Development and Application of Biomarker-Based Algorithms for Prediction of NAFLD and other Liver Conditions​


Medical algorithms are utilized by physicians and other health care professionals on a daily basis.   ​

Some are as simple as providing a calculated BMI from a patient’s height and weight, while others may be more complex and critical, e.g. providing specific medication recommendations for psychiatric conditions. ​

There are literally hundreds of medical algorithms available for diagnosis and treatment recommendations available today – and their continued improvement is an important target of researchers and all providers of quality health care delivery from patient care to health economics.  ​

Regardless of the target, healthcare professionals employ them to assure that their practices adhere to evidence-based principles while allowing for the influence of their own medical judgment.​

Addressing liver health status is a prime example of this balance, and researchers have been creating computational assessments of liver status for some time, and for good reason.  ​

As with other health diagnostic algorithms, computerized liver diagnoses like LiverFASt™ can provide rapid, non-invasive, inexpensive, and (most importantly) early alerts well before a patient’s condition demands the expense and added risk of liver biopsy​.

Imbert-Bismut, et al, 2001 is generally credited with the development of the first liver fibrosis algorithm combining biochemical markers to predict levels of liver fibrosis comparable to biopsy analyses.¹

Since then, refinements have been reported and the application of the predictive capability to chronic liver infections, NAFLD, and other liver disorders has been well-studied.  Similar algorithms provide liver fibrosis as well as steatosis and inflammation assessment. ​

Non Invasive Liver Screen

The report of the LiverFASt™ algorithm reproduces these earlier algorithms with the added value of requiring fewer biomarkers in the prediction of liver disease status and creates an agile platform that will allow refinement and refocus of the predictions as the population to be assessed varies.²​

For example, a European Association for the Study of the Liver (EASL)  Guidelines document points out that the NAFLD biomarker-derived fibrosis score seems to perform better in Caucasians than Asians.³

Algorithm specificity studies addressing their usefulness in pediatric liver diseases also indicate an absence of universality in this population. ⁴, ⁵

When examining tissue provided from liver biopsy, pathologists frequently use a scoring system known as the METAVIR score to determine liver disease progression. 

METAVIR identifies degree of  inflammation (the sum of hepatocyte ballooning and lobular inflammation) from A0 to A4 and fibrosis from F0 to F4.⁶

Subsequently the Goodman score was added to provide the ability to grade the degree of steatosis from S0 to S3 based on the percent of hepatocytes containing visible macrovesicular steatosis.⁷

The combination of these three criteria provides a more accurate, more comprehensive, and less subjective description for the diagnosis through the concept of an SAF score, and provided differential diagnoses of “no NAFLD”, “NAFLD”, or “NASH” from biopsies. ​

Non invasive liver fibrosis tests

Of course, biopsies are not the ideal first line of investigation for patients who may be at risk for NAFLD.  It is well known that the procedure for tissue collection is not without risk to the patient’s health and may not provide enough material to adequately assess disease progression. ​

Additionally, repeated biopsies over time are not a practical way to follow progression of disease or liver recovery. ​

With that in mind in 2015, the EASL3 offered revisions to their clinical practice guidelines.
Non-invasive diagnostic methods (either well-established combined biomarkers or device-delivered liver stiffness measurement) ​

  • Can be used instead of liver biopsy to assess liver disease severity prior to treatment at a safe level of predictability (Recommendation A1, EASL 2015). Non-invasive tests have no known complications and no sampling errors, and they also have the potential to be used in the staging of fibrosis, if they are accurate.

​The guideline goes on to note: ​

  • The practical advantages of analyzing serum biomarkers to measure fibrosis include their high applicability (>95%), their good inter-laboratory reproducibility, and their potential widespread availability.​
liver disease diagnostic tests

Good medical care provides that non-invasive test reports are best interpreted by specialists and take into account the clinical context and any possible downfalls or short comings of a particular test. ​

According to Castera in 2018⁸, simple steatosis does not increase mortality. Fibrosis, however, is the most important prognostic factor in NAFLD and is correlated with liver-related outcomes and mortality. Not all NAFLD patients will develop advanced fibrosis, but in studies where its diagnosis has been indicated by non-invasive biomarker algorithms, the results are consistent with the risk for liver-related complications and mortality. ​

Clearly a patient whose biomarker assessment demonstrates significant risk for advanced liver disease should be evaluated by a specialist and considered for biopsy follow-up.​

In summary, the current EASL recommendations read as follows:​

Screening of liver fibrosis for NAFLD patients is recommended, particularly among patients with metabolic syndrome or type 2 diabetes mellitus who have higher risk of liver fibrosis (Evidence level A1) ​

Non-invasive assessment including serum biomarkers or transient elastography can be used as first line procedure for the identification of patients at low risk of severe fibrosis/ cirrhosis (Evidence level A1) ​

The identification of significant fibrosis is less accurate with non-invasive tests as compared to liver biopsy and may necessitate, according to the clinical context, histological confirmation (Evidence level A1)

Follow-up assessment by either serum biomarkers or transient elastography for progression of liver fibrosis should be performed among NAFLD patients at a 3 year interval (Evidence level B1) ​

Additionally, there are similar recommendations regarding the comparable use of non-invasive markers for patients identified with HBV and HCV infections.​


1. Imbert-Bismut, F., Ratziu, V., Pieroni, L., Charlotte, F., Benhamou, Y., & Poynard, T. (2001). Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: a prospective study. The Lancet, 357(9262), 1069-1075.​

2. Byiringiro, A.  and  Letourneur, M. (2018) Improvements  to First Generation Non-Invasive Biomarker-Based Liver Disease Diagnostics Using Machine Learning (LiverFASt™) (submitted)​

3. European Association for Study of Liver. (2015). EASL-ALEH Clinical Practice Guidelines: Non-invasive tests for evaluation of liver disease severity and prognosis. Journal of hepatology, 63(1), 237.​

4. Hermeziu, B., Messous, D., Fabre, M., Munteanu, M., Baussan, C., Bernard, O., ... & Jacquemin, E. (2010). Evaluation of FibroTest–ActiTest in children with chronic hepatitis C virus infection. Gastroenterologie clinique et biologique, 34(1), 16-22.​

5. Pokorska-Śpiewak, M., Kowalik-Mikołajewska, B., Aniszewska, M., Pluta, M., & Marczyńska, M. (2017). Non-invasive evaluation of the liver disease severity in children with chronic viral hepatitis using FibroTest and ActiTest–comparison with histopathological assessment. Clinical and experimental hepatology, 3(4), 187.​

6. Munteanu M, Tiniakos D, Anstee Q, Charlotte F, Marchesini G, Bugianesi E, Tauner M, Gomez MR, Oliveira C, Day C, Dufour JF, Bellentani S, Ngo Y, Traussnig S, Perzaao H, Deckmyn O, Bedossa P, Ratziu V, Poynard T (2016) Diagnostic performance of FibroTest, SteatoTest and ActiTest in patients with NAFLD using the SAF score as histological reference. Alimentary pharmacology & therapeutics 44(8): 877-889​

7. Poynard, T., Ratziu, V., McHutchison, J., Manns, M., Goodman, Z., Zeuzem, S., ... & Albrecht, J. (2003). Effect of treatment with peginterferon or interferon alfa‐2b and ribavirin on steatosis in patients infected with hepatitis C. Hepatology, 38(1), 75-85.​

8. Castera, L. (2018). Diagnosis of non‐alcoholic fatty liver disease/non‐alcoholic steatohepatitis: Non‐invasive tests are enough. Liver International, 38(S1), 67-70.​


We are a team of life science experts, physicians, mathematicians, health IT professionals and software architects.