Nonalcoholic fatty liver disease and elastography: Incremental advances but work still to be done

Potential conflict of interest: Dr. Harrison advises Galectin, Genfit, Fibrogen, Intercept, NGM, Pfizer, Zafgen, and CLDF. He is on the speakers' bureau for Gilead and Merck. See article on Page 1817 The views expressed are those of the authors and should not be construed as official views of the United States Air Force, United States Army, or Department of Defense. Nonalcoholic fatty liver disease (NAFLD) is the leading cause of chronic liver disease in developed countries, with an estimated prevalence of 20%‐30% and increasing to 70% in obese individuals and as high as 90% in diabetics.1 Given the pathophysiological link with metabolic syndrome, rates of NAFLD are likely to continue to rise with the obesity pandemic. Within a community cohort, prevalence of patients with NAFLD who have nonalcoholic steatohepatitis (NASH) has been estimated at 3%‐4% and rising as high as 12.2% in middle‐aged patients.3 For diabetic patients, prevalence of NASH may be as high as 56%.4 Whereas advanced fibrosis has the highest risk of mortality, patients with NASH are 6 times more likely to develop liver‐related mortality over 20 years than those with non‐NASH NAFLD.5 It is precisely these individuals with NASH and fibrosis who have the greatest risk of progression to advanced disease, cirrhosis, hepatocellular carcinoma, and death. Therefore, it is critical to distinguish between patients with fibrosis, NASH, and non‐NASH NAFLD. The current gold standard for diagnosis and histological assessment of NAFLD is a liver biopsy. Unfortunately, liver biopsy is invasive, subject to complications, costly, and limited by sampling errors and intraobserver variability. Furthermore, given the high prevalence of NAFLD, it is neither practical nor cost effective to perform liver biopsies in all patients who are at risk for NASH and fibrosis. Recently, several noninvasive imaging modalities based on elastography have been developed. To date, these imaging modalities have proven promising in the assessment of liver fibrosis in chronic liver disease, though rigorous comparative studies in patients with NAFLD are lacking. Furthermore, distinction of NASH from non‐NASH NAFLD using noninvasive imaging remains elusive because the diagnosis of NASH historically relies on the observation of ballooning degeneration of hepatocytes as well as steatosis and inflammation—not simply steatosis and fibrosis. The most validated, commonly used elastography method is vibration controlled transient elastography (VCTE), which is also known as transient elastography (TE) and marketed as the FibroScan (Echosens, Paris, France). VCTE is ultrasound based, but is not used with image guidance. The velocity of a low‐amplitude shear wave is measured in a region of interest and is converted into a measure of liver stiffness. Both point quantification shear wave elastography (pSWE) and two‐dimensional shear wave elastography (SWE) use acoustic radiation force impulse (ARFI) technology to generate shear waves in the liver, and the shear wave speed is calculated in meters per second.6 pSWE is most commonly referred to as ARFI in the hepatology literature, whereas two‐dimensional SWE is referred to as supersonic shear imaging (SSI) based on one of the four systems available clinically. Kwok et al. completed a systematic review of VCTE in patients with NAFLD involving nine studies and 1,047 patients.7 In this review, VCTE was excellent in diagnosing fibrosis stages 3 (75% sensitivity, 82% specificity) and 4 (92% sensitivity, 92% specificity), but only moderate accuracy for fibrosis stage 2 (79% sensitivity, 75% specificity). To evaluate the diagnostic efficacy of ARFI in patients with NAFLD, Liu et al. completed a systematic review of seven studies for a total of 723 patients who underwent ARFI.8 From their meta‐analysis, the summary sensitivity was 80.2% in detecting significant fibrosis, which was defined as fibrosis stage ≥2. To date, no study has evaluated the efficacy of SSI in the subset of chronic liver disease patients with NAFLD. Therefore, it was with great interest that we read the article by Cassinotto et al. in this issue of Hepatology, which is the first trial to compare the effectiveness of SSI, FibroScan, and ARFI imaging in patients exclusively with biopsy‐proven NAFLD.9 The current trial adds substantially to our understanding of elastography‐based imaging modalities in the evaluation of fibrosis in NAFLD patients, while also establishing potential cut‐off values for these imaging studies in this patient demographic (Table 1). The results are similar to previous studies that have looked at specific individual imaging methods in NAFLD patients,7 though this is the first study to evaluate the effectiveness of SSI in NAFLD. Overall, SSI had statistically better diagnostic performance for fibrosis stage 2 or greater than ARFI with an area under the receiver operating curve (AUROC) of 0.85 versus 0.76 (P = 0.004), but similar diagnostic accuracy to VCTE (AUROC, 0.85 vs. 0.83; P = 0.5). The study did not include the use of the FibroScan XL probe, which may have significantly improved the number of reliable results for VCTE and potentially affected the diagnostic performance. For other stages of fibrosis, all of the imaging modalities performed equally well. Coupled with a similar comparative trial in patients with chronic liver disease,11 SSI appears to have a slight advantage for diagnosing earlier‐stage fibrosis, though more rigorous study is needed to confirm this finding. Table 1 - Discriminative Cut‐off Values of Elastography‐Based Imaging in NAFLD Fibrosis Stage VCTE‐Kwok(7) VCTE‐Cassinotto(9) ARFI‐Fierbinteanu Braticevici(10) ARFI‐Cassinotto(9) SSI‐Cassinotto(9) F ≥2 6.7‐7.7 kPa 6.2‐9.8 kPa 1.165 m/s 0.95 ‐ 1.32 m/s 6.3 ‐ 8.7 kPa AUROC 0.79‐0.87 0.76‐0.87 0.891‐0.997 0.70‐0.83 0.79‐0.90 Sensitivity, % 67‐94 60‐90 84.8 56‐90 71‐90 Specificity, % 61‐84 45‐90 90.3 36‐91 50‐90 F ≥3 8.0‐10.4 kPa 8.2‐12.5 kPa 1.48 m/s 1.15‐1.53 m/s 8.3‐10.7 kPa AUROC 0.76‐0.98 0.80‐0.90 0.956‐1.000 0.78‐0.89 0.83‐0.92 Sensitivity, % 65‐100 57‐90 86.4 59‐90 71‐91 Specificity, % 75‐97 61‐90 95.2 63‐90 71‐90 F4 10.3‐17.5 kPa 9.5‐16.1 kPa 1.635 m/s 1.3‐2.04 m/s 10.5‐14.4 kPa AUROC 0.91‐0.99 0.79‐0.92 0.958‐1.000 0.78‐0.89 0.82‐0.92 Sensitivity, % 78‐100 65‐92 91.7 44‐90 58‐90 Specificity, % 82‐98 62‐90 92.3 67‐90 72‐90 Abbreviations: kPa, kilopascals; m/s, meters per second. Cassinotto et al. have demonstrated that elastography can successfully identify significant fibrosis in the majority of patients. Fortunately, fibrosis is the histopathological feature most predictive of poor outcomes in NAFLD. On the other hand, NAFLD patients with NASH are also at substantially increased risk for progression of liver disease. In fact, patients with NASH have fibrosis progression rates that have been estimated to be one stage every 7 years, significantly higher than patients with non‐NASH NAFLD.12 In the Cassinotto et al. study, 29% of patients had fibrosis stage <2. Of this cohort, it is not possible by elastography to identify which patients had NASH from those with non‐NASH NAFLD, and thus one third of patients in this study will require additional risk stratification. Further evaluation is needed to assess whether any elastography‐based imaging modality can accurately discriminate those patients with NASH from those with non‐NASH NAFLD. Limitations of elastography‐based imaging have been identified. The most common limitations include obesity, intercostal wall thickness, and operator inexperience, but others include recent ingestion of food, ascites, and heart failure. Recent work by Petta et al. assessed the impact of steatosis severity on liver stiffness measurement (LSM) values and on VCTE accuracy for diagnosis of fibrosis in patients with NAFLD.13 Interestingly, patients with steatosis ≥66% at liver biopsy had higher LSM values, which led to higher false‐positive LSM results. Thus, in patients with a high degree of steatosis, VCTE may be less accurate in diagnosing severe fibrosis in NAFLD, and additional evaluation may be warranted in order to avoid overestimation of fibrosis. In the current study, LSM was mainly associated with fibrotic change. Steatosis, inflammation, and hepatocyte ballooning did not play a significant role in LSM results on multivariate analysis. However, on univariate analysis, VCTE had a slight correlation with steatosis and NAFLD activity score. Though 27% of patients had >66% steatosis, we are unable to determine from the data how many of these patients had higher LSM values and may represent false positives. Given these contradictory results, additional study on the effects of steatosis on LSM is needed. The controlled attenuation parameter (CAP) has the potential to provide steatosis quantification simultaneously with VCTE, though further data in NAFLD patients are needed to provide accurate cutoffs that correlate with moderate and severe steatosis on biopsy. Theoretically, CAP might be incorporated with VCTE to provide a corrected LSM in patients with severe steatosis. With the increasing prevalence of NAFLD worldwide, there is a continuing need to develop noninvasive methods to assess fibrosis, NASH, and non‐NASH NAFLD. In this issue of Hepatology, Cassinotto et al. added substantially to our understanding of the role of elastography‐based imaging modalities on assessment of fibrosis in NAFLD. In the current state, elastography imaging modalities appear most useful to exclude significant fibrosis and cirrhosis with a high negative predictive value. Unfortunately, based on current data, elastography alone cannot stratify those patients with NASH and minimal fibrosis. Subsequently, further study is needed to be able to identify this group of patients with high‐risk clinical features for disease progression.