Date of Award

6-3-2009

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)

First Advisor

Jeffrey Weinreb

Abstract

The purpose of this study was to chart unenhanced CT liver and spleen attenuation data on patients with normal and fatty livers to demonstrate a relationship between liver fat content and attenuation on unenhanced CT using MR chemical shift imaging as a reference standard. Retrospective data was gathered on 116 patients from NYU Tisch Hospital who had undergone both an unenhanced CT and MR opposed-phase imaging of the liver within 2 months. MR fat fraction (MR FF) was calculated using signal decrease on in]phase and opposed]phase imaging, with a cutoff of 0.2 considered significant fatty infiltration. Three CT liver attenuation indices were obtained: liver attenuation, liver minus spleen attenuation, and liver to spleen ratio. For each index, linear regression was performed and 95% confidence intervals calculated. Receiver operating characteristic (ROC) analyses were performed to determine the accuracy and optimal cutoff values for the best balance between sensitivity and specificity, and the highest cutoff value that had 100% specificity was determined. ROC analysis showed good accuracy of all CT indices, with sensitivities and specificities between 91 and 93% for cutoff values of 50, 0, and 1.0 for CT liver attenuation, liver minus spleen, and liver to spleen ratio, respectively. Correlation of hepatic steatosis between unenhanced CT and MR opposed-phase imaging was excellent, but as MR imaging is not the gold standard for determining liver fat content, conclusions drawn from this study depend on the accuracy of MR imaging as determined in other studies.

Share

COinS