Racial, Ethnic and Age Disparities in Liver Fibrosis Screening Using Fibrosis Score Measures: A Critical Review of Diagnostic Equity in Liver Disease

Abstract

Metabolic-associated steatotic liver disease is currently one of the most common causes of liver disease in the world, affecting a large portion of the global population; these patients are at risk of developing advanced liver fibrosis and cirrhosis. Noninvasive tests (NITs), including lab tests such as FIB-4, NAFLD Fibrosis Score and the Aspartate Aminotransferase-to-Platelet Ratio Index, are widely used for fibrosis risk stratification, but their accuracy across various racial, ethnic, and age groups remains poorly characterized. This review examines disparities in NIT performance across these populations and the need for tailored screening strategies. A comprehensive, narrative literature review highlighted significant variability in NIT performance, with studies in African American, Hispanic and Asian patients all revealing mixed results when the performance of NITs was used to assess fibrosis levels. Additionally, the age of patients may influence fibrosis testing, as older adults tend to have higher false-positive rates due to age-based biases. Although imaging modalities like VCTE and MRE may offer superior accuracy in the noninvasive assessment of hepatic fibrosis, they face accessibility limitations and have rarely been validated in specific racial groups. This review concludes that current NITs for MASLD risk stratification needs to be recalibrated with population-specific and age-adjusted thresholds, and future research should focus on inclusive validation studies and integrating clinical judgment to improve screening accuracy.

1. Introduction

Metabolic-associated steatotic liver disease (MASLD), formerly known as non-alcoholic fatty liver disease (NAFLD), refers to the presence of excess fat in the liver associated with metabolic syndrome. MASLD is diagnosed when liver steatosis, confirmed via biopsy or noninvasive tools, is present in addition to one or more metabolic risk factors, such as a BMI ≥ 25 kg/m², insulin resistance, hypertension, hypertriglyceridemia, and low HDL [1]. MASLD imposes a growing global health burden, affecting approximately one-third of adults worldwide, with escalating morbidity and mortality [2]. A recent Global Burden of Disease database analysis showed that in 2021, there were 1.26 billion cases of MASLD globally, with 138 thousand deaths and 3.67 million disability-adjusted life years (DALYs) [3].

The overwhelming global burden of MASLD necessitates early detection and treatment of hepatic steatosis to prevent its progression to metabolic dysfunction-associated steatohepatitis (MASH), which is marked by liver inflammation. As this disease process can lead to the development of fibrosis, cirrhosis, portal hypertension and hepatocellular carcinoma, risk stratification is essential. Early and accurate assessment of fibrosis is beneficial to allow clinicians the ability to target specific populations who are at a higher risk of developing more advanced MASLD. Racial and ethnic differences in the prevalence of MASLD exist. For example, in the US, the prevalence of steatosis is highest in Hispanic individuals, especially those of Mexican descent, and lowest among African American (AA) individuals [4,5]. A deeper understanding of these differences will allow targeted screening and enable risk stratification and structured therapy to minimize the risk of disease progression.

Hepatic fibrosis can be assessed via several different modalities, including liver biopsy; however, due to sampling error, risk of complications, as well as inter- and intra-observer differences in assessing fibrosis, noninvasive tests (NITs), including blood-based scoring systems and imaging-based modalities, are becoming the preferred option. Some of the blood-based scoring systems widely used in clinical practice include the Fibrosis-4 Index (FIB-4) score, the NAFLD fibrosis score (NFS), and the Aspartate Aminotransferase-to-Platelet Ratio Index (APRI) score. Imaging modalities include transient elastography, ultrasound with elastography, and magnetic resonance elastography (MRE). Noninvasive lab tests such as FIB-4 are recommended by multiple societies as initial tests due to their simplicity and because the laboratory components required to perform the test are readily available (i.e., age, AST, ALT and platelet count) [6]. However, lab-based NITs have limitations and can miss early fibrosis [7] and thresholds for fibrosis assessment may differ based on patient age. Transient elastography, US with elastography and MRE add value as NITs and have been found to be superior to serum-based tests with regard to sensitivity and specificity [8,9,10].

Emerging evidence suggests that the performance and interpretation of both lab and imaging NITs may vary significantly across different racial and ethnic groups, resulting in inequities in diagnosis, referral, and access to treatment. For example, the FIB-4 score under-predicts advanced liver fibrosis in AA adults [11]. The NFS, on the other hand, incorporates metabolic syndrome markers and is therefore often elevated in African American patients, who usually have a higher burden of metabolic risk factors [12]. Similarly, studies have shown that noninvasive scores for advanced MASLD fibrosis have moderate discriminatory ability in Hispanic patients [13]. Additionally, scores like FIB-4 and NFS incorporate age, and evidence suggests that patients over 60 may not be accurately screened using these scores, and the thresholds for minimal fibrosis may be different in different age groups. As an example, a FIB-4 of less than 1.3 excludes advanced fibrosis in patients under 60; however, the cut-off is <2 for patients over the age of 60 [14].

Our review will examine racial and age disparities in the screening and risk stratification of MASLD, focusing primarily on noninvasive scoring systems (such as FIB-4, NFS, and APRI scores). This review aims to inform more equitable screening practices by critically evaluating these disparities.

2. Noninvasive Tools for MASLD Risk Assessment

Given the limitations and risks associated with liver biopsy, the presumed “gold” standard for histologic evaluation, a concerted effort toward the validation and implementation of noninvasive diagnostic modalities has emerged. These modalities—spanning serum biomarkers, elastography technologies, composite scoring systems, and imaging-based quantification—constitute an evolving set of tools in the stratification of MASLD risk.

2.1. Blood-Based Scoring Systems

Over the past few decades, several noninvasive lab tests have been developed that utilize routine laboratory results, as well as specialized tests for predicting the risk of MASLD. These scores typically help predict liver fibrosis, a well-known stage in MASLD progression. They are now a key component of diagnostic algorithms and are endorsed by experts and major Endocrine and Gastroenterology/Hepatology societies. Their cost-effectiveness and ease of use have led to these scores becoming common for initial assessments.

Some of the more commonly used scores in clinical practice are listed below.

• FIB-4 (Fibrosis-4 Index):

The FIB-4 index is the most commonly used noninvasive, blood-based scoring system used for risk assessment of advanced fibrosis in MASLD. It was developed in 2007 and has since been endorsed by leading expert societies, including the AASLD and AGA [15]. The formula can be expressed as follows:

[Age × AST]/[Platelet count × √ALT]

Clinical thresholds have been defined as follows:

• <1.3: low risk for advanced fibrosis (≥F3).
• 1.3–2.67: indeterminate risk.
• >2.67: high risk for advanced fibrosis.

The FIB-4 score has a high negative predictive value (NPV) for establishing the presence or absence of advanced fibrosis. Standards are highly accurate (sensitivity ~66–71%, specificity ~77–86%), while intermediate scores may require additional lab, biopsy, or imaging studies to accurately assess fibrosis [16,17,18,19,20]. In adults aged ≥ 65, a higher low-risk cutoff (1.9–2.0) is suggested due to reduced specificity [21]. While FIB-4 performs reasonably well, it has limitations, including reduced accuracy in patients under 35 or over 65 and lower sensitivity for moderate fibrosis, necessitating additional testing.

• NAFLD Fibrosis Score (NFS):

NFS considers age, BMI, presence of diabetes or impaired blood glucose levels, AST/ALT ratio, platelet count, and albumin. It is interpreted as <−1.455 (low risk), −1.455 to 0.676 (indeterminate), >0.676 (high risk) [22]. Its formula is

NFS = −1.675 + (0.037 × age) + (0.094 × BMI) + (1.13 × diabetes) + (0.99 × AST/ALT) − (0.013 × platelet count) − (0.66 × albumin)

Comparative studies and meta-analyses show that NFS and FIB-4 have similar overall diagnostic performances for assessing advanced fibrosis in MASLD, with area under the receiver operator curve (AUROC) values typically in the 0.72–0.82 range. However, NFS is more complicated to calculate compared to FIB-4 and may be limited in its global applicability for assessing fibrosis as a first-line tool [23].

• APRI (AST to Platelet Ratio Index):

APRI is a simple score utilizing only two laboratory values, calculated using the following formula (ULN − Upper Limit of Normal of AST for that laboratory)

APRI = (AST/ULN)/Platelet count (10⁹/L) × 100

APRI is less accurate than FIB-4 and NFS for detecting advanced fibrosis in MASLD and is not recommended as a first-line tool, though it remains useful in resource-limited or primary care settings due to its simplicity. While common cutoffs range from 0.5 (to rule out) to 1.0 or 1.34 (to rule in) [21,24], recent studies suggest variable optimal thresholds across populations, emphasizing the need for further standardization [25,26,27].

• BARD score:

The BARD score is a simple tool calculated by assigning points for three variables: Body Mass Index (BMI ≥ 28 kg/m², 1 point), AST/ALT ratio (≥0.8, 2 points), and the presence of diabetes mellitus (1 point), for a total score ranging from 0 to 4.

In MASLD, the BARD score is used to help exclude advanced fibrosis, but its diagnostic performance is limited. Studies consistently show that the BARD score has lower accuracy for detecting advanced fibrosis compared to other noninvasive scores. In a cohort of patients with MASLD, the AUROC for BARD was 0.609, which is significantly lower than FIB-4 (0.736) and NFS (0.724). Similarly, a large European study found that FIB-4 and NFS both outperformed BARD for both cross-sectional identification of advanced fibrosis and prediction of long-term outcomes [20,28].

• Other Noninvasive lab tests:

The Enhanced Liver Fibrosis (ELF) test measures direct markers of fibrosis such as Hyaluronic Acid (HA), Tissue Inhibitor of Metalloproteinase-1 (TIMP-1), and N-terminal Propeptide of type III Procollagen (PIIINP). It has demonstrated superior diagnostic accuracy (AUROC 0.80–0.90) compared to simpler scores like FIB-4 and NFS, though its use is limited by cost, issues with insurance coverage, and proprietary access [29,30]. The Hepamet Fibrosis Score (HFS) has been shown to be particularly useful in assessing hepatic fibrosis in patients with obesity, diabetes, or older age and demonstrates higher accuracy and fewer indeterminate results than FIB-4 or NFS, though it is less commonly used [24,31].

• Emerging Technologies—MicroRNAs and NIS4^®/NIS2+™:

MicroRNAs (miRNAs) are smaller RNA molecules that can interact with messenger RNA (mRNA) and regulate gene expression. A study by Jampoka et al. observed the ability of serum miR-29a and miR-122 to serve as biomarkers for NAFLD. They established that miR-29a levels were considerably decreased in NAFLD patients, while miR-122 levels were elevated and related to disease severity [32]. Another work by Abdel-Al et al. focused on using miRNAs for identifying liver fibrosis in patients with chronic hepatitis C. They discovered that miR-221 and miR-222 exhibited high sensitivity and specificity in recognizing both early and advanced stages of fibrosis [33]. These results point to the potential of miRNAs, especially miR-221 and miR-222, as encouraging biomarkers for monitoring liver disease advancement.

The NIS4^® (Genfit-Loos, Loos, France) technology is a blood-based diagnostic test designed to classify patients at risk for metabolic dysfunction-associated steatohepatitis (MASH), using four biomarkers: miR-34a-5p, alpha-2 macroglobulin, YKL-40, and glycated hemoglobin. It has displayed superior performance to other NITs such as FIB-4, NAFLD fibrosis score, and ELF in detecting at-risk MASH patients. Recently, an improved version called NIS2+™ was developed, combining miR-34a-5p and YKL-40, with adjustments for sex differences in miR-34a-5p. NIS2+™ has demonstrated strong clinical results across various patient subgroups and could be particularly useful in clinical trials to accurately identify at-risk MASH patients and minimize screen failure rates [34,35].

While miRNA-based assays are both highly sensitive and specific for catching early- and late-stage fibrosis, they remain largely experimental and less practical for routine clinical use. These tests often require specialized laboratory equipment, stringent sample handling, and complex data analysis, which can limit scalability and rapid adoption in typical clinical settings [36]. In contrast, NIS4^® and NIS2+™ are further advanced in their clinical validation and are designed for integration into standard blood draws, making them more feasible for implementation in primary care and hepatology clinics. Additionally, these assays offer standardized protocols and quicker turnaround times, reducing variability and logistical barriers compared with individualized miRNA panels. Cost-effectiveness is also more favorable for NIS4^® and NIS2+™ in the near term, as they may decrease the need for invasive biopsies and repeated testing [37]. Together, these differences highlight that while miRNA testing has strong research potential, NIS4^® and NIS2+™ are currently better positioned for practical use in patient management and clinical trials.

2.2. Imaging-Based Modalities

As imaging techniques and accuracy have improved, there is increasing reliance on noninvasive imaging modalities, due to the increased ability to reliably detect fibrosis in patients with MASLD. Some of the commonly used modalities include the following:

• Ultrasound and Shear Wave Elastography (SWE):

Shear wave elastography (SWE) is an ultrasound-based technique that quantifies liver stiffness as a surrogate for fibrosis. It is noninvasive, rapid, and can be performed during a standard ultrasound exam. In MASLD, SWE demonstrates good accuracy for advanced fibrosis (AUROC~0.85–0.89), comparable to VCTE and superior to simple serum scores. Limitations include reduced performance in obesity, operator dependency, and lack of cross-platform standardization [38,39].

• Transient Elastography (FibroScan^®):

Transient elastography, also known as FibroScan^® (Echosens, Paris, France), and VCTE are widely used for noninvasive fibrosis and steatosis assessment, providing liver stiffness measurement (LSM) and a controlled attenuation parameter (CAP). An LSM < 8 kPa rules out advanced fibrosis, while >12 kPa suggests a high likelihood of advanced disease. Its diagnostic accuracy is comparable to SWE but lower than MRE (AUROC 0.92–0.95). Its strengths include accessibility and utility in sequential algorithms with FIB-4 [8,9,10]. VCTE has limitations, including reduced accuracy in obese patients, operator dependency, and difficulty assessing early-stage fibrosis. It may also be influenced by liver inflammation, ascites, or anatomical issues, making it less reliable in these populations [40].

• Quantitative Multiparametric Magnetic Resonance Elastography (MRE):

MRE provides the highest diagnostic accuracy for fibrosis (AUROC ~0.95–0.96 for advanced fibrosis), is less affected by obesity, and is considered the reference standard among noninvasive imaging. However, its limited availability and higher cost restrict its routine use [24].

• Computed Tomography (CT):

Computed tomography (CT) has a limited role in the assessment of liver fibrosis and MASLD. CT can detect moderate-to-severe hepatic steatosis by measuring liver attenuation in Hounsfield units, with greater steatosis resulting in lower attenuation. However, CT is less sensitive than ultrasound when it comes to assessing mild steatosis, exposes patients to ionizing radiation, and is not recommended for routine screening for hepatic steatosis or fibrosis staging due to these limitations and cost considerations [24].

2.3. Combined Modalities

• FAST (FibroScan-AST) Score:

This score combines liver stiffness measurement (LSM), controlled attenuation parameter (CAP), and AST. FAST outperforms FIB-4 and APRI in identifying high-risk steatohepatitis with significant fibrosis (AUROC~0.82) but requires access to transient elastography. A meta-analysis found that a FAST score ≥ 0.67 yielded a pooled specificity of 0.87 (0.82–0.90) while a FAST score ≤ 0.35 yielded a summary sensitivity of 0.88 (0.83–0.91) [41,42]. The score has a low PPV, and therefore, sequential testing with other methods becomes indicated.

• Agile 3+, Agile 4 score:

Agile 3+ and Agile 4 are noninvasive scoring systems that were developed in 2023 to identify advanced fibrosis (≥F3) and cirrhosis (F4), respectively, in patients with NAFLD/MASLD [43]. Both scores use logistic regression models incorporating LSM by VCTE, AST/ALT ratio, platelet count, sex, and diabetes status; Agile 3+ also includes age. The Agile 3+ formula is as follows:

Agile 3+ = logistic regression of (LSM, AST/ALT, platelets, sex, diabetes, age)

Agile 4 is similar but omits age. Agile 3+ demonstrates an AUROC of 0.85–0.91 for advanced fibrosis, with sensitivity and specificity at rule-out and rule-in cutoffs of approximately 88% and 87%, respectively. Agile 4 achieves an AUC of 0.93–0.97 for cirrhosis, with sensitivity and specificity at rule-out and rule-in cutoffs of 90% and 93%, respectively. Both scores outperform FIB-4 and LSM alone and reduce indeterminate results [44,45,46].

The most commonly used NITs are outlined in

Table 1. Comparison of commonly used NITs.

Test	Components	Strengths	Limitations	Performance
FIB-4	Age, AST, ALT, Platelet count	High NPV, high accuracy when using standard cutoffs	Reduced accuracy in patients under 35 or over 65, lower sensitivity for moderate fibrosis	Using standard cutoffs (<1.3 low risk, >2.67 high risk), sensitivity ~66–71%, specificity ~77–86%
NAFLD Fibrosis Score (NFS)	Age, BMI, Diabetes, AST/ALT ratio, Platelet count, Albumin	High NPV, incorporates metabolic risk factors	Complex formula, low specificity in patients over 65	AUROC values typically range from 0.72 to 0.82
APRI	AST, Platelet count	Useful in resource-limited/primary care settings due to its simplicity	Less accurate than FIB-4 and NFS, not recommended as a first-line tool	AUROC value 0.80 for predicting significant fibrosis
BARD Score	BMI ≥ 28, AST/ALT ratio ≥ 0.8, Diabetes	Helpful in excluding advanced fibrosis	Low accuracy for detecting advanced fibrosis when compared to other noninvasive scores.	AUROC value of 0.609 for detecting advanced fibrosis
VCTE	Liver stiffness determined via ultrasound	Superior diagnostic accuracy when compared to other ultrasound and blood-based scores, CAP adds measurement of steatosis	Less accurate than MRE, accuracy of assessment can be affected by obesity, presence of ascites or inflammation and operator expertise	AUROC value of 0.77–0.89 depending on degree of fibrosis.
Shear Wave Elastography (SWE)	Liver stiffness determined via ultrasound	Highly accessible, rapid results, can be performed during standard ultrasound exam, able to detect steatosis	Limited accuracy of assessment of fibrosis, can be affected by obesity and operator expertise, less standardization across radiologic platforms when compared to VCTE or MRE	AUROC for advanced fibrosis between 0.85 and 0.89
Magnetic Resonance Elastography (MRE)	Liver stiffness determined via MRI	Highest diagnostic accuracy among imaging modalities	Limited availability, high cost	AUROC 0.95–0.96 for advanced fibrosis
CT	Liver attenuation measured in Hounsfield units	Limited role in the assessment of liver fibrosis and MASLD but can be used in combination with laboratory assessments for tumor markers to screen for HCC	Less sensitive than ultrasound for assessing steatosis and exposes patients to ionizing radiation, high cost	Role is limited in assessment of fibrosis and is not recommended for screening
FAST	LSM (VCTE) + CAP + AST	Outperforms FIB-4 and APRI for identifying high-risk steatohepatitis with significant fibrosis	Low PPV, requires access to transient elastography	AUROC value of ~0.82 for identifying high-risk steatohepatitis with significant fibrosis
AGILE ¾	LSM (VCTE) + AST/ALT ratio, platelet count, sex, diabetes, age	Outperforms FIB-4 and LSM alone, reduces indeterminate results	Developed in Caucasian cohorts and not yet widely validated in diverse populations	AUROC value for AGILE 3 predicting advanced fibrosis 0.85–0.91 and for AGILE predicting cirrhosis 0.93–0.97

.

3. Disparities in Performance of Noninvasive Tests by Racial/Ethnic Group

Noninvasive scoring systems for screening and risk stratification of MASLD are commonly used in clinical practice. However, many of these diagnostic tools were developed primarily in non-Hispanic White populations. For instance, a study by Angulo et al. described that 90% of the patients included in the verification of the NFS were Caucasian [22]. The FIB-4 index was initially developed for patients with HIV and hepatitis C coinfection, with a majority of Caucasian participants in both the training and validation sets. When used to predict advanced fibrosis in a large NAFLD cohort, most patients were non-Hispanic White, with no information provided on the distribution of other racial groups [47].

There have been limited studies over the last 25 years describing the efficacy of NITs across racial and ethnic groups. The vast majority of this research has focused on AA, Hispanic, and Asian patient populations, and evaluating the accuracy of NITS in these populations. In the following narrative review, we will outline the major disparities in the accuracy of the NITs among various racial and ethnic groups.

3.1. Methods

Relevant literature was identified through systematic searches of PubMed and Google Scholar using keywords and Boolean operators including “metabolic dysfunction–associated steatotic liver disease,” “MASLD,” “noninvasive fibrosis testing,” “FIB-4,” “NAFLD fibrosis score,” “APRI,” “transient elastography,” “age,” “race,” and “ethnicity.” Searches were limited to articles published between 2000 and 2025 and written in English. Studies were included if they evaluated adult patients and reported outcomes stratified by race or ethnicity. Editorials, non-human studies, and articles lacking original data were excluded. Reference lists of all selected articles were manually reviewed to identify further relevant studies. Search results were screened in two stages. Initial screening involved review of titles and abstracts to assess relevance, followed by full-text review for eligibility. Data were extracted using a standardized template capturing study design, population characteristics, NIT evaluation, racial/ethnic composition, diagnostic performance metrics (sensitivity, specificity, AUROC), and key findings. Findings were summarized using a narrative synthesis approach, highlighting differences in test performance across racial and ethnic groups, age-related considerations, and emerging diagnostic technologies. Meta-analysis was not performed due to heterogeneity in study designs, populations, and outcome measures.

3.2. Hispanic/Latino Populations

Among Hispanic and Latino populations, the prevalence of MASLD is higher when compared to non-Hispanic populations due to a combination of hereditary, environmental, financial, and societal factors [48]. However, Kallwitz et al. identified that NAFLD prevalence also varies significantly when comparing individuals of different Hispanic/Latino heritages [49]. In their analysis, individuals of Cuban, Puerto Rican, and Dominican descent were found to have lower rates of suspected NAFLD compared to those of Mexican heritage. Meanwhile, individuals of Central American and South American heritage had a similar prevalence of suspected NAFLD compared to Mexicans. This suggests the standardized scoring systems for screening NAFLD/MASLD may not equally predict fibrosis risk across different ethnic populations.

In a critical 2023 study by Bril and Gray, the authors aimed to evaluate the effectiveness of NITs in different racial and ethnic groups by comparing the serum-based scores to VCTE fibrosis measures. Data for the analysis were obtained from the National Health and Nutrition Examination Survey (NHANES) conducted between 2017 and 2020. In Hispanic populations, noninvasive scoring systems, such as FIB-4, NFS, and APRI, had variable performance in detecting fibrosis and cirrhosis with significant inconsistencies depending on fibrosis stage and age group [47]. Tincopa et al. found that FIB-4 and NFS had limited utility in accurately identifying advanced fibrosis in Hispanic populations [50]. Balakrishnan et al. found that NITs (NFS, BARD, FIB-4, and APRI) showed moderate ability to detect advanced fibrosis in Hispanic patients in Texas, with NFS performing slightly better. While the NFS had the highest sensitivity and NPV overall, the overall ability of these NITs to rule out advanced fibrosis was lower in this predominantly Hispanic population compared to previous findings in Caucasian groups [13]. Conversely, Xu et al. found no statistically significant difference in the performance of standardized scoring systems, including FIB-4, NFS, APRI, and BARD, between Latino and White populations [51]. Singh et al. noticed that lowering the threshold for FIB-4 improved diagnostic accuracy in Hispanic patients with diabetes, suggesting that standard cutoffs may not be universally appropriate [48]. In regard to imaging modalities, such as MRE and VCTE, Tincopa et al. note that they had reduced accuracy in Hispanic patients when compared to non-Hispanic patients [50].

These studies clearly indicate that the performance and accuracy of noninvasive tools for assessing fibrosis in MASLD is variable among Hispanic populations. Careful interpretation is required when using these modalities, as they were developed and standardized in mostly White populations. This reinforces the need for race- and ethnicity-specific validation and potentially revised thresholds to ensure accurate risk stratification and equitable liver disease care.

3.3. Black/African American Populations

Among AAs and non-Hispanic Black populations, the accuracy of NITs for liver fibrosis is problematic. It has been observed that in AAs, the prevalence of MASLD and advanced liver fibrosis is significantly decreased despite increased rates of obesity and diabetes in this population [11,52,53]. These observations may contribute to lower sensitivity and PPV of NITs in these populations, which could lead to under-diagnosis.

The NFS has been shown to perform poorly in the detection of advanced liver fibrosis in non-Hispanic Black patients in multiple studies [11,12,47]. Studies assessing the accuracy of FIB-4, APRI and AST/ALT ratio in AA patient populations have demonstrated varied results. FIB-4 and APRI overall have been shown to have moderate utility in AAs; however, their performance is still poor when compared to White and Hispanic populations [11,12,47]. The AST/ALT ratio performed poorly in a cohort of AA patients compared to White patients [12]. However, Akpoigbe et al. found that the AST/ALT ratio had higher sensitivity in AA patients compared to Hispanic patients [54]. Wang et al. found that FIB-4 was less accurate in non-Hispanic Black patients compared to other ethnic groups, and suggested that this population has lower AST and ALT levels in the setting of advanced fibrosis compared to other races/ethnicities, which may affect the results of standardized lab NITs [52]. Conversely, all four tests, NFS, FIB-4, APRI, and AST/ALT ratio, have a high NPV, making them reliable for ruling out advanced fibrosis [11]. Overall, these findings indicate that race-specific calibration or alternative tools to improve diagnostic accuracy to assess hepatic fibrosis in AA patients are required.

3.4. Asian Populations

Asian populations are at a higher risk for the progression of liver fibrosis compared to non-Asian populations, highlighting the importance of evaluating the performance of current NITs in this demographic [55]. Furthermore, studies have shown that Asian populations tend to develop diabetes, an important risk factor in the development of MASLD, at younger ages and at lower BMIs than non-Hispanic White populations [56,57]. These differences suggest potential variability in the disease process between Asian and non-Asian groups and may necessitate the development of population-specific screening strategies.

In the work by Bril and Gray, among NITs used to detect advanced liver fibrosis, non-Hispanic Asian patients exhibited the highest diagnostic performance, with APRI achieving an AUROC of 0.85 and FIB-4 an AUROC of 0.77. It was also noted that FIB-4, APRI, and NFS demonstrated comparable performance within the non-Hispanic Asian population [47]. Similarly, Wong et al. reported no significant performance differences in NFS, FIB-4, the ELF test and VCTE when comparing Asian, and non-Hispanic White patients, suggesting that uniform cutoff values for these tests may be appropriate across these populations [58].

However, it has been documented that regional variation may potentially exist among Asian populations. In a Chinese cohort of obese patients, Chen et al. found that NFS and FIB-4 lacked predictive value for MASLD diagnosis, while SWE, APRI, and modified APRI (m-APRI) demonstrated greater diagnostic utility. Among these NITs, SWE and APRI were superior to m-APRI, with no significant difference observed between SWE and APRI in predictive performance [59]. This contrasts with the study conducted by Sumida et al., which found that in Japanese individuals, the FIB-4 score was the most effective at ruling out advanced fibrosis compared to other NITs, including NFS, APRI, and AST/ALT ratio [60]. An additional consideration worth noting is a study by Cheung et al., which included Chinese, Malaysian, and Indian individuals. An assessment of a modified NIT incorporating INR, age, platelet count, GGT, presence of diabetes, AST, and BMI outperformed FIB-4, NFS, and APRI in predicting advanced fibrosis [61].

3.5. Other and Mixed Populations

Despite the growing burden of MASLD globally, the assessment of Native American, Middle Eastern, and mixed-ethnicity populations remains poorly characterized. These groups are underrepresented in large cohort studies used to validate NITs, and few studies have stratified outcomes in these populations. Due to the absence of robust, population-specific studies, it remains unclear how well current scoring systems such as FIB-4, NFS, APRI, as well as imaging modalities perform in these groups. One particular study by van Dijk et al. investigated a multi-ethnic population in the Netherlands and found that the correlation between liver stiffness measurements obtained via TE versus FIB-4 and APRI was weak or absent across several ethnic groups, including Ghanaian, African-Surinamese, and Turkish participants. For example, the study found that while Ghanaian individuals had the lowest liver stiffness measurements assessed via VCTE, they also exhibited some of the highest FIB-4 and APRI scores, perhaps indicating a lack of correlation between lab and radiologic NITs in this patient population. Similarly, African-Surinamese participants had elevated FIB-4 and APRI values despite the absence of high liver stiffness assessed with VCTE, raising concerns about the reliability of these tests in these populations [62].

Another study found that in a cohort of Emirati patients with type 2 diabetes, the FIB-4 and APRI indices showed comparable performance in detecting severe fibrosis. However, FIB-4 was considerably more sensitive, identifying 88% of cases, while APRI captured only 12%. In contrast, APRI demonstrated greater specificity at 96% compared with 72% for the FIB-4 index [63].

Individuals of mixed ancestry may exhibit unique combinations of risk factors such as varying genetic polymorphisms, differences in body composition, metabolic traits, or environmental exposures, which may not allow accurate assessment of hepatic fibrosis with standard lab and radiologic NITs. Future research should prioritize inclusion of these populations to better characterize disease risk, evaluate the accuracy of existing tools, and inform the development of tailored strategies for effective screening and risk stratification.

Results analyzing the racial and ethnic differences are summarized in

Table 2. Comparison of the performance of noninvasive tests (NITs) across racial and ethnic groups.

	Fib-4	NFS	APRI	Imaging (VCTE)	Key Observations	Recommended Adjustments
Caucasian	Well-established performance, moderate sensitivity; moderate specificity	Generally effective with high accuracy	Moderate utility; less accurate than FIB-4 and NFS	Widely used with strong diagnostic capabilities; gold standard in clinical use	Well-established performance but limitations exist in age-based scores	Current thresholds appropriate for most; ensure use of age-adjusted NFS and FIB-4 cutoffs in older adults
Hispanic	Moderate ability; high false-negative rates in advanced fibrosis	Variable, especially in age and fibrosis stage; lower accuracy at advanced stages	Moderate ability; less accurate compared to FIB-4 and NFS	Reduced accuracy in Hispanic patients	Inconsistencies in accuracy due to metabolic and genetic factors; NITs require tailored thresholds for Hispanics	Consider heritage-specific validation (e.g., Mexican vs. Caribbean ancestry); incorporate metabolic risk factors (diabetes; obesity) into stratification.
African American	Lower accuracy in detecting advanced fibrosis; high false negatives	Elevated due to higher metabolic risk factors; performs better in early stages	Moderate utility	Continued investigation is necessary to clarify these findings.	Underdiagnosis due to lower sensitivity; requires race-specific calibration for better diagnostic accuracy	Adjust thresholds downward given lower LFTs; use NITs primarily for rule-out (high NPV); explore novel biomarkers beyond aminotransferase-based scoring
Asian	High sensitivity and specificity in advanced fibrosis; especially in non-Hispanic Asians	Moderate accuracy; useful but with regional variations (e.g., in Chinese patients)	Performance varies by region	Continued investigation is necessary to clarify these findings.	Higher risk for fibrosis but NITs perform well; however, regional differences should be considered	Apply region-specific approaches: FIB-4 reliable in Japanese; APRI/SWE superior in Chinese cohorts; consider earlier screening thresholds due to younger age/lower BMI diabetes onset; use modified scores incorporating metabolic markers (e.g., INR, GGT, diabetes status)

.

4. Disparities in Performance of Noninvasive Tests by Age Group

Age is an important factor to consider when assessing fibrosis risk using serum-based scores, as it can significantly influence diagnostic accuracy. The FIB-4 score and NFS both use age to calculate fibrosis risk. A 2017 study by McPherson et al. analyzed how age can be a confounding factor when assessing fibrosis using serum-based scores. A key finding of this study was that while the FIB-4 and NFSs showed similar AUROCs across all age groups over 35 years, their specificity for identifying advanced fibrosis decreased significantly in older patients. Specifically, the specificity dropped to an unacceptable 35% for FIB-4 and 20% for NFS in patients aged 65 and older. Essentially, factoring in age with these tests causes an artificial increase in the score for patients over 65, leading to more individuals being classified as intermediate-risk and contributing to a higher rate of false positives. The authors proposed alternate methods to circumvent this bias primarily by increasing the threshold for the FIB-4 score and NFS in patients aged ≥ 65 years, which decreased the number of patients with an indeterminate score. The new age-specific thresholds, FIB-4 > 2.0 and NFS > 0.12, improved specificity to 70% without adversely affecting sensitivity. Another key finding was that the AST/ALT ratio, NFS, and FIB-4 scores performed poorly in patients under the age of 35. The AUROCs for advanced fibrosis were not statistically significant, suggesting that these tests are unreliable in this age group. Only 11% had stage F3 fibrosis, and none were cirrhotic, which may explain the poor performance of these scores [14]. Moreover, a paper published by Ishiba et al. validated new age-specific FIB-4 cutoffs in 1050 biopsy-confirmed NAFLD patients across 14 hepatology centers in Japan. They found that optimal thresholds increased with age, with proposed cutoffs of 1.95 and 2.67 for patients ≥ 70 years, which improved diagnostic accuracy compared with conventional cutoffs. Their findings supported that incorporating age-adjusted thresholds can reduce false positives in the elderly while preserving sensitivity for advanced fibrosis [64].

Additionally, a study published in 2020 by Pitisuttithum reported that while FIB-4 and NFS had lower specificity in elderly patients, applying the new age-specific cutoffs proposed by McPherson et al. improved diagnostic accuracy. However, the study also noted that these adjustments might reduce sensitivity, indicating a trade-off between specificity and sensitivity when modifying thresholds for age [65].

A 2024 paper by Sung et al. examined how age influences the FIB-4 score’s ability to predict liver fibrosis severity, using VCTE as the reference standard. The results show that while older adults (≥65 years) have higher FIB-4 scores, the commonly used threshold of 1.3 remains effective at predicting liver fibrosis, with age having a minimal effect. However, for higher thresholds (e.g., 2.7), age significantly alters the FIB-4 score’s predictive value, with older individuals showing lower specificity for advanced fibrosis. The findings suggest that additional testing may be required in adults > 65 years old if a lab and radiologic NITs provide discordant results [66].

Overall, adjusting age-specific thresholds improves the performance of these tests in older populations, but it may require a balance between specificity and sensitivity to avoid overdiagnosis.

5. Discussion

Noninvasive fibrosis scoring systems such as FIB-4, NFS and APRI are widely used for risk stratification for MASLD. However, these tools provide different diagnostic accuracy when comparing different races, ethnicities and ages, which may result in misdiagnosis and delays in referral practitioners with expertise in managing patients with more advanced hepatic fibrosis.

In AA patients, studies have shown that FIB-4 and APRI have lower diagnostic accuracy for advanced fibrosis compared to White cohorts [11]. Similarly, Hispanic populations are at substantial risk for under-recognition of advanced fibrosis with regard to common NITs. For example, a recent study performed in Hispanic patient populations found high false-negative rates using FIB-4 and NFS, and recommended the use of optimized cut-offs to better align with their higher burden and severity of disease [50]. Clinician awareness of test limitations and diagnostic gaps in underserved ethnic groups is needed to ensure timely referral and intervention.

Conversely, use of static NIT cut-offs may also overestimate disease severity in certain populations, leading to unnecessary referrals, imaging, and invasive testing. While considered a group at risk for underdiagnosis, Hispanic patients, who often have a high burden of metabolic comorbidities such as diabetes and obesity, often have elevated NIT values driven by these comorbidities rather than fibrosis. As a result, Hispanic patients may be disproportionately classified as high-risk. A recent study demonstrated that incorporation of diabetes status with updated thresholds for patients with and without diabetes can improve diagnostic performance of FIB-4 scores [48].

In regard to imaging-based modalities, the study by Tincopa et al. demonstrated that MRE and VCTE had reduced accuracy in Hispanic patients compared to non-Hispanic patients [50]. However, there is limited research elucidating how hepatic fibrosis in other races is stratified using noninvasive imaging techniques. Thus, more rigorous research focused on the accuracy of imaging modalities to assess fibrosis in different races and ethnicities is required. Further efforts should be made to examine the efficacy of these imaging tests in different groups.

Additionally, as more NITs are becoming available, there must be adequate screening for reliability in a diverse set of patients. For example, newer scoring systems such as the AGILE 4, AGILE 3+ and FAST score were all tested in a cohort of Caucasian patients [44]. As mentioned previously, other races and ethnicities have clear differing risks when it comes to MASLD and hence would benefit from further research elucidating the accuracy of both lab and radiologic NITs in various populations. Developing technologies using biomarkers like miRNAs, NIS4^®, and NIS2+™, must also eventually be investigated in these populations.

Moreover, the increased use of NITs is occurring proactively, as pharmacotherapies for MASLD with liver fibrosis are anticipated in the near future [47]. It remains unclear whether subspecialty care or biopsy will be necessary to access these treatments, but it is hoped that FDA approval of new therapies as well access to these therapies will be based on NITs rather than liver biopsy. If that becomes the case, it becomes even more crucial to provide fair and equal access to these medications based on equitable and accurate NITs.

The disparities in noninvasive fibrosis testing across racial, ethnic, and age groups carry significant clinical implications. Misclassification—whether overestimating or underestimating fibrosis risk—can directly affect patient care, leading to either unnecessary interventions or delayed treatment for those with advanced disease. Over-referral can result in unwarranted imaging, invasive procedures, increased healthcare costs, and heightened patient anxiety, particularly among older adults and those with comorbid conditions that elevate NIT values. Conversely, under-recognition of advanced fibrosis in African American, Hispanic, or younger patients may postpone timely referral to specialists, enrollment in clinical trials, or initiation of emerging MASLD therapies.

These clinical risks are not experienced in isolation; they are often magnified by social and structural barriers that further impede equitable diagnosis and management. Socioeconomic factors, insurance coverage, and geographic proximity to centers with advanced imaging modalities such as MRE and transient elastography intersect to limit equitable access. Patients from lower socioeconomic backgrounds or with public insurance are less likely to receive specialist referrals or undergo imaging-based fibrosis assessment, increasing the risk of delayed diagnosis and under-treatment [67]. Regional disparities in the availability of advanced imaging exacerbate these gaps, particularly in rural areas, leaving underserved populations reliant on serum-based NITs that may have lower accuracy in certain ethnic groups [68]. These findings underscore the need for population-specific validation of NITs, policy interventions to expand coverage of noninvasive fibrosis testing, and targeted programs to improve access to imaging for marginalized groups.

Beyond these systemic barriers, race/heritage and metabolic comorbidities warrant tailored adjustments to NIT interpretation to enhance diagnostic accuracy. Within Hispanic and Latino populations, heritage-specific validation is needed, as studies suggest that individuals of Mexican descent carry a higher MASLD burden compared to Caribbean subgroups. In these groups, lowering FIB-4 thresholds and applying diabetes-stratified cutoffs may reduce false negatives, particularly in patients with obesity or metabolic syndrome. In African American patients, where lower baseline aminotransferase levels may impair the performance of serum-based scores, NITs may serve best as rule-out tools, with confirmatory imaging or emerging biomarkers reserved for high-risk cases. Among Asian populations, regional variability should be considered—FIB-4 has shown particular reliability in Japanese cohorts, whereas APRI and shear wave elastography appear more accurate in Chinese populations. Furthermore, modified composite scores that incorporate metabolic markers have demonstrated promise in multi-ethnic Asian cohorts and may provide more equitable risk stratification.

Complementing these population-specific refinements, age is a major modifier of serum-based scores, necessitating age-adjusted thresholds in older adults. The American Association for the Study of Liver Diseases (AASLDs) [6] and the European Association for the Study of the Liver (EASL) [69], have both recommend using a corrected FIB-4 threshold of 2.0 over the age of 65, but these have not yet been adopted universally into practice [14]. Additionally, patients with non-liver-related thrombocytopenia or elevated AST/ALT due to other causes may be misclassified as being at elevated risk, triggering increasing concern about advanced liver disease. Despite the growing evidence, the effects on referral patterns, treatment eligibility, and healthcare utilization are not well defined. One potential refinement is differential weighting of age within scoring algorithms, allowing physiologic rather than chronological aging to be reflected in fibrosis risk. This may better distinguish true fibrosis from age-related changes in AST, ALT, or platelets, thereby reducing unnecessary referrals. Prospective studies should also test two-tiered approaches—combining recalibrated scores with imaging—and incorporate comorbidity indices such as diabetes, obesity, or cardiovascular disease to ensure high-risk older adults are appropriately prioritized. Such strategies could improve diagnostic accuracy while balancing resource allocation in geriatric populations.

Finally, large, prospective multicenter studies are needed to validate and refine NIT performance in real-world populations. Notably, international and multi-ethnic studies directly evaluating NIT performance in MASLD are lacking, representing a critical evidence gap that should be addressed through future large, prospective, multicenter investigations. Prospective studies incorporating liver biopsy confirmation, longitudinal follow-up, and novel biomarkers could help determine whether recalibrated scores lead to earlier and more precise identification of advanced fibrosis. Such research would clarify how these adjustments affect clinical decisions, including referrals to specialists and eligibility for emerging MASLD therapies. Additionally, evaluating the impact of demographic-specific thresholds on treatment outcomes and health equity would provide critical insights into optimizing MASLD management across diverse patient populations.

Table 3. Summary of key discussion points on NIT performance.

Topic	Observations
Noninvasive Tests (NITs)	NITs like FIB-4, NFS, and APRI are widely used for risk stratification, but have limitations in certain populations.
Racial and Ethnic Disparities	Significant differences in NIT accuracy across racial/ethnic groups; underdiagnosis in African Americans, overdiagnosis in Hispanics.
Age-Specific Considerations	Age impacts the accuracy of NITs; older adults (>65) have higher false-positive rates. Age-adjusted thresholds may improve accuracy.
Imaging Modalities (VCTE, MRE)	Imaging modalities like VCTE and MRE show better accuracy than serum-based tests, but still have limitations based on ethnicity and obesity.
Bias and Calibration	NITs may require race-specific calibration or updated cutoffs to improve diagnostic accuracy and reduce biases in underrepresented populations.
Social and Structural Barriers	Social factors like limited healthcare access and language barriers impact diagnosis and management, particularly in underserved populations.
Emerging Therapies and NITs	New therapies for MASLD require equitable NITs for access and treatment, highlighting the need for reliable, inclusive screening tools.

highlights the essential topics discussed in the review.

6. Conclusions

NITs such as FIB-4, NFS, APRI have become invaluable tools for risk stratification in patients with MASLD. However, growing evidence suggests that these tools do not have similar diagnostic accuracy across all racial, ethnic, and age groups. Standard NIT cutoffs often misclassify fibrosis risk, leading to a high false-negative rate in the African American patient population, as well as a high false-positive rate in the non-White Hispanic population. Age has also been found to be a confounding variable for risk stratification, as older adults frequently exhibit falsely elevated risk scores, while younger patients may go undetected due to low test sensitivity. Though imaging modalities like VCTE and MRE offer improved accuracy, disparities in performance persist, and validation across diverse populations remains limited.

These limitations underscore the urgent need for recalibrated, population-specific thresholds and a more individualized approach to MASLD risk assessment. Future research should prioritize inclusive cohort studies to refine existing models and ensure diagnostic equity. Clinicians should be aware of the variability in NIT performance and interpret results within the broader clinical context, considering patient demographics and comorbidities. As new MASLD therapies emerge, accurate and equitable screening will be essential for guiding treatment access and improving outcomes across all populations.