Radiologic evaluation of nonalcoholic fatty liver disease

Abstract

Nonalcoholic fatty liver disease (NAFLD) is a frequent cause of chronic liver diseases, ranging from simple steatosis to nonalcoholic steatohepatitis (NASH)-related liver cirrhosis. Although liver biopsy is still the gold standard for the diagnosis of NAFLD, especially for the diagnosis of NASH, imaging methods have been increasingly accepted as noninvasive alternatives to liver biopsy. Ultrasonography is a well-established and cost-effective imaging technique for the diagnosis of hepatic steatosis, especially for screening a large population at risk of NAFLD. Ultrasonography has a reasonable accuracy in detecting moderate-to-severe hepatic steatosis although it is less accurate for detecting mild hepatic steatosis, operator-dependent, and rather qualitative. Computed tomography is not appropriate for general population assessment of hepatic steatosis given its inaccuracy in detecting mild hepatic steatosis and potential radiation hazard. However, computed tomography may be effective in specific clinical situations, such as evaluation of donor candidates for hepatic transplantation. Magnetic resonance spectroscopy and magnetic resonance imaging are now regarded as the most accurate practical methods of measuring liver fat in clinical practice, especially for longitudinal follow-up of patients with NAFLD. Ultrasound elastography and magnetic resonance elastography are increasingly used to evaluate the degree of liver fibrosis in patients with NAFLD and to differentiate NASH from simple steatosis. This article will review current imaging methods used to evaluate hepatic steatosis, including the diagnostic accuracy, limitations, and practical applicability of each method. It will also briefly describe the potential role of elastography techniques in the evaluation of patients with NAFLD.

Keywords: Computed tomography; Elastography; Liver steatosis; Magnetic resonance imaging; Magnetic resonance spectroscopy; Nonalcoholic fatty liver disease; Nonalcoholic steatohepatitis; Ultrasonography.

Figure 1

Ultrasonography evaluation of hepatic steatosis. A: Ultrasonography (US) image of a normal liver, showing that the echogenicity of liver parenchyma (L) and kidney cortex (K) is similar; B: US image of a steatotic liver, showing increased echogenicity of the liver parenchyma (L) which is clearly brighter than the kidney cortex (K).

Figure 2

Computed tomography evaluation of hepatic steatosis using computed tomography_L-S index. A: Computed tomography (CT) image of a normal liver, showing that its attenuation (65 HU) measured using regions-of-interest (white circles) was higher than that of the spleen (50 HU), and the CT_L-S value was 15 HU, which lies within the normal reference range; B: CT image of a steatotic liver, showing hepatic attenuation (10.5 HU) much lower than that of the spleen (51 HU), making the CT_L-S value -40.5 HU, far below the normal reference range and indicating moderate-to-severe hepatic steatosis.

Figure 3

Magnetic resonance spectroscopy spectrum of hepatic fat. Water and fat peaks are displayed at different frequencies; water appears as a single peak at 4.7 ppm, whereas fat appears as four peaks, including the dominant methylene (CH₂) peak at 1.3 ppm (3), a methyl (CH₃) peat at 0.9 ppm (4), an α-olefinic and α-carboxyl peak at 2.1 ppm (2), and a diacyl peak at 2.75 ppm (1); the areas of these four fat peaks and the water peak can be measured by spectral tracing. Proton density fat fraction can be calculated as (sum of fat peaks) ÷ (sum of fat peaks + water peak)[45,82].

Figure 4

Dual-echo opposed-phase and in-phase chemical shift images of steatotic liver. A: At opposed-phase (OP) (echo time = 2.3 ms at 1.5T), the protons in water and those in methylene (the largest fat moiety) are placed in opposite directions, so that the signals of these two components cancel each other. Therefore, the liver appears dark (i.e., decreased signal); B: At in-phase (IP), the protons in water and those in methylene are positioned in the same direction so that their signals are added. Liver fat fraction can be calculated based on signal intensities on OP and IP images as (signal at IP - signal at OP) ÷ 2 × signal on IP; the signal fat fraction calculated with dual-echo chemical shift images was not corrected for the T2* effect, and therefore may not accurately determine proton density fat fraction.

Figure 5

Supersonic shearwave elastography of simple steatosis vs nonalcoholic steatohepatitis. A: Supersonic shearwave elastography image of the liver with simple steatosis shows a mean liver stiffness value of 2.9 kPa, which lies within the normal reference range; B: Supersonic shearwave elastography image of the liver with nonalcoholic steatohepatitis shows an elevated mean liver stiffness value of 11.6 kPa.