Radiology_Oct04 Focal fatty infiltration and focal fatty sparing of the liver may be con- fused with multiple liver metastases on both ultrasound and computed tomography (CT) imaging. Radiolo- gists need to be aware of this benign condition so that unnecessary investi- gations and liver biopsies can be avoided. Introduction Focal fatty infiltration of the liver is caused by focal deposition of intracel- lular fat within hepatocytes. Although more commonly detected in adults it has also been described in children.1 The most common cause is alcohol abuse, although diabetes, obesity, cer- tain drugs, toxins, protein energy mal- nutrition and anoxia are recorded as causing fatty change. At a cellular level there is accumulation of fat in cyto- plasmic vacuoles. The significance of fatty change is dependent on the severity and cause. When mild it has no effect on liver function and is usu- ally reversible. Severe fatty change, however, impairs liver function and can cause cell death.2 Focal fatty infiltration can cause difficulty in the interpretation of CT and ultrasound studies of the liver and can easily be confused with liver metastases.3 This is especially true of regions where there is no fatty infiltra- tion in an otherwise diffusely infiltrat- ed liver, called focal fatty sparing.4 This misdiagnosis may lead to unnecessary invasive procedures such as biopsies, and further anxiety on the part of the patient. This diagnostic problem can be avoided by performing an initial unenhanced computed tomography (CT) scan of the liver in all suspected cases and then measuring the density of the affected region in Hounsfield units. Two representative cases in which the correct diagnosis of fatty infiltration was not initially consid- ered are reported. Case reports Case 1 A 60-year-old woman presented with a 3-month history of right upper quadrant pain. Liver function test results were within normal limits. Ultrasound examination demonstrat- ed an unhomogeneous echogenic liver in which multiple liver metas- tases were suspected. An unenhanced CT study demonstrated multiple geo- graphical areas of low density in the both the left and right lobes. Measure- ment of the liver density confirmed that the lower density areas had a mean density reading of 20 Hounsfield units (HU) while the nor- mal parenchyma had a mean density measurement of 40 HU (Fig. 1a). The contrast-enhanced study demonstrat- ed decreased perfusion in those regions of low density (Fig. 1b). A diagnosis of focal fatty infiltration of the liver was made. The patient was subsequently found to have type 2 diabetes, and her pain settled on oral hypoglycaemic agents. Case 2 A 32-year-old woman known to have type 1 diabetes presented with right upper quadrant pain. Gallstones were suspected on the basis of the clinical symptoms, and an ultrasound examination confirmed their pres- ence. In addition multiple echogenic lesions resembling liver metastases were noted in both lobes. An ultra- CASE REPORT 25 SA JOURNAL OF RADIOLOGY • October 2004 Focal fatty infiltra- tion and focal fatty sparing of the liver P Corr FFRad (D), FRCR, MMed, MD Department of Radiology Nelson Mandela School of Health Sciences Durban Fig. 1a. Unenhanced CT of the liver in case 1 demonstrates multiple focal low density regions in both lobes of the liver. Region of interest 1 over the fatty left lobe measured 10 HU while region of interest 2 over the right lobe measure 40 HU in keeping with focal fatty infiltration. Fig.1b. Contrast-enhanced CT at the same level demonstrates the hepatic and portal vein branch- es passing normally through the regions of fatty infiltration. 26 SA JOURNAL OF RADIOLOGY • October 2004 sound-guided percutaneous biopsy of a left lobar lesion demonstrated focal fatty infiltration of otherwise normal hepatocytes. CT of the liver demon- strated multiple areas of low density in both lobes in keeping with diffuse fatty infiltration, with ‘pseudo-lesions’ of foci of normal enhancing liver parenchyma (Figs 2a and 2b). A diag- nosis of focal fatty sparing of the liver was made. Discussion In both these cases an initial diag- nosis of liver metastases was consid- ered on the basis of on the ultrasound findings. This is understandable, as the echogenicity of these pseudole- sions differs from that of the sur- rounding parenchyma.5 In both cases measurement of the parenchymal density of the liver on CT suggested the correct diagnosis. The normal density of the liver on unenhanced CT is between 40 and 50 HU. After con- trast injection the density reaches 80 - 120 HU. A good comparison of nor- mal hepatic enhancement is compari- son with the enhanced spleen, which has a similar enhanced density. Fatty change or infiltration will measure anywhere from 10 to 20 HU. An important diagnostic pointer to the correct diagnosis is the fact that nor- mal hepatic venules and portal venules run through these pseudole- sions and are not displaced as would be expected by a focal mass. The true prevalence of focal fatty change or infiltration is unknown in adults, although a recent study in San Francisco demonstrated a prevalence of 25.6% in adolescents aged between 15 and 19 years of age.1 Although the causation of fatty change or infiltra- tion is well known, the distribution within the liver of focal fatty infiltra- tion and focal fatty sparing is not clearly understood. It is postulated that the distribution of fat deposition is related to regional differences in liver perfusion. Studies using CT por- tography have demonstrated that regions of fat deposition are associat- ed with decreased perfusion.6 Yoshimuitsu et al.7 have demonstrated that these hepatic pseudolesions are due to a ‘third inflow’ of blood to the liver via the cholecystic, parabiliary or epigastric-paraumbilical veins.7 The cholecystic vein drains segments IV and V. The parabiliary veins originate in the hepatoduodenal ligament and pancreas to supply the porta hepatis and segment IV. The epigastric-para- umbilical veins drain the abdominal wall into the liver in the region of the falciform ligament.7 The region around the falciform ligament, the porta hepatis and gallbladder fossa bed are usually affected, while the medial segment of the left lobe, espe- cially segment 4, is often spared.5,6 There appears to be an association between the presence of focal fatty sparing in the medial segment of the left lobe and anomalous gastric venous drainage.8 Magnetic resonance imaging (MRI) has been reported to be diag- nostic in confirming the presence of fatty infiltration using opposed phase gradient echo sequences or fat satura- tion sequences.9,10 These two cases demonstrate the importance of complementary imag- ing investigations in assessing focal liver pathology. Radiologists should always remember to include an unen- hanced CT study in their liver imaging protocols and to measure the CT den- sity of focal liver lesions if they are unsure of the diagnosis. References 1. Kammen BF, Pacharn P, Thoeni RF, et al. Focal fatty infiltration of the liver: analysis of preva- lence and CT findings in children and young adults. AJR Am J Roentgenol 2001; 177: 1035- 1039. 2. Cotran RC, Kumar V, Collins T. Cellular patho- logy II. In: Cotran RC, Kumar V, Collins T, eds. Pathologic Basis of Disease. 6th ed. Philadelphia: WB Saunders, 1999: 31-49. 3. Kemper J. CT and MRI findings of multifocal liver steatosis mimicking malignancy. Abdom Imaging 2002; 27: 708-710. 4. McKenzie A. Computed tomographic and ultrasound appearances of focal fatty infiltra- tion. Aust Radiol 1991; 35: 166-168. 5. Chong VF, Tan YF. Ultrasonic hepatic pseudole- sions: normal parenchyma mimicking mass lesions in fatty liver. Clin Radiol 1994; 49: 326- 329. 6. Paulson EK. Focal fatty infiltration: a cause for nontumorous defects in left lobe during during CT portography. J Comput Assist Tomogr 1993; 17: 590-595. 7. Yoshimuitsu K, Honda H, Kuroiwa T, et al. Unusual haemodynamics and pseudolesions of the noncirrhotic liver at CT. Radiographics 2002; 21: S81-S96. 8. Kawamori Y. Focal hepatic fatty infiltration in the posterior edge of the medial segment of the liver associated with anomalous gastric venous drainage. J Comput Assist Tomogr 1996; 20: 356- 359. CASE REPORT Fig. 2a and 2b. Contrast-enhanced CT in case 2 demonstrates multiple pseudolesions in both lobes from focal fatty sparing. Note that the nor- mally enhancing parenchyma measures 80 HU (region of interest 2) while the region of fatty infil- tration (region of interest 1) measures 15 HU.