Noninvasive liver fibrosis assessment in metabolic dysfunction-associated steatotic liver disease

Abstract

Metabolic-dysfunction-associated steatotic liver disease (MASLD) is a leading cause of chronic liver disease worldwide, with global prevalence estimated at about 32%. The main predictor of clinical outcome is liver fibrosis severity, making an accurate diagnosis essential. This study evaluated the role and accuracy of noninvasive methods (NIMs) for assessing liver fibrosis in MASLD. A systematic literature search was conducted in the PubMed, Cochrane, SciELO, and LILACS databases using the keywords “metabolic steatotic liver disease”, “noninvasive markers”, “liver fibrosis”, and related terms. A total of 76 articles were selected for review. NIMs offer advantages over liver biopsy due to their simplicity, accessibility, and reproducibility, enabling effective risk stratification at lower costs. Despite some limitations in defining intermediate fibrosis stages, a stepwise approach enhances diagnostic accuracy. Although liver biopsy remains the gold standard, NIMs are increasingly recognized for reliably excluding advanced fibrosis in patients with MASLD, reducing the need for invasive procedures.

Key Words: Metabolic-dysfunction-associated steatotic liver disease; Nonalcoholic steatohepatitis; Nonalcoholic fatty liver disease; Steatotic liver disease; Noninvasive markers; Liver cirrhosis; Liver fibrosis

Core Tip: Metabolic-dysfunction-associated steatotic liver disease affects approximately 30% of the global population, with a rising prevalence. Although liver biopsy is the gold standard for assessing liver fibrosis and inflammation, it has inherent limitations, leading to the search for noninvasive methods with high reproducibility for evaluating and monitoring patients with metabolic-dysfunction-associated steatotic liver disease. This systematic review summarizes key data on the noninvasive methods for assessing liver fibrosis, highlighting their advantages over liver biopsy in terms of accessibility, cost-effectiveness, and patient safety.

INTRODUCTION

Metabolic-dysfunction-associated steatotic liver disease (MASLD) has emerged as a leading cause of chronic liver disease (CLD) worldwide. Its global prevalence, initially estimated at approximately 25%, has markedly increased over the past decade, now reaching 32% of the world population and up to 33% in Latin America when the new terminology is applied[1-3]. This upward trend is primarily attributed to the recent epidemic of obesity[4-7]. Currently, MASLD represents the commonest etiology of liver malignancy and liver transplantation among women in the United States, and it is projected to become the primary cause of advanced chronic liver disease (ACLD) worldwide within the next two decades[8]. In 1980, Ludwig et al[9] introduced the term “nonalcoholic steatohepatitis” to describe histological findings similar to those observed in the liver tissue of patients with alcoholic hepatitis but in individuals without a history or clinical evidence of significant alcohol consumption. In 2020, Eslam et al[10] proposed a change in nomenclature to metabolic dysfunction associated with fatty liver disease (MAFLD), aiming to provide a more inclusive and accurate term by removing the exclusive “nonalcoholic” descriptor as well as to emphasize the central role of metabolic stress in disease pathogenesis. Moreover, concerns regarding the stigmatizing nature of the term “fatty” and the requirement to exclude other potential causes (e.g., alcohol consumption) led to an expert consensus panel to review the existing terminology. Currently, the term “steatotic liver disease” (SLD) is the preferred nomenclature to encompass the various etiologies of hepatic steatosis, being endorsed by multiple societies[11]. Therefore, the term MASLD has substituted the previous MAFLD, proposed by Eslam et al[10]. Table 1 summarizes the differences and definitions in terminology used to describe SLD.

Table 1 Steatotic liver disease nomenclature.

Nomenclature	Definition
NAFLD	Nonalcoholic fatty liver disease is defined by the presence of ≥ 5% of macrovesicular steatosis and the absence of an identified alternative diagnosis, as well as a significant ongoing or recent alcohol consumption (defined as < 30 g/day for men and < 20 g/day for women)
NASH	Presence of inflammation and cellular injury expressed by ballooning
MAFLD	Presence of steatosis evidenced by histology, imaging, or blood biomarker in addition to one of the following criteria: Overweight/obesity, type 2 diabetes mellitus, or evidence of metabolic syndrome
SLD	Presence of steatosis diagnosed histologically or by imaging techniques
MASLD	Presence of SLD associated with at least one cardiometabolic risk factor and no other identified cause, alcohol consumption should be less than 20 g/day for women and 30 g/day for men, previously called NAFLD
MASH	Presence of MASLD and cellular inflammation and/or injury, previously mentioned as NASH
Met-ALD	Met-ALD was recently proposed, defined as the presence of SLD and alcohol consumption between 20-50 g/day in women and 30-60 g/day in men

MAFLD: Metabolic dysfunction-associated fatty liver disease; MASH: Metabolic steatohepatitis; MASLD: Metabolic steatotic liver disease; Met-ALD: Metabolic steatotic liver disease with moderate alcohol consumption; NAFLD: Nonalcoholic liver disease; NASH: Nonalcoholic steatohepatitis; SLD: Steatotic liver disease.

Given the limited data specifically addressing the occurrence of SLD compared to nonalcoholic FLD (NAFLD), the possibility of extrapolating findings from studies based on the former definitions of the newly proposed classification has been evaluated. Although SLD appears to be associated with a slightly increased mortality risk, it shares similar clinical profiles and noninvasive test performance characteristics of the former terminology, suggesting the potential interchangeability of these terminologies[12]. Although the pathophysiology of MASLD is not yet fully elucidated, insulin resistance associated with obesity is believed to play a central role. A complex interplay between genetic and environmental factors, including rich processed foods diets and sedentary lifestyles, contributes to the accumulation of triglycerides and lipids within hepatocytes. Additionally, adipose tissue hypertrophy and ectopic fat deposition - frequently observed in affected patients - may promote hepatic lipogenesis, oxidative stress, and the activation of hepatic stellate cells, leading to subsequent liver fibrosis through inflammatory signaling pathways[10-14]. The presence of low-grade inflammation and insulin resistance is the cornerstone of the pathogenesis of MASLD. The intrinsic relationship between MASLD and metabolic disorders such as obesity, insulin resistance, type 2 diabetes mellitus (T2DM), arterial hypertension, and dyslipidemia is well established[11].

MASLD encompasses a broad clinical spectrum, ranging from simple hepatic steatosis to more severe forms characterized by hepatic inflammation and hepatocyte ballooning, referred to as metabolic dysfunction-associated steatohepatitis (MASH). Studies have demonstrated that the risk of MASH is two to threefold higher in individuals with obesity and/or T2DM[11,13-15]. Therefore, the presence of MASH also carries a higher risk of liver fibrosis progression, cirrhosis, and even hepatocellular carcinoma (HCC)[1,5,14,15]. It is estimated that approximately 10%-20% of patients with MASLD will develop MASH and subsequently progress to ACLD and its complications[15]. Therefore, early diagnosis and accurate staging are crucial to reduce the morbidity and mortality associated with ACLD due to MASLD, as well as other related outcomes such as cardiovascular disease and malignancies[16]. Liver biopsy remains the gold standard for assessing hepatic fibrosis; however, several limitations restrict its widespread indication[17-21]. In addition to its considerable cost, even an adequate biopsy sample represents 1/50000 of the liver parenchyma, rendering the procedure susceptible to sampling error mainly due to disease heterogeneity. Significant inter- and intra-observer variability in histopathological interpretation has also been documented. Moreover, the procedure comes with certain risks, such as pain, bleeding, infection, perforation, and, in rare cases, death[1,16,22].

Despite these limitations, the development and validation of noninvasive methods (NIMs) for liver fibrosis assessment - characterized by safety, reproducibility, and applicability for disease monitoring - have become the focus of extensive research[23-25]. Various NIMs employing different diagnostic strategies have been evaluated in MASLD, targeting either the quantification of hepatic steatosis or the determination of hepatic fibrosis. These include clinical and laboratory-based biomarkers, as well as scoring systems derived from intrinsic properties of hepatic tissue. Among these, liver elastography[22,24,26], the aspartate aminotransferase (AST) to platelet ratio index (APRI)[27], the fibrosis-4 index (FIB-4)[28-30], the NAFLD fibrosis score (NFS)[31], the body mass index (BMI), AST/alanine aminotransferase (ALT) ratio and diabetes (BARD) score[31], the enhanced liver fibrosis (ELF) test[32], and other emerging biomarkers such as microRNAs[33], proteases[5], myeloperoxidase[34], and hyaluronic acid (HA)[21,33-35], have shown promise. However, the utility and accuracy of these methods remain to be fully established, considering factors such as diagnostic performance (biomarkers vs imaging methods, combinations or sequential strategies), clinical context (primary care vs referral centers), and target population (general groups vs high-risk groups). This review summarizes the findings on the NIMs for AF assessment in patients with MASLD, emphasizing their accuracy, strengths, and limitations.

LITERATURE REVIEW

A comprehensive systematic literature search was conducted in the PubMed, Cochrane Library, SciELO, and LILACS databases to identify relevant studies addressing MASLD, noninvasive biomarkers, and liver fibrosis. The search strategy included the following terms and their synonyms: “metabolic steatotic liver disease”, “noninvasive markers”, and “liver fibrosis”. Studies published in English or Portuguese were considered eligible for inclusion. The inclusion criteria comprised original articles, systematic reviews, and meta-analyses involving adult human subjects that focused on the diagnosis, staging, or risk stratification of liver fibrosis using NIMs. Exclusion criteria included editorials, case reports, conference abstracts, animal studies, and articles unrelated to the scope of MASLD or liver fibrosis evaluation. After screening titles and abstracts for relevance and assessing full texts according to the eligibility criteria, a total of 76 articles were selected for qualitative synthesis. The study selection process followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, ensuring transparency and methodological rigor.

LITERATURE DISCUSSION

The rates of progression from MASLD to MASH and the development of hepatic fibrosis or hepatic decompensation are multifactorial. It depends on the initial severity of the disease, genetic predisposition, environmental factors, as well as the presence of comorbidities, such as T2DM[1,11]. Fibrosis progresses at an estimated rate of 0.09 stages per year, indicating that it takes approximately 14.3 years to advance one stage of fibrosis in MASLD and 7.1 years in MASH[3]. Based on the initially proposed classification by Kleiner et al[36], four distinct patterns of liver fibrosis have been described, with stage F4 corresponding to liver cirrhosis. Key concepts for disease progression include the recognition of advanced fibrosis (AF), corresponding to Metavir stages F3-F4, and significant fibrosis, defined as fibrosis greater than stage F2. The latter group has recently been identified as comprising patients at higher risk of disease progression (at-risk MASH) who would benefit most from early diagnosis, inclusion in clinical trials, and interventions aimed at preventing adverse outcomes. T2DM and obesity are other important drivers of disease progression[36-39].

The search for less invasive and more feasible methods to screen for hepatic fibrosis has emerged in response to the limitations of liver biopsy previously mentioned and the high prevalence of patients at risk of developing advanced forms of the disease, highlighting the importance of NIMs for diagnostic purposes[40-43]. Therefore, several criteria are essential when selecting a diagnostic test. In addition to safety and reproducibility, consideration must be given to the clinical setting in which the test will be applied, as well as the local disease prevalence, to define the most appropriate diagnostic goals[21,22,40]. In low-prevalence settings, such as primary care, tests with high negative predictive value are preferred to rule out AF. Conversely, in specialized centers with a higher prevalence of MASLD, confirmatory tests with high specificity and positive predictive value are desirable[1,21,40]. Thus, in primary care, the optimal strategy involves an initial simple and inexpensive blood-based test (e.g., FIB-4) to exclude AF, followed by a confirmatory test with high specificity (e.g., vibration-controlled transient elastography [VCTE] or patented serum biomarkers) if the initial test is positive[1,40,44-46]. This sequential strategy was evaluated in a United Kingdom cohort of patients with suspected MASLD, where 81% were identified as low-risk and managed within primary care. Therefore, only 19% required referral to hepatology services[47]. The application of this algorithm significantly reduced unnecessary specialist referrals and improved the precision of identifying patients in need of advanced care. In a separate study, Perera et al[48] assessed the clinical and cost-effectiveness of implementing NIMs in AF detection in primary care. Their results demonstrated that the use of NIMs led to a meaningful reduction in inappropriate referrals and achieved cost savings ranging from 3% to 17%, reinforcing the value of integrating NIMs into routine primary care pathways for the early identification and management of ACLD[48]. Liver fibrosis assessment methods can be categorized into three distinct groups: Indirect tests (markers of hepatic inflammation or dysfunction, such as aminotransferases and liver function tests); direct tests (serum biomarkers involved in extracellular matrix turnover, including collagens I-IV, type III procollagen, hyaluronic acid); and imaging techniques (evaluation of liver morphology and stiffness degree).

Indirect methods

The indirect methods, which involve assessing liver biochemistry parameters such as aminotransferases and liver function tests (including bilirubin, albumin, and coagulation tests), are insufficient for the diagnosis of hepatic fibrosis, as these could remain normal despite advanced disease or conversely be abnormal only in late stages of cirrhosis. Therefore, several clinical scores combining laboratory, anthropometric, and clinical data have been proposed and extensively evaluated to optimize the prediction of disease severity. The APRI, initially developed by Wai et al[28] for predicting AF in patients with chronic hepatitis C, has since been validated for other LDs and the prediction of liver-related outcomes[49,50]. Its main advantage lies in the broad availability of its components and the simplicity of calculation. It also has a potentially reliable role in predicting cardiovascular risk in patients with metabolic syndrome, as an APRI > 1.5 was associated with an elevated cardiovascular risk[51]. Although APRI showed high diagnostic accuracy in chronic hepatitis C populations, its performance in MASLD is modest and inferior to other tests[52]. The NFS, developed by Angulo et al[31], is based on age, BMI, diabetes status or glucose intolerance, albumin, platelet count, and AST/ALT ratio. Though slightly more complex, these variables are widely accessible. Suggested cutoffs of < -1.455 and > 0.676 enable the reliable exclusion and confirmation of AF, with negative and positive predictive values of 93% and 90%, respectively[31]. While widely validated across ethnicities with reproducible accuracy, its performance in MASLD cohorts is suboptimal [negative predictive value (NPV): 89%; net present value (PPV): 25%][41], with moderate performance. Therefore, NFS is no longer the primary NIM recommended for screening in the latest American Association for the Study of Liver Diseases and European Association for the Study of the Liver guidelines, being substituted by the FIB-4 score[1,44].

The BARD score, proposed by Harrison et al[32], integrates BMI ≥ 28, AST/ALT ratio ≥ 0.8, and diabetes, producing a weighted score where 2-4 points correlate with an odds ratio of 17 for AF (95% confidence interval: 9.2-31.9) and an NPV of 96%. However, subsequent studies, especially after the adoption of MASLD terminology, revealed a suboptimal performance of BARD (area under the receiver operating characteristic curve [AUROC]: 0.59-0.61)[50,42].

Sterling et al[29] introduced the FIB-4 index in a cohort of hepatitis C virus/human immunodeficiency virus co-infected patients, which can be calculated using age, AST, ALT, and platelet count. Suggested cutoffs are < 1.3 to exclude AF (NPV: 90%) and > 2.67 to confirm its presence (PPV: 65%). The FIB-4 has since been validated across various etiologies for predicting liver complications, HCC, and all-cause mortality[52,53]. In MASLD, Shah et al[54] found FIB-4 to be superior to other scores, with subsequent meta-analyses supporting its high accuracy (AUROC: 0.80-0.85). FIB-4 is the endorsed NIM for fibrosis initial screening by American, European, and Brazilian guidelines, with a cutoff of less than 1.3 to rule out AF and more than 2.67 to rule it in[1,21,44-46,54]. However, limitations include age (under 35 or over 65) and the presence of T2DM[53,54]. A cutoff of 2 has been proposed for AF exclusion in patients older than 65. Nonetheless, data from a recently published study in the primary care setting suggest that this threshold may result in a significant rate of false negatives[55,56]. A multicenter study (14 tertiary centers) evaluated the optimal FIB-4 cutoff for predicting AF in 1050 patients with biopsy-proven MASLD. The lower and upper cutoff points suggested by the authors, based on age range, are as follows: 1.05 and 1.21 ≤ 49 years, 1.24 and 1.96 in the 50-59 years age range, 1.88 and 2.67 in the 60-69 years age range, and 1.95 and 2.67 in the ≥ 70 years age range. These cutoff points were defined to achieve a sensitivity and specificity of 90%[57]. Therefore, the FIB-4 score should be used with caution, mainly in patients older than 65, and further studies are necessary to establish the optimal cutoff for this subgroup of patients.

Recently, a novel multi-targeted machine learning framework (FIB-9, FIB-11, and FIB-12) has been developed, designed to enhance the performance of the FIB-4 index in predicting AF[58]. FIB-9 is based on nine routinely available blood biomarkers - AST, ALT, gamma-glutamyl transferase, alkaline phosphatase, bilirubin, albumin, platelets, prothrombin index (or international normalized ratio), and urea - adjusted for demographic and metabolic variables such as age, weight, height, and diabetes status. It is freely accessible via an online calculator and is particularly suitable for population-level screening. FIB-11 builds upon this model by incorporating two additional fibrosis-specific markers, hyaluronate and alpha2-macroglobulin, achieving diagnostic accuracy exceeding 80%. FIB-12 further integrates liver stiffness measurement (LSM), demonstrating superior performance with an AUROC of 0.91 and the lowest indeterminate rate (16.4%) in ternary classification among the evaluated tools[58]. All three models offer a direct estimation of liver fibrosis stage (F0-F4), enhancing clinical interpretation and supporting more informed decision-making. Compared to conventional NIMs, these models demonstrate superior diagnostic accuracy, particularly when stratified into binary, ternary, or quaternary classification schemes. However, the increased complexity of the formulas, potential limitations in clinical accessibility due to future patenting, and diminished accuracy in patients with extreme biomarker values represent notable drawbacks; further data are required for their validation[58].

Direct tests

Direct biomarkers reflect serum levels of substances associated with fibrogenesis and extracellular matrix turnover. However, these markers are not liver-specific, representing a limitation since they can also be elevated in the presence of extrahepatic fibrosis, such as pulmonary fibrosis. Procollagen III (Pro-C3) has been proposed as a direct biomarker for predicting AF, either alone or in combination with demographic characteristics such as age or the presence of DM (ADAPT score composed by age, diabetes status, platelet count and Pro-C3 levels). It demonstrates good diagnostic accuracy for AF, with AUROC values of 0.75 and 0.81 when used alone or with the ADAPT score, respectively, in both tertiary settings for AF confirmation and primary care for its exclusion[56,57]. However, its clinical applicability remains limited due to high costs and low availability.

The ELF score was initially derived from a cohort of patients with CLD of various etiologies and incorporates age along with serum levels of tissue inhibitor of metalloproteinases-1, amino-terminal propeptide of type III procollagen, and hyaluronic acid[1,50,51]. A threshold of 7.7 or lower categorizes patients as low risk, while values above 7.7 indicate intermediate to high risk[1,21]. ELF values above 9.8 reliably identify patients with MASLD at increased risk of cirrhosis progression and liver-related events[46,59,60]. Additionally, in patients with suspected or confirmed AF, an ELF ≥ 11.3 predicts future liver-related outcomes[1]. ELF testing is recognized and recommended by both the American Association for the Study of Liver Diseases and European Association for the Study of the Liver societies as a second-line method for further evaluating patients at intermediate risk based on FIB-4 levels (> 1.3), guiding specialist referral[1,44,45]. However, this method also suffers from limited availability and high costs.

Imaging-based methods

Imaging modalities enable not only the assessment of anatomic features of the liver but also the estimation of liver stiffness, serving as a surrogate marker of fibrosis. Liver stiffness can be measured using various elastography techniques, including VCTE - FibroScan, ultrasound-based elastography (point shear wave or acoustic radiation force impulse [ARFI] and two-dimensional shear wave), and magnetic resonance elastography (MRE)[22,61-63]. VCTE is the most validated technique for assessing liver stiffness. It has been shown to accurately predict fibrosis stages (primarily for excluding AF), liver-related outcomes, and both overall and liver-specific mortality[1,44,64]. Different cutoff values can be used depending on the clinical objective. In routine clinical practice, a cutoff of less than 8 kPa effectively excludes AF with good diagnostic accuracy[1,21,22,63]. Conversely, values above 12 kPa indicate a high risk of AF, with PPVs ranging from 76% to 88% in hepatology or endocrinology clinic populations, where the prevalence of MASLD is high[1,21,44,45].

For other ultrasound-based elastography techniques (ARFI and two-dimensional shear wave), optimal cutoff values are less well-established compared to VCTE[65]. However, their diagnostic performance has been considered comparable[23,66-68]. A recent meta-analysis involving 29 studies demonstrated that ARFI performs well in differentiating simple steatosis from significant fibrosis and offers excellent accuracy in identifying AF, with sensitivity and specificity of 92% and 85%, respectively, and an AUROC of 0.94 when using a cutoff value of 2.06 m/second[69]. The integration of ARFI technology into conventional ultrasound systems is one of the major advantages over VCTE, as it enables a comprehensive evaluation of the liver, including assessment of cirrhosis, the presence of portal hypertension, and even the identification of focal lesions. A liver stiffness value of less than 9 kPa (1.7 m/second), in the absence of other known clinical signs, has recently been proposed as a criterion to rule out compensated ACLD (cACLD)[68]. Conversely, values greater than 13 kPa (2.1 m/second) are highly suggestive of cACLD, and when greater than 17 kPa (2.4 m/second), the probability of clinically significant portal hypertension should be considered. However, additional studies may be required[22,68].

Although transient elastography (TE) and ultrasound-based elastography are highly effective diagnostic tools for assessing liver fibrosis[1,22,44,68], they present several limitations, including restricted use in obese patients, increased skin-to-liver capsule distance, narrow intercostal spaces, and the presence of ascites. Operator expertise, the selection of the appropriate probe (e.g., M or XL for TE), and adherence to high-quality criteria - such as obtaining at least 10 valid measurements with an interquartile range less than 30% of the median - are essential for reliable LSMs. Additionally, transient increases in liver stiffness caused by acute inflammation, cholestasis, or hepatic congestion can lead to overestimation of fibrosis. Standardization of measurement protocols, control of confounding clinical factors, and cross-validation with other noninvasive markers or histology remain crucial to enhance the diagnostic performance and clinical utility of elastography[24,70].

MRE quantifies liver stiffness by measuring the propagation velocity of shear waves generated by an external mechanical driver through the hepatic tissue. The shear wave speed increases proportionally with tissue stiffness. MRE is considered the most accurate noninvasive diagnostic tool for detecting and staging liver fibrosis[1,22,44,71-73]. MRE demonstrates high sensitivity and specificity in quantifying steatosis, as well as in predicting AF, with area under the receiver operating characteristic curve values ranging from 0.92 to 0.96[1,44,74]. LSM with MRE is also helpful for HCC risk prediction in patients with cACLD[75].

However, despite its high diagnostic accuracy, several limitations prevent its broader clinical implementation. Cost is a significant barrier. Adding MRE capability to a magnetic resonance imaging scanner requires substantial investment, and per-examination costs remain higher than those of ultrasound-based techniques[76,77]. Furthermore, MRE is primarily available only in specialized centers, reducing accessibility in primary care and low-resource settings[77,78]. Technical and patient-related limitations to MRE are also relevant. The technique depends on breath-hold compliance and is sensitive to motion artifacts, reducing its utility in patients with ascites, dyspnea, or cognitive impairment[79]. Moreover, the presence of severe hepatic iron overload may impair MRE performance by attenuating signal intensity, and factors such as claustrophobia or obesity may preclude adequate image acquisition[80,81]. Lastly, there is no universal consensus on optimal cutoff values for MRE. Notably, MRE-based LSM values are not directly comparable to TE, to which the commonly cited “rule of four” applies. Some of the proposed thresholds are: < 2.5 kPa as normal; 3.0-3.5 kPa indicating fibrosis stages 1-2; 3.5-4.0 kPa for stages 2-3; 4.0-5.0 kPa for stages 3-4; and > 5.0 kPa as indicative of stage 4 fibrosis or cirrhosis. Furthermore, MRE can reliably distinguish between mild (F0-F2) and significant (F3-F4) fibrosis, with sensitivity ranging from 86% to 91% and specificity between 80% and 85%[64]. The main advantages of MRE are its noninvasive nature, high accuracy, and the ability to examine a much larger hepatic volume than ultrasound elastography. With a technical success rate of 94%, it outperforms TE (FibroScan), which has a rate of 84% and is less affected by obesity, ascites, and anatomical variability[64,65]. MRE’s superior performance in detecting early-stage fibrosis and its integration into routine liver magnetic resonance imaging make it a powerful tool for both fibrosis staging and surveillance of other liver pathologies, such as HCC. Conversely, the economic, logistical, and physiological barriers limit the scalability of MRE, despite its inclusion in guideline-based fibrosis assessment algorithms. A summary of suggested cutoffs, strengths, and limitations of NIMs for AF assessment in MASLD is presented in Table 2.

Table 2 Parameters for noninvasive assessment of advanced fibrosis in metabolic dysfunction-associated steatotic liver disease.

NIM	Rule-in cutoff	Rule-out cutoff	AUROC1	Strengths and limitations
Biomarkers
APRI	< 0.5	> 1.5	0.74	False positive in patients with alternative causes of thrombocytopenia
FIB-4	< 1.32; < 1.67	≥ 2.67; ≥ 3.483	0.84	No added cost; low accuracy in patients younger than 35 years; low specificity in those older than 65 years
NFS	< -1.44	≥ 0.672	0.82	No added cost; not accurate in age < 35 years, people with obesity and type 2 diabetes Restricted use to MASLD
ELF	< 7.7; < 7.73	≥ 9.8; ≥ 11.33	0.83	Cost and not widely available; less useful for early fibrosis
Imaging techniques
VCTE	< 8 kPa; < 8 kPa3	≥ 12 kPa; ≥ 20 kPa3	0.9	Point of care, possibility of steatosis degree evaluation; necessity of specific device; confounded by active hepatitis, food intake, congestive heart failure, biliary obstruction, amyloidosis, ascites; less applicable and reliable in severe obesity (especially with M probe)
pSWE (ARFI)	< 1.3 m/second	> 2.1 m/second	0.8-0.9	Integration into conventional ultrasound systems, which enables a comprehensive evaluation of the liver at the same time; cut points not well validated; Probably affected by the same confounders as VCTE, though the success rate is higher than VCTE in obese patients
2D-SWE	< 8 kPa	> 12 kPa	0.80-0.98	Integration into conventional ultrasound systems, which enables a comprehensive evaluation of the liver at the same time; cut points not well validated; probably affected by the same confounders as VCTE
MRE	< 2.55 kPa; < 3 kPa3	≥ 3.63kPa; ≥ 5 kPa3	0.89-0.96	Superior performance in detecting early-stage fibrosis; integration into routine liver MRI allows both fibrosis staging and surveillance of other liver pathologies, such as HCC; cost, not widely available; probably affected by the same confounders as VCTE and iron content; some patients may have contraindications to magnetic resonance imaging or have claustrophobia

¹For F3-F4.

²A lower cutoff of 2 can be used for patients older than 65 years.

³To rule in or rule out cirrhosis.

2D-SWE: Two-dimensional shear wave elastography; APRI: Aspartate aminotransferase to platelet ratio index; ARFI: Acoustic radiation force impulse; AUROC: Area under the curve; ELF: Enhanced liver fibrosis; FIB-4: Fibrosis-4 index; MRE: Magnetic resonance elastography; NFS: Nonalcoholic fatty liver disease fibrosis score; NIM: Noninvasive method; pSWE: Point shear-wave; VCTE: Vibration-controlled transient elastography; HCC: Hepatocellular carcinoma.

Sequential diagnostic approach

While NIMs are increasingly recommended, awareness of their limitations remains crucial. A significant drawback is the “indeterminate zone” (40%-60% of cases), where results remain inconclusive, which can be mitigated by sequential strategies that combine different modalities of tests, as recently proposed by major liver societies[1,21,44-46]. A stepwise approach has been widely endorsed by international guidelines and real-world studies, aiming to balance diagnostic performance, cost-effectiveness, and accessibility. A frequently recommended strategy involves initial screening with the FIB-4 index, a simple, widely available score, followed by LSM using TE in patients with intermediate or indeterminate FIB-4 values[44-46,82]. For patients with a FIB-4 score < 1.3 (under 65 years) or < 2.0 (≥ 65 years), AF can often be ruled out, whereas those between 1.3 and 2.67 warrant further imaging evaluation[1]. In selected populations - such as those with obesity, inconclusive TE results, or suspected confounders like inflammation - MRE may be considered as a second-line modality due to its superior reproducibility and accuracy[83]. For settings where imaging is not immediately available, a biomarker-based second-line test, such as the ELF score, may be an alternative[44,46]. Combining noninvasive tools in this way reduces unnecessary biopsies, increases specificity for AF, and supports individualized patient care[84,85]. A proposed algorithm for the sequential approach might follow the sequence: FIB-4 TE MRE or ELF, depending on local resources, patient comorbidities, and pretest probability of fibrosis (Figure 1).

Figure 1 Proposed algorithm for noninvasive assessment of the risk of advanced fibrosis in patients suspected of metabolic dysfunction-associated steatotic liver disease. A: T2 diabetes mellitus (T2DM); B: Abdominal obesity and ≥ 1 additional cardiometabolic risk factor or persistently elevated enzymes. ¹Fibrosis-4 index (FIB-4) threshold valid for age ≤ 65 years (for age > 65 years: The lower FIB-4 cutoff is 2.0). AF: Advanced fibrosis; ELF: Enhanced liver fibrosis; GI: Gastrointestinal; MRE: Magnetic resonance elastography; SWE: Shear-wave elastography; VCTE: Vibration-controlled transient elastography.

Emerging biomarkers

Heredity plays a crucial role in fat accumulation and the progression of liver injury. However, genetic predisposition alone does not fully account for the wide phenotypic variability observed in MASLD. Epigenetics, a heritable yet reversible phenomenon, modulates gene expression without altering the DNA sequence and represents a promising new avenue for understanding MASLD development and progression, with special emphasis on microRNAs (miRNAs)[29]. Dysregulation of these post-transcriptional regulators of complementary target miRNAs has demonstrated high prognostic and predictive value in MASLD[29,74]. These molecules can be detected in serum, plasma, saliva, urine, and formalin-fixed tissues. In humans, miRNA expression appears to be tissue-specific. When released into circulation via cellular microvesicles, these miRNAs function as signaling molecules, and several miRNAs have been identified in various studies[86,87].

Among the studies, notable examples include miR-122[88], miR-192[89], miR-21, miR-34a, miR-451[90,91], miR-1290, miR-27b-3p, miR-192-5p[92], miR-301a-3p, miR-34a-5p, miR-375, lncRNA-RABGAP1 L-DT-206[93]. Given the need for reliable and innovative biomarkers, circulating miRNAs have emerged as attractive candidates for the diagnosis and staging of LDs[30]. Beyond genetic factors, inflammation is recognized as a central contributor to MASLD pathogenesis and progression, with the pro-inflammatory cytokine interleukin-1β serving as a key mediator in initiating and sustaining hepatic inflammation. Interleukin-1β is secreted as an inactive precursor and activated through the action of caspase-1, neutrophil serine proteases, proteinase 3, and neutrophil elastase. Several studies suggest that NE and proteinase 3 are important drivers of chronic inflammation and that imbalances between these enzymes and their inhibitor, alpha-1 antitrypsin, may contribute to obesity-induced insulin resistance, type 2 diabetes, and MASLD[4]. This systematic review has potential limitations that should be acknowledged. The heterogeneity in the patient population, as well as in the diagnostic criteria related to the disease nomenclature modification, and the different technologies and NIMs may limit the generalizability of the findings. Additionally, the risk of publication bias cannot be excluded, particularly given the reliance on studies with positive results. Besides that, the cross-sectional design of most studies precludes assessment of longitudinal outcomes and limits conclusions regarding prognostic utility.

CONCLUSION

Given the high prevalence of MASLD and the risks associated with biopsy for the diagnosis of liver fibrosis, NIMs have emerged as promising diagnostic alternatives. Liver elastography, particularly VCTE for assessing liver stiffness and MRE for measuring fibrosis, is considered the noninvasive gold standard. Among serum scores, FIB-4 and NFS are valuable for ruling out AF and selecting candidates for liver biopsy, although their low positive predictive value requires complementary tests. The combination of methods, such as FIB-4 and elastography, represents a practical strategy for monitoring, determining the need for referral to a specialist, and guiding the indication for liver biopsy. Despite the relevance and applicability of these biomarkers, further studies are necessary to enhance the accuracy of NIMs, establish cutoff points ideally outside indeterminate ranges, and integrate multiple assessment modalities to predict early liver fibrosis risk more accurately.