T1/T2 mapping as a non-invasive method for evaluating liver fibrosis based on correlation of biomarkers: a preclinical study

Abstract

Background

Given the inherent limitations of invasive biopsy and the insufficient accuracy of liver-related serum biomarkers, there is an urgent need for the development of reliable, non-invasive imaging techniques for the diagnosis of liver fibrosis. This study aims to investigate the correlation between magnetic resonance imaging (MRI) T1/T2 mapping sequences and biomarkers of collagen deposition and ongoing systemic inflammation, and to evaluate the potential of T1/T2 mapping as a non-invasive method for the accurate diagnosis of liver fibrosis.

Methods

A mouse model of carbon tetrachloride (CCl₄)-induced liver fibrosis was established and T1/T2 mapping were performed at different weeks of treatment. The histopathological analysis, collagen quantification, and inflammatory factors measurements (IL-1, IL-6, TNF-α) were conducted to correlate MRI parameters with collagen deposition and inflammation. Statistical analysis was performed using IBM SPSS Statistics (version 22.0, Chicago, IL, USA) and Origin 2018 (OriginLab Corporation, Northampton, MA, USA).

Results

The principal findings indicated that T1 and T2 values exhibited a progressive increase with the severity of fibrosis, demonstrating a positive correlation with collagen deposition and inflammatory factors, especially the hydroxyproline content (r = 0.880, P < 0.001). The HYP content exhibited a progressive increase with advancing fibrosis stages (ρ = 0.914, P < 0.001). Similarly, T1 values increased significantly across fibrosis stage(ρ = 0.854, P < 0.001). Statistical comparison of these coefficients revealed no significant difference (Z = 1.031, P = 0.303). ROC curve analysis showed that T1 mapping was more accurate than T2 mapping in detecting collagen deposition and inflammation.

Conclusions

This study highlighted the potential of T1/T2 mapping as non-invasive and quantitative biomarkers for diagnosing and staging liver fibrosis, providing new insights into the onset and progression of liver fibrosis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-025-03701-9.

Introduction

Liver fibrosis (LF) is a pathological change caused by the excessive deposition of collagen-based extracellular matrix (ECM) resulting from chronic liver injury [1]. The fibrotic response disrupts normal hepatic architecture and function, ultimately impairing liver regeneration and leading to organ dysfunction [2]. Timely intervention can potentially reverse liver damage and prevent progression to cirrhosis and hepatocellular carcinoma [3, 4]. Therefore, the development of accurate diagnostic and monitoring strategies for LF is essential [5].

Currently, invasive biopsy has several inherent limitations, including the potential for complications and the risk of sampling bias, which could lead to an inaccurate representation of the overall liver condition, especially for heterogeneous diseases such as LF [6–8]. While liver-related serum biomarkers, including but not limited to interleukin-1 (IL-1), interleukin-6 (IL-6), tumor necrosis factor -α (TNF -α), as well as other markers such as hyaluronic acid and tissue inhibitor of metalloproteinases-1, have shown practical application value, they often exhibit limited sensitivity and specificity in clinical practice [9–12]. This can result in false-negative or false-positive results, which may hinder the accurate diagnosis. With the development of non-invasive magnetic resonance imaging (MRI), T1/T2 mapping sequences have become an important tool for evaluating fibrosis.

The T1/T2 mapping sequences provide a quantitative assessment of tissue relaxation properties, offering superior precision compared with conventional qualitative imaging techniques [13, 14]. Qiu [15] concluded that there is a robust correlation between the T1 mapping and the severity of fibrosis in cholestatic liver disease. Alexander [16] indicated that the T2 mapping may prove more efficacious in specific scenarios such as patients with hepatitis C. In the worsening of LF, changes in the molecular environment of liver tissue, including alterations in water content, tissue architecture, and the presence of macromolecules like collagen, significantly influence T1/T2 relaxation times [17–19]. Specially, the hydroxyproline (HYP) content is a well-established biomarker for evaluating collagen deposition, serving as a direct indicator of fibrosis severity [20–22].

Despite aforementioned researches about the use of T1/T2 mapping for staging LF, there is a necessity for further explore its potential for capturing the complex pathological changes. This study aims to investigate the association between T1/T2 mapping relaxation times with ongoing systemic inflammation and collagen deposition, and evaluate the potential of T1/T2 mapping as a non-invasive method for accurately diagnosing and staging LF.

Materials and methods

Animals

All experimental protocols were conducted within Anhui Medical University guidelines for animal research and were approved by the Animal Experiment Ethics Review of Anhui Medical University (Hefei, China) (Approval ID: LLSC20200977). After referring to the AVMA animal euthanasia guidelines (2020), mice were sacrificed by cervical dislocation. Healthy male C57BL/6 mice (6 weeks of age, ~ 12 g) were purchased from Animal Experiment Center of Anhui Medical University (Approval ID: SCXK (wan) NO.202010418). The mice were maintained under a controlled light–dark cycle in air-conditioned rooms at a temperature of 26 ℃, with free access to food and water.

Liver fibrosis mouse model

The specific experimental process is depicted in Fig. 1. mice were randomly allocated into experimental groups using a computer-generated randomization table to ensure unbiased distribution. CCl4 0.5 ml/kg, 10% dissolved in olive oil was administered i.p. 2 times per week of age for up to 6 weeks. Subsequent MRI scans were performed at predefined intervals (2 weeks, 4 weeks and 6 weeks post-induction) to monitor disease progression. Immediately after the MRI scan, mice were euthanized, and liver tissues were harvested for histopathological analysis and hydroxyproline (HYP) quantification. The time points of 2, 4 and 6 weeks were selected on the basis of previous studies [23] indicating a dynamic transition from initial inflammatory response to more severe chronic fibrosis changes in CCl₄-induced models. The control mice were administered equivalent doses of saline solution and proceeded through the same imaging and histopathological analysis (n = 3 per time point).

Fig. 1.

Experimental design: liver fibrosis was induced by administering intraperitoneal injections of a 10% CCl₄-olive oil suspension at a dose of 0.05 ml/10 g body weight, twice weekly for 6 weeks. The time points for analysis were set at weeks 2, 4, and 6 (n = 10 per time point), with pathological changes, extracellular protein expression, and imaging of different fibrosis stages observed accordingly. These intervals allow us to capture the dynamic changes in T1/T2 values and the biomarkers of ongoing systemic inflammation and collagen deposition in different stages of liver fibrosis

MRI protocol

Mice were fasted prior to imaging, and their chest and abdominal regions were secured with tape to minimize respiratory motion artifacts. Anesthesia was induced with an intraperitoneal injection of 0.5 mg/10g sodium pentobarbital. MRI was performed on a 1.5 T scanner (Ingenia, Philips Healthcare, Best, The Netherlands) using an 8-channel, 3 cm mouse coil. To ensure consistent slice positioning, a three-step anatomical localization protocol was followed: initial scout imaging (coronal and sagittal T2-weighted) to identify the hepatic dome and inferior liver margin; plane alignment with the axial imaging plane parallel to the portal vein bifurcation; and target slice selection at the confluence of the middle hepatic vein and inferior vena cava, verified with orthogonal plane (coronal and sagittal) alignment. The MRI sequences used were:T2-weighted imaging: Repetition time (TR) = 660 ms, echo time (TE) = 80 ms, FA = 90°, Slices = 10, slice thickness = 2mm. T1 mapping: TR = 4.1 ms, TE = 2 ms, FA = 90°, Slices = 5, slice thickness = 2mm. T2 mapping: TR = 2000, TE=13 ms, FA = 25°, Slices = 5, slice thickness = 2mm. Details of the MRI sequences and parameters are provided in Additional file 1. All the mice were humanely sacrificed after the scan.

Imaging analysis

Quantitative T1/T2 mapping was reconstructed using the IntelliSpace™ Portal (Philips) workstation. T1 maps were generated using the Modified Look-Locker Inversion recovery (MOLLI) sequence, and the pixel-wise T1 values were calculated based on the Messroghli equation [24]. Three-parameter nonlinear curve fitting using a Levenberg–Marquardt algorithm [24] was performed for

$$\text{y}=\text{A}-\text{B exp}\left(-\text{t}/\text{T}1_{}^{\ast }\right)$$ 1

for corresponding ROIs and for corresponding pixels. In Eq. (1), y denotes signal intensity, and T1* corresponds to the apparent, modified T1 in an LL experiment. The algorithm iteratively minimizes the residuals between the observed signal and the modeled inversion recovery curve. Outlier rejection was used to exclude pixels with poor signal-to-noise ratios or motion artifacts, and a confidence threshold was applied to reduce the impact of partial volume effects and magnetic field inhomogeneities.

T2 maps were calculated by fitting multi-echo TSE signals to a mono-exponential model using a least squares nonlinear algorithm [25]. The analysis focused on the largest homogeneous liver region surrounding the middle hepatic vein-inferior vena cava junction. Three standardized ROIs (8.0 mm² each) were positioned according to: The left lobe ROI is centered in the left lateral lobe; the middle lobe ROI is placed bilaterally to the middle hepatic vein; the right lobe ROI is located in the right lateral lobe. The ROIs were delineated independently by two radiologists (one with 15 years of experience in abdominal imaging and the other with four years of experience, respectively) who were blinded to the histopathological results. The ROIs exclude hepatic blood vessels, biliary structures, liver edges, and artifacts. The average T1/T2 values from three ROIs were used for final analysis.

Histopathological analysis

Liver tissues were sectioned in the axial plane using the bread-loafing technique, mimicking MRI slice appearances. All sections were fixed in 4% paraformaldehyde and embedded in paraffin. Corresponding liver regions were identified using anatomical landmarks, and wedge-shaped sections were obtained. Samples were stained with hematoxylin and eosin (H&E), Masson trichrome, and Sirius Red, following by analyzing via light microscopy. Two pathologists, blinded to imaging results and tissue acquisition time points, independently evaluated the specimens. Pathological staging was determined according to the METAVIR fibrosis score [25, 26], a 5-point scale ranging from F0 (no fibrosis) to F4 (cirrhosis).

Biomarkers quantification

The hepatic HYP content was quantified from liver hydrolysate using spectrophotometry, according to the HYP Content Detection Kit protocol (Solarbio, Beijing, China). Results are expressed in micrograms per gram of liver tissue (μg/g). Total protein was extracted from liver tissue using RIPA lysis buffer and separated by SDS-PAGE. All blot images were cropped before hybridization with antibodies to highlight the target protein and its corresponding signal. Inflammatory factors(IL-1, IL-6, and TNF-α) were quantified using Real-time quantitative PCR (qRT-PCR), and their protein levels were validated by Western blot analysis [27] using antibodies against α-SMA, IL-1, IL-6, and TNF-α, with β-actin serving as an internal control for calculation of fold changes.

Statistical analysis

A power analysis was conducted prior to the study to determine the appropriate sample size needed to detect significant differences in T1/T2 mapping and the HYP content across different fibrosis stages. Statistical analysis was performed using IBM SPSS Statistics (version 22.0, Chicago, IL, USA) and Origin 2018 (OriginLab Corporation, Northampton, MA, USA). All parameters were expressed as mean ± standard deviation. Measurement repeatability was evaluated using the intraclass correlation coefficient (ICC), with values ≥ 0.75 indicating good consistency. For normally distributed data with homogeneous variance, one-way ANOVA followed by an LSD post hoc test was used. For non-normally distributed data, the Kruskal–Wallis test followed by the Nemenyi test was performed. The correlation between MRI parameters, fibrosis staging, and biomarkers were assessed using Pearson and Spearman correlation coefficients, with statistical differences calculated via Fisher r-to-z transformation, providing a more accurate measure of the relationship between these variables. The receiver operating characteristic (ROC) curve analysis was conducted to evaluate the diagnostic efficiency of MRI parameters for fibrosis staging, with the area under the curve (AUC) values for T1/T2 mapping compared using the DeLong test.

Results

Repeatability

The ICC for T1 values between the two readers was 0.980 (P < 0.001, 95% CI: 0.935–0.994), indicating excellent repeatability. Similarly, the ICC for T2 values was 0.869 (P < 0.001, 95% CI: 0.627–0.958). The high ICCs demonstrate robust inter-reader reliability, ensuring the repeatability of the measurements for subsequent study of fibrosis assessment.

Dynamic changes of collagen content and inflammatory factors

The representative images of liver tissues stained with H&E, Masson, and Sirius red in the CCl₄-induced fibrosis model are illustrated Fig. 2a. Liver tissues section showed a progressive increase in inflammatory cell infiltration, alongside the development of incomplete fibrous septa. The qPCR and Western blot analysis showed that this was paralleled by a continuous increase in collagen content and inflammatory factors with the progression (Fig. 2b-e). The main biological parameters for mice are shown in Table 1. A total of three mice were excluded from the study due to mortality. It can be assumed that the cause of death in these mice was the toxicity of CCl₄. To mitigate the potential impact of exclusion, further statistical analysis was conducted using pathological staging. The number of mice exhibiting F0-4 pathological stages was 9, 5, 7, 7, and 8, respectively.

Fig. 2.

Dynamic changes of inflammatory biomarkers and collagen content. The representative images of liver tissues stained with H&E, Masson, and Sirius red in the CCL₄ model of fibrosis (light microscopy, × 200 original magnification) (a). The qRT-PCR results of IL-1, IL-6, and TNF-α (fold change) (b). The Western blot results of α-SMA, IL-1, IL-6, and TNF-α. Due to the lack of full-length membrane images in the experiment, we provided all cropped images in the Supplementary information and ensured that the edges of the membrane were visible. All of liver inflammatory biomarkers illustrate increased trends with the prolongation of CCL₄ injection. # P < 0.05 vs. 2 weeks. *P < 0.05 vs. 4 weeks (c-d). The relative HYP levels. *P < 0.05, **P < 0.01, ***P < 0.001 (e)

Table 1.

The main characteristics and biological parameters for mice

Variable	Normal (n = 9)	Week 2 (n = 10)	Week 4 (n = 10)	Week 6 (n = 10)	P-value
T1 values (ms)	498.12 ± 36.27	552.92 ± 34.16$	645.27 ± 32.26$#	682.68 ± 37.39$#	< 0.001
T2 values (ms)	38.19 ± 2.58	42.13 ± 1.88	44.15 ± 5.20$	47.86 ± 6.03$#	0.002
Collagen protein expression
HYP content (μg/g)	194.44 ± 7.53	352.65 ± 9.35$	420.14 ± 2.08$#	549.06 ± 12.83$#&	< 0.001
α-SMA (fold change)	1.05 ± 0.05	1.58 ± 0.10$	1.67 ± 0.08$	1.75 ± 0.09$#	< 0.001
Inflammatory factors protein expression (fold change)
IL-1	1.08 ± 0.08	1.48 ± 0.10$	1.62 ± 0.16$	1.88 ± 0.23$#	0.001
IL-6	1.03 ± 0.04	1.24 ± 0.02$	1.40 ± 0.09$#	1.57 ± 0.05$#&	< 0.001
TNF-α	1.09 ± 0.09	1.37 ± 0.03$	1.58 ± 0.04$#	1.69 ± 0.06$#&	< 0.001

P-value represents the comparison between the model and normal groups

HYP hydroxyproline, α-SMA α-smooth muscleactin, IL-1 interleukin-1, IL-6 interleukin-6, TNF-α tumor necrosis factor-α

^$P < 0.05 versus. normal

^# P < 0.05 versus. 2 weeks

^& P < 0.05 versus. 4 weeks

Correlations between T1/T2 values and fibrogenesis biomarkers

The linear regression analysis showed a strong positive correlation between T1 values and the HYP content (r = 0.880, P < 0.001; Fig. 3a), indicating that T1 values effectively reflect collagen deposition in LF. Furthermore, the T1 values exhibited superior performance in reflecting collagen deposition compared with T2 values, as confirmed by Fisher r-to-z transformation (Z = 4.92, P < 0.001; Table 2). Additionally, T1 values showed stronger correlations with α- SMA and inflammatory markers (IL-1, IL-6, TNF—α; P < 0.05), as shown in Table 2.

Fig. 3.

Correlations between T1/T2 values and fibrogenesis biomarkers. The linear regression analysis between the HYP content and T1 values (a). Box charts of the HYP content (b), T1 values (c), and T2 values (d) in different fibrosis scores. ρ represents Spearman correlation coefficient. *P < 0.05, **P < 0.01, ***P < 0.001

Table 2.

Correlation between T1/T2 values and liver fibrotic biomarkers

	Collagen associated protein				Inflammatory markers
	HYP content		α-SMA		IL-1		IL-6		TNF-α
Parameters	r	Pa-Value	r	P-Value	r	P-Value	r	P-Value	r	P-Value
T1(ms)	0.880	< 0.001	0.874	< 0.001	0.841	< 0.001	0.665	< 0.001	0.835	< 0.001
T2(ms)	0.595	0.001	0.610	< 0.001	0.698	< 0.001	0.446	0.020	0.649	< 0.001
Fisher Z	4.92		10.43		2.54		2.28		3.09
Pb-value	< 0.001		< 0.001		0.011		0.022		< 0.001

The Fisher Z is used to determine whether two Pearson correlation coefficients in same r column are statistically significant.

CI confidence interval

P^a value < 0.05 indicated statistical significance in Pearson correlation coefficients. 

P^b value < 0.05 indicated statistical significance in Fisher Z. 

The HYP content exhibited a progressive increase with advancing fibrosis stages in Fig. 3b. Spearman analysis confirmed a significant correlation between HYP content and METAVIR fibrosis stages (ρ = 0.914, P < 0.001). Similarly, T1 values increased significantly across fibrosis stages (Fig. 3c), with a Spearman correlation coefficient of ρ = 0.854 (P < 0.001). Statistical comparison of these coefficients revealed no significant difference (Z = 1.031, P = 0.303), indicating that the strength of the correlation between T1 values and fibrosis stages is comparable to that between HYP content and fibrosis stages.

For T2 values, a moderate correlation with fibrosis stages was observed (ρ = 0.697, P < 0.001; Fig. 3d), which was significantly weaker than the T1-HYP correlation (Z = 2.535, P = 0.011). These results further support the superiority of T1 values in quantifying collagen deposition.

Diagnostic efficiency of T1/T2 mapping

The axial images of T2WI, T1 mapping, and T2 mapping were illustrated in Fig. 4, which represent the fibrosis control and model groups. These images demonstrated the changes in MRI signals that occur as fibrosis progresses. In accordance with the values indicated in Table 1, an increase in MRI signals can be observed. This may be attributed to its enhanced sensitivity in detecting alterations in collagen deposition during the progression of fibrosis. The ROC curve analysis was conducted on data labelled with binary classifications, which revealed that T1 and T2 values exhibited strong diagnostic efficiency for fibrosis stages (Fig. 5). Furthermore, the Delong test revealed that, despite no significant difference in diagnostic efficiency (P > 0.05), the AUC values of T1 were slightly greater than those of T2. Notably, the highest AUC for T1 mapping was observed in discriminating between F0-2 and F3-4 (Table 3), a critical distinction for diagnosing advanced LF. The AUC range and cut-off values are shown in Table 3, which can facilitate the determination of the diagnosis and stage of LF.

Fig. 4.

The representative axial images of T2WI, T1 maps and T2 maps in the control (F0) and model (F1-4) groups. Three ROIs measuring 8.0 mm² are located on the largest slice of liver area of MR images. ROI = region of interest; F = fibrosis score

Fig. 5.

The ROC curve analysis of T1/T2 mapping in diagnostic efficiency on data labelled with binary classifications. The ROC analysis of T1/T2 mapping in F0 versus F1-4 (a), F0-1 versus F2-4 (b), F0-2 versus F3-4 (c), and F0-3 versus F4 (d). The P-values shown in the figure represent the differences between T1 and T2 mapping. For detailed AUC values and their corresponding P-values, please refer to Table 3

Table 3.

ROC analysis of T1/T2 mapping on data labelled with binary classifications

Parameters	AUC	95%CI	Pa-value	Sensitivity (%)	Specificity (%)	Cutoff value(ms)	Pb-value
F0 vs. F1-4
T1	0.951	0.792–0.997	< 0.001	77.78	100	551.08	0.374
T2	0.883	0.701–0.974	0.001	77.78	88.89	40.00
F0-1 vs. F2-4
T1	0.962	0.808–0.999	< 0.001	92.86	100	575.89	0.280
T2	0.863	0.676–0.964	< 0.001	78.57	100	42.85
F0-2 vs. F3-4
T1	0.994	0.862–0.999	< 0.001	100	93.75	610.10	0.127
T2	0.869	0.684–0.967	< 0.001	81.82	87.50	42.85
F0-3 vs. F4
T1	0.842	0.646–0.954	< 0.001	100	70.00	610.10	0.693
T2	0.808	0.607–0.935	0.006	83.33	90.00	44.60

ROC receiver operating characteristic, AUC area under the curve, CI confidence interval, F fibrosis stage

P^a-value calculated for comparing variables between different fibrosis stages; 

P^b-value calculated for comparing the AUC between T1 and T2. 

Discussion

The study revealed strong associations between T1/T2 mapping and biomarkers of collagen deposition and ongoing systemic inflammation, especially the HYP content. The T1 mapping demonstrated superior diagnostic efficiency in detecting collagen deposition and inflammation in CCl₄-induced LF.

To our knowledge, this is the first study to report on the correlation between T1/T2 mapping and HYP content and inflammatory factors in LF. HYP was selected as the primary biomarker due to its specificity to collagen, its stability in tissue samples, and its direct correlation with total collagen deposition [28, 29]. While other markers, such as collagen type I or TIMP-1, could provide additional insights into specific collagen subtypes or the dynamic balance of extracellular matrix remodeling, they may also introduce complexity due to their dependence on antibody specificity, metabolic turnover, or multifactorial regulation [30, 31]. The progressive accumulation of collagen in the liver leads to changes in tissue composition, which are reflected in MRI relaxation times. In advanced stages of fibrosis, inflammatory cytokines (IL-1, IL-6, and TNF-α) typically peak, potentially altering tissue composition and MRI relaxation times by influencing water content, cellularity, and extracellular matrix properties. Our results of the correlation between T1/T2 values and the HYP content are in accordance with existing literature, which indicates that an increase in collagen deposition is associated with prolonged T1 and T2 relaxation times due to alterations in water content and tissue architecture in fibrotic liver tissue [13, 31, 32]. Further, we comprehensively explored the correlation between T1/T2 and HYP content, indicating that T1 values may provide a more accurate reflection of the extent of collagen deposition in the liver, providing a novel insight for the use of T1 mapping in the early detection of fibrosis.

Another core factor in LF is inflammation [33], which involves nuclear factor kappa B (NF-κB) and other signal pathways. In damaged liver cells, these pathways are activated, inducing the release of pro-inflammatory cytokines such as TNF-α, IL-1, and IL-6, which in turn lead to inflammatory responses [34]. In order to avoid selection bias in the analysis of sample inflammation, liver slices were subjected to histopathological and Western Blot analysis in this study. Furthermore, the elevation in inflammatory factors showed a positive correlation with T1/T2 mapping, indicating that inflammatory responses may play a crucial role in LF. From a molecular perspective, T1/T2 mapping are reliable non-invasive methods.

This study highlighted the use of T1/T2 mapping in the evaluation of LF induced by CCl₄. We observed a dose-dependent increase in T1/T2 values with increasing severity of fibrosis. This observation is consistent with the expected pathophysiological changes in fibrotic tissue, where increased extracellular matrix deposition and changes in tissue hydration alter the local magnetic environment, leading to prolonged relaxation times. The sensitivity of T1/T2 mapping to these changes makes them powerful tools for detecting and quantifying fibrosis in its early stages (Table 3). Notably, T1 mapping achieved the highest AUC in distinguishing between F0-2 and F3-4 fibrosis stages, a critical threshold for identifying advanced fibrosis. This finding highlights the potential of T1 mapping as a key tool for clinical decision-making, particularly in cases where early detection of advanced fibrosis is essential for patient management. The lack of significant difference between T1 and T2 mapping in distinguishing F0-2 from F3-4 fibrosis (P > 0.05) may be attributed to a limited sample size, an uneven distribution of fibrosis stages, and inherent limitations of the CCL₄-induced model. Differences in collagen deposition and inflammatory processes between these stages could mask true differences in the effectiveness of T1 and T2 mapping. The CCL₄-induced fibrosis model may exhibit variability in fibrosis progression and local tissue micro-environment, which could affect the relationship between MRI parameters and pathological changes, thus influencing the statistical outcomes. T1 mapping showed a higher accuracy in detecting collagen deposition compared with T2 mapping, since the fact that T1 mapping values exhibit a higher degree of correlation with collagen content, while T2 mapping is predominantly influenced by inflammatory processes and oedema [19].

Models established at different time points have been shown to dynamically reflect the various stages of human LF, from early to advanced stages. It has been determined that shorter time (2 weeks) models are appropriate for the study of early lesions, while longer time (6 weeks) models have been demonstrated to effectively simulate late fibrosis or cirrhosis [23]. However, it is important to note that the ultimate staging of LF models is contingent upon pathological analysis. Although the results are promising, there are several limitations. Firstly, this study was performed on an animal model of CCl₄-induced LF, which, although widely used, may not fully replicate the complexity of human LF. Specifically, CCl₄-induced fibrosis primarily results from direct hepatotoxicity, whereas human LF often arises from chronic inflammation, metabolic dysfunction, or viral infections. These differences may lead to variations in inflammation patterns, fibrosis distribution, and progression speed, which are critical factors in human disease. Additionally, the model does not account for lipid degeneration and iron deposition processes commonly observed in human liver diseases, potentially limiting its clinical settings. Secondly, the heterogeneity of LF can result in significant regional variation in collagen and extracellular matrix composition, leading to inconsistencies in T1 and T2 values. This variability may reduce accuracy, especially in early fibrosis, as global mapping values often fail to reflect localized pathological changes. Thirdly, it is not possible to eliminate the impact of abdominal respiratory artefacts in mice entirely, which has the potential to impair image quality and introduce measurement bias. It is therefore imperative to exercise caution when applying the aforementioned mapping techniques in a direct manner within the context of practical clinical practice.

In conclusion, this study highlighted the potential of T1/T2 mapping as non-invasive, quantitative biomarkers for staging LF, and providing new insights into the onset and progression of LF. In the future work, our studies will aim to validate these findings in clinical practice and explore the applicability of T1/T2 mapping in different causes of LF, and employ multi-parametric MRI or the integration of other biomarkers, to enhance the precision of fibrosis assessment.

Abbreviations

LF

Liver fibrosis

ECM

Extracellular matrix

MRI

Magnetic resonance imaging

HYP

Hydroxyproline

ROIs

Regions of interest

H&E

Hematoxylin and eosin

ICC

Intraclass correlation coefficient

ROC

Receiver operating characteristic

AUC

Area under the curve

CI

Confidence interval

Authors’ contributions

Shuqin Xue and Yujie Zhu: study conception and design, analysis and interpretation of data, drafting of manuscript; Jing Rong: study conception and design, acquisition of data; Kun Zhu and Min Shao: acquisition of mice imaging data; Xiujuan Yin and Tongtong Liu: acquisition of data, analysis and interpretation of data; Saisai Zhang and Likang Yin: acquisition of data; Xiao Wang: critical revision. All authors reviewed the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (82370641) and Major Research Foundation of Higher Education of Anhui Province (2023AH040370).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Animal Ethics Committee of Anhui Medical University and the informed consents were obtained from Animal Experiment Center of Anhui Medical University.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.