Patients with MASH experience worse health-related quality of life: a multinational real-world study using LiverMultiScan

Abstract

Purpose

Prevalence of metabolic dysfunction-associated steatotic liver disease (MASLD) and its progressive form, metabolic dysfunction-associated steatohepatitis (MASH), is increasing globally. Assessment of health-related quality of life (HRQoL), alongside non-invasive tests (NITs), may support identification of patients requiring closer clinical management. This multinational real-world randomised controlled trial evaluated associations between HRQoL (using the EQ-5D) and multiparametric MRI (mpMRI) markers of liver disease.

Methods

RADIcAL-1 included 802 patients (453 male) with suspected MASLD/MASH from Germany, Netherlands, Portugal and UK. Participants completed the EQ-5D-5 L at baseline, 2, 6 and 12-months. EQ-VAS was derived to assess overall health. Between-group comparisons (imaging vs. standard-of-care) were assessed using t-tests, and multivariable logistic regression adjusted for age, obesity, diabetes and hyperlipidaemia. MASH was defined as corrected T1 (cT1) ≥800ms and liver fat as proton density fat fraction (PDFF) ≥ 10%. High-risk MASH was cT1 ≥ 875ms.

Results

Within the imaging arm, patients with MASH reported lower HRQoL across all EQ-5D dimensions and EQ-VAS compared with MASLD. Baseline utilities ranged from 0.81 (Coimbra) to 0.90 (Ulm) and this range decreased at month 12 from 0.82 (Coimbra and UK) to 0.90 (Ulm). These findings remained after adjustment for comorbidities. Patients with high-risk MASH reported up to 12% more complaints in all EQ-5D dimensions compared to all others. Increasing cT1 was associated with 29% higher odds of reporting anxiety/depression. An 80ms increase in cT1 was associated with 0.023-point decrease in EQ-5D utility (p = 0.039). Due to COVID-19, 12-month response rates varied across sites (22–81%).

Conclusion

mpMRI-defined MASH is associated with worse HRQoL independent of key metabolic comorbidities. These findings support the use of imaging biomarkers for disease stratification and for identifying patients with greater clinical burden, with potential implications for management and monitoring strategies.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11136-026-04420-8.

Plain language summary

Currently, assumed utility decrements may be used in economic modelling and decision making. Therefore, there is a need to establish relationships between non-invasive test outputs and health related quality of life (HRQoL) in liver disease. This study presents the HRQoL from a large, randomised control trial that took place in several European countries. Across all jurisdictions, self-care and usual activities had the lowest HRQoL scoring, whilst pain/discomfort had the highest. There were no differences between reported HRQoL in the study arms, however within the imaging arm, those with metabolic dysfunction-associated steatohepatitis (MASH) and therefore higher corrected T1 values, reported a lower HRQoL compared to those with metabolic dysfunction-associated steatotic liver disease (MASLD). These findings provide important evidence linking non-invasive tests to patient reported outcomes that may be incorporated into economic evaluations and clinical risk stratification pathways in the MASLD/MASH disease space.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11136-026-04420-8.

Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD), steatotic liver disease with the presence of at least one cardiometabolic risk factor without harmful alcohol consumption, encompasses a spectrum of diseases ranging from simple steatosis, to metabolic dysfunction-associated steatosis (MASH), and MASH with advanced fibrosis and cirrhosis [1]. MASLD is associated with major non-communicable diseases, including obesity, type 2 diabetes (T2DM), cardiovascular disease and hepatocellular carcinoma [1], and is increasingly recognised as a leading indication for liver transplantation [2, 3]. Patients with MASLD are typically asymptomatic until advanced stages, creating challenges for early diagnosis and potentially resulting in negative implications for patient quality of life [4]. Consequently, the economic burden of MASLD is projected to reach €334 billion by 2034 in Europe alone [5].

An important consideration for notified bodies and organisations responsible for national guidance, technology appraisal and clinical guideline development is cost-effectiveness. Cost-effectiveness is typically expressed as the incremental cost-effectiveness ratio (ICER) per quality-adjusted life year (QALY), which incorporates both survival and health-related quality of life (HRQoL) [6, 7]. Generic instruments such as the EQ-5D-3L and EQ-5D-5L are commonly used to derive QALYs [8], providing a standardised approach to comparing the value of interventions across disease areas. As a result, robust and disease-relevant HRQoL data are critical inputs for health technology assessment and reimbursement decision-making.

In parallel, there has been increasing adoption of non-invasive and imaging-based technologies to support the diagnosis and management of MASLD. MRI-based approaches, including magnetic resonance elastography (MRE) and multiparametric MRI (mpMRI), are increasingly used in clinical practice to characterise liver disease and support risk stratification [9–11]. Recent evidence has shown that incorporation of mpMRI into clinical pathways is cost-effective with an ICER of €4,968 per QALY gained [12]. However, previous studies, including Stevenson et al. (2012) [13], which informed National Institute for Health and Care Excellence (NICE) assessments such as NG49 for non-invasive evaluation of NAFLD (now MASLD), have used assumed utility decrements rather than direct measurement. While appropriate for modelling purposes, this approach may not fully capture the relationship between disease activity and patient-reported outcomes. There is therefore a need to establish direct relationships between non-invasive test (NIT) outputs and HRQoL in real-world populations. Such evidence would enable empirically derived utility inputs, strengthening future health economic evaluations and technology appraisals.

In this study, we evaluate data from RADIcAL-1, a large multinational European cohort of patients with suspected MASLD/MASH, to investigate the association between mpMRI-derived biomarkers and HRQoL. As part of the study objectives, we characterised the HRQoL across different jurisdictions and explored the relationship between HRQoL and disease severity.

Methods

Study population and design

RADICAL-1, a multinational European prospective, randomised controlled clinical trial (RCT) included 802 patients aged between 18 and 75 years with suspected MASLD between September 2017 and June 2020 from specialist MASLD outpatient clinical centres and T2DM specialist centres in Ulm (Germany), Leiden (Netherlands), and Coimbra (Portugal) and the UK (Dundee, Glasgow, Liverpool, London, and Southampton) [14]. Patients were randomised using a 1:1 allocation to either the control arm (SoC) or intervention arm (SoC + mpMRI). Randomisation was automatically calculated using a random number generator based on the inclusion criteria for the study. Based on the nature of randomisation (MRI vs. SoC), allocation concealment could not take place.

Study criteria were as described by Tonev and colleagues [14]. To be eligible for inclusion, patients needed to have either (1) increased liver enzyme tests (ALT, AST, or GGT ≥ 1.5 x upper limit of normal, and ALT, AST ≤ 5 x upper limit of normal) up to 1 year prior to patient recruitment, or, (2) imaging suggestive of steatotic liver disease up to three years prior to patient recruitment (no limitations were put on imaging modality), or, (3) presence of three or more of the following: T2DM, obesity (BMI > 30 kg/m²), hypertension (≥ 130/85 mmHg), elevated triglycerides (≥ 1.7 mmol/l), or low HDL-cholesterol (< 1.05 mmol/l for men and < 1.25 mmol/l for women). Exclusion criteria included contraindications to MRI, pregnancy, liver transplant, other proven liver disease, alcohol-related liver disease, heart failure and severe mental illness [14].

Multiparametric MRI acquisition and imaging analysis

All centres implemented a standardised multiparametric MRI protocol (LiverMultiScan, Perspectum Ltd, Oxford, UK) which is standardised across scanners and field strengths [15]. Non-contrast abdominal scans (15 min) were acquired, including shMOLLI and multi-echo sequences to quantify cT1, T2*, and PDFF. Images were anonymised and analysed centrally by trained analysts blinded to clinical data. Whole liver cT1 and PDFF maps were generated using semi-automated whole slice segmentation with exclusion of non-parenchymal structures and artefacts [14].

Health related quality of life assessment

All participants, irrespective of randomisation arm, had their HRQoL assessed using the EuroQol-five-dimensions, five-level (EQ-5D-5L) questionnaire [16] at four time points (baseline, 2, 6, and 12-months). For the EQ-5D-5L scores range from having no problems (1), slight problems (2), moderate problems (3), severe problems (4), and inability to perform tasks or having extreme pain, anxiety/depression (5). For the EQ-VAS, scoring ranges between 0 (worst health) to 100 (best health). The intended patient follow-up was 12 months, however due to the impact of COVID-19 and associated restriction, some patients recruited during this period were followed up for a minimum of six months. Due to hospital-based restrictions, follow up had to be done remotely over the telephone or by mail as opposed to the in person follow up that had been taken place originally.

To allow for cross-country comparisons of self-reported HRQoL, EQ-5D-5L responses were converted into utility scores, that represent the overall HRQoL of a patient. Additionally, EuroQol visual analogue scale (EQ-VAS) was also used alongside the EQ-5D questionnaire to report overall HRQoL. For direct calculation of utility scores between jurisdictions, official EQ-5D-5L tariffs were used for Germany, the Netherlands, and Portugal [17–19], whilst official EQ-5D-3L tariffs were used for the UK.

Statistical analysis

Descriptive statistics are presented as mean and standard deviations for continuous variables, median and interquartile ranges for ordinal variables, and as counts with percentages for categorical variables. Between-group comparisons (imaging vs. SoC) were performed using Student’s t-test, Wilcoxon rank-sum test or Kruskal–Wallis test for continuous variables, as appropriate, and Pearson’s chi-squared test for categorical variables. Disease burden was characterised by the presence of MASLD-related comorbidities, with prevalence reported across all jurisdictions.

To enable cross-country comparisons, EQ-5D-5L responses were mapped to EQ-5D-3L index values (UK) using the van Hout crosswalk method [20]. The UK has recently (March 2026) released official UK EQ-5D-5L tariffs [21]. For comparison, utility scores were also calculated using these recently introduced tariffs. EQ-VAS scores were used as a measure of self-reported overall health status, with comparisons between jurisdictions conducted using t-tests across all timepoints. Missing data can introduce bias and reduce the same size available for analysis. Where appropriate, sensitivity analyses were conducted when more than 10% of the data were missing. Data missing due to being unreachable during the COVID-19 pandemic or because of changing contact details would be unlikely to introduce bias [22].

Covariates were selected based on clinical relevance within the study population [5]. Commonly occurring characteristics were prioritised because they had greater potential to influence population-level HRQoL estimates and were sufficiently represented to allow reliable adjustment, while limiting model complexity and reducing the risk of overfitting [23, 24]. Sex, obesity (OBS), other chronic diseases (OCD), treatment status (TRT), hyperlipidaemia (HPL), hypertension (HPT), diabetes (T2DM), age and liver fat content (PDFF) were included to adjust for their potential independent effects on the outcome and thereby reduce confounding of the associations between the primary markers of interest and the outcome. Average marginal effects (AMEs) were derived to quantify the change in predicted EQ-5D utility associated with a one-unit increase in each covariate. Exploratory analysis evaluating the association between cT1 and EQ-5D utility index for each jurisdiction were performed using a beta regression model.

A multivariable logistic regression model was used to identify predictors of diagnostic certainty for MASLD/MASH, including obesity, cT1, PDFF, and sex. Adjusted odds ratios (ORs) with corresponding 95% confidence intervals were derived.

In the imaging group, MASH was defined as cT1 ≥ 800 ms and PDFF ≥ 10%, whilst high-risk MASH was defined as cT1 ≥ 875 ms. This is aligned with studies investigating the clinical utility of mpMRI for identifying patients with MASH [25] A PDFF threshold of 10% was used in-line with identification of more sever disease [26, 27].

All statistical tests were two-sided, with a significance threshold of p < 0.05. Analyses were conducted using RStudio (version 2025.05.0 + 496).

Results

Patient characteristics and co-morbidity profile

Participants recruited into this study were aged 53 years (ranging 18–75) with the majority (> 50%) being male (Table 1). There were significant differences in the biochemical serum markers of liver and metabolic health across all jurisdictions. Participants recruited in Portugal (Coimbra) had significantly lower liver function test markers (Albumin, ALT and AST) but the highest HbA1C concentration when compared to those recruited from Germany (Ulm), the Netherlands (Leiden) and the UK. In addition to having the highest prevalence of T2DM, those recruited in Portugal also had the highest prevalence of hypertension and hyperlipidaemia (Table 1; Fig. 1). In this cohort, patients were on a wide range of treatment with Coimbra (Portugal) and Leiden (Netherlands) having the highest proportions (88% and 85% respectively), whilst UK (64%) and Ulm (Germany, 41%) had the lowest (Supplementary Table 2).

Table 1.

Demographic summary and baseline health related quality of life

	Ulm	Leiden	Coimbra	UK
	N = 216	N = 177	N = 154	N = 255	p-value
Demographics
Age (years)	50 (13)	51 (13)	61 (9)	53 (13)	< 0.001
Sex (n, Female (%))	97 (45%)	73 (41%)	79 (51%)	100 (39%)	0.10
On treatment (n, %)	88 (41%)	150 (85%)	135 (88%)	162 (64%)	< 0.001
Biochemistry
Albumin (g/L)	47.4 (19.1)	46.2 (4.7)	38.9 (16.3)	43.6 (7.0)	< 0.001
AST (U/L)	36.3 (20.4)	35.9 (20.2)	26.8 (15.0)	37.8 (21.2)	< 0.001
ALT (U/L)	55.6 (37.9)	53.5 (34.0)	34.5 (27.2)	55.5 (47.2)	< 0.001
HbA1C (mmol/mol)	41.2 (10.1)	42.0 (12.4)	56.0 (18.9)	51.8 (17.6)	< 0.001
Comorbidities
Type 2 diabetes (n, %)	48 (22%)	36 (20%)	103 (67%)	35 (14%)	0.009
Hypertension (n, %)	92 (43%)	63 (36%)	112 (73%)	148 (58%)	< 0.001
Obesity (n, %)	71 (33%)	53 (30%)	87 (56%)	139 (55%)	< 0.001
Hyperlipidaemia (n, %)	144 (67%)	75 (42%)	112 (73%)	82 (32%)	< 0.001
Other chronic diseases (n, %)	167 (77%)	80 (45%)	132 (86%)	99 (39%)	< 0.001
Health-Related Quality of Life
EQ-5D-5L dimension levels
Mobility	1 (1, 1)	1 (1, 1)	1 (1, 3)	1 (1, 2)	< 0.001
Self-Care	1 (1, 1)	1 (1, 1)	1 (1, 1)	1 (1, 1)	< 0.001
Usual Activities	1 (1,1)	1 (1, 2)	1 (1, 2)	1 (1, 2)	< 0.001
Pain/Discomfort	2 (1, 2)	2 (1, 2)	2 (1, 3)	2 (1, 3)	0.007
Anxiety/Depression	1 (1,1)	1 (1, 2)	1 (1, 2)	1 (1, 2)	< 0.001
EQ-VAS scores	75 (16)	73 (14)	70 (15)	72 (17)	0.008

Fig. 1.

A Prevalence of reported comorbidities in those with suspected MASLD/MASH. B Demographic differences in age, imaging and blood markers between sites. cT1: Corrected T1, LFC: Liver fat content, AST: aspartate aminotransferase, ALT: alanine aminotransferase

Overall, across jurisdictions, T2DM had the lowest prevalence (ranging 23% – 67%) whilst hyperlipidaemia had the highest (ranging 32% – 73%) (Fig. 1). Those recruited from Portugal and the UK had the highest prevalence of hypertension and obesity, whilst those from Portugal and Germany had the highest prevalence of hyperlipidaemia and other chronic diseases (including hyperuricaemia [52%], inflammatory disease [8%], coronary heart disease [8%], cancer [6%], renal disease [5%], stroke [2%], and celiac disease [1%]).

Health-related quality of life

As official EQ-5D-5L tariffs are available for Germany, the Netherlands, and Portugal [17–19], but tariffs for the EQ-5D-3L were available for the UK, to allow for comparisons, all EQ-5D responses were mapped to the England EQ-5D-3L tariff using the van Hout cross-walk method [20].

Evaluation of EQ-5D dimensions showed that across all jurisdictions, self-care and usual activities consistently had the lowest scores whilst pain/discomfort had the highest (Table 1). Between jurisdiction analyses showed significant differences in baseline utilities across all jurisdictions, with the most differences observed between Portugal and other jurisdictions (Fig. 2). Overall, 95% of patients completed the HRQoL questionnaire at baseline (range across jurisdictions: 91%-99%), with 56%-95% completing it at subsequent time points (Table 2). The response rate declined over the course of the study, with the lowest response rates occurring in month 12 (22%-81%).

Fig. 2.

Baseline (A) EQ-5D (mapped using English tariffs) and (B) Mean EQ-VAS HRQoL scores across jurisdictions

Table 2.

EQ-5D utilities, EQ-VAS scores mapped to English tariffs (EQ-5D-3 L) and response rates across jurisdictions

	Coimbra N = 154	Leiden N = 177	UK N = 255	Ulm N = 216	p-value
EQ-5D utilities
Baseline	0.81 (0.19)	0.86 (0.18)	0.83 (0.21)	0.90 (0.12)	< 0.001
Month 2	0.82 (0.19)	0.84 (0.18)	0.85 (0.20)	0.89 (0.12)	0.002
Month 6	0.81 (0.19)	0.86 (0.16)	0.87 (0.17)	0.88 (0.14)	0.008
Month 12	0.82 (0.19)	0.85 (0.17)	0.82 (0.21)	0.90 (0.18)	0.07
EQ-VAS scores
Baseline	70 (16)	75 (15)	71 (20)	76 (17)	0.003
Month 2	70 (18)	72 (17)	73 (20)	74 (19)	0.16
Month 6	70 (17)	73 (16)	72 (20)	76 (19)	0.06
Month 12	70 (16)	74 (17)	72 (20)	80 (16)	0.005
Health-related quality of life response rate
Baseline (%)	153 (99)	163 (92)	253 (99)	197 (91)
Month 2 (%)	132 (86)	139 (79)	149 (58)	171 (79)
Month 6 (%)	128 (83)	146 (82)	157 (62)	100 (46)
Month 12 (%)	124 (81)	128 (72)	124 (49)	48 (22)

Although the number of patients with available HRQoL data decreased over time, reported HRQoL remained broadly similar across the assessed time points. Similar to that observed at baseline, across all subsequent timepoints (2, 6 and 12-months) EQ-5D utilities were significantly different across all jurisdictions (Table 2). Assessment of longitudinal changes in quality of life over a 12-month period showed that HRQoL utilities remained broadly stable with minor non-statistically significant fluctuations across all jurisdictions (Table 2). However, HRQoL utilities remained lowest in those recruited from Portugal. Likewise, when looking at the overall self-reported health status, EQ-VAS, HRQoL was lowest in Portugal (Table 2; Fig. 2). Investigation using the recently released UK EQ-5D-5L tariffs showed similar results to those obtained using the EQ-5D-3L, with only minor differences (increase of ~ 0.01) between scores from the different sites for each timepoint (Supplementary Table 1).

Relationship between markers and HRQoL

Logistic regression analyses evaluating associations between the covariates and the odds of reporting problems in each EQ-5D domain showed that higher cT1 was significantly associated with impairment in the anxiety/depression and usual activities domains. Specifically, each 80-ms increase in cT1 was associated with 29% higher odds of reporting anxiety/depression problems (OR 1.29, 95% CI 1.01–1.61; p = 0.039) and 46% higher odds of reporting problems with usual activities (OR 1.46, 95% CI 1.13–1.90; p = 0.004) (Fig. 3).

Fig. 3.

Relationship between corrected T1 (cT1) and change in health-related quality of life (HRQoL) assessed using the EQ-5D questionnaire. cT1: Corrected T1, MASH: metabolic dysfunction-associated steatohepatitis

Obesity was also associated with higher odds of reporting problems with self-care (OR 11.06, 95% CI 2.18–56.15; p = 0.004) and pain/discomfort (OR 1.90, 95% CI 1.02–3.52; p = 0.043), although the confidence interval for the self-care estimate was wide, suggesting potential instability. Treatment status was associated with anxiety/depression, with participants receiving any form of treatment having 69% lower odds of reporting problems than those not receiving treatment (OR 0.31, 95% CI 0.16–0.60; p < 0.0001). Participants with type 2 diabetes also had higher odds of reporting anxiety/depression problems than those without type 2 diabetes (OR 2.37, 95% CI 1.14–4.91; p = 0.020) (Fig. 4).

Fig. 4.

Logistic regression showing Odds ratios of investigated markers for each EQ-5D domain

Multivariable linear regression, adjusted for age, PDFF, sex, obesity, hyperlipidaemia, hypertension, type 2 diabetes, other chronic diseases, and treatment status, showed that higher cT1 was independently associated with lower EQ-5D utility. Specifically, each 80-ms increase in cT1 was associated with a 0.023-point reduction in EQ-5D utility (p = 0.039) (Fig. 5). Treatment status was also significantly associated with utility, with participants receiving no treatment reporting utility scores that were, on average, 0.055 points lower than those receiving treatment (p = 0.047). No significant associations were observed for sex, obesity, other chronic diseases, hyperlipidaemia, hypertension, type 2 diabetes, age, or PDFF.

Fig. 5.

Forest plot showing impact of variables on overall self-reported quality of life using the EQ-5D utility. cT1: Corrected T1, PDFF: Proton Density Fat Fraction

High-risk MASH and HRQoL

The greatest reductions in HRQoL were observed across progressive disease categories, particularly between low-risk MASH and MASH, and between MASH and high-risk MASH. Although HRQoL varied across jurisdictions (Fig. 2), mixed-effects modelling demonstrated minimal between-jurisdiction variation in the association between cT1 and EQ-5D utility after adjustment for age, PDFF, obesity, hypertension, hyperlipidaemia, and other chronic diseases (Supplementary Fig. 1).

Patients with high-risk MASH, who comprised 31% of the study population, represent the principal target group for therapeutic intervention. Compared with patients without high-risk MASH, those with high-risk MASH reported consistently poorer HRQoL across all EQ-5D domains (Fig. 6). Specifically, they reported 4.7% more anxiety/depression complaints, 11.1% more mobility problems, 8% more pain/discomfort, 4% more self-care difficulties, and 12% more impairment in usual activities (Fig. 6; Supplementary Fig. 2). These differences were statistically significant in the mobility and usual activities domains.

Fig. 6.

Differences in prevalence of reported issues in health-related quality of life EQ-5D domains between those classified as high-risk MASH (cT1 ≥ 875ms) compared to all others

Discussion

In this multinational European cohort of patients with suspected MASLD/MASH, we observed significant associations between imaging-derived markers of disease activity and health-related quality of life. Specifically, higher cT1 values, representing elevated liver disease activity, were associated with lower EQ-5D utility scores and increased impairment across several HRQoL domains, particularly mobility, anxiety/depression, and usual activities. Importantly, these associations were observed across jurisdictions, despite differences in baseline utility scores and cardiometabolic burden. Collectively, these findings suggest that imaging biomarkers may capture clinically meaningful aspects of disease burden that extend beyond histological or biochemical abnormalities alone.

Many economic evaluations in MASLD rely on assumed utility decrements across fibrosis stages rather than empirically measured utility data [28, 29]. By directly evaluating the relationship between quantitative MRI biomarkers and utility-based HRQoL measures suitable for health economic modelling, this study provides important evidence for future economic evaluations, particularly as non-invasive imaging technologies become increasingly integrated into MASLD care pathways [1]. This is particularly relevant as studies have consistently reported reduced HRQoL in patients with MASLD, especially in those with worsening fibrosis, metabolic burden, and adverse body composition, reinforcing the systemic nature of the disease beyond liver pathology alone [30, 31]. These observations align with the present study, where obesity, hypertension, and T2DM were highly prevalent across jurisdictions and likely contributed to reduced HRQoL, particularly in the Portuguese cohort where the multimorbidity burden was greatest.

The observed relationship between cT1 and HRQoL is clinically important. Although MASLD is often considered asymptomatic in its early stages, increasing evidence suggests that disease progression is associated with meaningful physical and psychological burden even prior to decompensated cirrhosis [31, 32]. In this study, increasing cT1 was associated with worsening anxiety/depression and impairment in usual activities, suggesting that liver disease activity may reflect broader systemic processes affecting patient wellbeing. Furthermore, given that cardiovascular disease and extrahepatic malignancy remain the leading causes of mortality in patients with MASLD [33, 34], the relationship between increasing cT1 and worsening HRQoL may reflect broader systemic disease activity rather than liver disease severity alone. Population-based studies support this interpretation and have demonstrated that elevated cT1 is associated with adverse cardiovascular outcomes, including a 1.3-fold increased risk of cardiovascular hospitalisation and a 2.2–2.3-fold increased risk of major adverse cardiovascular events, alongside increased all-cause mortality [35, 36]. Collectively, these findings suggest that cT1 captures underlying fibro-inflammatory disease activity and metabolic processes linked to multisystem disease burden. Therefore, mpMRI likely has utility beyond liver disease characterisation alone, supporting broader cardiometabolic risk stratification in high-risk populations.

From a public health perspective, the observed impact on HRQoL in patients with MASLD without cirrhosis reinforces the concept that clinically meaningful patient burden occurs earlier in the disease spectrum and may carry important economic implications. According to NHS England health trend data, ~ 16,500 cardiovascular deaths in individuals aged under 75 years were considered preventable in 2024 [37], alongside ~ 345 premature deaths from MASLD in the same population [38]. Earlier identification of high-risk patients using non-invasive biomarkers may therefore have implications extending beyond liver-specific outcomes, particularly in populations with obesity, T2DM, and metabolic syndrome where cardiovascular risk is substantial. This is particularly relevant in the emerging therapeutic era of MASLD/MASH, where patients with high-risk MASH represent the primary target population for pharmacological intervention and recently approved therapies such as resmetirom now play a role in patient management [39]. In the MAESTRO-NASH clinical trial, improvements in HRQoL outcomes were observed alongside histological improvements, suggesting that successful treatment and disease regression may translate into measurable patient-reported benefits [40]. In parallel, budget impact models comparing different NITs have demonstrated the potential advantages of incorporating MRI technologies into management pathways [41], whilst previous studies have shown that screening for MASH in obese and T2DM populations may achieve favourable cost-effectiveness, particularly when targeting patients at higher risk of advanced disease [42]. The present study adds further evidence by demonstrating that an imaging biomarker associated with disease activity is also associated with measurable differences in HRQoL, strengthening the biological and economic plausibility of incorporating such technologies into care pathways. Supporting this, real-world studies have demonstrated the utility of mpMRI in improving diagnostic confidence, informing clinical management decisions, supporting biopsy avoidance, and facilitating risk stratification in patients with MASLD/MASH [9, 10, 43].

HRQoL values in the general population vary by age and geographical setting but typically range from 0.85 to 0.95 [44]. In the present study, baseline EQ-5D utility values among participants with MASLD ranged from 0.81 to 0.90, while EQ-VAS scores ranged from 70 to 76. These utility values are comparable to those reported in patients with type 2 diabetes for whom a mean utility of 0.73–0.85 has been reported [45], and higher than those reported in patients receiving treatment for stable metastatic breast cancer, where utility values range from 0.72 with a corresponding gain of + 0.07 following a treatment response and a decline by 0.27 for disease progression [46]. However, poorer HRQoL was significantly associated with greater MASLD disease activity. Specifically, each 100-ms increase in cT1 was associated with a 0.02 reduction in EQ-5D utility, indicating that greater fibro-inflammatory disease activity was associated with poorer HRQoL. To contextualise the magnitude of this association, a study assessing recovery in HRQoL following different types of fracture reported utility improvements of approximately 0.005 following wrist fracture, 0.05 following hip fracture, and 0.11 following vertebral fracture over 12–18 months [47]. Similarly, Stevenson et al. assumed a QALY decrement of 0.03 associated with a missed diagnosis in their analyses [13]. Although these estimates are not directly equivalent to the utility difference observed in the present study, they provide context for its potential clinical and economic relevance. These findings suggest that increasing MASLD disease activity may be associated with a clinically relevant deterioration in HRQoL across multiple EQ-5D dimensions. Compared with participants at low risk of disease, those with at-risk MASH reported 4.7% more anxiety/depression problems, 11.1% more mobility problems, 8.0% more pain/discomfort, 4.0% more self-care difficulties, and 12.0% more impairment in usual activities.

This study has several strengths. First, it included a large, well-characterised multinational cohort recruited across multiple European jurisdictions, improving the generalisability of the findings across different healthcare settings and patient populations. Second, the use of quantitative MRI biomarkers alongside validated preference-based HRQoL measures enabled the relationship between objective disease activity and patient-reported outcomes to be evaluated directly. Furthermore, the consistency of the associations across jurisdictions and after adjustment for relevant clinical and cardiometabolic covariates supports the robustness of the findings. However, there are some limitations to consider. First, MASLD/MASH is a chronic condition with gradual evolution over many years, therefore, the follow-up period of 12 months may not fully capture long-term changes in HRQoL associated with disease progression and would need to be externally validated in real-world conditions. Second, the study was conducted prior to the availability of approved pharmacological therapies for MASH. Consequently, the HRQoL data presented here primarily reflect the natural history of disease and standard clinical management using non-invasive imaging rather than treatment-associated improvements that may be investigated using both histological and non-invasive endpoints. Third, historically many UK studies relied on EQ-5D-3L crosswalk approaches due to the absence of an approved EQ-5D-5L valuation set at the time of the study. In our analysis, although the impact was minimal, application of the newly released UK EQ-5D-5L value set resulted in only modest increases in utility values. Fourth, the study took place in European countries only, and therefore the generalisability of results may vary based on geographic differences in clinical practice. The impact of COVID-19 may also impact the generalisability of results given the response rate to study specific questionnaires.

In summary, in this large multinational European cohort, higher cT1 was associated with poorer HRQoL across multiple domains. Higher cT1 values were independently associated with lower EQ-5D utility scores and greater impairment in mobility, anxiety/depression, and usual activities. These findings provide important evidence linking a quantitative imaging biomarker with patient-reported outcomes. Whilst there is the need for external validation and real world clinical evidence, these findings support the use and incorporation of imaging-derived biomarkers within health economic evaluations and clinical risk stratification pathways in MASLD/MASH.

Authors’ contributions

Recruitment, data collection and data entry oversight: MHM, CDB, JP, SA, KA, JB, MJC, JJS, AF, MCB, SB, DJC, MDY, DF, HJL. Data analysis: ES, MU. Statistical Methodology: ES, MU, AJT, MH. Manuscript drafting: ES. Manuscript review and editing: ES, MH, MU, AJT, DT, MHM, CDB, JP, SA, KA, JB, MJC, JJS, AF, MCB, SB, DJC, MDY, DF, HJL, AJ. All authors reviewed, discussed, and agreed with the manuscript.

Funding

This paper presents independent research titled Non-invasive Rapid Assessment of NAFLD Using Magnetic Resonance Imaging with LiverMultiScan (RADIcAL1) supported by a European Union’s Horizon 2020 SME Instrument Phase 2 Program under grant agreement No 719445.

Data availability

The data and analytic methods used in this study remain the property of the study sponsors. All deidentified participant data may be made available to other researchers upon request following permission, investigator support and following a signed data access agreement.

Declarations

Compliance with ethical standards

As this study was a multicentre randomised controlled trial with human participants, ethics approval was required in each country involved in the study. Relevant institutional approvals were obtained from the following committees with associated reference numbers prior to study commencement:

Data sharing statement

The data and analytic methods used in this study remain the property of the study sponsors. All deidentified participant data may be made available to other researchers upon request following permission, investigator support and following a signed data access agreement.

Competing interests

ES, MH, AJ, are employees of Perspectum Ltd. AJT is a consultant for Perspectum. CDB has received research funding from Echosens and was supported in part by the Southampton National Institute for Health Research Biomedical Research Centre (NIHR203319), U.K.All other co-authors have no conflicts of interest to declare relevant to this work.