Abstract
Background & Aims
Corrected T1 (cT1), measured by liver MRI, enables non-invasive assessment of hepatic water content. Although a threshold of ≥800 ms has been proposed to identify individuals at risk of metabolic dysfunction-associated steatohepatitis (MASH) and adverse outcomes, it identifies only a small proportion of individuals (∼5%) and may fail to capture risk below this threshold. We aimed to evaluate whether cT1 values below 800 ms are associated with mortality and whether hepatic steatosis, assessed by MRI-derived proton density fat fraction (PDFF), modifies this association.
Methods
We analyzed 29,597 UK Biobank participants with liver MRI-derived cT1 measurements. Participants were categorized into three groups (<700, 700–799, and ≥800 ms) based on penalized spline analysis. Outcomes included all-cause mortality, cause-specific mortality, and liver-related events. Cox proportional hazards models were adjusted for demographic, lifestyle, and cardiometabolic factors.
Results
Approximately 40% of participants had cT1 values of 700–799 ms, and 5% had cT1 values ≥800 ms. Over a median follow-up of 5.3 years, participants with cT1 values of 700–799 ms had higher risks of all-cause mortality (adjusted hazard ratio [aHR] 1.21; 95% CI 1.02–1.43) and cardiovascular mortality (aHR 1.54; 95% CI 1.04–2.28) than those with cT1 values <700 ms. Risks were even higher among participants with cT1 values ≥800 ms. Clinically significant hepatic steatosis (PDFF ≥10%) significantly modified the association between cT1 and all-cause mortality (p for interaction <0.05). Higher cT1 was associated with increased all-cause mortality only among participants with PDFF <10% (aHR 1.25 [95% CI 1.06–1.49] for cT1 700–799 ms and 2.20 [95% CI 1.40–3.48] for cT1 ≥800 ms, compared with cT1 <700 ms), but not among those with PDFF ≥10%.
Conclusion
cT1 values between 700 and 799 ms were independently associated with an increased risk of mortality. Notably, this association was modified by the presence of clinically significant hepatic steatosis as assessed by PDFF.
Impact and implications
Using liver MRI to measure cT1, we demonstrated that cT1 values of 700–799 ms, below the conventional threshold of 800 ms, were independently associated with an increased risk of all-cause and cardiovascular mortality. We also found that this association was present only among participants with PDFF <10%, indicating that clinically significant hepatic steatosis modifies the relationship between cT1 and mortality. These findings suggest that lower cT1 thresholds may improve risk stratification and that the prognostic significance of cT1 should be interpreted in the context of hepatic steatosis.
Keywords: Magnetic resonance imaging, Corrected T1, Mortality, Liver-related outcomes, hepatic steatosis
Graphical abstract
Highlights
-
•
Liver MRI-derived cT1 values below the traditional threshold of 800 ms were associated with an increased risk of all-cause mortality.
-
•
Liver MRI-derived cT1 values between 700 and 799 ms were associated with an increased risk of cardiovascular-related mortality.
-
•
Clinically significant hepatic steatosis evaluated by MRI-PDFF significantly modified the association between hepatic cT1 and all-cause mortality.
-
•
Higher cT1 was associated with increased all-cause mortality only in participants with PDFF <10%.
Introduction
Hepatic cT1 was originally developed as a non-invasive MRI biomarker of hepatic fibro-inflammation and fibrosis, reflecting extracellular water accumulation associated with liver injury after correction for iron-related signal effects.1,2 cT1 has been shown to correlate with histological features of liver disease activity, including ballooning, lobular inflammation, and fibrosis,3,4 and a cT1 threshold of 800–825 ms has been recommended to identify the transition from the upper limit of normal to histologically confirmed metabolic dysfunction-associated steatohepatitis (MASH). This threshold has been validated in several studies as a marker of increased liver disease activity.5 Moreover, recent population-based research has shown that individuals with cT1 ≥800 ms are at an increased risk of all-cause mortality.6,7 However, the prognostic significance of intermediate cT1 values below the conventional 800 ms threshold remains unclear. Given that cT1 represents a continuous measure of hepatic tissue injury, risk may increase even within ranges currently considered subthreshold.
Hepatic steatosis may modify the prognostic significance of cT1, as fat deposition within the liver can independently affect T1 relaxation properties and confound quantitative T1 measurements.8 MRI-derived proton density fat fraction (PDFF), a non-invasive quantitative measure of hepatic steatosis, may therefore act as an important effect modifier in the association between cT1 and clinical outcomes. This issue is particularly relevant in metabolic dysfunction-associated steatotic liver disease (MASLD) and MASH, in which substantial hepatic steatosis is common. However, whether the association between cT1 and mortality varies according to steatosis burden has not been systematically investigated.
Therefore, the aim of this study was to assess whether cT1 values below 800 ms are also associated with an increased risk of mortality, and whether this association is modified by hepatic steatosis, as quantified by MRI-derived PDFF.
Patients and methods
Study population
We used a dataset from the UK Biobank, a large, ongoing prospective cohort that consists of over 502,506 participants aged 37-73 years, recruited from across the UK between 2006 and 2010. Since 2014, the UK Biobank imaging project has been collecting brain, heart, bone, and abdominal scans from 100,000 participants. This study was conducted on 69,605 individuals who had undergone a liver MRI scan. Among them, 33,988 individuals with cT1 values indicative of liver disease activity were initially included in the study. Fig. 1 shows the study design as a flow chart.
Fig. 1.
Flowchart of the study design.
CMRF, cardiometabolic risk factor; HCC, hepatocellular carcinoma; LT, liver transplantation; PDFF, proton density fat fraction.
This study complies with the 1975 Declaration of Helsinki and was conducted under the ethical approval granted for UK Biobank studies by the National Health Service (NHS) National Research Ethics Service. The UK Biobank has approval from the North West Multi-Center Research Ethics Committee (ref: 11/NW/0382) and written informed consent was obtained from all participants prior to the study. We downloaded the UK Biobank data in February 2025; the access application number for this study is 132578. As it used a publicly available, de-identified dataset, IRB review was exempted. No identifiable private information was included, and participants cannot be re-identified.
Liver MRI protocol
As part of the UK Biobank imaging project, participants underwent abdominal multiparametric MRI examinations using Siemens Aera 1.5T scanners (Syngo MR D13). The LiverMultiScan image acquisition protocol from Perspectum Ltd (UK) was used as part of the UK Biobank abdominal imaging protocol. Liver MRI data including cT1 and proton density fat fraction (PDFF) were reported by the LiverMultiScan software.1,6 The quantitative assessment of hepatic fat content was evaluated using PDFF,9 with clinically significant hepatic steatosis being defined as a PDFF measurement of ≥10%.10,11 cT1 (Perspectum, Oxford) is a more specific marker of hepatic inflammation or fibrosis that has been developed to standardize for iron levels, magnetic field strength, and manufacturers in the conventional MOLLI (modified look-locker inversion recovery) sequence.1
Clinical, anthropometric, and laboratory examinations
Smoking status was categorized as ever smoker (including current and former smokers), or never smoker. Alcohol consumption was categorized into three groups according to the amount of alcohol intake as follows: 1) mild as alcohol intake <30 g/day for men and <20 g/day for women; 2) moderate as 30-60 g/day for men and 20-50 g/day for women; and 3) severe (>60 g/day for men and >50 g/day for women). Cardiometabolic risk factors (CMRFs) were defined according to the modified NCEP-ATP III criteria:12 (1) elevated blood pressure, defined as blood pressure ≥130/85 mmHg, a diagnosis of hypertension (ICD-10 codes I10–I15), or self-reported use of antihypertensive medication. Blood pressure was measured twice at the time of the MRI scan using either manual or automated methods, and the mean value was used for analysis;2 abdominal obesity, defined as a waist circumference >102 cm in men and >88 cm in women;3 elevated triglycerides, defined as ≥1.7 mmol/L (150 mg/dl);4 low HDL cholesterol, defined as <1.03 mmol/L (40 mg/dl) in men and <1.29 mmol/L (50 mg/dl) in women; and5 elevated fasting glucose, defined as ≥6.1 mmol/L (110 mg/dl). Because triglycerides, HDL cholesterol, and fasting glucose were measured only at the initial UK Biobank assessment, we additionally incorporated medication use at the time of the MRI scan (antihypertensive, lipid-lowering, or insulin therapy) and ICD-10 diagnoses recorded by the time of the MRI scan, including dyslipidemia (E78) and diabetes mellitus (E10–E14).
Height, weight, waist circumference, and systolic and diastolic blood pressure were measured using standardized methods. BMI was calculated as body weight (kg)/height2 (m2). Details on how the UK Biobank collected and processed blood samples are described elsewhere.13
Primary and secondary outcomes
The primary outcome was all-cause mortality, and the secondary outcome was cause-specific mortality, including cardiovascular, cancer-related mortality, and liver-related events. Mortality data were obtained from linked death registries, which are from NHS England for participants in England and Wales, and from the NHS Central Register (part of the National Records of Scotland) for participants in Scotland. We used the mortality dates from these death registries, and defined the censoring date as 31 December 2023 or the earlier date among these two death data sources. For participants lost to follow-up, the censoring date was set as the date of their last follow-up. The cause of death was determined according to the primary cause of death recorded using ICD-10 codes. Cardiovascular death was defined as death due to heart disease (I00–I09, I11, I13, I20–I51) or cerebrovascular disease (I60–I69); cancer-related death as death due to malignant neoplasms (C00–C97), excluding hepatocellular carcinoma (HCC; C22.0); and liver-related death as death due to liver disease (K70–K76) or HCC (C22.0). The liver-related outcome was defined as a composite of incident cirrhosis, liver transplantation, HCC and liver-related mortality. The detailed definition of each event is presented in Table S1. The censoring date for liver-related outcome was defined as the date of disease diagnosis, death, or loss to follow-up, whichever occurred first.
Statistical analysis
Continuous variables were expressed as means (standard deviations) and categorical variables as frequencies (%). To minimize potential bias and retain sample size, variables with missing data were modelled by introducing an explicit 'missing' category, as presented in Table 1. Comparisons between groups were performed using ANOVA for continuous variables and chi-squared tests for categorical variables. The cT1 variable was categorized by applying multiple cut-offs at 20 ms intervals (640, 660, …, 860 ms), creating a binary indicator for values above vs. below each cut-off. For each cut-off point, we estimated the hazard ratios (HRs) and their corresponding 95% CIs for all-cause mortality using Cox proportional hazards models. To model and visualize the relationship between cT1 and mortality risk across the full range of cut-offs, penalized spline (P-spline) functions were used within the Cox regression framework.14,15 The P-spline curve showed that all-cause mortality started to increase significantly when cT1 values approached 700 ms, with a more pronounced rise observed above the previously suggested threshold of 800 ms (Fig. 2). Harrell's C-statistic was also slightly higher at 700 ms, although it did not differ substantially across the values (Table S2). Therefore, 700 ms, which was also close to the median cT1 value, was used to assess overall discriminative ability.
Table 1.
Baseline characteristics according to the cT1 values in all participants.
| Total | cT1 <700 ms | cT1 700-799 ms | cT1 ≥800 ms | p value | |
|---|---|---|---|---|---|
| Prevalence, % | 29,597 | 16,613 (56.1) | 11,474 (38.8) | 1,510 (5.1) | |
| Age, years | 64.0 (7.6) | 63.7 (7.5) | 64.7 (7.7) | 63.2 (7.6) | <0.001 |
| Men, % | 13,601 (46.0) | 6,745 (40.6) | 5,997 (52.3) | 859 (56.9) | <0.001 |
| Smoking status, % (n = 29,375) | <0.001 | ||||
| Never smoker | 18,696 (63.7) | 10,808 (65.6) | 7,038 (61.8) | 850 (56.8) | |
| Former or current smoker | 10,648 (36.3) | 5,659 (34.4) | 4,343 (38.2) | 646 (43.2) | |
| Alcohol consumption, % | <0.001 | ||||
| None or mild | 17,200 (58.1) | 10,217 (61.5) | 6,318 (55.1) | 665 (44.0) | |
| Moderate | 5,340 (18.0) | 3,028 (18.2) | 2,032 (17.7) | 280 (18.5) | |
| Severe | 1,120 (3.8) | 606 (3.7) | 430 (3.8) | 84 (5.6) | |
| Missing | 5,937 (20.1) | 2,762 (16.6) | 2,694 (23.5) | 481 (31.9) | |
| BMI, kg/m2 | 26.3 (4.3) | 25.0 (3.4) | 27.4 (4.4) | 31.5 (5.1) | <0.001 |
| Waist circumference, cm | 87.7 (12.5) | 83.9 (10.7) | 91.3 (12.5) | 102.2 (12.5) | <0.001 |
| SBP, mmHg | 138.9 (18.6) | 137.2 (18.6) | 140.6 (18.5) | 144.4 (17.5) | 0.014 |
| DBP, mmHg | 78.6 (10.0) | 77.7 (9.9) | 79.3 (10.0) | 82.8 (9.8) | 0.668 |
| Number of CMRFs | <0.001 | ||||
| 0-2 | 19,631 (66.3) | 12,822 (77.2) | 6,362 (55.5) | 447 (29.6) | |
| 3-5 | 9,966 (33.7) | 3,791 (22.8) | 5,112 (44.6) | 1,063 (70.4) | |
| Townsend deprivation index∗ | -1.89 (2.73) | -1.97 (2.69) | -1.82 (2.74) | -1.51 (2.92) | <0.001 |
| IPAQ questionnaire | <0.001 | ||||
| Mild | 2,698 (9.1) | 1,194 (7.2) | 1,225 (10.7) | 279 (18.5) | |
| Moderate to high | 22,478 (76.0) | 13,063 (78.6) | 8,456 (73.7) | 959 (63.5) | |
| Missing | 4,421 (14.9) | 2,356 (14.2) | 1,793 (15.6) | 272 (18.0) | |
| Fasting glucose, mmol/L∗ | 5.0 (0.9) | 4.9 (0.8) | 5.0 (1.0) | 5.2 (1.3) | <0.001 |
| HbA1c, mmol/mol∗ | 34.9 (5.1) | 34.2 (4.2) | 35.6 (5.7) | 37.3 (7.0) | <0.001 |
| Total cholesterol, mmol/L∗ | 5.75 (1.07) | 5.78 (1.04) | 5.73 (1.10) | 5.64 (1.15) | <0.001 |
| Triglyceride, mmol/L∗ | 1.61 (0.93) | 1.43 (0.80) | 1.79 (1.01) | 2.22 (1.16) | <0.001 |
| HDL, mmol/L∗ | 1.49 (0.38) | 1.58 (0.38) | 1.40 (0.35) | 1.24 (0.29) | <0.001 |
| AST, U/L∗ | 25.5 (9.6) | 24.8 (8.9) | 26.1 (10.3) | 29.3 (11.7) | <0.001 |
| ALT, U/L∗ | 22.6 (13.4) | 20.4 (11.4) | 24.4 (14.1) | 33.1 (20.8) | <0.001 |
| PDFF, % | 4.8 (4.7) | 3.1 (1.9) | 5.8 (4.7) | 16.3 (8.2) | <0.001 |
| PDFF ≥10% | 3,197 (10.8) | 229 (1.4) | 1,835 (16.0) | 1,133 (75.0) | <0.001 |
Data are presented as means (standard deviation) for continuous variables or number (%) for categorical variables. ANOVA for continuous variables and chi-squared test for categorical variables were used in this analysis.
ALT, alanine aminotransferase; AST, aspartate aminotransferase; CMRF, cardiometabolic risk factor; DBP, diastolic blood pressure; HDL, high-density lipoprotein; SBP, systolic blood pressure.
Variables were assessed at UK Biobank enrollment because corresponding data were unavailable at the image evaluation visit.
Fig. 2.
P-spline curve for all-cause mortality.
HRs for all-cause mortality according to cT1 values estimated using a penalized spline (P-spline) Cox regression model. The blue line represents the estimated HR, and the shaded area indicates the 95% CI. The red dashed line denotes HR = 1 (reference). cT1, corrected T1; HR, hazard ratio.
Survival analyses for three groups were performed using the Cox proportional hazards model, which was adjusted for confounders including age, sex, smoking, alcohol consumption, CMRF, Townsend deprivation index, and physical activity. The proportional hazards assumption was evaluated using Schoenfeld residual-based tests. The performance of the three-category cT1 classification (<700, 700–799, and ≥800 ms) was compared with that of the conventional binary classification (<800 vs. ≥800 ms) using Harrell's C-index, likelihood ratio tests, and the integrated discrimination improvement (IDI).16,17 Sensitivity analyses were also performed with a different cut-off value for cT1 and with cT1 as a continuous variable. All analyses were conducted using Stata version 17.0 (StataCorp, College Station, TX, USA), and a p value <0.05 was considered statistically significant.
Results
Baseline characteristics
The baseline characteristics of the three groups in the study population are presented in Table 1. The mean age was 64.0 years, and 46.0% of participants were male. Of the 29,597 participants, 11,474 (38.8%) had cT1 values of 700–799 ms and 1,510 (5.1%) had cT1 values of ≥800 ms. Participants with higher cT1 values were more likely to be male and to have ever smoked. They also had a higher BMI and waist circumference, and a significantly greater number of accompanying CMRFs. Baseline characteristics according to PDFF values are presented in Table S3.
All-cause, cause-specific mortality and liver-related events according to cT1 value
The proportional hazards assumption was not violated (global test p = 0.855 for all participants; p = 0.583 for participants with PDFF <10%; and p = 0.398 for participants with PDFF ≥10%). The numbers of all-cause deaths, cause-specific deaths and liver-related events are presented in Table S4. Over a median follow-up of 5.3 years, the cT1 700–799 ms group exhibited a significantly higher risk of all-cause mortality (adjusted hazard ratio [aHR] 1.21, 95% CI 1.02–1.43, p = 0.028), cardiovascular mortality (aHR 1.54, 95% CI 1.04–2.28, p = 0.030), and a marginally increased risk of liver-related events (aHR 1.94, 95% CI 0.97–3.89, p = 0.063) compared to the <700 ms group (see Fig. 3 and Table 2). The cT1 ≥800 ms group showed higher risks of all-cause mortality (aHR 1.84, 95% CI 1.31–2.59, p <0.001) and liver-related event (aHR 3.09, 95% CI 1.01–9.44, p = 0.047) compared to those with cT1 <700 ms.
Fig. 3.
Kaplan-Meier curve for all-cause mortality according to the cT1 value.
cT1, corrected T1.
Table 2.
Multivariate analysis for all-cause, cause-specific mortality, and liver-related outcome risk according to the cT1 values and PDFF values.
| Total | All-cause mortality |
Cardiovascular mortality |
Cancer mortality |
Liver-related outcome |
||||
|---|---|---|---|---|---|---|---|---|
| aHR (95% CI) | p value | aHR (95% CI) | p value | aHR (95% CI) | p value | aHR (95% CI) | p value | |
| cT1 <700 ms | 1 (Reference) | 1 (Reference) | 1 (Reference) | 1 (Reference) | ||||
| cT1 700-799 ms | 1.21 (1.02-1.43) | 0.028 | 1.54 (1.04-2.28) | 0.030 | 1.21 (0.97-1.51) | 0.084 | 1.94 (0.97-3.89) | 0.063 |
| cT1 ≥800 ms | 1.84 (1.31-2.59) | <0.001 | 1.84 (0.86-3.94) | 0.116 | 2.02 (1.27-3.22) | 0.003 | 3.09 (1.01-9.44) | 0.047 |
| PDFF <10% | ||||||||
| cT1 <700 ms | 1 (Reference) | 1 (Reference) | 1 (Reference) | 1 (Reference) | ||||
| cT1 700-799 ms | 1.25 (1.06-1.49) | 0.010 | 1.63 (1.08-2.45) | 0.019 | 1.23 (0.98-1.54) | 0.070 | 2.04 (0.98-4.23) | 0.056 |
| cT1 ≥800 ms | 2.20 (1.40-3.48) | 0.001 | 3.88 (1.66-9.08) | 0.002 | 1.68 (0.84-3.34) | 0.139 | 4.84 (1.05-22.24) | 0.043 |
| PDFF ≥10% | ||||||||
| cT1 <700 ms | 1 (Reference) | 1 (Reference) | 1 (Reference) | 1 (Reference) | ||||
| cT1 700-799 ms | 0.58 (0.28-1.21) | 0.148 | 0.42 (0.12-1.51) | 0.184 | 1.24 (0.28-5.40) | 0.778 | 0.72 (0.08-6.17) | 0.764 |
| cT1 ≥800 ms | 0.80 (0.37-1.70) | 0.560 | 0.26 (0.06-1.14) | 0.075 | 2.31 (0.52-10.14) | 0.269 | 1.01 (0.11-9.13) | 0.990 |
Values in bold indicate statistical significance (p <0.05). cT1, corrected T1; HR, hazard ratio; MASH, metabolic dysfunction-associated steatohepatitis; MASLD, metabolic dysfunction-associated steatotic liver disease; PDFF, proton density fat fraction.
Model was adjusted for age, sex, smoking, alcohol, cardiometabolic risk factors, Townsend deprivation index, physical activity (and PDFF 10% in total model, not in subgroup model).
Effect modification of hepatic steatosis on the association of cT1 and all-cause mortality
Clinically significant hepatic steatosis (PDFF ≥10%) significantly modified the association between hepatic cT1 and all-cause mortality (p for interaction = 0.039 for cT1 700–799 ms and 0.036 for cT1 ≥800 ms). Subgroup analysis according to the presence of clinically significant hepatic steatosis produced distinct associations. In the group with PDFF <10%, cT1 was associated with a higher risk of all-cause mortality in a dose-dependent manner (aHR 1.25, 95% CI 1.06–1.49, p = 0.010 in the cT1 700-799 ms group and aHR 2.20, 95% CI 1.40-3.48, p = 0.001 in the cT1 >800 ms group compared to the cT1 <700 ms group). However, cT1 was not associated with all-cause mortality in the PDFF ≥10% group (p >0.05 in both the cT1 700-799 ms and cT1 ≥800 ms groups; Table 2). Similarly, cT1 was associated with an increased risk of cardiovascular mortality in participants with PDFF <10% (aHR 1.63, 95% CI 1.08–2.45, p = 0.019 in the cT1 700-799 ms group and aHR 3.88, 95% CI 1.66-9.08, p = 0.002 in the cT1 ≥800 ms group compared to the cT1 <700 ms group), but not in participants with PDFF ≥10%. The risk of liver-related outcomes was approximately twofold higher in participants with cT1 values of 700–799 ms and 4.8-fold higher in those with cT1 values ≥800 ms, but only among participants with PDFF <10%, although these associations were only marginally significant.
Comparison of prognostic performance between 2- and 3-category models and sensitivity analysis
The three-category model improved model fit (likelihood ratio test p = 0.028) and discrimination compared with the binary model (IDI = 0.09), despite minimal change in C-index. The intermediate group (cT1 700–799 ms) showed a higher 5-year mortality risk compared with the low-risk group (<700 ms), corresponding to an absolute risk difference of 0.72% and a number needed to screen of 139. Subgroup analyses stratified by PDFF revealed a consistent pattern of effect modification (Table S5).
Sensitivity analyses using cT1 in three different categories (cT1 <680, 680–799 and ≥800 ms) and as a continuous variable (Tables S6 and S7) showed similar results. cT1 was only associated with increased all-cause mortality in participants with PDFF <10% and was not associated with increased all-cause mortality in participants with PDFF ≥10%. Compared with cT1 <680 ms, both the 680–799 ms and ≥800 ms groups demonstrated a higher risk of all-cause mortality, indicating a dose–response relationship across increasing cT1 categories.
Discussion
Our large population-based cohort study demonstrated that elevated liver MRI-derived cT1 values (700-799 ms), even below the conventional threshold of 800 ms, were associated with increased risks of all-cause mortality and cardiovascular mortality. Clinically significant hepatic steatosis, defined as MRI-PDFF ≥10%, modified the association between cT1 and all-cause mortality. cT1 was associated with all-cause mortality in a dose-dependent manner in participants with PDFF <10%, however cT1 was not associated with all-cause mortality in participants with PDFF ≥10%.
cT1 by multi-parametric MRI has been studied as an imaging biomarker to identify patients with MASH at high risk of disease progression.7 A cT1 cut-off value of 800-825 ms, suggested as an optimal rule-out threshold for MASH, and as a predictor of failure to maintain remission in autoimmune hepatitis,18 was associated with increased all-cause mortality and cardiac events.7 However, there had been no previous study evaluating the long-term prognosis of individuals not meeting the criteria of cT1 >800 ms. Our study showed that individuals with cT1 700-799 ms are at an increased risk of all-cause mortality and cardiovascular mortality, suggesting that cT1 may be useful for risk stratification of adverse clinical outcomes. Regarding liver-related outcomes, cT1 700-799 ms was associated with a marginally increased risk of liver-related events. However, liver-related outcomes should be interpreted cautiously. Although we used a composite liver-related outcome as an alternative endpoint because of the very low rate of liver-related mortality, the number of liver-related events remained limited, precluding definitive conclusions regarding liver-specific risk.
In this study, the prognostic value of cT1 was predominantly observed among individuals with low hepatic fat content (PDFF <10%), suggesting that clinically significant hepatic steatosis modifies the relationship between cT1 and all-cause mortality. cT1 is thought to reflect extracellular water content, and in PDFF <10%, it may provide a relatively specific measure with minimal confounding. In contrast, in individuals with PDFF ≥10%, the association between cT1 and mortality was attenuated, which may be explained by several mechanisms. First, PDFF ≥10% may introduce biological heterogeneity, encompassing a spectrum from simple steatosis to steatohepatitis, thereby reducing the specificity of cT1 for clinically meaningful liver disease activity. Second, steatosis itself may influence T1 relaxation properties,8 potentially leading to signal contamination and diminished discriminatory capacity of cT1. Finally, in PDFF ≥10%, mortality risk may be more strongly driven by systemic cardiometabolic factors rather than liver-specific changes, further diluting the prognostic contribution of cT1. Several recent reports have also shown that metabolic risk burden, rather than hepatic steatosis itself, is associated with long-term outcome.[19], [20], [21], [22] Collectively, these findings support the notion that hepatic fat content acts as an effect modifier, with cT1 providing greater prognostic discrimination of all-cause and cardiovascular mortality in those with PDFF <10%. In contrast, the prognostic performance of cT1 was limited in those with PDFF ≥10%. Although approximately 90% of the study population had PDFF <10%, the use of cT1 in individuals with clinically significant hepatic steatosis (PDFF ≥10%) should be interpreted with caution, particularly in patients with MASLD or MASH.
This study has several limitations. First, the follow-up period for mortality and liver-related outcomes was relatively short (median 5.3 years), which may limit the ability to fully capture longer-term outcomes. Second, the generalizability of our findings is limited by the predominantly White and healthier-than-average UK Biobank population, as well as by potential selection bias arising from the MRI sub-cohort. External validation is needed. Third, this study lacks histologic data; thus, we were unable to show how cT1 values of 700-799 ms correlate with specific histologic changes in the liver. Fourth, cT1 measurements may be subject to residual variability across scanners and acquisition protocols despite standardization, particularly in multicenter settings. Also, cT1 reflects a composite of extracellular water changes and is not biologically specific, with elevations arising from inflammation, fibrosis, or hepatic congestion rather than a single process. However, the use of cT1 reduces the confounding effect of hepatic iron, improving the reliability of tissue characterization. Furthermore, evaluation of cT1 in conjunction with PDFF also enhances interpretability and partially addresses its limited specificity for individual pathological processes.2
In conclusion, this study demonstrated that liver MRI-derived cT1 values of 700–799 ms, falling below the traditional threshold of 800 ms, are independently associated with increased risks of all-cause mortality and cardiovascular mortality. Hepatic steatosis significantly modified this association: in participants with PDFF <10%, cT1 was associated with all-cause mortality in a dose-dependent manner, whereas no significant association was observed in those with PDFF ≥10%. Further studies are warranted to elucidate the underlying mechanisms, and to validate these findings across diverse populations and clinical settings.
Abbreviations
cT1, corrected T1; CMRF, cardiometabolic risk factor; HR, hazard ratio; HCC, hepatocellular carcinoma; IDI, integrated discrimination improvement; LR, likelihood ratio; MASH, metabolic dysfunction-associated steatohepatitis; MASLD, metabolic dysfunction-associated steatotic liver disease; NHS, National Health Service; PDFF, proton density fat fraction; P-spline, penalized spline; WC, waist circumference.
Authors' contributions
MK, JDY contributed to the study conception and design, acquisition, interpretation of the data and drafting the manuscript. AMA, HK and YW contributed to the data acquisition, analysis, interpretation of the data and critical revision of manuscript for important intellectual content. MSR, YW, LW, WA, AK, PM, HT, AV, YW, SP, DL, SCL contributed to the interpretation of data and critical revision of the manuscript.
Data availability
This research has been conducted using the UK Biobank Resource under application number 132578. The data are available upon application to the UK Biobank (http://www.ukbiobank.ac.uk/), subject to approval.
Financial support
No financial support was received to produce this manuscript.
Conflicts of interest
Dr. Ju Dong Yang reports the following conflicts of interest: consulting service for AstraZeneca, Eisai, Exact Sciences, and Fujifilm Medical Sciences. The other authors have no conflicts of interest to declare.
Please refer to the accompanying ICMJE disclosure forms for further details.
Acknowledgements
This research has been conducted using the UK Biobank Resource under application number 132578.
Footnotes
Author names in bold designate shared co-first authorship
Supplementary data to this article can be found online at https://doi.org/10.1016/j.jhepr.2026.101927.
Supplementary data
The following are the Supplementary data to this article:
References
- 1.Alkhouri N., Beyer C., Shumbayawonda E., et al. Decreases in cT1 and liver fat content reflect treatment-induced histological improvements in MASH. J Hepatol. 2025;82:438–445. doi: 10.1016/j.jhep.2024.08.031. [DOI] [PubMed] [Google Scholar]
- 2.Banerjee R., Pavlides M., Tunnicliffe E.M., et al. Multiparametric magnetic resonance for the non-invasive diagnosis of liver disease. J Hepatol. 2014;60:69–77. doi: 10.1016/j.jhep.2013.09.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Dennis A., Kelly M.D., Fernandes C., et al. Correlations between MRI biomarkers PDFF and cT1 with histopathological features of non-alcoholic steatohepatitis. Front Endocrinol (Lausanne) 2020;11 doi: 10.3389/fendo.2020.575843. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Imajo K., Tetlow L., Dennis A., et al. Quantitative multiparametric magnetic resonance imaging can aid non-alcoholic steatohepatitis diagnosis in a Japanese cohort. World J Gastroenterol. 2021;27:609–623. doi: 10.3748/wjg.v27.i7.609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Andersson A., Kelly M., Imajo K., et al. Clinical utility of magnetic resonance imaging biomarkers for identifying nonalcoholic steatohepatitis patients at high risk of progression: a multicenter pooled data and meta-analysis. Clin Gastroenterol Hepatol. 2022;20:2451–2461 e2453. doi: 10.1016/j.cgh.2021.09.041. [DOI] [PubMed] [Google Scholar]
- 6.Roca-Fernandez A., Banerjee R., Thomaides-Brears H., et al. Liver disease is a significant risk factor for cardiovascular outcomes - a UK biobank study. J Hepatol. 2023;79:1085–1095. doi: 10.1016/j.jhep.2023.05.046. [DOI] [PubMed] [Google Scholar]
- 7.Jackson E., Dennis A., Alkhouri N., et al. Cardiac and liver impairment on multiorgan MRI and risk of major adverse cardiovascular and liver events. Nat Med. 2025;31:2289–2296. doi: 10.1038/s41591-025-03654-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Roberts N.T., Tamada D., Muslu Y., et al. Confounder-corrected T(1) mapping in the liver through simultaneous estimation of T(1) , PDFF, R2∗, and B1+ in a single breath-hold acquisition. Magn Reson Med. 2023;89:2186–2203. doi: 10.1002/mrm.29590. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Gnatiuc Friedrichs L., Trichia E., Aguilar-Ramirez D., et al. Metabolic profiling of MRI-measured liver fat in the UK Biobank. Obesity (Silver Spring) 2023;31:1121–1132. doi: 10.1002/oby.23687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Caussy C., Alquiraish M.H., Nguyen P., et al. Optimal threshold of controlled attenuation parameter with MRI-PDFF as the gold standard for the detection of hepatic steatosis. Hepatology. 2018;67:1348–1359. doi: 10.1002/hep.29639. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.EASL-EASD-EASO Clinical Practice Guidelines on the management of metabolic dysfunction-associated steatotic liver disease (MASLD) J Hepatol. 2024;81:492–542. doi: 10.1016/j.jhep.2024.04.031. [DOI] [PubMed] [Google Scholar]
- 12.Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III) JAMA. 2001;285:2486–2497. doi: 10.1001/jama.285.19.2486. [DOI] [PubMed] [Google Scholar]
- 13.UK Biobank . Companion document to accompany serum biomarker data [Internet] 2019. https://biobank.ctsu.ox.ac.uk/ukb/ukb/docs/serum_biochemistry.pdf Oxford: UK Biobank. [cited 2026 Jul 20]. Available from: [Google Scholar]
- 14.Strasak A.M., Lang S., Kneib T., et al. Use of penalized splines in extended Cox-type additive hazard regression to flexibly estimate the effect of time-varying serum uric acid on risk of cancer incidence: a prospective, population-based study in 78,850 men. Ann Epidemiol. 2009;19:15–24. doi: 10.1016/j.annepidem.2008.08.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Palmer C.R., Bellinge J.W., Dalgaard F., et al. Association between vitamin K(1) intake and mortality in the Danish Diet, Cancer, and Health cohort. Eur J Epidemiol. 2021;36:1005–1014. doi: 10.1007/s10654-021-00806-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Harrell F.E., Jr., Lee K.L., Mark D.B. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996;15:361–387. doi: 10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4. [DOI] [PubMed] [Google Scholar]
- 17.Pencina M.J., D'Agostino R.B., Sr., Steyerberg E.W. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med. 2011;30:11–21. doi: 10.1002/sim.4085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Heneghan M.A., Shumbayawonda E., Dennis A., et al. Quantitative magnetic resonance imaging to aid clinical decision making in autoimmune hepatitis. EClinicalMedicine. 2022;46 doi: 10.1016/j.eclinm.2022.101325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kwak M., Kim H.S., Jiang Z.G., et al. MASLD/MetALD and mortality in individuals with any cardio-metabolic risk factor: a population-based study with 26.7 years of follow-up. Hepatology. 2025;81:228–237. doi: 10.1097/HEP.0000000000000925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Huang S.C., Su T.H., Tseng T.C., et al. All-cause and cause-specific mortality in patients with chronic hepatitis B and concurrent steatotic liver disease. J Hepatol. 2025;83:43–51. doi: 10.1016/j.jhep.2024.12.009. [DOI] [PubMed] [Google Scholar]
- 21.Ioannou G.N. MASLD and non-liver-related mortality: association, independent association and causality. J Hepatol. 2025;83:611–614. doi: 10.1016/j.jhep.2025.04.020. [DOI] [PubMed] [Google Scholar]
- 22.Younossi Z.M., Paik J.M., Stepanova M., et al. Clinical profiles and mortality rates are similar for metabolic dysfunction-associated steatotic liver disease and non-alcoholic fatty liver disease. J Hepatol. 2024;80:694–701. doi: 10.1016/j.jhep.2024.01.014. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
This research has been conducted using the UK Biobank Resource under application number 132578. The data are available upon application to the UK Biobank (http://www.ukbiobank.ac.uk/), subject to approval.