Novel artificial intelligent/neural network system for staging of nonalcoholic steatohepatitis

Application of artificial intelligence techniques for non-alcoholic fatty liver disease diagnosis: A systematic review (2005-2023)

BACKGROUND AND OBJECTIVES

Non-alcoholic fatty liver disease (NAFLD) is a common liver disease with a rapidly growing incidence worldwide. For prognostication and therapeutic decisions, it is important to distinguish the pathological stages of NAFLD: steatosis, steatohepatitis, and liver fibrosis, which are definitively diagnosed on invasive biopsy. Non-invasive ultrasound (US) imaging, including US elastography technique, and clinical parameters can be used to diagnose and grade NAFLD and its complications. Artificial intelligence (AI) is increasingly being harnessed for developing NAFLD diagnostic models based on clinical, biomarker, or imaging data. In this work, we systemically reviewed the literature for AI-enabled NAFLD diagnostic models based on US (including elastography) and clinical (including serological) data.

METHODS

We performed a comprehensive search on Google Scholar, Scopus, and PubMed search engines for articles published between January 2005 and June 2023 related to AI models for NAFLD diagnosis based on US and/or clinical parameters using the following search terms: "non-alcoholic fatty liver disease", "non-alcoholic steatohepatitis", "deep learning", "machine learning", "artificial intelligence", "ultrasound imaging", "sonography", "clinical information".

RESULTS

We reviewed 64 published models that used either US (including elastography) or clinical data input to detect the presence of NAFLD, non-alcoholic steatohepatitis, and/or fibrosis, and in some cases, the severity of steatosis, inflammation, and/or fibrosis as well. The performances of the published models were summarized, and stratified by data input and algorithms used, which could be broadly divided into machine and deep learning approaches.

CONCLUSION

AI models based on US imaging and clinical data can reliably detect NAFLD and its complications, thereby reducing diagnostic costs and the need for invasive liver biopsy. The models offer advantages of efficiency, accuracy, and accessibility, and serve as virtual assistants for specialists to accelerate disease diagnosis and reduce treatment costs for patients and healthcare systems.

Artificial intelligence/neural network system that accurately diagnoses hepatocellular carcinoma in nonalcoholic steatohepatitis

BACKGROUND AND AIM

The aim of this study was to develop a novel noninvasive test using an artificial intelligence/neural network system (called HCC-Scope) to diagnose early-stage hepatocellular carcinoma (HCC) on the background of nonalcoholic steatohepatitis (NASH).

METHODS

In total, 175 patients with histologically proven nonalcoholic fatty liver disease and 55 patients with NASH-HCC were enrolled for training and validation studies. Of the 55 patients with NASH-HCC, 27 (49.1%) had very early-stage HCC, and six (10.9%) had early-stage HCC based on the Barcelona Clinic Liver Cancer staging system. Diagnosis with HCC-Scope was performed based on 12 items: age, sex, height, weight, AST level, ALT level, gamma-glutamyl transferase level, cholesterol level, triglyceride level, platelet count, diabetes status, and IgM-free apoptosis inhibitor of macrophage level. The FMVWG2U47 hardware (Fujitsu Co. Ltd, Tokyo, Japan) and the originally developed software were used.

RESULTS

HCC-Scope had sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 100% for the differential diagnosis between non-HCC and HCC in a training study with gray zone analysis. It was also excellent in the validation study (95.0% sensitivity, 100% specificity, 100% PPV, and 97.1% NPV with gray zone analysis and 95.2% sensitivity, 100% specificity, 100% PPV, and 97.1% NPV without gray zone analysis). HCC-Scope had a significantly higher sensitivity (85.3%) and specificity (85.1%) than alpha-fetoprotein (AFP) level, AFP-L3 level, des-gamma-carboxy prothrombin (DCP) level, and the gender-age-AFP-L3-AFP-DCP (GALAD) score.

CONCLUSIONS

HCC-Scope can accurately differentially diagnose between non-HCC NASH and NASH-HCC, including very early-stage NASH-HCC.

SWOT analysis of noninvasive tests for diagnosing NAFLD with severe fibrosis: an expert review by the JANIT Forum

Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease. Nonalcoholic steatohepatitis (NASH) is an advanced form of NAFLD can progress to liver cirrhosis and hepatocellular carcinoma (HCC). Recently, the prognosis of NAFLD/NASH has been reported to be dependent on liver fibrosis degree. Liver biopsy remains the gold standard, but it has several issues that must be addressed, including its invasiveness, cost, and inter-observer diagnosis variability. To solve these issues, a variety of noninvasive tests (NITs) have been in development for the assessment of NAFLD progression, including blood biomarkers and imaging methods, although the use of NITs varies around the world. The aim of the Japan NASH NIT (JANIT) Forum organized in 2020 is to advance the development of various NITs to assess disease severity and/or response to treatment in NAFLD patients from a scientific perspective through multi-stakeholder dialogue with open innovation, including clinicians with expertise in NAFLD/NASH, companies that develop medical devices and biomarkers, and professionals in the pharmaceutical industry. In addition to conventional NITs, artificial intelligence will soon be deployed in many areas of the NAFLD landscape. To discuss the characteristics of each NIT, we conducted a SWOT (strengths, weaknesses, opportunities, and threats) analysis in this study with the 36 JANIT Forum members (16 physicians and 20 company representatives). Based on this SWOT analysis, the JANIT Forum identified currently available NITs able to accurately select NAFLD patients at high risk of NASH for HCC surveillance/therapeutic intervention and evaluate the effectiveness of therapeutic interventions.

Comparison of thrombocytopenia between patients with non-alcoholic fatty liver disease and those with hepatitis C virus-related chronic liver disease

AIM

Thrombocytopenia is widely recognized as a simple surrogate marker of liver fibrosis in non-alcoholic fatty liver disease (NAFLD). Thrombocytopenia of NAFLD has not been compared with that of hepatitis C virus-related chronic liver disease (CLD-C). Here, we examined whether there is any difference in the platelet counts between patients with NAFLD and CLD-C and investigated the underlying mechanisms.

METHODS

A total of 760 biopsy-confirmed NAFLD and 1171 CLD-C patients were enrolled. After stratification according to the liver fibrosis stage, platelet counts between NAFLD and CLD-C patients were compared. The platelet count, spleen size, serum albumin level, serum thrombopoietin level, and immature platelet fraction (IPF) value were also compared after covariate adjustment using propensity score (PS) matching.

RESULTS

The median platelet counts (×10⁴ /μL) of NAFLD and CLD-C patients were 20.2 and 18.7 (p = 2.4 × 10^-5 ) in F1; 20.0 and 14.5 (p = 2.1 × 10^-12 ) in F2; 16.9 and 12.3 (p = 8.1 × 10^-10 ) in F3; and 11.1 and 8.1 (p = 0.02) in F4, respectively. In the F3 group, NAFLD patients had a significantly higher platelet count and significantly smaller spleen volume than CLD-C patients. Although the serum thrombopoietin levels were comparable between NAFLD and CLD-C patients, the IPF value of NAFLD patients was significantly higher than that of CLD-C patients.

CONCLUSIONS

NAFLD patients had a significantly higher platelet count than CLD-C patients following stratification according to the liver fibrosis stage. The milder hypersplenism and higher platelet production in NAFLD than CLD-C may have contributed to this difference.

Development and multicenter validation of FIB-6: A novel, machine learning, simple bedside score to rule out liver cirrhosis and compensated advanced chronic liver disease in patients with chronic hepatitis C

BACKGROUND

Non-invasive tests (NITs), such as Fibrosis-4 index (FIB-4) and the aspartate aminotransferase-to-platelet ratio index (APRI), developed using classical statistical methods, are increasingly used for determining liver fibrosis stages and recommended in treatment guidelines replacing the liver biopsy. Application of conventional cutoffs of FIB-4 and APRI resulted in high rates of misclassification of fibrosis stages.

AIM

There is an unmet need for more accurate NITs that can overcome the limitations of FIB-4 and APRI.

PATIENTS AND METHODS

Machine learning with the random forest algorithm was used to develop a non-invasive index using retrospective data of 7238 patients with biopsy-proven chronic hepatitis C from two centers in Egypt; derivation dataset (n = 1821) and validation set in the second center (n = 5417). Receiver operator curve analysis was used to define cutoffs for different stages of fibrosis. Performance of the new score was externally validated in cohorts from two other sites in Egypt (n = 560) and seven different countries (n = 1317). Fibrosis stages were determined using the METAVIR score. Results were also compared with three established tools (FIB-4, APRI, and the aspartate aminotransferase-to-alanine aminotransferase ratio [AAR]).

RESULTS

Age in addition to readily available laboratory parameters such as aspartate, and alanine aminotransferases, alkaline phosphatase, albumin (g/dl), and platelet count (/cm³ ) correlated with the biopsy-derived stage of liver fibrosis in the derivation cohort and were used to construct the model for predicting the fibrosis stage by applying the random forest algorithm, resulting in an FIB-6 index, which can be calculated easily at http://fib6.elriah.info. Application of the cutoff values derived from the derivation group on the validation groups yielded very good performance in ruling out cirrhosis (negative predictive value [NPV] = 97.7%), compensated advance liver disease (NPV = 90.2%), and significant fibrosis (NPV = 65.7%). In the external validation groups from different countries, FIB-6 demonstrated higher sensitivity and NPV than FIB-4, APRI, and AAR.

CONCLUSION

FIB-6 score is a non-invasive, simple, and accurate test for ruling out liver cirrhosis and compensated advance liver disease in patients with chronic hepatitis C and performs better than APRI, FIB-4, and AAR.

Application of artificial intelligence in non-alcoholic fatty liver disease and liver fibrosis: a systematic review and meta-analysis

BACKGROUND

The global prevalence of non-alcoholic fatty liver disease (NAFLD) continues to rise. Non-invasive diagnostic modalities including ultrasonography and clinical scoring systems have been proposed as alternatives to liver biopsy but with limited performance. Artificial intelligence (AI) is currently being integrated with conventional diagnostic methods in the hopes of performance improvements. We aimed to estimate the performance of AI-assisted systems for diagnosing NAFLD, non-alcoholic steatohepatitis (NASH), and liver fibrosis.

METHODS

A systematic review was performed to identify studies integrating AI in the diagnosis of NAFLD, NASH, and liver fibrosis. Pooled sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and summary receiver operating characteristic curves were calculated.

RESULTS

Twenty-five studies were included in the systematic review. Meta-analysis of 13 studies showed that AI significantly improved the diagnosis of NAFLD, NASH and liver fibrosis. AI-assisted ultrasonography had excellent performance for diagnosing NAFLD, with a sensitivity, specificity, PPV, NPV of 0.97 (95% confidence interval (CI): 0.91-0.99), 0.98 (95% CI: 0.89-1.00), 0.98 (95% CI: 0.93-1.00), and 0.95 (95% CI: 0.88-0.98), respectively. The performance of AI-assisted ultrasonography was better than AI-assisted clinical data sets for the identification of NAFLD, which provided a sensitivity, specificity, PPV, NPV of 0.75 (95% CI: 0.66-0.82), 0.82 (95% CI: 0.74-0.88), 0.75 (95% CI: 0.60-0.86), and 0.82 (0.74-0.87), respectively. The area under the curves were 0.98 and 0.85 for AI-assisted ultrasonography and AI-assisted clinical data sets, respectively. AI-integrated clinical data sets had a pooled sensitivity, specificity of 0.80 (95%CI: 0.75-0.85), 0.69 (95%CI: 0.53-0.82) for identifying NASH, as well as 0.99-1.00 and 0.76-1.00 for diagnosing liver fibrosis stage F1-F4, respectively.

CONCLUSION

AI-supported systems provide promising performance improvements for diagnosing NAFLD, NASH, and identifying liver fibrosis among NAFLD patients. Prospective trials with direct comparisons between AI-assisted modalities and conventional methods are warranted before real-world implementation.

PROTOCOL REGISTRATION

PROSPERO (CRD42021230391).