Noninvasive strategies for metabolic dysfunction-associated steatotic liver disease assessment and referral in Japan

Abstract

To establish practical, evidence-based strategies for noninvasive assessment and referral of patients with metabolic dysfunction-associated steatotic liver disease (MASLD) in Japan, we must address the urgent clinical need for accurate risk stratification and timely specialist intervention. A panel of 11 Japanese hepatology experts conducted a modified Delphi process to evaluate consensus recommendations regarding the use of noninvasive tests (NITs), including the fibrosis-4 index, enhanced liver fibrosis test, Mac-2-binding protein glycosylation isomer, type IV collagen 7S, cytokeratin-18 fragments, and imaging modalities such as ultrasound elastography and magnetic resonance elastography, for MASLD assessment and clinical referral. Practical algorithms were developed based on current Japanese data and panel consensus. The expert panel validated the utility of NITs as reliable tools for identifying patients with MASLD at risk for advanced fibrosis. Sequential use of NITs improved diagnostic accuracy and referral appropriateness while minimizing unnecessary specialist consultations. The proposed algorithms offer stepwise guidance for primary care physicians, supporting efficient, evidence-based decision-making. However, prospective longitudinal studies remain necessary for full prognostic validation of NITs in MASLD management. Noninvasive testing algorithms enable effective risk stratification and referral for MASLD in real-world Japanese practice with anticipated benefit for patient outcomes and healthcare systems. Broader adoption and further validation are warranted.

Key Words: Fibrosis-4 index; Enhanced liver fibrosis test; Mac-2-binding protein glycosylation isomer; Type 4 collagen 7S; Cytokeratin-18 fragment; Elastography

Core Tip: MASLD is highly prevalent, progressing to MASH, fibrosis, cirrhosis, and hepatocellular carcinoma. Although liver biopsy is traditional, noninvasive tests (NITs) such as the fibrosis-4 index, enhanced liver fibrosis test, Mac-2-binding protein glycosylation isomer, type IV collagen 7S, cytokeratin-18 fragments, and imaging (elastography/magnetic resonance elastography) now enable accurate and accessible assessment of liver fibrosis. This review summarized expert consensus on the appropriate use of NITs in Japan, providing evidence-based referral strategies and practical algorithms for primary care and non-hepatologists. Delphi panel results confirmed the utility of NITs but also highlighted the need for further longitudinal studies to establish their prognostic value.

INTRODUCTION

Metabolic dysfunction-associated steatotic liver disease (MASLD) represents the most prevalent chronic liver disease globally. Its progression to metabolic dysfunction-associated steatohepatitis (MASH) and hepatic fibrosis substantially contributes to the increasing burden of cirrhosis and hepatocellular carcinoma (HCC)[1,2]. In Japan MASLD affects approximately one in four individuals[3]. Histological evaluation by liver biopsy has long been considered the gold standard for assessing disease severity and stage. However, its application is limited by several drawbacks, including invasiveness[4], high cost, sampling variability[5], and inter-observer inconsistency. Given the large number of affected individuals, identifying which patients should be referred to hepatology specialists remains a major challenge in routine clinical practice.

In recent years with the evolution of diagnostic imaging methods and blood biomarkers, accurate diagnosis of liver fibrosis using noninvasive tests (NITs) has become possible (Table 1)[6-8]. In Japan various blood biomarkers and imaging diagnostics are covered by insurance[6]. Among NITs the fibrosis-4 (FIB-4) index is a very simple and useful scoring system that combines clinical data. Many of the proposed algorithms that combine various NITs use the FIB-4 index as the first stage and other NITs as the second stage[9-11]. We are now able to use many liver fibrosis markers, hepatocyte apoptosis markers, elastography, and liver fat quantification methods for the evaluation of liver disease progression (Table 1)[6]. By combining these NITs we have reached the next generation of liver pathophysiological evaluation. For reference the cutoff values for each NIT according to the stages of liver fibrosis covered in this review are shown in Table 2.

Table 1 Noninvasive tests for the diagnosis of liver fibrosis and activity used in Japan.

Classification of NITs		Name of NITs
Scoring system		FIB-4 index
		NAFLD fibrosis score
Blood biomarkers	Liver-specific fibrosis markers	ELF test
		COL4-7S
		M2BPGi
		ATX
		HA
	Hepatocyte apoptosis marker	CK-18F (M30)

NITs: Noninvasive tests; FIB-4: Fibrosis-4; NAFLD: Non-alcoholic fatty liver disease; ELF: Enhanced liver fibrosis; COL4-7S: Type IV collagen 7S; M2BPGi: Mac-2-binding protein glycosylation isomer; ATX: Autotaxin; HA: Hyaluronic acid; CK-18F: Cytokeratin 18 fragment.

Table 2 Cutoff values for each noninvasive test according to liver fibrosis stage.

	F2	F3	F4
ELF score	10.1	11.1	11.5
ELF score1 (for prediction of event)	-	9.8	11.3
M2BPGi (C.O.I./AU/mL)	1.15	0.99	1.71
COL4-7S (ng/mL)	3.8	4.8	6.6
VCTE1 (kPa)	-	Rule-in: 12 kPa. Rule-out: 8 kPa	Rule-in: 20 kPa. Rule-out: 8 kPa
MRE (kPa)	3.14	3.53	4.45

¹According to guidelines from the American Association for the Study of Liver Diseases and the European Association for the Study of the Liver-European Association for the Study of Diabetes-European Association for the Study of Obesity.

ELF: Enhanced liver fibrosis; M2BPGi: Mac-2-binding protein glycosylation isomer; C.O.I.: Cutoff index; COL4-7S: Type IV collagen 7S; VCTE: Vibration-controlled transient elastography; MRE: Magnetic resonance elastography.

Although NITs are widely available in Japan, guidance for primary care physicians on referral thresholds remains limited. This review therefore aimed to fill this void by providing the first comprehensive, expert-led consensus from Japan, offering tailored algorithms for the Japanese healthcare context. This review clearly showed how to use NITs in terms of what MASLD cases family doctors should refer to hepatologists and is expected to be very useful in daily clinical practice. In particular, the following NITs are noted in this review: FIB-4 index; liver fibrosis markers [enhanced liver fibrosis (ELF) test, Mac-2-binding protein glycosylation isomer (M2BPGi), type IV collagen 7S (COL4-7S)]; a hepatocyte apoptosis marker [cytokeratin-18 fragment (CK-18F)]; and imaging diagnosis [ultrasound elastography, magnetic resonance elastography (MRE)]. We also focused on the interactions between NITs and genetic factors and differentiated use of NITs by cohort. Furthermore, we proposed diagrams of simplified referral indicators for non-hepatologists and primary care physicians (Figures 1 and 2).

Figure 1 Algorithm for referral of patients with metabolic dysfunction-associated steatotic liver disease using noninvasive tests for liver fibrosis. The algorithm for referral from a primary care physician to a hepatologist to narrow down cases of metabolic dysfunction-associated steatotic liver disease with advanced liver fibrosis is shown. First, the fibrosis-4 index is evaluated (step 1), and if the score is 1.30 or equal to or greater than 2.00 age 65 or older, the patient is evaluated using one of the various noninvasive tests (step 2). Patients with high values are referred to a hepatologist (refer to hepatologist). In cases followed by a family physician or a regular physician (primary care or follow-up), it is recommended to repeat the above procedure every 2-3 years. FIB-4: Fibrosis-4; US: Ultrasound; MRI: Magnetic resonance imaging; ELF: Enhanced liver fibrosis; M2BPGi: Mac-2-binding protein glycosylation isomer; C.O.I.: Cutoff index; COL4-7S: Type IV collagen 7S; VCTE: Vibration-controlled transient elastography; MRE: Magnetic resonance elastography.

Figure 2 Stepwise approach for evaluating metabolic dysfunction-associated steatotic liver disease using the fibrosis-4 index and the cytokeratin 18 fragment biomarker. A: High-risk group (orange). Patients with fibrosis-4 (FIB-4) > 2.67 should be referred to a hepatologist regardless of cytokeratin 18 fragment (CK-18F) levels due to the high risk of advanced fibrosis; B: Low-risk group (green). Patients with FIB-4 < 1.3 and CK-18F < 260 U/L are considered to have minimal fibrosis and inflammation. Follow-up with repeat FIB-4 and CK-18F testing every 1-2 years is recommended; C: Referral consideration group (pink). Patients with CK-18F ≥ 260 U/L may have progressive inflammation. Consider referring to a hepatologist for pathologic evaluation; D: Intermediate-risk group (yellow). Patients with FIB-4 of 1.30-2.67 and CK-18F < 260 U/L may include cases of full-blown metabolic dysfunction-associated steatohepatitis with advanced fibrosis. ¹Indicates further evaluation with the enhanced liver fibrosis test, Mac-2-binding protein glycosylation isomer, type IV collagen 7S, vibration-controlled transient elastography, or magnetic resonance elastography will be necessary to confirm fibrosis status. ²Indicates a fibrosis-4 index of 2.0 is considered intermediate risk for patients age 65 or older. CK-18F: Cytokeratin 18 fragment; FIB-4: Fibrosis-4.

DELPHI PROCESS AND THE USEFULNESS OF NIT IN THE TREATMENT OF MASLD

Main statements and their summaries (Table 3): 1-2: Known noninvasive MASLD markers and imaging tests lack longitudinal validation to confirm their prognostic value. 1-3: Blood and imaging tests help assess liver fibrosis and should guide referrals to hepatologists.

Table 3 Statements and agreement degree in the Delphi round,.

Number	Statements	Agreement 1st round	Agreement 2nd round
The usefulness of NIT in the treatment of MASLD
1-1	Known noninvasive markers and imaging tests that have been validated for MASLD have been verified in various cross-sectional studies, and their clinical application in Japanese MASLD cohorts would be highly useful	11 (100)	11 (100)
1-2	Known noninvasive markers and imaging tests that have been validated for MASLD have a lack of longitudinal studies, and further validation is needed to determine whether they can serve as prognostic markers or tests	11 (100)	11 (100)
1-3	Blood biomarkers and imaging diagnostics are useful for evaluating liver fibrosis, and if possible, non-hepatologists and even family physicians should use these to guide referral to a hepatologist	11 (100)	11 (100)
1-4	The occurrence of liver disease-related events in patients with MASLD depends on the degree of progression of liver fibrosis. Evaluation of liver activity (inflammation, necrosis, and fatty changes) in addition to evaluation of liver fibrosis is useful for evaluating the rate of progression of the disease	11 (100)	11 (100)
FIB-4 index
2-1	The FIB-4 index can be recommended as a first-line tool to screen for advanced fibrosis in patients with MASLD	11 (100)	11 (100)
2-2	A FIB-4 index > 1.3 is a reasonable criterion for referral to a hepatologist	4 (36.4)	0 (0)
2-3	The FIB-4 index is useful for predicting the prognosis of MASLD	11 (100)	11 (100)
ELF test
3-1	ELF is recommended in international guidelines to narrow down the range of cases of MASLD with advanced fibrosis, and should also be recommended in Japan	11 (100)	11 (100)
3-2	Evidence from cross-sectional studies of ELF in the assessment of MASLD liver fibrosis has been established for the diagnosis of fibrosis in Japanese patients with MASLD	7 (63.6)	5 (45.5)
M2BPGi
4-1	Evidence for the use of M2BPGi in the assessment of MASLD liver fibrosis has been established through a cross-sectional study of Japanese patients with MASLD	9 (81.8)	8 (72.7)
4-2	Longitudinal studies to predict the prognosis of patients with MASLD using M2BPGi are still lacking, and further validation is required	11 (100)	11 (100)
COL4-7S
5-1	COL4-7S can be used to diagnose advanced liver fibrosis in place of liver biopsy	11 (100)	11 (100)
5-2	Because COL4-7S is a single marker, it is less affected by fatty changes or ballooning hepatocytes, and reflects the state of fibrosis better than other markers	10 (90.9)	11 (100)
5-3	As the second step of FIB-4, COL4-7S is useful for identifying MASLD risk groups	11 (100)	11 (100)
5-4	For COL4-7S, it is necessary to set and verify a cutoff value for the current measurement method, the CLEIA method	11 (100)	11 (100)
CK-18F
6-1	MASL/MASH can be diagnosed using CK-18F	2 (18.2)	0 (0)
6-2	CK-18F is a valid marker for evaluating MAS as an index of liver activity	10 (90.9)	10 (90.9)
6-3	CK-18F can reflect the effects of various treatments by assessing liver activity	7 (63.6)	8 (72.7)
6-4	CK-18F can diagnose at-risk MASH	6 (54.5)	1 (9.1)
6-5	The combination of CK-18F and FIB-4 indexes is useful for diagnosing MASH	10 (90.9)	10 (90.9)
US elastography
7-1	Liver stiffness measured by VCTE can be used as the standard for evaluating liver fibrosis	9 (81.8)	7 (63.6)
7-2	2D-SWE measurements may be compatible with VCTE measurements	9 (81.8)	10 (90.9)
7-3	The effectiveness of treatment can be assessed based on changes in liver stiffness measured by VCTE	10 (90.9)	9 (81.8)
MRE
8-1	Liver stiffness measured by MRE may be used as the first line of treatment in routine clinical practice	1 (9.1)	0 (0)
8-2	Liver stiffness measurement by MRE could become the gold standard for clinical trials and research	11 (100)	11 (100)
8-3	It is possible to predict prognosis based on liver stiffness measured by MRE	11 (100)	11 (100)
8-4	The MEFIB score can efficiently diagnose F > 2	11 (100)	11 (100)
8-5	The MAST score can diagnose at-risk MASH	10 (90.9)	11 (100)
Genetic polymorphism
9-1	PNPLA3 gene polymorphism is involved in the progression of liver fibrosis	11 (100)	11 (100)
9-2	Measurement of gene polymorphisms is useful for assessing the risk of progression of liver fibrosis	10 (90.9)	11 (100)
Differentiated use by cohort
10-1	It is better for the NIT cutoff values to be the same in the health checkup cohort and the hospital cohort	5 (45.5)	2 (18.2)
10-2	If FIB-4 is < 1.3, observation is sufficient regardless of the cohort	9 (81.8)	4 (36.4)

Data are presented as n (%). NIT: Noninvasive test; MASLD: Metabolic dysfunction-associated steatotic liver disease; FIB-4: Fibrosis-4; ELF: Enhanced liver fibrosis; M2BPGi: Mac-2-binding protein glycosylation isomer; COL4-7S: Type IV collagen 7S; CLEIA: Chemiluminescent enzyme immunoassay; CK-18F: Cytokeratin 18 fragment; MASL: Metabolic dysfunction-associated steatotic liver; MASH: Metabolic dysfunction-associated steatohepatitis; MAS: Macrophage activation syndrome; US: Ultrasound; VCTE: Vibration-controlled transient elastography; 2D-SWE: Two-dimensional shear wave elastography; MRE: Magnetic resonance elastography; MEFIB: Magnetic resonance elastography combined with the fibrosis-4; MAST: Magnetic resonance imaging-aspartate aminotransferase; PNPLA3: Patatin-like phospholipase domain-containing 3.

A total of 11 hepatologists with more than 10 years of practice experience were invited to participate in a modified Delphi panel to respond to the statements in this review with an agreement threshold of 70% or higher. The basis for creating the statement was a combination of systematic literature review and expert knowledge. The Delphi panel asked participants to answer whether they agreed or disagreed with each statement. The first-round survey was administered from May 14, 2025 to May 20, 2025, followed by the second-round survey from June 12, 2025 to June 20, 2025 (Table 3). Our Delphi consensus study confirmed 100% expert agreement that NIT was useful in the treatment of MASLD (statements 1-1, 2, 3, and 4 in Table 3). However, noninvasive markers and imaging tests for MASLD lack longitudinal studies and require further validation to determine their usefulness as prognostic markers or tests (statement 1-2 in Table 3).

FIB-4 index

Main statements and their summaries (Table 3): 2-1: The FIB-4 index can be recommended as a first-line tool to screen for advanced fibrosis in patients with MASLD. 2-3: The FIB-4 index is useful for predicting the prognosis of MASLD.

The FIB-4 index, calculated using age, aspartate aminotransferase (AST) and alanine aminotransferase levels, and platelet count, offers a noninvasive alternative to liver biopsy for fibrosis assessment due to its simplicity and noninvasive nature. Its primary strength lies in its high negative predictive value (NPV, 99%) for advanced fibrosis (≥ F3) at a cutoff of < 1.3, effectively identifying patients with low risk. Major hepatology societies, including the Japan Society of Gastroenterology, Japan Society of Hepatology[12], American Association for the Study of Liver Diseases (AASLD)[13], American Diabetes Association[14], and European Association for the Study of the Liver (EASL)-European Association for the Study of Diabetes (EASD)-European Association for the Study of Obesity (EASO)[15], recommend the FIB-4 index as a primary screening tool for fibrosis risk stratification for inpatients with MASLD.

The FIB-4 index has evolved beyond its original role as a noninvasive diagnostic tool for liver fibrosis, emerging as a critical biomarker for predicting clinical outcomes and monitoring therapeutic responses in MASLD. The FIB-4 index powerfully predicts future liver complications for liver-related morbidity and mortality. In a United Kingdom population study of 20433 individuals with obesity or type 2 diabetes (T2D), those with high baseline FIB-4 (> 2.67) had a 24-fold increased risk of liver-related events (LREs) compared with low-risk groups (< 1.3)[16]. In a Japanese study of 1398 individuals with biopsy-proven MASLD (Clinical Outcome Nonalcoholic Fatty Liver Disease study), the low-risk groups (< 1.3) had a decreased risk of LREs and liver-related mortality[17].

The index also correlates with cardiovascular outcomes and all-cause mortality, independent of traditional cardiovascular risk factors[16,18]. High FIB-4 scores (> 2.67) are associated with a 2.5-3.0-fold increased mortality risk compared with low scores[16,19]. Time-varying models show even stronger associations with hazard ratios (HRs) reaching 13.26 for liver-related outcomes[19]. A 12-month increase in FIB-4 predicts higher risks of LREs (adjusted HR = 2.70) and mortality (HR = 3.01)[16,20]. Annual FIB-4 increases ≥ 0.2 units predict fibrosis progression [odds ratio (OR) = 3.1][19,20]. Conversely, a 12-month decrease reduces liver event risk by 63% (HR = 0.37)[20]. This temporal association supports its use for real-time risk stratification. Annual FIB-4 increases ≥ 0.2 units predict fibrosis progression (OR = 3.1)[19,20].

In the PIVENS trial FIB-4 changes paralleled improvements in histological activity scores [nonalcoholic fatty liver disease activity score (NAS)] following pioglitazone or vitamin E therapy (r = 0.237, P < 0.001)[21]. A reduction in FIB-4 correlated with decreased steatosis and lobular inflammation[21]. A Delphi consensus study confirmed 100% expert agreement that the FIB-4 index is recommended as a first-choice screening tool for fibrosis progression in patients with MASLD and useful for predicting prognosis in MASLD (statements 2-1 and 3 in Table 3).

The accuracy of the FIB-4 index is significantly influenced by age and the existence of T2D. In elderly patients the age factor can lead to overestimation of fibrosis risk, whereas in younger patients (< 35 years), it may underestimate the risk. To address age-related variations, different cutoff values are recommended[22]: (1) < 35 years: Underestimates fibrosis risk (NPV = 89%); and (2) ≥ 65 years: Overestimates risk unless adjusted to a cutoff of > 2.0 (NPV improves from 83% to 100%). A multicenter study by the Japan Study Group of nonalcoholic fatty liver disease demonstrated improved diagnostic accuracy with age-specific cutoffs[23]: (1) ≤ 49 years: 1.05 (low risk)/1.21 (high risk); (2) 50-59 years: 1.24/1.96; (3) 60-69 years: 1.88/2.67; and (4) ≥ 70 years: 1.95/2.67.

This adjustment reduced false positives in elderly patients from 43% to 18%. Conversely, the application of a FIB-4 threshold of 1.3 is only minimally altered by age, whereas higher values of FIB-4, such as 2.7, are associated with lower risk in older people with MASLD[20]. These results suggest that the FIB-4 1.3 threshold for referral to subsequent testing should not be altered in people ≥ 65 as suggested previously. In contrast, the finding of a FIB-4 value over 2.7 should not be used to classify elderly people with MASLD as high risk but should trigger further risk stratification with other NITs. Notably, a Delphi study revealed limited expert consensus [36.4% (1^st round), 0.0% (2^nd round) agreement] on using FIB-4 > 1.3 as a referral criterion to hepatologists, reflecting concerns about its application in indeterminate risk cases (statement 2-2 in Table 3).

Patients with prediabetes/T2D or two or more metabolic risk factors are at higher risk for hepatic fibrosis and should have sequential or consecutive testing with a second NIT even if they show FIB-4 index < 1.3[24]. A study involving 1489 patients with MASLD from Japan, Korea, and Taiwan demonstrated that the presence of T2D reduced the diagnostic performance of noninvasive liver fibrosis markers, including the FIB-4 index[25]. In patients with MASLD with T2D, COL4-7S is more accurate than the FIB-4 index[26]. When using the FIB-4 index in aged patients or those with T2D, it is crucial to recognize its limitations and combine the results with other diagnostic methods for a comprehensive evaluation of liver fibrosis. Patients with FIB-4 scores between 1.3 and 2.67 fall into an indeterminate risk category. Approximately 30% of this group may have advanced fibrosis (≥ F3), necessitating additional tests such as elastography or hepatic fibrosis markers. Combinations with other NITs and utilizing additional NITs, such as the ELF test, M2BPGi[27], or COL4-7S, can improve overall diagnostic accuracy.

In conclusion, the FIB-4 index has transitioned from a static diagnostic tool to a dynamic prognostic biomarker in MASLD. Its ability to predict liver/cardiovascular outcomes and track therapeutic responses makes it indispensable for risk-adapted management. As MASLD prevalence escalates globally, optimizing FIB-4-based monitoring protocols will be critical for effective disease management and resource allocation in healthcare systems.

Liver fibrosis markers

ELF test main statements and their summaries (Table 3): 3-1: ELF is globally recommended to identify advanced MASLD and should be adopted in Japan.

The ELF test was developed in 2004 and consists of a logarithmic algorithm that combines the concentrations of three serum markers of hepatic extracellular matrix metabolism: Hyaluronic acid; tissue inhibitor of metalloproteinase-1; and N-terminal peptide of procollagen III[28]. The ELF test predicts pathological liver fibrosis based on other serum biomarkers in chronic liver disease. Because several lines of clinical evidence for the ELF test are available for patients with MASLD, the test is recommended for use under several guidelines and practice guidance for identifying patients with advanced liver fibrosis when the FIB-4 index ≥ 1.3[13-15] (statement 3-1 in Table 3 and Table 4).

Table 4 Cutoff scores in guidelines and clinical guidance for the enhanced liver fibrosis test.

Guidelines/clinical guidance	Cutoffs for the ELF test
EASL-EASD-EASO Clinical Practice Guidelines[15]	Rule-out: 7.7
	Rule-in: 9.8
AASLD Practice Guidance[13]	Low risk: < 7.7
	Intermediate risk: 7.7-9.8
	High risk: > 9.8
ADA Guidelines[14]	Single cutoff: 9.8

ELF: Enhanced liver fibrosis; EASL: European Association for the Study of the Liver; EASD: European Association for the Study of Diabetes; EASO: European Association for the Study of Obesity; AASLD: American Association for the Study of Liver Disease; ADA: American Diabetes Association.

According to these guidelines and practice guidance, the cutoff for ELF test scores to suspect advanced fibrosis (≥ F3) was 9.8. According to the EASL-EASD-EASO guidelines, 7.7 was also suggested as a rule-out cutoff[15]. In the AASLD Clinical Practice Guidelines, an ELF test result between 7.7 and 9.8 indicates intermediate risk of advanced fibrosis to be referred to a gastroenterologist or hepatologist[13]. Guidelines from the American Diabetes Association simply suggest that patients with ELF test scores higher than 9.8 should be referred to a gastroenterologist or hepatologist[14]. However, the cutoff scores for advanced fibrosis and individual fibrosis stages and cirrhosis are confusing and vary according to the study cohort (Table 5).

Table 5 Cutoffs for the enhanced liver fibrosis test.

Ref.	Study design	Stage 1	Stage 2	Stage 3	Stage 4
Nobili et al[29], 2009	Cross-sectional	9.28	10.18	10.51	-
Inadomi et al[30], 2020	Cross-sectional	-	9.86	10.38	-
Seko et al[31], 2023	Cross-sectional	9.1	10.11	11.1	11.54
Sanyal et al[35], 2023	Cross-sectional	-	8.8/10.01	9.2/10.41	9.7/10.91
Hinkson et al[32], 2023	Meta-analysis	-	9.5	9.6	-
Vali et al[33], 2020	Meta-analysis	-	9.37	9.43	-
Sanyal et al[34], 2019	Longitudinal	-	-	9.82	11.32

¹Cutoff with sensitivity ≥ 90% and specificity ≥ 90%.

²Cutoff for the prediction of clinical event.

To our knowledge Nobili et al[29] first reported cutoff scores for the ELF test for MASLD in 2009; these scores were 9.28 for fibrosis stage 1, 10.18 for fibrosis stage 2, and 10.51 for fibrosis stage 3 according to Youden’s index. Several cross-sectional studies and meta-analyses have demonstrated the cutoff values of 9.10 for stage 1 fibrosis, 9.37-10.11 for stage 2 fibrosis, 9.43-11.10 for stage 3 fibrosis, and 11.54 for cirrhosis according to Youden’s index[30-33]. The key evidence was published by Sanyal et al[34] in 2019. In this study a cohort of simtuzumab trials was longitudinally analyzed, including patients with fibrosis stage 3 and compensated cirrhosis. The optimal cutoff of the baseline ELF test was 9.76, which was associated with an increased risk of disease progression for patients with fibrosis stage 3. In the patients with compensated cirrhosis (fibrosis stage 4), the optimal threshold for baseline ELF to predict clinical events, including ascites, hepatic encephalopathy, newly diagnosed varices, esophageal variceal bleed, ≥ 2-point increase in the Child-Pugh score and/or the Model for End-Stage Liver Disease score ≥ 15, and death was 11.27. Therefore, cutoff scores of 9.76 and 11.27 were not for the diagnosis of stage 3 and stage 4 fibrosis but for the prediction of clinical events. In the current guidelines and clinical guidance, the cutoff score of 9.76 for stage 3 fibrosis is rounded off to one decimal place to 9.8[13-15].

Recently, another line of key evidence was published by the FNIH-NIMBLE project in 2023 and showed dual cutoff values of the ELF test with a sensitivity ≥ 90% or a specificity ≥ 90%; those values were 8.8 or 10.0 for fibrosis stage 2, 9.2 or 10.4 for fibrosis stage 3, and 9.7 or 10.9 for fibrosis stage 4[35]. In terms of the implementation of surveillance without omission to detect patients who are high risk with advanced fibrosis, a cutoff of 9.2 rather than 9.8 would be useful. Regarding a cutoff score of 7.7, several studies denied its utility as too many patients would be diagnosed as positive[36,37]. The ELF test correlates better than other NITs with changes in fibrosis stage regardless of the treatment arm[38], suggesting that the test reflects histological changes regardless of the modes of action of any drugs.

The ELF test is also associated with the clinical outcome of MASLD; a change in the ELF score of 0.5 is related to clinical outcomes[34]. Moreover, the diagnostic performance of the ELF test is maintained better than those of other NITs in patients with T2D. Whereas several cross-sectional studies confirmed the performance of ELF test for diagnosis of liver fibrosis in MASLD[9,29,30], these studies did not reach consensus (statement 3-2 in Table 3). A possible explanation is the varied cutoff value in the study of the Japanese cohort (Table 5). Moreover, the ELF test is a relatively new fibrosis marker in Japan and usage in clinical practice might not be enough to reach a high agreement rate. However, according to the mature global and Japanese studies, the ELF test is among the most evident biomarkers representing liver fibrosis and clinical outcomes (Figure 1).

M2BPGi main statements and their summaries (Table 3): 4-1: Evidence for M2BPGi in assessing MASLD fibrosis comes from a Japanese cross-sectional study. 4-2: Longitudinal studies for its prognostic use are still needed.

M2BPGi is a liver fibrosis biomarker that originated in Japan and has been covered by health insurance for the past 10 years. M2BPGi is recognized by Fuji lectin (also known as WFA+-Mac-2bp)[39]. M2BPGi is useful for not only predicting the stage of liver fibrosis but also evaluating the degree of liver inflammation and predicting the onset of HCC; many reports have shown its usefulness as a biomarker of not only liver fibrosis but also liver activity. Outside of Japan there have been many reports from Korea, Taiwan, Hong Kong, China, Indonesia, and West Africa (Gambia)[27].

Initially, M2BPGi was measured as a cutoff index [C.O.I.: Negative, positive (1+), positive (2+)]; however, recently, a quantitative M2BPGi (M2BPGi-Qt) assay was developed[40]. In the M2BPGi-Qt test, the accuracy was improved by increasing the calibration curve points from two to five points and analyzing using a logistic curve. The M2BPGi-Qt uses AU/mL as the unit of measurement, adjusting the 1 C.O.I. in the current qualitative measurement method of M2BPGi to 1 AU/mL. Even in cases where the measured value was high using the qualitative method, the quantitative method allows for a more accurate measurement[40]. It is expected that the M2BPGi-Qt will be able to diagnose the progression of liver fibrosis based on more accurate test results and enable early detection, follow-up, and post-treatment monitoring of the diseases caused by chronic hepatitis.

More recently, we analyzed 1024 patients from a hospital cohort who had biopsy-confirmed MASLD and had their M2BPGi levels assessed. The patients’ M2BPGi levels were significantly negatively correlated with hepatic steatosis and positively correlated with lobular inflammation, ballooning, and fibrosis stage. Serum M2BPGi levels increased significantly with fibrosis progression (F0/F1/F2/F3/F4: 0.76 ± 0.54/0.90 ± 0.49/1.24 ± 0.89/1.73 ± 1.10/2.56 ± 1.55 C.O.I.). Advanced fibrosis (F3-4) was associated with higher M2BPGi levels than early fibrosis (F0-2). The cutoff value for F3 or greater was 0.99 and that for F4 was 1.71. The cutoff value for F2 or higher was 1.15, which was the opposite of the cutoff value for F3 or higher (Table 2). One possible reason for this is that M2BPGi values change not only depending on liver fibrosis but also on activity (steatosis, inflammation, necrosis, etc.).

Using receiver operating characteristic analyses, the areas under the receiver operating characteristic curve (AUROCs) were 0.774 for F3 or greater and 0.812 for F4. In our study the optimal cutoff value for advanced fibrosis was set at 1.0, achieving a sensitivity of 75%, a specificity of 66%, a positive predictive value (PPV) of 43%, and an NPV of 88%. This cutoff value was consistent with the results reported in a Korean health checkup cohort[41], further supporting the robustness of our data. Our two-step algorithm combining FIB-4 and M2BPGi reduced the need for additional testing by 31.5% while maintaining a false-negative rate (FN) of 13.9%. M2BPGi is a robust biomarker for advanced liver fibrosis in patients with MASLD. Integrating FIB-4 into our screening algorithm will improve diagnostic accuracy and minimize unnecessary liver biopsies and referrals to specialists.

Our Delphi consensus study confirmed high expert agreement that M2BPGi is effective for diagnosing liver fibrosis in MASLD (statements 4-1 and 2 in Table 3, Figure 1). Many cross-sectional studies from various countries, including Japan, have reported the usefulness of M2BPGi as a biomarker for evaluating liver fibrosis in MASLD. However, there have been no longitudinal studies reporting the usefulness of M2BPGi as a prognostic marker for MASLD, and this remains a topic for the future.

As a liver fibrosis biomarker for MASLD, M2BPGi has several problems. M2BPGi was originally developed as a glycan biomarker for patients with chronic hepatitis C[39]. Therefore, its predictive ability for liver fibrosis in other chronic liver diseases (MASLD, chronic hepatitis B, primary biliary cholangitis, and autoimmune hepatitis) is relatively low compared with chronic hepatitis C[39,42-45]. To that end a new M2BPGi-Qt has been established as a more accurate method than the qualitative measurement [M2BPGi (C.O.I.)], and M2BPGi-Qt is more accurate than the qualitative method at high M2BPGi levels[40]. The difference in the usefulness of M2BPGi (C.O.I.) depending on the disease may be improved using the M2BPGi-Qt. Another current problem is that it is not yet clear whether M2BPGi can accurately detect changes in MASLD pathology. We hope that future longitudinal studies, especially detailed disease investigations using M2BPGi-Qt, will clarify the relationship between MASLD pathology and M2BPGi.

COL4-7S main statements and their summaries (Table 3): 5-1: COL4-7S helps diagnose advanced fibrosis noninvasively. 5-2: As a single marker it reflects fibrosis better than others. 5-4: A cutoff for the chemiluminescent enzyme immunoassay (CLEIA) method needs validation.

COL4-7S is a fragment of type IV collagen primarily found in the basement membrane; it is one of the products of extracellular matrix turnover and derived from extensive dynamic extracellular matrix remodeling involving both the interstitial and basement membrane matrices that occurs during fibrogenesis. Proteolytic fragments of different collagen subtypes are released during fibrogenesis and/or fibrinolysis and can be used as noninvasive biomarkers[46]. Serum COL4-7S levels are increased during the progression of liver fibrosis in various types of chronic hepatitis, including MASH[47]. COL4-7S has been used to study fibrogenic activity in both animal models and humans[48]. It has been covered by insurance in Japan since 1989 and has become one of the most useful liver fibrosis markers in Japan[49].

According to our previous studies[26,50], the diagnostic performance of COL4-7S for liver fibrosis of MASLD is as follows. When COL4-7S is compared with FIB-4 and ELF, which are currently recommended in the guidelines, FIB-4 has the disadvantage that its fibrosis diagnosis is reduced in the presence of diabetes or age, whereas in the absence of diabetes, its diagnostic ability for F3 or higher is equivalent to FIB-4 (AUROC 0.883 of COL4-7S vs 0.879 vs FIB-4); in the presence of diabetes, its diagnostic ability is maintained without being affected by diabetes, and it has a better diagnostic ability than FIB-4 (AUROC 0.872 of COL4-7S vs 0.790 of FIB-4). When compared with ELF, which includes biomarkers of liver fibrosis progression such as tissue inhibitor of metalloproteinase-1, peptide of procollagen III, and hyaluronic acid, the diagnostic performance of COL4-7S was equivalent to that of ELF for F3 or higher (AUROC 0.844 of COL4-7S vs 0.82 of ELF, P = 0.096); for F2 or higher, COL4-7S is superior to ELF (the net reclassification improvement when comparing the diagnostic model of ELF to that of COL4-7S was 0.167, P = 0.018). When focusing on MASLD with diabetes, COL4-7S is even more superior to ELF in terms of its ability to differentiate between F2 or higher (AUROC 0.817 of COL4-7S vs 0.773 of ELF, P = 0.040) and above and at-risk MASH (AUROC 0.772 of COL4-7S vs 0.728 of ELF, P = 0.044). Our Delphi consensus study confirmed 100% expert agreement that COL4-7S is effective for assessing liver fibrosis in MASLD (statement 5-1 in Table 3). Also, in comparison with histological findings, ELF was positively influenced by steatosis, inflammation, ballooning, and fibrosis, whereas COL4-7S was influenced only by inflammation and fibrosis. This suggests that COL4-7S may more purely reflect the progression of liver fibrosis (statement 5-2 in Table 3).

In April 2021 the COL4-7S CLEIA (Fujirebio Inc.) was approved for use in Japan. However, the evidence of the clinical utility of the COL4-7S for the diagnosis of liver fibrosis is mainly in relation to COL4-7S measured by the radioimmunoassay (RIA) method. Therefore, the cutoff values of COL4-7S for F2 or F3 or higher were determined using the RIA method. In 2021 the cutoff value of CLEIA was calculated with reference to the correlation formula reported by Shima et al[51]. The equation is as follows: COL4-7S-CLEIA = 1.28 × COL4-7S-RIA - 1.15 (r = 0.888, P < 0.01). In our study[17] the diagnostic performance of COL4-7S for fibrosis was as follows for the two cutoff values of sensitivity 90% and specificity 90% (Table 6).

Table 6 Clinical utilities of the fibrosis-4 and type IV collagen 7S (radioimmunoassay) for the diagnosis of fibrosis stages 2 and 3 or higher[17].

		Cutoff value	PPV	NPV	Sensitivity	Specificity	FN	FP
FIB-4 index	Low	1.30	0.257	0.962	0.892	0.513	0.108	0.487
	High	2.67	0.382	0.093	0.547	0.833	0.453	0.167
COL4-7S (ng/mL), F0-1 vs F2-4)	Low	3.8	0.541	0.842	0.902	0.405	0.098	0.595
	High	6.0	0.775	0.676	0.446	0.900	0.554	0.100
FIB-4 and COL4-7S		1.3 and 3.8	0.690	0.792	0.752	0.737	0.248	0.263
COL4-7S, F0-2 vs F3-4	Low	4.8	0.315	0.969	0.892	0.634	0.108	0.366
	High	6.8	0.530	0.081	0.576	0.904	0.424	0.096
FIB-4 and COL4-7S (ng/mL)		1.3 and 4.8	0.366	0.953	0.806	0.737	0.194	0.263
F0-1 vs F2-4		1.3 and 4.3	0.706	0.787	0.730	0.766	0.270	0.234
F0-2 vs F3-4		1.3 and 4.3	0.307	0.963	0.871	0.630	0.129	0.370

Cutoff values were calculated with 90% sensitivity and 90% specificity. Type IV collagen 7S scores were measured using the radioimmunoassay method. FIB-4: Fibrosis-4; COL4-7S: Type IV collagen 7S; PPV: Positive predictive value; NPV: Negative predictive value; FN: False negative rate; FP: False positive rate.

In the algorithm to identify patients at risk of liver fibrosis, the diagnostic performance of the cutoff values of 3.8 or 4.0 for COL4-7S and 1.30 for FIB-4 as the first step is: The FN increased in the pick-up of F2 or higher compared with COL4-7S alone, but the false-positive rate was significantly improved (Table 6). In the pick-up of F3 or higher, both the FN and the false-positive rate were almost unchanged compared with COL4-7S alone. Considering the rate of fibrosis progression in MASLD[17], annual evaluation using this algorithm may be sufficient to identify high-risk MASLD groups (statement 5-3 in Table 3).

Kawanaka et al[52] recently suggested a new cutoff value of COL4-7S measured by CLEIA for detecting stage 2 or higher. Their study showed that COL4-7S demonstrated the highest diagnostic performance with an AUROC of 0.795 compared with FIB-4 and M2BPGi for stage 2 or higher diagnoses, and the cutoff value was 3.8 ng/mL (sensitivity 77%, specificity 68%). Furthermore, they proposed that using COL4-7S in the intermediate group of FIB-4 1.30-2.67 can improve diagnostic performance for identifying patients at risk of stage 2 or higher[52]. Referring to their study, we propose a cutoff value of 4.0 for COL4-7S measured by CLEIA as the second step (Figure 1).

Although there is a large amount of evidence on the evaluation of MASLD fibrosis by COL4-7S measured by RIA, there is still little evidence on the diagnostic performance of COL4-7S measured by CLEIA. In the future it will be necessary to validate whether there is a difference between the RIA and CLEIA methods for COL4-7S values and to verify the actual diagnostic performance of CLEIA for fibrosis (statement 5-4 in Table 3).

Liver apoptosis markers (CK-18F)

Main statements and their summaries (Table 3): 6-2: It serves as a marker of liver activity. 6-3: It reflects treatment effects on liver activity. 6-5: Combined with FIB-4, it improves MASH diagnosis.

CK-18F has been included in insurance coverage in Japan starting in 2024 as a supplementary diagnostic tool for MASLD. CK-18 itself serves as a structural component, functioning as an intermediate filament in epithelial cells where it plays crucial roles in maintaining cellular integrity and mechanical resilience. The apoptosis or necrosis (forms of cell death) of epithelial cells results in the release of CK-18 protein fragments into the bloodstream. These fragments are generated by caspase-mediated cleavage in apoptotic cells. In contrast, CK-18F released during necrosis differs in its fragmentation pattern compared with that released during apoptosis[53].

M30 and M65 antibodies are frequently utilized to quantify CK-18F: (1) M30 antibody. Caspase cleavage of CK-18 exposes a specific epitope that the M30 antibody specifically recognizes. This allows for the measurement of apoptosis-derived CK-18F in serum; and (2) M65 antibody. This antibody detects total CK-18, encompassing both intact CK-18F derived from total cell death, including necrosis. Consequently, it enables the evaluation of the overall CK-18F levels, reflecting both apoptosis and necrosis.

CK-18F is primarily measured in serum or plasma samples and is typically quantified using ELISA. In Japan the M30 antibody-based measurement of CK-18F is covered by insurance and is utilized as an adjunctive diagnostic tool for MASLD with a specific focus on liver apoptosis. It is important to note that CK-18F serves as an activity marker rather than a fibrosis marker. Therefore, accurate diagnosis of MASLD/MASH using CK-18F alone has not been established (statement 6-1 in Table 3). As liver fibrosis progresses to full-blown MASH, inflammation, ballooning, and lipid accumulation decline or disappear, resulting in decreased metabolic-associated steatosis and corresponding reductions in CK-18F levels[54-56]. This distinction emphasizes that CK-18F functions as an activity marker and supports its use in the insurance-approved adjunctive diagnosis of MASH.

For this reason Tada et al[55] evaluated CK-18F levels in patients with a FIB-4 index < 2.67 and established a CK-18F cutoff value of 260 U/L (Figure 2). Similarly, Liebig et al[57] proposed CK-18F levels > 200 U/L as indicative of MASH even in cases where FibroScan and the nonalcoholic fatty liver disease fibrosis score results were low. They further suggested that CK-18F could be diagnostic for MASH even in patients with minimal fibrosis based on NITs. It is widely agreed that combining fibrosis markers, the FIB-4 index, and CK-18F levels enables the diagnosis of at-risk MASH (stage ≥ 2, NAS ≥ 4) and the resolution of MASH (defined as a reduction in NAS ≥ 2 without worsening fibrosis or complete disappearance of MASH) in at-risk individuals[54]. However, the adequacy of CK-18F as a standalone biomarker for identifying individuals at risk for MASH has not been fully elucidated. Furthermore, the extent to which CK-18F captures therapeutic response and reflects improvements in disease activity in MASLD/MASH remains insufficiently characterized, underscoring the need for further prospective investigations (statements 6-2, 3, and 4 in Table 3). At present, CK-18F is useful for diagnosing at-risk MASH, but there is no evidence that it alone is sufficient (statement 6-4 in Table 3).

Evaluation of MASLD using FIB-4 and CK-18F: In the assessment of MASLD, a combination of the FIB-4 index and CK-18F can be utilized to determine the necessity of referral to hepatologists. The evaluation criteria are as follows. In the high-risk group (FIB-4 index ≥ 2.67), patients with FIB-4 ≥ 2.67 should be referred to hepatologists regardless of CK-18F levels as they are at a high risk of advanced fibrosis. In the low-risk group (FIB-4 index < 1.3 & CK-18F < 260 U/L), if FIB-4 < 1.3 and CK-18F < 260 U/L, fibrosis and inflammation are considered minimal and routine follow-up with repeat FIB-4 and CK-18F testing every 1-2 years is recommended. In the intermediate-risk group (FIB-4 index < 1.3 or 1.3-2.66 with CK-18F ≥ 260 U/L), elevated CK-18F suggests active inflammation, which could lead to fibrosis progression. In such cases referral to a hepatologist should be considered. For scores of FIB-4 index 1.30-2.66 with CK-18F < 260 U/L, advanced fibrosis cases such as full-blown MASH could be present although inflammation is low. A second-step evaluation using the ELF test, M2BPGi, COL4-7S, vibration-controlled transient elastography (VCTE), or MRE is necessary to confirm fibrosis status. This stepwise approach optimizes the use of hepatology referrals by identifying patients who require specialized care while ensuring appropriate follow-up for those at lower risk (statement 6-5 in Table 3 and Figure 2).

Imaging diagnosis

Elastography (ultrasound) main statements and their summaries (Table 3): 7-1: VCTE liver stiffness is the standard for fibrosis assessment. 7-3: VCTE stiffness changes can track treatment effects.

VCTE (FibroScan™) measures the propagation velocity of a 50 MHz shear wave emitted into the liver[58]. Due to its noninvasive nature, liver stiffness measurement (LSM) using VCTE has been employed in drug development for MASH[59] and in longitudinal studies to predict liver disease-related events[60].

A 2021 meta-analysis reported that the diagnostic performance for detecting fibrosis stage F2 or higher (37 studies, 2763 cases) had an AUROC of 0.83 with a sensitivity of 80% and specificity of 73%, using a cutoff range of 3.8-10.2 kPa[61]. For stage F3 or higher (44 studies, 4219 cases), the AUROC was 0.85 with a sensitivity of 80% and specificity of 77%, using a cutoff range of 6.8-12.9 kPa[61]. For cirrhosis (22 studies, 337 cases), the AUROC was 0.89 with a sensitivity of 76% and specificity of 88%, using a cutoff range of 6.9-19.4 kPa[61].

The FibroScan AST (FAST) score is a noninvasive index that combines LSM and controlled attenuation parameter from FAST blood level to identify patients with active, fibrotic MASH[62]. A meta-analysis of the FAST score demonstrated that its diagnostic performance for identifying MASH with a high risk of fibrosis progression had an AUROC of 0.79 [95% confidence interval (CI): 0.77-0.81][63]. The sensitivity at the rule-out cutoff (FAST ≤ 0.35) and the specificity at the rule-in cutoff (FAST ≥ 0.67) were both 0.89[63]. The 2023 AASLD guidelines addressed the role of FibroScan in MASLD management, highlighting its utility in excluding cases with significant fibrosis progression, identifying fatty liver disease using controlled attenuation parameter measurements, and assessing high-risk MASH cases through the FAST score[64]. The EASL-EASD-EASO guidelines established threshold values for VCTE in MASLD. The LSM cutoff for excluding advanced fibrosis was set at 8 kPa, whereas the threshold for diagnosing advanced fibrosis was set at 12 kPa (Figure 1). Furthermore, patients with MASLD with an LSM exceeding 10 kPa and a progressive increase in LSM over time have been identified as having a high risk of developing HCC[15].

Currently, VCTE is recognized as a crucial diagnostic tool in not only hepatology but also the management of diabetes, collaboration with primary care physicians[65], and the diagnosis of portal hypertension[66]. More than 160 international guidelines recommend its use. A large-scale multinational cohort study analyzing 16603 patients with MASLD reported that over a median follow-up of 51.7 months, 316 patients (1.9%) developed LREs, including HCC, liver failure, liver transplantation, and liver-related mortality[67]. The Agile score is a noninvasive index that combines LSM with clinical and laboratory variables such as AST, alanine aminotransferase, platelet count, age, sex, and diabetes status to assess the risk of advanced fibrosis or cirrhosis[68]. The study demonstrated that LRE prediction was comparable with or superior to histological diagnosis with Agile scores (Agile 3+ and Agile 4) utilizing VCTE proving highly useful[67]. These findings suggest that ultrasound elastography may surpass liver biopsy as a diagnostic tool.

Point shear wave elastography (pSWE) was first introduced in 2008 as Virtual Touch Quantification™. pSWE utilizes acoustic radiation force impulse technology to apply a focused acoustic push pulse to tissues, inducing a small displacement, and measures the propagation velocity of the resulting shear wave to assess tissue stiffness. A 2021 meta-analysis reported that the diagnostic performance for detecting fibrosis stage F2 or higher (nine studies, 805 cases) had an AUROC of 0.86 (sensitivity 69%, specificity 86%) with a cutoff range of 1.18-1.81 m/second[61]. For F3 or higher (11 studies, 1209 cases), the AUROC was 0.89 (sensitivity 80%, specificity 86%) with a cutoff range of 1.34-4.24 m/second, whereas for cirrhosis (F4) the AUROC was 0.90 (sensitivity 76%, specificity 88%) with a cutoff range of 1.36-2.54 m/second[61].

Two-dimensional shear wave elastography (2D-SWE) was first incorporated into the Aixplorer^® system by SuperSonic Imagine in 2010 and has since been adopted by various ultrasound manufacturers. 2D-SWE generates planar shear waves by continuously applying compression pulses at different depths, inducing vertical oscillations in the tissue. The measured shear wave velocity is overlaid on B-mode images as a color-coded map for liver stiffness evaluation. A 2021 meta-analysis on MASLD reported that the diagnostic performance for F2 or higher (four studies, 488 cases) had an AUROC of 0.75 (sensitivity 71%, specificity 67%) with a cutoff range of 8.3-11.6 kPa[61]. For F3 or higher (four studies, 488 cases), the AUROC was 0.72 (sensitivity 72%, specificity 72%) with a cutoff range of 9.3-13.1 kPa). For cirrhosis (F4) the AUROC was 0.88 (sensitivity 78%, specificity 84%) with a cutoff range of 14.4-15.7 kPa[61].

In the first round of the Delphi survey, the statement that “liver stiffness measured by VCTE can be used as the standard for evaluating liver fibrosis” was supported by a majority (81.8%). However, in the second round, it did not reach a super majority (63.6%) (statement 7-1 in Table 3). This may be attributed to the strong recommendation of MRE as the reference standard. Nevertheless, the use of VCTE for assessing treatment response as well as the suitability of 2D-SWE as an alternative to VCTE achieved super majority consensus in both the first and second rounds of the Delphi process (statements 7-2 and 3 in Table 3).

In conclusion, ultrasound elastography, represented by VCTE, is the most commonly used method for ruling out significant hepatic fibrosis in MASLD management. Additionally, measuring changes in liver stiffness has been reported as an effective method for monitoring disease progression. In August 2025 the United States Food and Drug Administration accepted a letter of intent for the qualification of LSM by VCTE as a reasonably likely surrogate endpoint in clinical trials for non-cirrhotic MASH with moderate-to-advanced fibrosis. This biomarker, which correlates with fibrosis severity and predicts liver-related outcomes, represents a major step toward replacing invasive liver biopsy with NITs in drug development. Currently, pSWE/2D-SWE has not been standardized, and future standardization is expected to provide reliable reference values.

MRE main statements and their summaries (Table 3): 8-2: MRE could become the research gold standard. 8-3: MRE stiffness can predict prognosis. 8-4: The MER combined with the FIB-4 (MEFIB) score efficiently detects F > 2. 8-5: The magnetic resonance imaging (MRI)-AST (MAST) score identifies at-risk MASH.

In 1995 the Mayo Clinic reported MRE as a method for measuring tissue elasticity using MRI. MRE generates shear waves within the liver using an external vibration device (magnetic resonance touch) and measures liver stiffness noninvasively by reflecting the phase changes of these waves as proton phase shifts in MRI results. Comparative studies evaluating the diagnostic performance of MRE and ultrasound elastography have confirmed that MRE is either equivalent or superior in accuracy across all fibrosis stages[69,70]. A 2021 meta-analysis involving 14609 cases reported MRE-LSM cutoff values for fibrosis staging: ≥ F1 at 2.5-3.14 kPa; ≥ F2 at 2.86-4.14 kPa; ≥ F3 at 2.99-4.80 kPa; and F4 at 3.35-6.70 kPa[61]. More recently, a meta-analysis of 798 cases reported even more precise cutoff values demonstrating extremely high diagnostic accuracy: ≥ F1 at 2.65 kPa (AUROC: 0.82); ≥ F2 at 3.14 kPa (AUROC: 0.92); > F3 at 3.53 kPa (AUROC: 0.92); and > F4 at 4.45 kPa (AUROC: 0.94)[71].

Although ultrasound elastography is convenient and has relatively high diagnostic capability, it has limitations such as a high failure rate in obese patients and the inability to assess the entire liver. Although MRE is less convenient, it provides the highest accuracy for liver fibrosis diagnosis and enables a comprehensive evaluation of the entire liver. Liver biopsy, often regarded as the gold standard, suffers from sampling errors. MRE is particularly useful when liver stiffness is heterogeneous as heterogeneity in stiffness can lead to discrepancies with histopathological diagnosis, making the whole-liver assessment provided by MRE advantageous[72]. In addition, MRE-LSM and its longitudinal change (ΔLSM) are both significant predictors of clinical outcomes in MASLD, highlighting the value of serial MRE assessments for prognostication and treatment monitoring[73].

Here, we would like to discuss the statements about MRE and MASLD. MRE-based LSM may become the gold standard in clinical trials and clinical research, but MRE is not suitable as a first-line method in routine clinical practice in my country (statements 8-1 and 2 in Table 3). MRE-based LSM is a useful prognostic factor (statement 8-3 in Table 3, consensus; 100% of respondents selected agree). Although MRE has excellent performance for assessing liver fibrosis, its use as a first-line method is not currently recommended due to the high cost of the equipment. The AASLD and the EASL recommend a two-step diagnostic approach incorporating the FIB-4 index with either ultrasound elastography-based VCTE or the ELF test as a fibrosis biomarker. Meanwhile, Tamaki et al[74] from the University of California, San Diego proposed the MEFIB score, which combines MRE and the FIB-4 index for fibrosis assessment. The MEFIB score is particularly useful in diagnosing significant fibrosis (> F2) with rule-in criteria of MRE-LSM ≥ 3.3 kPa and FIB-4 ≥ 1.6, achieving a PPV of 91.2%. Conversely, rule-out criteria of MRE-LSM < 3.3 kPa and FIB-4 < 1.6 yielded an NPV of 92.8% (Figure 1).

The MEFIB score can efficiently diagnose LSM of F2 or higher (statement 8-4 in Table 3, consensus; 100% of respondents selected agree). Additionally, for diagnosing “at-risk MASH” with significant fibrosis (≥ F2) and severe disease activity (NAS ≥ 4), the MAST and M-PAST scores, which integrate MRE-LSM, proton density fat fraction, and AST, have been reported as useful tools[75,76]. Furthermore, the MAST score has shown strong prognostic utility: Higher MAST values are associated with significantly increased risks of major adverse liver outcomes including decompensation events, HCC, liver transplantation, or liver-related death with HRs of approximately 7.8 for intermediate scores (0.165-0.242) and approximately 22 for high scores (> 0.242) and a concordance statistic of approximately 0.92[77].

The MAST score can efficiently diagnose “at-risk MASH” (statement 8-5 in Table 3, consensus; 100% of respondents selected agree). Due to the Delphi method associated with MRE, there is consensus that its diagnostic ability is extremely high, but there is also strong agreement that its limited accessibility makes it unsuitable as a first-line treatment. Both ultrasound and MRE have their respective advantages and limitations. Therefore, selecting the appropriate modality based on clinical circumstances is crucial to optimizing patient care.

Interactions between NITs and genetic factors main statements and their summaries (Table 3): 9-1: Patatin-like phospholipase domain-containing 3 (PNPLA3) polymorphism promotes fibrosis progression. 9-2: Gene testing helps assess fibrosis risk.

Genetic factors are known to play roles in the histological progression and development of HCC in individuals with MASLD. The single nucleotide polymorphisms (SNPs) in various genes are involved in the onset and progression of MASLD. Genetic testing can be considered one of the NITs that is useful for predicting the prognosis of patients with MASLD. The SNP that results in the substitution of isoleucine with methionine at residue 148 (rs738409 C>G) in the PNPLA3 gene is associated with MASLD through mechanisms likely mediated by hepatocytes and hepatic stellate cells[78].

In hepatic stellate cells the normal function of PNPLA3 (CC) is to hydrolyze retinyl esters to promote extracellular retinol release. However, its activity is reduced in the presence of the I148M mutant, referred to as the PNPLA3 G allele (homozygous GG; heterozygous CG), which leads to profibrotic activity in stellate cells[79,80]. The PNPLA3 GG genotype has allelic ORs of approximately two to three for risks of MASLD, MASH, and HCC[81]. In a meta-analysis of 13817 patients, the allelic OR for MASH was 2.54 (95%CI: 2.03-3.16), and the genotypic ORs were 1.75 (95%CI: 1.24-2.46) for heterozygotes and 4.44 (95%CI: 2.92-6.76) for homozygotes[82].

Thus, although the PNPLA3 gene polymorphism is thought to contribute to histological progression in MASLD, it alone is insufficient to predict cases of advanced fibrosis (statement 9-1 in Table 3). Furthermore, no study has yet reported whether the cutoff values of NITs for predicting advanced fibrosis vary according to genetic polymorphisms. Instead, several risk classification systems for severe liver conditions, including HCC, have been proposed, incorporating combinations of multiple SNPs as well as genetic and metabolic factors. The polygenic risk score, which includes PNPLA3, transmembrane 6 superfamily 2 (TM6SF2), membrane-bound O-acyltransferase domain-containing 7), and glucokinase regulatory protein, has been shown to improve diagnostic accuracy for severe liver disease when combined with NITs such as the FIB-4 index[83]. Additionally, a combined classification system incorporating the FIB-4 index, diabetes status, and the PNPLA3 genotype has been reported[84]. In that study patients with intermediate FIB-4 index scores (1.30-2.67), diabetes, and the PNPLA3 GG genotype had cirrhosis incidence rates similar to those with high FIB-4 index scores (> 2.67).

Seko et al[85] compared these methods and risk classification based on the FIB-4 index combined with PNPLA3/TM6SF2 genotypes in a large Japanese cohort. Classification using PNPLA3/TM6SF2 and polygenic risk score with FIB-4 > 2.67 demonstrated adequate predictive ability for LREs and outcomes. However, in models that included diabetes, there was no significant difference in the risk of developing LREs or prognosis between the low-risk and intermediate-risk groups.

As demonstrated in these studies, genetic polymorphisms are currently used to identify additional high-risk groups for LREs among patients suspected of advanced fibrosis based on NITs. The combination of the FIB-4 index and SNP analysis may provide a means to identify cases with a high likelihood of current fibrosis development as well as those at increased risk for future disease progression (statement 9-2 in Table 3).

Tailoring NITs for MASLD fibrosis assessment: Strategies for health checkups and clinical settings

The following two statements were discussed in this session, neither of which led to agreement. 10-1: NIT cutoffs should be unified across cohorts. 10-2: FIB-4 < 1.3 warrants observation only. NITs have revolutionized the diagnosis and management of MASLD. These tests offer a safer, more accessible alternative to liver biopsy for assessing liver fibrosis. However, the application of NITs must be tailored to different clinical contexts, particularly when comparing health checkup cohorts to hospital-based populations.

The prevalence of advanced fibrosis in MASLD populations varies dramatically between health checkup cohorts and hospital-based cohorts, necessitating distinct diagnostic approaches. Population-based studies, such as Japanese health checkups utilizing VCTE, report advanced fibrosis (stages 3-4) rates of 0.8%-1.3%[86,87], contrasting sharply with the 16.1% prevalence observed in tertiary care cohorts like the Clinical Outcome Nonalcoholic Fatty Liver Disease study (a Japanese biopsy-proven MASLD cohort study)[88]. This disparity affects diagnostic performance. In low-prevalence settings NITs tend to yield higher specificity but lower PPV, making them particularly suitable for ruling out rather than confirming advanced fibrosis[89].

Below are the statements and our thoughts on the results of the discussion. At first we would like to discuss statements that achieved consensus for strategies for health checkups and clinical settings (statement 10-1 in Table 3, consensus; 18.2% of respondents selected agree). A multinational Asian biopsy-proven cohort study (n = 759) by Chan et al[90] demonstrated the feasibility of a tiered approach that sequentially applies a simple serum-based score followed by VCTE to only inpatients falling outside the clearly low-risk range. By modeling various real-world prevalence scenarios (3.7%, 10%, 24%, and 50%), the study showed that such a strategy minimized the number of patients requiring elastography while maintaining acceptable accuracy and that optimal LSM cutoffs differed between screening and clinical populations[90]. These findings illustrate how diagnostic pathways can be adapted to setting-specific disease prevalence, thereby optimizing resource use.

International guidelines, such as those from the AASLD and EASL, propose structured algorithms that integrate NITs into broader metabolic risk assessment frameworks[15,91]. While they converge on the principle of longitudinal monitoring for patients who are low risk, and further assessment for higher-risk cases, specific recommendations on interval, secondary testing choice, and management of indeterminate-risk individuals remain heterogeneous. In Japan expert consensus has yet to fully establish whether low-scoring individuals in community settings should be managed by observation alone or undergo additional noninvasive evaluation, highlighting an important area for future investigation[15,91]. Our study suggests that COL4-7S should be used to diagnose liver fibrosis in this population[26].

Secondly, we would like to discuss the statements that achieved consensus for strategies for health checkups and clinical settings (statement 10-2 in Table 3, consensus; 36.4% of respondents selected agree). This is thought to be due to the poor diagnostic ability of the FIB-4 index in patients with MASLD and T2D as mentioned above, and the fact that patients with MASLD and T2D are a high-risk group. In summary, NIT use in MASLD should be adapted to the setting (whether health screening or hospital care) to improve detection and management of advanced fibrosis, which may help enhance patient outcomes.

FUTURE PERSPECTIVES OF NIT FOR MASLD DIAGNOSIS

Although various NITs are available in Japan, we have shown that the cutoff values for each NIT vary depending on the cohort for which it is used and that there is still insufficient evidence for prognostic prediction. The availability of M2BPGi and COL4-7S as NITs is currently mainly limited to Japan as these tests are only covered by the National Health Insurance System and are not widely available in other countries. It is expected that the accuracy of imaging diagnosis can be improved with the advancement of machines and technology, but the cutoff values of ultrasound elastography vary depending on the model; therefore, standardization will be essential in the future.

It is also expected that diagnostic methods combining NITs will be attempted. To accomplish this it will be necessary to develop methods that can more accurately evaluate pathology and predict prognosis using artificial intelligence. It is expected that a cost-effective diagnostic process will be established when combining these methods. Regarding prognostic prediction, the development of highly accurate NITs is awaited not only for LREs but also for cardiovascular disease events and cancers of other organs. Combination with gene polymorphism analysis, including PNPLA3, is also important.

In addition, the use of mobile health technologies will be necessary to popularize diagnosis using NITs. For example, it is expected that early detection and ongoing management will be facilitated by integrating NITs with daily health management using smartphone apps and wearable devices. In August 2025 the Food and Drug Administration accepted a letter of intent to consider using LSM (VCTE) as a surrogate end point in clinical trials of patients with MASH. This noninvasive method has the potential to replace liver biopsy, leading to more efficient new drug development and reduced patient burden.

Cost-effectiveness is an important consideration in the implementation of NIT-based assessment strategies for MASLD, particularly in settings in which certain biomarkers or imaging modalities are not covered by insurance or are limited in availability due to resource constraints. In such contexts, initial risk stratification using simple, inexpensive tests such as the FIB-4 index, which relies on commonly available laboratory parameters, should be prioritized. In the Japanese biopsy validation cohort, the AUROC of FIB-4 was 0.76-0.85, and in the recent Egyptian ultrasound cohort[92], the AUROC was also high at 0.83, demonstrating the consistent utility of FIB-4 as a “gatekeeper for risk stratification” despite differences in reference standards. However, because the Japanese biopsy cohort uses advanced fibrosis (F3 or higher) as the criterion and the Egyptian cohort uses ultrasound-based steatotic liver diagnosis as the criterion, it is difficult to directly compare the absolute diagnostic performance of the two. When more sophisticated NITs (e.g., M2BPGi, COL4-7S, ELF, or imaging modalities) are unavailable or not reimbursed, a stepwise approach using readily accessible tools is recommended. This strategy ensures broad applicability and cost efficiency, maximizing the number of patients who can benefit from early risk assessment and appropriate referral. The adoption and continuous evaluation of locally feasible NIT algorithms are essential to optimize resource utilization while maintaining diagnostic accuracy.

CONCLUSION

We outlined the currently used NITs for MASLD. While various NITs are available, we recommend a two-step algorithm “starting with the FIB-4 index and then using other NITs” (Figures 1 and 2). With the improvement of the accuracy of NITs, the development of long-term prognostic prediction models will progress and may be useful for early intervention and determining treatment policies. If these prospects are realized, it will be possible to diagnose MASLD early, accurately evaluate its pathology, and provide effective treatment, which is expected to improve patients’ quality of life and reduce medical costs.