The prevalence of non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH), both potentially progressive diseases that can lead to cirrhosis and liver cancer,1 is increasing.2 This is not wholly surprising given that NAFLD/NASH with and without fibrosis are associated with lifestyle (e.g., poor diet and exercise habits) and metabolic risk factors (e.g., obesity, type 2 diabetes, etc.).3

As the gold standard for evaluating and classifying stages of liver disease, liver biopsy unfortunately has some drawbacks: it is invasive, can be painful, is limited by undersampling (represents only 1/50,000 of the entire liver volume), and is dependent on the pathologist’s expertise for accurate interpretation and scoring. Further, subject retention can be affected by repeated measurements and use of biopsy for disease monitoring, which are both indicated for drug development.

Alternatives to Biopsy

Considerable progress has been made in developing and validating blood- and imaging-based biomarkers – both promising non-invasive methods for accurate detection, staging, and monitoring of liver disease and fatty change. What we’re lacking for broad use of these biomarkers in research and clinical settings is systematic comparisons of these methods against each other and across technology manufacturers as well as validation against biopsy – not only for inclusion criteria but also to determine treatment efficacy.

I recently attended the EASL International Liver Congress where I had the chance to discuss using ultrasound (US) methods before embarking on expensive magnetic resonance imaging (MRI) testing in clinical trials. The EchoSens FibroScan® system provides a controlled-attenuation parameter (CAP) value for liver steatosis and vibration-controlled transient elastography (VCTE™) value for liver stiffness/fibrosis, and both seem to be favorable quantitative biomarkers. CAP and TE values can also be obtained using US elastography technology from other manufacturers (e.g., Siemens, Philips, Esaote) – this led me to wonder:

  • How do the measurements compare across manufacturers?
  • Can we use the same cut-off values indiscriminately in protocol inclusion criteria?
  • What are the positive predictive (PPV) and negative predictive value (NPV) of each imaging assessment?
  • What are the best respective roles of US and MRI in the detection and monitoring of liver fat and fibrosis in clinical trials?

Apples to Apples?

Some consortia have already undertaken the task of standardizing, comparing, and validating imaging biomarkers for liver disease diagnosis and staging as well as to measure the response to treatment – essentially to help us know whether we are achieving the same results independent of equipment and technology. These are the Non-Invasive Biomarkers of Metabolic Liver Disease (NIMBLE) and Liver Investigation: Testing Marker Utility in Steatohepatitis (LITMUS) projects.

Measurement reproducibility has been published for MRI – liver fat and stiffness were reproducible across field strengths, imager manufacturers, and reconstruction methods.4,5 If you want an introduction to fatty liver MRI, make sure to download the review by Reeder and Sirlin.6

The same measurement reproducibility has not yet been completed for US-based assessments, although cut-off values have been published. Within the total possible US measurement range of 100-400 dB/m, the following CAP score ranges have been established to grade steatosis: ≥248, stage 1; ≥268, stage 2; and ≥280, stage 3.7  

Given that we don’t know if performance between manufacturers varies, imagine a clinical trial whose inclusion criteria required stage 1 steatosis at the minimum. If the same patient (whose biopsy result indicated stage 3 steatosis) was measured using two different pieces of US equipment, one providing a result of 270 dB/m (site #1) and the other 280 dB/m (site #2), this patient would be excluded if they were screened at site #1. Therefore, in conjunction with setting the minimum CAP value, flexibility may be needed to customize the desired level of liver fat based on the US system parameters.

Clearly defined cut-offs for TE scores to grade fibrosis have not been established, although a recent meta-analysis of pooled data reported the following ranges (kPa): 6.7-7.7, ≥F2; 8.0-10.4, ≥F3; 10.3-17.5, F4.8 And, the FibroScan® system currently uses the following VCTE score ranges (kPa) to grade fibrosis: 2-7, F0 to F1; 7.5-10, F2; 10-14, F3; and ≥14, F4. This variability in results presents particular challenges for use in clinical studies.

US, MRI, Both?

As we learn more about the use of imaging in detecting fatty liver, the question becomes – what is the proper use of MRI and/or US in screening/eligibility and monitoring for treatment-related change?

Given its low cost, safety, quick execution, and wide availability, liver US is accepted as a first-line imaging technique to screen for liver steatosis. In addition, US elastography (CAP and TE) has demonstrated good accuracy in quantifying the levels of liver steatosis and fibrosis in patients with NAFLD.9 However, while US-TE is good at distinguishing non-significant fibrosis (F0-F1) from significant (F≥2) or advanced (F≥3-4) fibrosis, it is limited in its ability to differentiate individual fibrotic stages.10 Based on a meta-analysis of 40 studies, Tsochatzis et al. reported that, although pooled sensitivity and specificity were 0.79 and 0.78 for F2 and 0.83 and 0.89 for F4, no optimal cut-off for individual fibrosis stage could be determined because the cut-offs ranged widely – with significant overlap within and between fibrotic stages. These results indicate that the best cut-off values for each liver fibrosis stage are difficult to determine because they are highly dependent on the etiology of liver disease.11

Because of its good sensitivity, US-CAP and/or TE could be particularly useful to screen subjects for study entry – reducing the imaging cost of screening and saving MRI for patients who screen CAP-positive. A poster presented at this year’s NASH Summit reported the ability of US-CAP to predict MRI-PDFF. The AUROC of 0.77 suggested that a US-CAP of 328 should be used as the cut-off for 10% MRI-PDFF. Given the improved PPV of MRI-PDFF, screening subjects using US-CAP should increase the acceptance rate when you get to the more expensive MRI method.

To measure treatment efficacy, MRI is the more suitable choice. It can distinguish between stages – providing a more granular measure of the effect – something that US is not able to do.

For steatosis staging, compared with liver biopsy, MRI-PDFF had an AUROC of 0.989 (95% CI: 0.968, 1.000) for distinguishing patients with steatosis grade 0 from those with ≥grade 1, 0.825 (95% CI: 0.734, 0.915) to distinguish those with ≤grade 1 from those with ≥grade 2, and 0.893 (95% CI: 0.809, 0.977) to distinguish those with ≤grade 2 from those with grade 3.12

Regarding fibrosis staging, MRE can differentiate ≥F2 from F1 (sensitivity 89.7%, specificity 87.1%) at a stiffness cutoff value of 3.05 kPa and F3 from F4 (sensitivity 100%; specificity 92.2%) with a cutoff value of 5.32 kPa.13 In comparison, in a meta-analysis of 50 studies by Friedrich-Rust et al., US-TE demonstrated an AUROC of 0.84 with an optimal cut-off value of 7.6 kPa for detection of significant fibrosis (≥F2), while for cirrhosis (F4), the best cut-off value was 13 kPa with an AUROC of 0.94, as compared with liver biopsy.14

In addition to staging advanced fibrosis, MRE has shown high accuracy for discriminating NAFLD from NASH (cutoff value 2.74 kPa, AUROC 0.93, sensitivity 94%, specificity 73%) in a retrospective study of a patient cohort with a relatively high frequency of advanced disease. However, the performance of MRE to diagnose NASH in the absence of significant fibrosis is unknown but is thought to be limited.15 These studies indicate that MRE-measured hepatic stiffness has the potential to identify NASH before fibrosis onset. Although MRE shows superior diagnostic performance for liver fibrosis evaluation compared with US-TE, limited availability and high cost limit its widespread clinical adoption.16,17

Moving into the Future with Medical Imaging

Imaging of the liver for patients with NAFLD/NASH provides non-invasive options for screening and monitoring the disease in clinical trials. However, some consideration needs to occur during the trial planning phase to ensure that the most appropriate modality is being used for the intended outcomes – given the sites and their capabilities as well as the accuracy and validation of the chosen equipment.


1. Than NN, Ghazanfar A, Hodson J, et al. Comparing clinical presentations, treatments and outcomes of hepatocellular carcinoma due to hepatitis C and non-alcoholic fatty liver disease. QJM. 2017;110:73-81.

2. Perumpail BJ, Khan MA, Yoo ER, et al. Clinical epidemiology and disease burden of nonalcoholic fatty liver disease. World J Gastroenterol. 2017;23:8263-8276.

3. Leoni S, Tovoli F, Napoli L, et al. Current guidelines for the management of non-alcoholic fatty liver disease: a systematic review with comparative analysis. World J Gastroenterol. 2018;24:3361-3373.

4. Yokoo T, Serai SD, Pirasteh A, et al. Linearity, bias, and precision of hepatic proton density fat fraction measurements by using MR imaging: a meta-analysis. Radiology. 2018;286:486-498.

5. Trout AT, Serai S, Mahley AD, et al. Liver stiffness measurements with MR elastography: agreement and repeatability across imaging systems, field strengths, and pulse sequences. Radiology. 2016;281:793-804.

6. Reeder SB, Sirlin CB. Quantification of liver fat with magnetic resonance imaging. Magn Reson Imaging Clin N Am. 2010;18:337-357, ix.

7. Chan WK, Nik Mustapha NR, Mahadeva S, et al. Can the same controlled attenuation parameter cut-offs be used for M and XL probes for diagnosing hepatic steatosis? J Gastroenterol Hepatol. 2018;33:1787-1794.

8. Kwok R, Tse Y‐K, Wong GL-H, et al. Systematic review with meta‐analysis: non‐invasive assessment of non‐alcoholic fatty liver disease – the role of transient elastography and plasma cytokeratin‐18 fragments. Aliment Pharmacol Ther. 2014;39:254-269.

9. Mikolasevic I, Orlic L, Franjic N, et al. Transient elastography (FibroScan((R))) with controlled attenuation parameter in the assessment of liver steatosis and fibrosis in patients with nonalcoholic fatty liver disease – where do we stand? World J Gastroenterol. 2016;22:7236-7251.

10. Ferraioli G, Filice C, Castera L, et al. WFUMB guidelines and recommendations for clinical use of ultrasound elastography: part 3: liver. Ultrasound Med Biol. 2015;41:1161-1179.

11. Tsochatzis EA, Gurusamy KS, Ntaoula S, et al. Elastography for the diagnosis of severity of fibrosis in chronic liver disease: a meta-analysis of diagnostic accuracy. J Hepatol. 2011;54:650-659.

12. Tang A, Tan J, Sun M, et al. Nonalcoholic fatty liver disease: MR imaging of liver proton density fat fraction to assess hepatic steatosis. Radiology. 2013;267:422-431.

13. Kim BH, Lee JM, Lee YJ, et al. MR elastography for noninvasive assessment of hepatic fibrosis: experience from a tertiary center in Asia. J Magn Reson Imaging. 2011;34:1110-1116.

14. Friedrich-Rust M, Ong MF, Martens S, et al. Performance of transient elastography for the staging of liver fibrosis: a meta-analysis. Gastroenterology. 2008;134:960-974.

15. Chen J, Talwalkar JA, Yin M, et al. Early detection of nonalcoholic steatohepatitis in patients with nonalcoholic fatty liver disease by using MR elastography. Radiology. 2011;259:749-756.

16. Lee SS, Park SH. Radiologic evaluation of nonalcoholic fatty liver disease. World J Gastroenterol. 2014;20:7392-7402.

17. Imajo K, Kessoku T, Honda Y, et al. Magnetic resonance imaging more accurately classifies steatosis and fibrosis in patients with nonalcoholic fatty liver disease than transient elastography. Gastroenterology. 2016;150:626-637.e627.