Show Summary Details
Page of

Case 2.2 

Case 2.2
Case 2.2

Christine U. Lee

, and James F. Glockner

Page of

PRINTED FROM OXFORD MEDICINE ONLINE ( © Oxford University Press, 2016. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: null; date: 16 October 2019


52-year-old man with elevated AST and ALT levels

Imaging Findings:

Axial IP and OP T1-weighted 2D SPGR images (Figure 2.2.1) demonstrate diffuse homogeneous signal loss on OP images, consistent with the presence of intravoxel fat.


Diffuse fatty infiltration of the liver


This case provides a fairly typical example of diffuse fatty infiltration of the liver. Hepatic steatosis, fatty liver, and fatty infiltration of the liver are all terms used interchangeably by most radiologists. Some authors have objected to the term fatty infiltration since they believe that the term implies infiltration of fat between hepatocytes rather than accumulation within hepatocytes, but this expression is so widely used that it seems unlikely to disappear.

The most common causes for fatty liver are alcoholic and nonalcoholic liver disease, both of which result in accumulation of triglyceride within hepatocytes. The prevalence of fatty liver in the general population is estimated to be 15%, but it is much more common in patients with obesity (75%), hyperlipidemia (50%), and excess alcohol consumption (45%). In patients with both obesity and excessive alcohol consumption, the prevalence is 95%. Fatty liver associated with nonalcoholic liver disease ranges in increasing severity from steatosis, as described in this discussion, to steatohepatitis, in which the fat causes inflammation and cell injury, and eventually to hepatic fibrosis and cirrhosis, when the inflammation leads to damage and scarring of the liver. With a history of alcohol consumption, there is a similar progression from steatosis, which is reversible, to alcoholic hepatitis, and finally to cirrhosis. Alcohol alone is not sufficient to initiate alcoholic liver disease, and several additional factors, including genetic predisposition as well as multiple environmental stimuli, have a large role. It is probably true, however, that the incidence of alcoholic cirrhosis increases with increasing alcohol intake. Alcoholic hepatitis has considerable variability in its presentation, but the 30-day mortality rate remains 0% to 50%.

The Model for End-Stage Liver Disease (or MELD) score, which originally was the Mayo End-Stage Liver Disease score, is used to estimate the severity of chronic liver disease. A higher score indicates an increasing risk of death and also may cause the patient to move higher up on the transplantation list. The gold standard for determining the severity of hepatic fibrosis in the clinical setting of fatty infiltration is histologic evaluation through liver biopsy, although there is increasing acceptance of the use of MR elastography.

Visualization of hepatic steatosis with MRI relies on inherent differences in the microscopic environment of water protons and fat protons, which in turn leads to small differences in their resonant frequencies. This difference is termed chemical shift and reminds us of the spectroscopic origins of MRI, where water and fat peaks are separated by 3.5 ppm on the frequency spectrum. The chemical shift is directly proportional to field strength, so a chemical shift of 3.5 ppm translates to differences of about 220 Hz at 1.5 T and about 440 Hz at 3 T for water and fat protons.

At 1.5 T, fat protons precess about 220 Hz more slowly than water protons (fat is to the right of water on the spectrum), and if we use a double-echo GRE technique, we can set the TEs to coincide with times when fat and water are out of phase (eg, 2.2 msec, 6.7 msec, 11.2 msec) and in phase (eg, 4.5 msec, 9.0 msec, 13.4 msec). A standard IP and OP pulse sequence should be set up by the vendor so that the OP TE is shorter than the IP TE; this ensures that signal dropout from an IP image to an OP image is truly the result of fatty infiltration and not dependent on T2* relaxation. This almost always happens at 1.5 T; however, at 3 T, the first OP TE is fairly short and is technically challenging to obtain, and so some vendors (including ours) switched the order of the IP and OP TEs (apparently hoping that no one would notice). To be fair, it is true that T2* relaxation effects are stronger at 3 T, and therefore using the first IP TE and the second OP TE also has some inherent disadvantages. The solution most vendors have arrived at is to use a Dixon-based technique, where the IP and OP images can be reconstructed from data obtained at different TEs.

There is fairly extensive literature discussing MR methods for quantifying hepatic steatosis. Spectroscopy is probably considered the gold standard, although it is one of the least frequently used, since it’s relatively unfamiliar to most abdominal imagers and generally acquires only single voxels (and therefore is prone to sampling error in patients with heterogeneous fatty infiltration). A simple method, and one with excellent reported results, involves an estimation of the fat fraction (FF) from IP and OP images. The FF is defined as follows:

where SIP is the signal from the IP acquisition and SOP is the signal from the OP acquisition. The caveat is that this estimate does not take into consideration T1, T2, or T2* relaxation effects, which are problematic mainly in patients with iron overload (ie, those with hemochromatosis, hemosiderosis, or cirrhosis with siderotic nodules). A multiecho Dixon technique (IDEAL) is less widely available but allows simultaneous estimation of hepatic fat and iron content while accounting for the effects of each substance on measurement of the other.

Diffuse hepatic steatosis is not a diagnostic dilemma, nor is focal fatty infiltration when it occurs in one or more well-known locations, including the porta hepatis and gallbladder fossa and adjacent to the falciform ligament or ligamentum venosum. More challenging are cases with focal nodular or masslike fatty infiltration, and examples of these are illustrated in Cases 2.3, 2.4, and 2.5.