, NC. Electronic address: benjamin.wildman-tobriner@duke.edu. 2. Department of Radiology, Duke University Medical Center, Durham, NC.
Abstract
OBJECTIVE: When objectively measured on computed tomography (CT), does hepatic heterogeneity or overall liver attenuation predict the presence of shock? METHODS: This retrospective study included 73 patients (mean age 33 years) with the hypoperfusion shock complex (HSC) on CT (cases) and 100 patients (mean age 43 years) with negative trauma CT scans (controls). Liver heterogeneity was calculated by using consistently sized regions of interest (ROIs) to measure the 2 highest and the 2 lowest areas of hepatic density (in Hounsfield units [HU]). The difference between the means of the 2 highest and 2 lowest ROIs was considered the heterogeneity. Attenuation was calculated using the mean of 3 randomly placed ROIs. Both heterogeneity and attenuation were then compared between cases and controls. RESULTS: Median hepatic heterogeneity was 16.8 HU (IQR: 10.7-23.4) for the HSC group and 9.0 HU (IQR: 7.0-10.4) for the controls (P < 0.001). The area under the curve was 0.79, and a threshold of 30 HU yielded a specificity of 100%. Median hepatic attenuation was not significantly different between the HSC and the control groups, with an area under the curve of 0.56. CONCLUSIONS: Increased hepatic heterogeneity may represent an objective marker of the HSC that performs in a similar manner to other established signs. By comparison, overall hepatic hypoattenuation is a poor indicator of the HSC.
OBJECTIVE: When objectively measured on computed tomography (CT), does hepatic heterogeneity or overall liver attenuation predict the presence of shock? METHODS: This retrospective study included 73 patients (mean age 33 years) with the hypoperfusion shock complex (HSC) on CT (cases) and 100 patients (mean age 43 years) with negative trauma CT scans (controls). Liver heterogeneity was calculated by using consistently sized regions of interest (ROIs) to measure the 2 highest and the 2 lowest areas of hepatic density (in Hounsfield units [HU]). The difference between the means of the 2 highest and 2 lowest ROIs was considered the heterogeneity. Attenuation was calculated using the mean of 3 randomly placed ROIs. Both heterogeneity and attenuation were then compared between cases and controls. RESULTS: Median hepatic heterogeneity was 16.8 HU (IQR: 10.7-23.4) for the HSC group and 9.0 HU (IQR: 7.0-10.4) for the controls (P < 0.001). The area under the curve was 0.79, and a threshold of 30 HU yielded a specificity of 100%. Median hepatic attenuation was not significantly different between the HSC and the control groups, with an area under the curve of 0.56. CONCLUSIONS: Increased hepatic heterogeneity may represent an objective marker of the HSC that performs in a similar manner to other established signs. By comparison, overall hepatic hypoattenuation is a poor indicator of the HSC.