Noninvasive tests in the evaluation of nonalcoholic fatty liver disease: A review

Download video file ^{(22.3MB, mp4)}

INTRODUCTION

NAFLD is the most prevalent form of liver disease worldwide and affects about 25% of the general population.^1,2 NAFLD consists of a spectrum of liver diseases from NAFL, which features macro-vesicular accumulation of triglycerides in more than 5% of hepatocytes, to NASH, a more severe, and inflammatory subtype of NAFLD. NASH has hepatocyte injury (ballooning) and lobular inflammation, with or without fibrosis.^2,3 It can be associated with complications such as cirrhosis, portal hypertension, and HCC.⁴

The gold standard for diagnosis of NASH is liver biopsy with hepatic steatosis of more than 5% and evidence of hepatocyte injury, which is characterized by ballooning, lobular inflammation, and/or portal inflammation.² Fibrosis, although not required for the diagnosis of NASH, has been found to be strongly associated with the prognosis of NAFLD.^2–5 The stage of baseline fibrosis is predictive of overall mortality, disease progression, and development of clinical liver events.⁶ Liver biopsy is a near-universal requirement to differentiate simple steatosis from NASH and to monitor the progression of the disease. However, liver biopsy is limited in availability due to its cost and invasive nature.^2,7,8 Therefore, in the past decade, there has been increasing interest in the use of noninvasive tests (NITs) to diagnose and monitor the progression of NAFLD. This review article discusses the efficacy of the most common noninvasive tests for diagnosing and monitoring the stages of NAFLD.

Nonalcoholic fatty liver disease fibrosis score

In 2006, NAFLD fibrosis score (NFS) was developed to identify patients with NALFD at increased risk of developing cirrhosis, liver-related complications, or death.^9,10 NFS uses routinely available laboratory tests and patient characteristics, such as age, body mass index (BMI), impaired fasting glucose (IFG) or diabetes (score of 1 if present, 0 if absent), AST, ALT, platelet count, and albumin.¹⁰ NFS is derived from the following formula:

$$\mathrm{NFS}=-1.675+\left(0.037\times \mathrm{Age\; in\; years}\right)+\left(0.094\times \mathrm{BMI}\right)+\left(1.13\times \mathrm{Presence\; of\; IFG\; or\; Diabetes}\right)+\left(0.99\times \frac{\mathrm{AST}}{\mathrm{ALT}}\right)-\left(0.013\times \mathrm{platelet\; count}\right)-(0.66\times \mathrm{Albumin})$$

If a patient’s NFS is below −1.455, advanced fibrosis can be ruled out with a negative predictive value (NPV) of 93%, while a high cutoff score of 0.676 can be used to diagnose advanced fibrosis with a positive predictive value (PPV) of 90%.^5,9 The value of the area under receiver operating characteristic curve (AUROC) of NFS was demonstrated to be 0.85 (95% CI 0.81–0.90), which is indicative of high accuracy.^5,11–13 NFS has been shown to identify progression to advanced fibrosis with C-statistic of 0.80 (95% CI 0.71–0.88); a unit change in NFS is associated with a change in stage of fibrosis by 0.19 (95% CI 0.71–0.88).¹³ Furthermore, patients with higher baseline NFS, with a cutoff ≥ 0.67, are at a higher risk of disease progression.¹⁴ A point increase in NFS corresponded to an increased risk of developing adverse clinical events.¹⁴

For patients whose scores fall into the indeterminate area between the low cutoff score and the high cutoff score, different NITs can be used in a sequential manner to decrease the incidence of indeterminate results. However, the use of multiple NITs can increase the rate of misclassification.¹⁵ Nevertheless, research trials have demonstrated that the sequential use of NITs does reduce the need for liver biopsy and does not compromise the sensitivity and specificity of each NIT.^15,16

Fibrosis-4 index

Fibrosis-4 (FIB-4) index was developed in 2006 to assess the severity of liver disease and the presence of fibrosis in patients with HIV and/or HCV.¹⁷ FIB-4 index can be derived by the following:

$$\mathrm{FIB}-4=\frac{\mathrm{Age}\times \mathrm{AST}}{\mathrm{Platelets}\times \sqrt{\mathrm{ALT}}}$$

A FIB-4 score of less than 1.3 excludes advanced fibrosis with NPV 95%, while a score of higher than 2.67 indicates advanced fibrosis with 70% PPV.¹⁸ With an AUROC of 0.86 (95% CI 0.78–0.94)¹⁸ and C-statistic of 0.80 (95% CI 0.78–0.82),¹³ FIB-4 index can accurately identify patients in the NAFLD population who have advanced fibrosis. High baseline FIB-4 index, with a cutoff value ≥ 3.25, is correlated with the occurrence of adverse clinical liver events (Hazard ratio (HR) 1.46 (95% CI 1.35–1.57)).¹⁴ For individuals whose FIB-4 index scores lie between 1.3 and 2.67, repeating the FIB-4 index can improve the identification of patients at elevated risk of severe liver disease.⁶

Recent societies guidelines have recommended FIB-4 as the first NIT to be used to screen high-risk population (eg, type 2 diabetes or individuals with 2 risk metabolic factors) followed by transient elastography in those with fall within the indeterminate zone.^19–21

Pro-peptide of type III collagen

Fibrogenesis results in the release of extracellular matrix protein fragments into the circulation, one of which is pro-peptide of type III collagen (PRO-C3).²² It is an N-terminal pro-peptide that is released from the synthesis and deposition of type III collagen and can be used as a marker of fibrosis in a variety of liver diseases, including NAFLD.⁷ PRO-C3 was discovered to be an independent marker of advanced fibrosis (F3-F4) compared to mild-moderate fibrosis (F1-F2) or without fibrosis (F0) [OR 1.84 with 95% CI (1.05, 3.23), p value 0.0341].²³ With a cutoff value of ≥ 20.9 ng/mL, advanced fibrosis can be identified with an AUROC of 0.81 (95% CI 0.74–0.87).^23,24 The level of PRO-C3 also declined in patients with regression of fibrosis.^23,25–27 Thus, based on the available data, the level of PRO-C3 is correlated with the severity of fibrosis, and PRO-C3 might be a noninvasive biomarker for the longitudinal monitoring of patients with NASH. However, PRO-C3 and PRO-C3-based serum biomarkers have not yet been approved by the United States Food and Drug Association and are not yet available in the United States

ADAPT, FIBC3, and ABC3D

PRO-C3 can also be incorporated into multiple diagnostic algorithms to evaluate the stage and severity of NASH. A model named ADAPT can be derived with the following formula:

$$\mathrm{ADAPT}=\exp \left({\log }_{10}\left(\frac{\mathrm{Age\; x}\mathrm{PROC}3}{\sqrt{\mathrm{platelet\; count}}}\right)\right)+\mathrm{diabetes}$$

With a cutoff value of 6.3287, ADAPT was able to distinguish patients with advanced fibrosis from those without advanced fibrosis with an AUROC of 0.86 (95% CI 0.709–0.91).²⁴ Similarly, a new diagnostic model called FIBC3 was able to diagnose advanced fibrosis using a threshold value of >−0.4 with AUROC of 0.89 (95% CI 0.843–0.941).²⁸ The score can be calculated by the following formula:

$$\mathrm{FIBC}3=-5.939+(0.053\times \mathrm{Age})+(0.076\times \mathrm{BMI})+(1.614\times \mathrm{T}2\mathrm{DM})-(0.009\times \mathrm{platelets})+(0.071\times \mathrm{PROC}3)$$

ABC3D, also a PRO-C3-based score, consists of A = age > 50 years, B = BMI > 30, C = platelet Count < 200, 3 = PRO-C3 > 15.5 ng/mL, and D = presence of diabetes.²⁸ Each factor is scored 1 point, except for the presence of diabetes, which is scored 2 points. An ABC3D score above 3 is suggestive of advanced fibrosis, with an AUROC of 0.88 (95% CI 0.822–0.929).²⁸ The 3 above-mentioned scores are not yet approved in the United States, however.

Enhanced liver fibrosis test

The enhanced liver fibrosis (ELF) test consists of 3 components: N-terminal pro-peptide of type III procollagen, hyaluronic acid, and TIMP-1.²⁹ Various algorithms of the ELF test have been derived to calculate a score for the assessment of hepatic fibrosis. At a low cutoff value of 7.7, the ELF test has a sensitivity of 0.93 (95% CI 0.82–0.34) for excluding fibrosis, although the specificity is limited at 0.34 (95% CI 0.13–0.65).²⁹ On the other hand, a high cutoff at 9.80 can raise the specificity to 0.86 (95% CI 0.77–0.92), but the sensitivity is lowered to 0.65.²⁹ At the high cutoff value of the ELF test, even though the NPV remains high, the PPV tends to be much lower in low-prevalence settings (prevalence < 30%), such as primary care, than in high-prevalence settings, such as secondary and tertiary referral centers.²⁹ Thus, the performance of the ELF test can vary based on the prevalence of the target population. A cutoff threshold of ≥ 9.76 at baseline can predict progression to cirrhosis, with a sensitivity of 77% and specificity 66%.²⁵ The baseline ELF test also predicted the development of liver-related events, with 56% sensitivity and 75% specificity at a cutoff value of ≥ 11.27.²⁵ In contrast, the ELF test reliably excluded the occurrence of liver-related adverse events within 1 year at <9.8 cutoff, with NPV 90%.³⁰ Furthermore, a unit increase in the ELF test was associated with developing liver-related outcomes [HR 2.53 (95% CI 1.45–4.43)] or progression to advanced fibrosis [HR 4.34, (95% CI 2.4–7.8)].³¹ Overall, the ELF test has been an accurate and noninvasive diagnostic and prognostic marker and has been approved in the United States. for use as a prognostic tool for patients with NASH with advanced fibrosis. However, in the recent REGENERATE study where obeticholic acid was used in treating NASH, ELF did not correlate well with fibrosis improvement. Thus, more studies are needed to assess the role of ELF in assessing response to therapies.³²

Metabolomics and lipodomics

The metabolomics approach has been studied in recent years as a noninvasive diagnostic biomarker for NAFLD, based on the expansive array of substrates including lipids, carbohydrates, peptides, proteins, amino acids, and nucleic acids.³³ Multiple research studies have developed metabolomics-based models to differentiate normal liver from NAFLD and/or NASH from NAFL.³⁴

Lipodomics, a subfield of metabolomics, is a promising field of interest for developing a noninvasive diagnostic and monitoring tool for NAFLD and NASH. At physiological levels, hepatic steatosis occurs when the fatty acid synthesis and uptake from the bloodstream of the liver exceeds the hepatic secretion of triglycerides as very low-density lipoprotein (VLDL).³³ Accumulation of lipids in the liver can lead to lipotoxicity and can cause hepatocyte injury and inflammation. A BMI-dependent lipodomic test can distinguish patients with NAFLD from those without NAFLD, with AUROC 0.90, 0.98 sensitivity, and 0.78 specificity.⁴ In addition, the test recognizes patients with NASH from those with NAFL, with AUROC 0.95, sensitivity 0.83, and specificity 0.94.⁴ Other studies have since validated the utility of lipodomics among patients with NAFLD.³⁵

Vibration-controlled transient elastography

Using the speed of a mechanically induced shear wave in the liver, vibration-controlled transient elastography (VCTE) calculates liver stiffness measurement (LSM), which estimates the degree of hepatic fibrosis.³⁶ LSM identified patients with F ≥ F2, F ≥ F3, and F4, with AUROC of 0.77 (95% CI 0.72–0.82), 0.80 (95% CI 0.75–0.84), and 0.89 (95% CI 0.84–0.93), respectively.³⁶ The cutoff strata for LSM were 8.2 kPa for F ≥ F2, 9.7 kPa for F ≥ F3, and 13.6 kPa for F=F4. LSM results were not affected by steatosis.

Moreover, VCTE can quantify hepatic steatosis by measuring controlled attenuation parameter (CAP), which is a magnitude of ultrasonic attenuation of the echo wave.³⁶ CAP was discovered to be able to identify patients with steatosis with AUROC of 0.87 (95% CI 0.82–0.92) for steatosis grade 1 or above, 0.77 (95% CI 0.71–0.82) for grade 2 or above, and 0.70 (95% CI 0.64–0.75) for grade 3.³⁶ The cutoff values used were 302 dB/m, 331 dB/m, and 337 dB/m for S ≥ S1, S ≥ S2, and S=S3, respectively.³⁶ This study established VCTE as a reliable, noninvasive tool to accurately assess steatosis and fibrosis in patients with NAFLD.

Patients with higher levels of fibrosis at baseline, as determined by VCTE, had poor overall survival and survival free of death from liver-related complications.³⁷ Patients with LSM above threshold 21 kPa at baseline were also at 14% risk of developing hepatic decompensation compared to 2% risk for those with LSM below the threshold (HR 1.03, 95% CI 1.02–1.04).^38,39 Despite its longitudinal use in clinical trials, LSM has not been externally validated through long-term assessment of fibrosis ⁴⁰; therefore, correlation with histology is required, and further research is warranted.

MRI-derived proton density fat fraction

Although abdominal ultrasound has been the first-line assessment of hepatic steatosis, one of its limitations is the inability to differentiate degrees of steatosis.⁴¹ Magnetic resonance spectroscopy is an alternative imaging technique that can accurately quantify liver fat and stage steatosis.⁴² However, the use of Magnetic resonance spectroscopy is restricted by the small volume of sampled tissue and the technical difficulty to perform.⁴² Hence, MRI-derived measurement of hepatic fat fraction (MRI-PDFF) may be considered in patients with evidence of NAFLD on ultrasonography who warrant accurate staging of steatosis for diagnostic, therapeutic, or research purposes. In clinical practice, it is usually coupled with magnetic resonance elastography (MRE) in the same examination, especially in patients who do not want to undergo liver biopsy. With the information from both imaging modalities, patients can have accurate quantification of fat and fibrosis in the liver.

MRI-PDFF has been recognized to have a strong correlation with histologically graded steatosis with r _s (correlation coefficient) of 0.68.^42,43 Individuals with 30% or greater reduction in MRI-PDFF were observed to have higher odds of improvement in steatosis grade by 1 or more points (OR 14.9, 95% CI 3.8–57.7).¹⁶ MRI-PDFF results were also correlated with NAFLD Activity Score (NAS) (r _s = 0.61).¹⁶ NAS is the histological scoring system for NAFLD and comprises steatosis, lobular inflammation, and hepatocyte ballooning. A 30% relative reduction in MRI-PDFF represents a 2-point reduction of NAS, which is determined as histological improvement of steatosis grade.^44,45 Although MRI-PDFF is highly sensitive to even small (<5%) changes in liver fat content, its use has been limited due to cost and availability.

MRI-PDFF responders have been noted with higher odds of achieving a 1-point or greater improvement in ballooning compared with that measure in nonresponders (OR = 2.9).^16,46 While an association between MRI-PDFF and lobular or portal inflammation has not been found, reductions of ballooning and inflammation were only seen in MRI-PDFF responders (~40% of responders).⁴⁷ Furthermore, the resolution of NASH was observed in 37% of MRI-PDFF responders compared to 4% in nonresponders.⁴⁷ Hence, MRI-PDFF is not only strongly correlated with hepatic steatosis, but it is also associated with inflammation and hepatocyte ballooning to a lesser degree. Longitudinal improvements in MRI-PDFF were associated with histological response in patients with NASH.⁴⁴ However, changes in MRI-PDFF have not been found to be significantly correlated with changes in fibrosis.

Magnetic resonance elastography

Two-dimensional magnetic resonance elastography (2D-MRE) can also be used as a noninvasive diagnostic tool for advanced fibrosis in NASH.⁴⁵ With a cutoff of >3.63 kPa, MRE yielded an AUROC of 0.924, sensitivity of 0.86, specificity of 0.91, PPV of 0.68, and NPV of 0.97 for diagnosis of advanced fibrosis.⁴⁵ Hepatic fat content has no effect on the 2D-MRE assessment of fibrosis.⁴⁸ MRE was able to maintain the sensitivity and specificity above 80% and AUROC above 0.9 for detecting patients with F ≥ 3 or F ≥ 4.^49,50 MRE is one of 2 tests that meet the minimum acceptable performance of diagnostic accuracy for advanced fibrosis; the second test is point shear wave elastography.⁴⁹

MRE can be a predictor of prognosis for patients with NAFLD. An MRE threshold of 4.39 kPa has been demonstrated to distinguish patients with cirrhosis from those without cirrhosis with an AUROC of (95% CI 0.854–0.989), 81.8% sensitivity, and 91.8% specificity.⁵¹ Moreover, the MRE cutoff at 6.48 kPa differentiated compensated cirrhosis from decompensated cirrhosis, with AUROC of 0.71 (95% CI 0.511–0.902), 66.7% sensitivity, and 80.8% specificity. Moreover, a unit increase in MRE-measured liver stiffness translates to 3.28 times higher risk of hepatic decompensation (OR 3.28, 95% CI 2.04–5.28).⁵¹ The risk of developing liver-related events increased per 1 kPa increment with HR 1.32 (95% CI 1.13–1.56).⁵² Moreover, regression in MRE-based liver stiffness was accompanied by a decline in fibrosis markers such as hyaluronic acid.⁵³ Such findings indicate that MRE is a potential noninvasive longitudinal monitoring tool of fibrosis. However, the validity of monitoring longitudinal changes in fibrosis with MRE should be tested in more research trials, especially in those correlating MRE with histology.

Ultrasound shear wave elastography

Ultrasound shear wave elastography is also another noninvasive method available to determine liver fibrosis. The AUROC for detecting F2, F3, and F4 was 0.871, 0.895, and 0.937, respectively, with corresponding cutoff values at 8.6 kPa, 10.7 kPa, and 13.8 kPa.⁵⁴ It indicates that ultrasound shear wave elastography is an effective noninvasive method for assessing liver fibrosis. Interestingly, the diagnostic accuracy, sensitivity, and specificity values were found to be higher with advanced fibrosis than F2 or below.^54–56

Iron-corrected T1 mapping

Iron-corrected T1 mapping (cT1) is an MRI-derived biomarker that reflects regional tissue water content. The cT1 was positively correlated with the overall NAS score (r _s = 0.36) and with the individual components of steatosis, ballooning, and inflammation (r _s = 0.54, 0.36, and 0.17, respectively).⁴³ A moderate correlation between cT1 and fibrosis was demonstrated (r _s = 0.33), and a significant association between cT1 and MRI-PDFF was also noted (r _s = 0.66).⁴³ An increase in cT1 by 88ms was associated with a 2-unit increase in NAS and a 21% relative change in MRI-PDFF.⁴³ These findings mark cT1 as a possible noninvasive tool for assessing the severity of fibrosis in patients with NAFLD. At threshold ≥ 825 ms, cT1 could also be used as a predictor of liver-related clinical outcomes with HR of 9.9 (95% CI 1.29–76.4).⁵⁷ The results implied that cT1 may be a sustainable test for screening NAFLD and for monitoring longitudinal changes.

FibroScan-AST score

The FibroScan-AST score (FAST score) is an algorithm that is being used to identify patients at risk of clinically significant NASH. This model, which combines LSM, CAP, and AST, can be derived by the following algorithm:

$$\mathrm{FAST}=\frac{{\mathrm{e}}^{-1.65+1.07\times \mathrm{In}\left(\mathrm{LSM}\right)+2.66*{10}^{-8}\times {\mathrm{CAP}}^3-63.3\times {\mathrm{AST}}^{-1}}}{1+{\mathrm{e}}^{-1.65+1.07\times \mathrm{In}\left(\mathrm{LSM}\right)+2.66*{10}^{-8}\times {\mathrm{CAP}}^3-63.3\times {\mathrm{AST}}^{-1}}}$$

FAST score uses a dual cutoff approach with cutoffs of ≤ 0.35 and ≥ 0.67.⁵¹ AUROC, PPV, and NPV were 0.80 (95% CI 0.76–0.85), 0.83, and 0.85, respectively, which is indicative of satisfactory performance.^51,58

However, up to 30%–39% of patients could fall within the gray zone between the 2 cutoff values. Moreover, FAST score relies heavily on AST values, and many patients with advanced fibrosis have normal aminotransferase values. To address such diagnostic challenges, Noureddin et al⁵⁹ devised a “traffic-light” approach to FAST score; if the FAST score is 0.67 or above, the patient can be enrolled in clinical trials.⁵⁹ Conversely, for patients who have FAST score ≤ 0.35, components of the predictive model should be considered individually, especially the LSM of VCTE, which is an accurate assessment of fibrosis.^36,59 On the other hand, if a patient’s score is ≤ 0.35 on FAST but LSM values meet the threshold for F2 or above, sequential testing with ELF or MRE is recommended.³⁶ Sequential testing should also be considered for patients whose scores fall within the gray zone between 0.67 and 0.35.

MRE combined with FIB-4 index (MEFIB)

In 2020, MRE combined with FIB-4 index (MEFIB) was developed as a potential noninvasive method to identify patients with NASH with significant fibrosis (F ≥ 2).⁶⁰ The cutoff thresholds used for MRE and FIB-4 index were ≥ 3.3 kPa and ≥ 1.6, respectively.¹⁹ Based on these parameters, MEFIB demonstrated PPV of 97.1% (p<0.02) and AUROC of 0.90 (95% CI 0.85–0.95) to rule in significant fibrosis.¹⁹ More recently, MEFIB has demonstrated correlation with clinical liver events.⁶¹ Positive MEFIB score (defined as MRE ≥ 3.3 kPa and FIB-4 index ≥ 1.6) is associated with liver-related events with HR 20.6 (95% CI: 10.4–40.8, p<0.001), while negative MEFIB score has high NPV at 99.1%.⁶¹

NIS-4 score

NIS-4 is a novel blood-based biomarker panel that consists of 4 independent NASH-associated biomarkers: miR-34a-5p, alpha-2 macroglobulin, YKL-40, and glycated hemoglobin.⁶² The test evaluated at-risk NASH or fibrotic NASH, which is defined as NAS ≥ 4 and fibrosis stage ≥ F2.⁶² It uses cutoffs of 0.36 and 0.63 to rule out or rule in those NAS ≥ 4 and fibrosis stage ≥ F2.⁶² The test has an AUROC of 0.80 (95% CI 0.73–0.85), which suggests that NIS-4 is a way to noninvasively rule in or rule out at-risk NASH in patients with NAFLD who have metabolic risk factors.

The MAST score

The MAST (MRI and AST) score is also a predictive model that uses MRI-PDFF, MRE, and AST values to rule in or rule out fibrotic NASH (those with NASH with NAS ≥ 4 and fibrosis stage ≥ F2).⁶³ The MAST score is derived from the results of MRE, MRI-PDFF, and AST with the following formula:

$$\mathrm{MAST}=-12.17+7.07\log \mathrm{MRE}+0.037\mathrm{PDFF}+3.55\log \mathrm{AST}$$

The MAST score uses dual cutoff values; low cutoff threshold of 0.165 represented 90% sensitivity and 98.1% NPV, while high cutoff threshold of 0.242 indicated 90% specificity and 50% PPV.⁶³ Overall AUROC of the MAST score was 0.93 (95% CI 0.88–0.97) and was higher than those of NFS, Fib-4 index, and FAST scores.⁶³ Moreover, fewer patients measured by the MAST score have indeterminate scores in the gray zone compared to previously established models.⁶³ Although the MAST score is currently primarily being used in research trials for recruiting subjects, it has shown promising potential to reduce requirement of liver biopsies especially for selection of patients in phase II trials and requires further research for use in the public community. In addition, recent data showed that MAST score predicts liver clinical events reach c-statistics of 0.92.⁶⁴ A logit unit increase in MAST is associated with adverse clinical liver events at HR 2.01 (95% CI: 1.59–2.54, p<0.0001).⁶⁴

MASEF score

Noureddin et al⁶⁵ developed a serum-based score to distinguish at-risk NASH patients with at least one point each on steatosis, lobular inflammation, and ballooning on NAS, NAS ≥ 4, and fibrosis stage ≥ 2, from patients without those features.⁶⁵ The score, named the Metabolomics Advanced Steatohepatitis Fibrosis Score, is a serum-based test that includes 12 lipids, BMI, AST, and ALT.⁶⁵ The AUROC was 0.81, as well as similar specificity, sensitivity, PPV, and NPV.⁶⁵ Therefore, the Metabolomics Advanced Steatohepatitis Fibrosis Score score may be an accurate, serum-based tool that can identify patients with clinically significant NASH for enrollment in clinical trials.

Summary and conclusions

The rising prevalence of NAFLD/NASH cases in the United States and worldwide has led to an increasing demand for noninvasive tests (NITs) for screening, diagnosing, and monitoring the disease.⁶⁶ However, no noninvasive test to diagnose or monitor NAFLD has been accepted for wide-scale use. Based on the results of multiple research trials, several NITs have shown diagnostic accuracy in including or excluding fibrosis. Certain NITs such as NFS, FIB-4 index, ProC3, ABC3D, metabolomics, VCTE, cT1, and MRI-PDFF may also be used to monitor longitudinal changes. It has been noted that patients with increased NIT scores at baseline and/or with longitudinal increase in scores are at increased risk of disease progression or decompensation.¹⁴

The Division of Gastroenterology of the Food and Drug Administration defines the target histologic improvement end points for patients with NASH as the resolution of steatohepatitis without worsening of liver fibrosis, improvement of fibrosis by at least one stage without worsening of steatohepatitis, and both resolution of steatohepatitis and improvement in fibrosis.⁶⁷ Fibrosis is an independent predictor of morbidity and mortality.¹⁴ In lieu of the current gold standard liver biopsy, several NITs can be used alone or in a sequential manner to identify at-risk NASH patients (or fibrotic NASH), screen for advanced fibrosis, or monitor the aforementioned end points. According to the practice guidance from the American Association for the Study of Liver Diseases, patients with clinically suspected NAFLD or hepatic steatosis should undergo FIB-4 as a primary screening step.⁶⁸ Afterward, patients with at-risk NASH (patients with NASH with NAS ≥ 4 and ≥ F2) or high-risk individuals such as those with type 2 diabetes mellitus, obesity, family history of cirrhosis, or significant alcohol consumption should receive additional screening for advanced fibrosis with tests. Patients diagnosed with advanced fibrosis, advanced NASH, or discordant NITs should be referred to a specialist for further evaluation and management.⁶⁸

In addition to the NITs discussed above, Agile 3+ and Agile 4 are emerging VCTE-based scoring systems that had shown promising potential in identifying advanced fibrosis based on retrospective analyses.^69,70 Further prospective studies are warranted to validate their efficacy.

For long-term monitoring, MRI-PDFF can be used in conjunction with at least 1 test that measures fibrosis such as VCTE, or MRE. Patients with two or more metabolic risk factors or imaging evidence of hepatic steatosis should also continue monitoring with FIB-4 index every 1-2 years.⁶⁸ The use of FAST, MAST, cT1, Pro-C3, or ELF as a monitoring tool for longitudinal changes is currently being evaluated. An example is illustrated in Figure 1. Since there are several NITs available, this is only an example of several different ways that NITs can be combined and personalized in accordance with the patient’s needs and availability of resources. However, MRI-based studies may not be available in all clinical settings and may have limited accessibility by the patients.

FIGURE 1.

An example of how different NITs can be used in combination to diagnose and monitor NASH patients with advanced fibrosis.

Figure 1 illustrates an example of how different NITs can be used in combination to diagnose and monitor patients with NASH with advanced fibrosis in a personalized fashion based on the clinical presentation of and resources available to the patient.

In summary, the results of several studies of NASH have documented the effectiveness of various NITs in assessing disease severity, monitoring longitudinal disease changes, and correlating results with liver-related events. The use of NITs in phase 3 trials as primary end points is being evaluated. With increasing use of and research on NITs, the tests have the potential for reducing the use of liver biopsy to diagnose NAFLD in clinical settings with limited resources (Tables 1 and 2).

TABLE 1.

Summary of cutoff thresholds of various NITs

NIT	AUROC	Cutoff threshold	Description
Serum-based biomarkers
NFS	0.85	0.676	Includes advanced fibrosis with PPV 90%
FIB-4 index	0.86	2.67	Includes advanced fibrosis with PPV 70%
ELF	N/A	9.80	Includes significant fibrosis with specificity of 0.86
PRO-C3	0.81	20.9	Includes advanced fibrosis OR 1.84 95% CI (1.05–3.23)
ADAPT	0.86	6.3287	Includes advanced fibrosis
FIBC3	0.89	−0.4	Includes advanced fibrosis
ABC3D	0.88	3	Includes advanced fibrosis
NIS-4	0.80	0.63	Includes patients with NAS ≥ 4 and ≥ F2
Imaging-based biomarkers
LSM by VCTE	0.77 0.80 0.89	8.2 kPa for F ≥ 2 9.7 kPa for F ≥ 3 13.6 kPa for F = 4	Diagnoses different stages of fibrosis
MRE	0.924	3.63 kPa	Includes advanced fibrosis with sensitivity of 0.86, specificity of 0.91, PPV of 68%, and NPV of 97%.
Ultrasound shear wave elastography	0.871 0.895 0.937	8.6 kPa for F ≥ 2 10.7 kPa for F ≥ 3 13.8 kPa for F = 4	Diagnoses different stages of fibrosis
MEFIB	0.90	MRE ≥ 3.3 kPa and FIB-4 index ≥ 1.6	Includes patients with significant fibrosis (F ≥ 2) with PPV 97%.
FAST	0.80	0.67	Includes patients with NAS ≥ 4 and ≥ F2 with PPV 83%.
MAST	0.93	0.242	Includes patients with NAS ≥ 4 and ≥ F2 with 0.90 specificity and 50% PPV.
CAP by VCTE	0.87 0.77 0.70	302 dB/m for S ≥ S1 331 dB/m for S ≥ S2 337 dB/m for S = S3	Diagnoses different stages of steatosis
MRI-PDFF	N/A	≥ 30% reduction	≥ 30% reduction in MRI-PDFF correlates with improvement in steatosis grade by 1 or more points with (OR 14.9, 95% CI 3.8–57.7).

Note: Advanced fibrosis includes fibrosis stages 3 and 4. Significant fibrosis includes fibrosis stage 2.

Abbreviations: ELF, enhanced Liver Fibrosis; FIB-4, Fibrosis-4 Index; LSM, liver stiffness measurement; MEFIB, MRE combined with FIB-4 index; MRE, magnetic resonance elastography; PPV, positive predictive value; VCTE, vibration-controlled transient elastography.

TABLE 2.

Summary of NITs whose per U change is associated with longitudinal outcomes

NIT	Description
NFS	A U change correlates with a change in the stage of fibrosis by 0.19 (95% CI 0.71–0.88).
FIB-4 index	FIB-4 index ≥ 3.25 correlates with the occurrence of adverse clinical liver events HR 1.46 (95% CI 1.35–1.57).
ELF	A unit increase was associated with developing liver-related outcomes [HR 2.53 (95% CI 1.45–4.43)] or progression to advanced fibrosis [HR 4.34, (95% CI 2.4–7.8)].
VCTE	Patients with LSM above threshold 21 kPa at baseline are at 14% higher risk of developing hepatic decompensation compared to those below the threshold.
MRE	The risk of developing liver-related events increased per 1 kPa increment with HR 1.32 (95% CI 1.13–1.56).
cT1	An increase in cT1 by 88ms was associated with a 2 U increase in NAS and a 21% relative change in MRI-PDFF.
MEFIB	Positive MEFIB score (defined as MRE ≥ 3.3 kPa and FIB-4 index ≥ 1.6) is associated with liver-related events with HR 20.6 (95% CI: 10.4–40.8, p<0.001) while negative MEFIB score has high NPV at 99.1%.
MAST	A logit unit increase in MAST is associated with adverse clinical liver events at HR 2.01 (95% CI: 1.59–2.54, p<0.0001).

Abbreviations: ELF, enhanced Liver Fibrosis; FIB-4, Fibrosis-4 Index; LSM, liver stiffness measurement; MEFIB, MRE combined with FIB-4 index; MRE, magnetic resonance elastography; NAS, NAFLD Activity Score; PPV, positive predictive value; VCTE, vibration-controlled transient elastography.