Noninvasive evaluation of hepatic fibrosis using acoustic radiation force-based shear stiffness in patients with nonalcoholic fatty liver disease

Abstract

Background & aims: Nonalcoholic fatty liver disease (NAFLD), the most common form of chronic liver disease in developed countries, may progress to nonalcoholic steatohepatitis (NASH) in a minority of people. Those with NASH are at increased risk for cirrhosis and hepatocellular carcinoma. The potential risk and economic burden of utilizing liver biopsy to stage NAFLD in an overwhelmingly large at-risk population are enormous; thus, the discovery of sensitive, inexpensive, and reliable noninvasive diagnostic modalities is essential for population-based screening.

Methods: Acoustic Radiation Force Impulse (ARFI) shear wave imaging, a noninvasive method of assessing tissue stiffness, was used to evaluate liver fibrosis in 172 patients diagnosed with NAFLD. Liver shear stiffness measures in three different imaging locations were reconstructed and compared to the histologic features of NAFLD and AST-to-platelet ratio indices (APRI).

Results: Reconstructed shear stiffnesses were not associated with ballooned hepatocytes (p=0.11), inflammation (p=0.69), nor imaging location (p=0.11). Using a predictive shear stiffness threshold of 4.24kPa, shear stiffness distinguished low (fibrosis stage 0-2) from high (fibrosis stage 3-4) fibrosis stages with a sensitivity of 90% and a specificity of 90% (AUC of 0.90). Shear stiffness had a mild correlation with APRI (R(2)=0.22). BMI>40kg/m(2) was not a limiting factor for ARFI imaging, and no correlation was noted between BMI and shear stiffness (R(2)=0.05).

Conclusions: ARFI imaging is a promising imaging modality for assessing the presence or absence of advanced fibrosis in patients with obesity-related liver disease.

Figure 1

B-mode image from one of the NAFLD study subjects. The yellow box represents the ROI for reconstructing the shear stiffness, which is chosen to be free of vessels and other liver substructures. The radiation force excitation was applied at a lateral position of 0, and propagation was monitored to the right (positive lateral locations).

Figure 2

Reconstructed shear moduli in 135 patients being evaluated for NAFLD as a function of their biopsy-proven fibrosis stage. The boxes represent the interquartile range (IQR), while the whiskers represent 1.5 times the respective IQR level over the mean stiffnesses for each study subject (the mean of 3 replicate measures in the 3 imaging locations with an IQR/mean < 0.3). The number of study subjects for each fibrosis stage is indicated in parentheses. The horizontal dashed line represents the threshold (4.24 kPa) to distinguish fibrosis stages F0–2 from F3–4.

Figure 3

(a) APRI as a function of fibrosis score; (b) APRI as a function of reconstructed log shear stiffness (R2 = 0.22, p < 0.001).

Figure 4

(a) BMI as a function of fibrosis score; (b) mean reconstructed log shear stiffness for each study subject as a function of study subject BMI. No correlation between mean liver stiffness and BMI was noted (R2 = 0.05, p < 0.01).

Figure 5

Distribution of study subject BMIs; successful shear stiffness reconstructions are indicated in gray, while unsuccessful shear stiffness reconstructions are indicated in black. The percentages above each bar represent the percentage of successful reconstructions.

Figure 6

Box plots of mean reconstructed shear stiffness for different biopsy-proven hepatocyte ballooning scores; no significant trends in shear stiffness exist as a function of hepatocyte ballooning (p = 0.11).

Figure 7

Box plots of mean reconstructed shear stiffness for different biopsy-proven inflammation scores; no significant trends in shear stiffness exist as a function of inflammation (p = 0.69).