Comparative study of Gd-EOB-DTPA-enhanced magnetic resonance imaging signal intensity parameters and albumin-bilirubin grade for liver function assessment based on early and middle-stage of liver dysfunction Authors: Ruike Xia, Xiaoxiao Wu, Jun Zheng, Xiaoyang Wang, Wenxian Peng Published: 2026 ABSTRACT AbstractBackgroundGadolinium-ethoxy benzyl-diethylenetriamine penta-acetic acid (Gd-EOB-DTPA)-enhanced magnetic resonance imaging (MRI) enables not only liver lesion detection but also a one-stop assessment of hepatic function. Among the commonly used liver function scoring systems, the albumin-bilirubin (ALBI) grade is more objective than the Child-Turcotte-Pugh (CTP) grade and is particularly suitable for evaluating liver function in patients with early and middle-stage chronic liver disease. However, although various Gd-EOB-DTPA-enhanced MRI signal intensity (SI) parameters have been employed for liver function assessment, a comprehensive comparison of their efficacy is lacking, and no consensus exists on the optimal parameter. This study aimed to compare the efficacy of eight Gd-EOB-DTPA-enhanced MRI SI parameters in assessing liver function based on the ALBI grading system and identify the optimal imaging biomarker for clinical application.MethodsA total of 99 high-risk patients with hepatitis B-related hepatocellular carcinoma (HCC) in the early or middle stages (61 with ALBI grade 1 and 38 with ALBI grades 2–3) underwent Gd-EOB-DTPA-enhanced MRI. Eight parameters were analyzed, including the relative enhancement rate of hepatic parenchyma, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-portal vein contrast ratio, liver-muscle ratio growth rate, liver-spleen ratio growth rate, relative SI of biliary tract, and portal vein-spleen contrast ratio. Statistical analyses included receiver operating characteristic (ROC) curves evaluation, Spearman correlation, and independent-samples t-test to compare diagnostic performance and correlate parameters with ALBI scores.ResultsLiver-portal vein contrast ratio showed the numerically highest discriminative power for distinguishing ALBI grade 1 from grades 2–3 in all MRI parameters [area under the curve (AUC): 0.806, 95% confidence interval (CI): 0.715–0.897; sensitivity: 85.2%, specificity: 63.2%, cutoff value: 1.4763], followed by liver-spleen contrast ratio (AUC: 0.799, cutoff value: 1.7536) and relative enhancement rate of hepatic parenchyma (AUC: 0.765, cutoff value: 0.7874). Liver-muscle ratio growth rate exhibited moderate efficacy (AUC: 0.730, cutoff value: 0.9116). Parameters showed significant correlations with ALBI scores (|r|: 0.368–0.532, P<0.001), except for portal vein-spleen contrast ratio (r=0.116, P=0.253). The AUC of relative SI of biliary tract was 0.678. The cutoff value of relative SI of biliary tract was 4.3656. The assessment of portal vein-spleen contrast ratio did not show a significant association with functional differences (P=0.589).ConclusionsLiver-portal vein contrast ratio showed the numerically best performance among all MRI parameters for differentiating ALBI grades in this study population, offering superior sensitivity for early-stage liver dysfunction assessment. Gd-EOB-DTPA-enhanced MRI not only enables a more accurate prediction of postoperative functional liver reserve but also provides a comprehensive foundation for clinicians to formulate optimal, individualized treatment strategies, thereby potentially reducing the incidence of postoperative complications. Meanwhile, we caution against reliance on portal vein-spleen contrast ratio or biliary signal measurements.Keywords: Gadolinium-ethoxy benzyl-diethylenetriamine penta-acetic acid (Gd-EOB-DTPA), magnetic resonance imaging (MRI), contrast-enhanced scan, liver function, albumin-bilirubin grade (ALBI grade) Introduction Gadolinium-ethoxy benzyl-diethylenetriamine penta-acetic acid (Gd-EOB-DTPA) is a hepatocellular contrast medium that is easily taken up by normal hepatocytes and secreted into the biliary system without any change in its chemical structure (1). Additionally, Gd-EOB-DTPA has the characteristics of both nonspecific extracellular space contrast mediums and hepatocyte-specific contrast mediums. Therefore, Gd-EOB-DTPA-enhanced magnetic resonance imaging (MRI) has been used not only to detect liver lesions (2), but also for a one-stop assessment of hepatic function (1). The degree of hepatocyte-specific contrast uptake has been linked to liver function in the progression of cirrhosis (3,4) and even to the staging of liver fibrosis (5). The greatest advantage of using Gd-EOB-DTPA-enhanced MRI for liver function assessment lies in its ability to provide regional and high spatial-resolution functional assessment at the segmental level. This allows for the evaluation of segmental liver function heterogeneity. Studies have shown that Gd-EOB-DTPA-enhanced MRI outperforms the indocyanine green (ICG) clearance test in evaluating both global and regional liver reserve function while also providing a superior estimation of the incidence of post-hepatectomy liver failure (6). Such functional mapping is valuable for complex cases; for example, the assessment of functional future liver remnant via Gd-EOB-DTPA-MRI can serve as a useful tool for evaluating regional functional liver reserve in hepatocellular carcinoma (HCC) patients with portal vein tumor thrombus (7). In addition to this critical functional information, all data required for surgical planning—including tumor volume and distribution, liver anatomy, vascular supply, and extrahepatic findings—can be obtained through a single, integrated examination. Therefore, Gd-EOB-DTPA-enhanced MRI not only enables a more accurate prediction of postoperative functional liver reserve but also provides a comprehensive foundation for clinicians to formulate optimal, individualized treatment strategies, thereby potentially reducing the incidence of postoperative complications (8). At present, the scoring systems commonly used to evaluate liver function include Child-Turcotte-Pugh (CTP) scoring, model for end-stage liver disease (MELD) scoring, and albumin-bilirubin (ALBI) scoring. Compared with the ALBI grade, the CTP grade tends to be overestimated (9). The ALBI grade provides a more objective and quantitative prognostic stratification for HCC patients (especially CTP A patients), and serves as a valuable complementary and refined tool for the CTP classification in HCC staging and liver function assessment, which is helpful for the individualized evaluation of liver function in HCC patients (10,11). The ALBI grade does not need imaging to evaluate subjective parameters such as ascites and hepatic encephalopathy. Although the MELD score avoids the influence of subjective factors and is well-validated for short-term mortality prediction in patients with advanced or decompensated liver disease, it has reduced sensitivity and limited discriminatory power for evaluating liver function in patients with preserved or mildly impaired liver function. The latest ALBI grade proposed by Johnson et al. (12) can be calculated with only two parameters, albumin and total bilirubin (TB), and it is highly sensitive to mild changes in liver function, exhibits a good ability to stratify liver function in patients with early and middle-stage chronic liver disease, and can effectively reflect subtle changes in liver reserve function that are difficult to identify by the CTP classification in such patients. According to the ALBI grading criteria established by Johnson et al. (12), liver function is classified into three grades based on the calculated ALBI score: ALBI grade 1 (score ≤−2.60) indicates the best-preserved liver function and most favorable prognosis; ALBI grade 2 (score >−2.60 and ≤−1.39) indicates moderately impaired liver function; and ALBI grade 3 (score >−1.39) indicates severely impaired liver function and the worst prognosis, with the value of −2.6 serving as the key cutoff distinguishing grade 1 from grade 2 within this grading system. This group of studies is mainly on patients with early and middle-stage liver disease, so it is more reasonable to use ALBI scoring as a standard for evaluating liver function. Previous studies (13-15) evaluated liver function by Gd-EOB-DTPA-enhanced MRI signal intensity (SI). Since using SI ratio is the simplest and most convenient way to evaluate liver function, including the relative enhancement rate of hepatic parenchyma, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-portal vein contrast ratio, liver-muscle ratio growth rate, liver-spleen ratio growth rate, relative SI of biliary tract, and portal vein-spleen contrast ratio. However, there is still a lack of comprehensive comparative research on the efficacy of these Gd-EOB-DTPA-enhanced MRI parameters in evaluating liver function. Additionally, there is still no consensus on which Gd-EOB-DTPA-enhanced MRI SI parameter is most suitable for evaluating liver function. Therefore, this study aimed to explore whether the Gd-EOB-DTPA-enhanced MRI parameters can be used to evaluate the liver function of patients with cirrhosis, and which parameter is the best. We present this article in accordance with the STARD reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2463/rc). Methods General data We selected all continuous Gd-EOB-DTPA-enhanced MRI studies conducted in clinically evaluated patients with focal liver lesions from April 2020 to June 2023. This study enrolled 99 high-risk patients with hepatitis B-related HCC in the early or middle stages, all of whom had a history of hepatitis, cirrhosis, or other hepatic disorders and were under follow-up at our hospital. The cohort comprised 67 males and 32 females, with a mean age of (56±13) years. They were divided into Group A (n=61) and Group B (n=38) according to ALBI scoring. They were stratified into three subgroups based on ALBI grades, with 61 patients in ALBI grade 1, 36 in grade 2, and 2 in grade 3. The scoring criteria were as follows: those below −2.6 were classified as Group A (ALBI 1), and those above −2.6 are classified as Group B (ALBI 2–3). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Shanghai University of Medicine and Health Sciences, and the requirement for individual consent for this retrospective analysis was waived. Patient inclusion criteria: (I) patients with hepatitis B virus were screened from the high-risk population of HCC who underwent Gd-EOB-DTPA-enhanced scanning, and these patients completed laboratory tests such as routine blood test, blood biochemistry test, and coagulation function test within 1 week; (II) they have the required indicators for ALBI scoring (including serum TB, serum albumin, prothrombin time, ascites, hepatic encephalopathy, etc.). Patient exclusion criteria: (I) cirrhosis caused by congenital factors or vascular malformations; (II) presence of diffuse lesions or excessively large tumor burden, which precluded accurate measurements in the hepatic parenchyma; (III) large image artifacts affect the evaluation; (IV) patients who have undergone splenectomy, partial hepatectomy, or who have received radiofrequency ablation, transcatheter arterial chemoembolization (TACE), radiotherapy and chemotherapy, or targeted drug therapy for lesions before evaluation; (V) patients with biliary stricture, obstruction, or embolism of the hepatic portal vein and its main branches; (VI) patients who are not suitable for contrast-enhanced imaging with Gd-EOB-DTPA contrast mediums (Figure 1). Flowchart of patient screening and ALBI grade distribution in HCC patients. This flowchart summarizes the screening process of eligible hepatitis B-related HCC patients, showing that 99 patients were finally included after exclusions, and categorized into ALBI grade 1 (n=61), grade 2 (n=36), and grade 3 (n=2). ALBI, albumin-bilirubin; HCC, hepatocellular carcinoma; TACE, transcatheter arterial chemoembolization. Examination method All examinations were performed using a 1.5 T MRI scanner (Avanto, Siemens, Erlangen, Germany) with a body phased-arrays surface coil and a spinal coil as the receiving coil. Patients were required to fast for 4–6 hours before the examination and undergo strict breath-holding training. Gd-EOB-DTPA (Xian’ai, Jiangsu Hengrui Medicine Co., Ltd., Jiangsu, China) was used as a hepatocellular contrast medium. Gd-EOB-DTPA-enhanced scanning received the contrast medium injection at a float rate of 1.0 mL/s (dose =0.1 mL/kg). The contrast was intravenously administered via a power injector followed by a 20–30 mL (float rate 2.0 mL/s) saline flush. The scanning sequences employed in this study are as follows: (I) T2-weighted imaging-coronal plane (T2WI-COR) sequence; (II) T1-weighted imaging-in-phase and out-of-phase (T1WI-IN-OPP) sequence; (III) T1-weighted imaging with fat saturation (T1WI-FS) sequence; (IV) dynamic contrast-enhanced (DCE)-MRI sequence (parameters identical to those of T1WI-FS); (V) T2-weighted imaging with fat saturation (T2WI-FS) sequence; (VI) diffusion-weighted imaging (DWI) sequence; (VII) T1WI-FS (hepatobiliary phase; HBP) sequence. Detailed parameters for each sequence, including repetition time (TR), echo time (TE), field of view (FOV), slice thickness, and other relevant settings, are provided in Table 1. Gd-EOB-DTPA-enhanced scanning received the contrast medium injection at a float rate of 1.0 mL/s (dose =0.025 mmoL/kg). Scans were performed at 15–20 seconds, 60 seconds, 180 seconds, and 300 seconds after the start of contrast medium injection, to obtain images of the arterial phase, portal venous phase, equilibrium phase, and transitional phase. We performed a T1WI-FS HBP scan at 20 minutes after Gd-EOB-DTPA injection, and this phase was defined as the key time point for all region of interest (ROI)-based quantitative SI measurements in this study. All other scanning sequences were the same as those mentioned above. Table 1Scanning parameters for each sequenceParametersT2WI-CORT1WI-IN-OPPT1WI-FSDCE-MRIT2WI-FSDWIT1WI-FS (hepatobiliary phase)TR (ms)3.691605.045.042,2002,7005.04TE (ms)1.852.222.312.3185692.31FOV (mm × mm)350×350350×260360×270360×270380×380400×400360×270Slice thickness (mm)5733663Slice gap (%)30202020203020Number of excitations1111171Multi-phase acquisition1211411Flip angle (°)807010101409020b-value (s/mm2)–––––0/600/800–DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; DWI, diffusion-weighted imaging; FOV, field of view; T1WI-FS, T1-weighted imaging fat saturation; T1WI-IN-OPP, T1-weighted imaging in-phase and out-of-phase; T2WI-COR, T2-weighted imaging coronal plane; T2WI-FS, T2-weighted imaging fat saturation; TE, echo time; TR, repetition time. Scanning parameters for each sequence DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; DWI, diffusion-weighted imaging; FOV, field of view; T1WI-FS, T1-weighted imaging fat saturation; T1WI-IN-OPP, T1-weighted imaging in-phase and out-of-phase; T2WI-COR, T2-weighted imaging coronal plane; T2WI-FS, T2-weighted imaging fat saturation; TE, echo time; TR, repetition time. Image analysis All the MRI images were evaluated by Picture Archiving and Communication System (PACS; Neusoft, Shenyang, China). All ROIs were pinpointed by two other radiologists with many years of experience in abdominal MRI. When discrepancies arose between two radiologists in ROI delineation (including the selection range of hepatic parenchymal ROIs and the definition of portal vein ROI size), a third senior radiologist with more than 20 years of clinical experience was invited to conduct an independent evaluation, with the delineation scheme determined by this senior radiologist serving as the unified standard for subsequent analysis. To avoid measurement bias, the radiologists were blinded of patients’ clinical, hematological, and radiological information. Each ROI was either a circle or an oval. The ROIs of the liver parenchyma was measured in four sections (left lateral, left medial, right anterior, and right posterior) (Figure 2). We then took the average value of the four sections. The ROI (size: 72.16 mm2, 73.38 mm2, 81.45 mm2, 84.30 mm2) was manually set by the observers, avoiding visible vessels and biliary ducts, focal lesions, and imaging artifacts. We referred to the method for delineating the ROIs in the liver area as described in the study by Peng et al. (16). We kept the size and position of ROI in different time phases as consistent as possible. In the same slice and at the same location during plain scan and HBP (20 minutes), we drew circular ROIs on the main portal vein, left portal vein branch, right portal vein branch, and the spleen. The diameter (unit: 12.37 mm) of the portal vein ROI was determined by the radiologists based on the scope of the portal vein. To minimize measurement errors, the ROI was placed as close to the center of the portal vein as possible (Figure 3). For the spleen ROI, we placed 3 ROIs, measured the SI of each, and calculated the average value. The average value was used as the final splenic SI (Figure 2). The maximum SI of the left hepatic duct, right hepatic duct, common hepatic duct, and common bile duct were measured (Figure 4). At the hepatic hilum level, the average SI of the right erector spinae muscle was be used to measure circular ROI (about 100 mm2) (Figure 2). We referred to the method described in the study by Nishie et al. for drawing ROIs (17). We calculated the relative enhancement rate of hepatic parenchyma, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-portal vein contrast ratio, liver-muscle ratio growth rate, liver-spleen ratio growth rate, relative SI of biliary tract, and portal vein-spleen contrast ratio. The calculation formulas are as follows: Example of liver parenchymal SI and splenic SI and erector spinae SI. (A) and (B) illustrate the placement of ROIs (encircled areas) for measuring the SI of liver parenchyma (left lateral lobe, left medial lobe, right anterior lobe, and right posterior lobe) and the SI of the erector spinae muscle, with images obtained before (A) and 20 minutes after (B) the administration of Gd-EOB-DTPA. (C) and (D) illustrate the placement of ROIs (encircled areas) for measuring the SI of the spleen, also with magnetic resonance images taken before (C) and 20 minutes after (D) the administration of Gd-EOB-DTPA. Gd-EOB-DTPA, gadolinium-ethoxy benzyl-diethylenetriamine penta-acetic acid; ROI, region of interest; SI, signal intensity. Example of portal vein signal intensity measurement. Pre-contrast images are shown in panels A, B, and C, and post-contrast images in panels D, E, and F. The ROIs in panels B, C, E, and F are placed within the main portal vein, whereas the ROIs in panels A and D are positioned within the left portal vein branch. ROI, region of interest. Example of bile duct signal intensity measurement. (A-D) Magnetic resonance images obtained 20 minutes after administration of Gd-EOB-DTPA. The ROI in (A) measures the SI of the left hepatic duct. The ROI in (B) measures the SI of the right hepatic duct. The ROI in (C) measures the SI of the common hepatic duct, and the ROI in (D) measures the SI of the common bile duct. Gd-EOB-DTPA, gadolinium-ethoxy benzyl-diethylenetriamine penta-acetic acid; ROI, region of interest; SI, signal intensity. Statistical methods We analyzed the statistical significance of the differences between the Gd-EOB-DTPA SI index in Group A (ALBI grade 1) and Group B (ALBI grade 2–3) (independent-samples t-test), and then analyzed the correlation with the ALBI scoring (Spearman correlation). Finally, the optimal cut-off value, sensitivity, and specificity of Gd-EOB-DTPA SI index (relative enhancement rate of hepatic parenchyma/liver-portal vein contrast ratio /liver-spleen contrast ratio, etc.) in each group were obtained through the receiver operating characteristic (ROC) curve. We analyzed the statistical significance of the differences in Gd-EOB-DTPA uptake indicators between groups and their correlation with AIBI scoring. At the same time, we analyzed the included patients’ various clinical indicators, and recorded the TB, aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatinine, albumin, prothrombin time international normalized ratio (PT-INR), prothrombin activity (PT), and platelet count (PLT). We used the Spearman correlation to analyze the correlation between various clinical laboratory indicators and Gd-EOB-DTPA uptake indicators. Liver function SPSS 22.0 software (IBM Corp., Armonk, NY, USA) was used for statistical analysis. First, we carried out the normality test (Shapiro-Wilk test) and the homogeneity of variance tests (Bartlett test) on the MRI parameters. If they were consistent with the normal distribution and the homogeneity of variance, independent-samples t-test was used to analyze the differences in MRI parameters between the ALBI-graded groups. MRI imaging parameters included relative enhancement rate of hepatic parenchyma, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-portal vein contrast ratio, portal vein-spleen contrast ratio, liver-muscle ratio growth rate, liver-spleen ratio growth rate, and relative SI of biliary tract. First, we carried out the normality test (Shapiro-Wilk test) and the homogeneity of variance tests (Bartlett test) on the MRI parameters. If they were consistent with the normal distribution and the homogeneity of variance, independent-samples t-test was used to analyze the differences in MRI parameters between the ALBI-graded groups. MRI imaging parameters included relative enhancement rate of hepatic parenchyma, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-portal vein contrast ratio, portal vein-spleen contrast ratio, liver-muscle ratio growth rate, liver-spleen ratio growth rate, and relative SI of biliary tract. We used the ROC curves to evaluate the above parameters to identify the effectiveness of adjacent liver function levels and the effectiveness of ALBI grading. We used the ROC curves to evaluate the above parameters to identify the effectiveness of adjacent liver function levels and the effectiveness of ALBI grading. If the above-mentioned MRI imaging parameters followed a normal distribution, the Pearson correlation test was adopted; if not, the Spearman correlation test was used. A p-value of less than 0.05 was considered statistically significant. If the above-mentioned MRI imaging parameters followed a normal distribution, the Pearson correlation test was adopted; if not, the Spearman correlation test was used. A p-value of less than 0.05 was considered statistically significant. No multiple comparison correction was performed for the correlation tests between multiple MRI parameters and laboratory indicators, as the sample size of this study is relatively small and these analyses are currently considered exploratory. No multiple comparison correction was performed for the correlation tests between multiple MRI parameters and laboratory indicators, as the sample size of this study is relatively small and these analyses are currently considered exploratory. Results Differences in MRI imaging indicators among different ALBI grades The results of the independent-samples t-test showed that there were statistically significant differences in relative enhancement rate of hepatic parenchyma, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-portal vein contrast ratio, liver-muscle ratio growth rate, liver-spleen ratio growth rate, and relative SI of biliary tract between the ALBI grade 1 group and the ALBI grades 2–3 group (Table 2). There was no significant difference in portal vein-spleen contrast ratio between the two groups. Table 2Comparison of the MRI imaging parameters between different groups (x¯±s)MRI imaging parametersGroup AGroup Bt-valueP valueRelative enhancement rate of hepatic parenchyma0.96±0.250.69±0.294.753<0.0001Liver-muscle contrast ratio2.86±0.602.29±0.554.751<0.0001Liver-spleen contrast ratio2.18±0.481.69±0.415.245<0.0001Liver-portal vein contrast ratio1.85±0.411.45±0.345.150<0.0001Portal vein-spleen contrast ratio1.19±0.141.17±0.110.5310.597Liver-muscle ratio growth rate1.03±0.380.73±0.353.871<0.0001Liver-spleen ratio growth rate0.62±0.320.32±0.284.654<0.0001Relative signal intensity of biliary tract5.37±1.644.32±1.852.9690.004Group A: ALBI grade 1; Group B: ALBI grade 2–3. ALBI, albumin-bilirubin; MRI, magnetic resonance imaging. Comparison of the MRI imaging parameters between different groups (x¯±s) Group A: ALBI grade 1; Group B: ALBI grade 2–3. ALBI, albumin-bilirubin; MRI, magnetic resonance imaging. The efficacy of MRI indices in differentiating liver function ROC curve analysis was used to evaluate the capacity of individual magnetic resonance parameters to discriminate between Group A (ALBI grade 1) and Group B (ALBI grade 2–3) in the cohort with impaired liver function (Figure 5 and Table 3). To visually demonstrate the changes in the liver-portal vein contrast ratio parameter across different ALBI grades, we provided pre- and post-contrast images at the HBP20min for representative patients with ALBI grades 1–3 (Figure 6). These results supported the declining trend of liver-portal vein contrast ratio with worsening liver function, which was consistent with the ROC curve analysis. ALBI grades 2 and 3 were combined for analysis due to the small number of ALBI grade 3 patients (only 2 cases) in this study cohort (mainly early and middle-stage liver disease patients). This grouping method limits the assessment of graded liver dysfunction. ROC curves of MRI parameters across ALBI grades. ALBI, albumin-bilirubin; MRI, magnetic resonance imaging; ROC, receiver operating characteristic. Table 3MRI parameters for ALBI grade discriminationMRI parametersAUC (95% CI)CutoffSensitivity (%)Specificity (%)P valueRelative enhancement rate of hepatic parenchyma0.765 (0.663–0.868)0.787475.476.3<0.0001Liver-muscle contrast ratio0.772 (0.675–0.870)2.62467.278.9<0.0001Liver-spleen contrast ratio0.799 (0.707–0.891)1.753680.368.4<0.0001Liver-portal vein contrast ratio 0.806 (0.715–0.897)1.476385.263.2<0.0001Portal vein-spleen contrast ratio0.532 (0.418–0.647)1.219242.676.30.589Liver-muscle ratio growth rate0.730 (0.624–0.835)0.911667.281.6<0.0001Liver-spleen ratio growth rate0.762 (0.664–0.860)0.393777.061.1<0.0001Relative signal intensity of biliary tract0.678 (0.567–0.790)4.365668.963.20.003ALBI, albumin-bilirubin; AUC, area under the curve; CI, confidence interval; MRI, magnetic resonance imaging. MRI parameters for ALBI grade discrimination ALBI, albumin-bilirubin; AUC, area under the curve; CI, confidence interval; MRI, magnetic resonance imaging. Pre-contrast and post-contrast images at the 20 min HBP in patients with ALBI grades 1–3. Panels A, C, E are pre-contrast images for patients with ALBI grades 1–3. Panels B, D, F are the corresponding HBP images acquired 20 minutes after contrast administration. Panels A and B correspond to patients with ALBI grade 1, with an ALBI score of −3.25 and a liver-portal vein contrast ratio of 1.93. Panels C and D correspond to patients with ALBI grade 2, with an ALBI score of −1.52 and a liver-portal vein contrast ratio of 1.57. Panels E and F correspond to patients with ALBI grade 3, with an ALBI score of −1.37 and a liver-portal vein contrast ratio of 1.18. ALBI, albumin-bilirubin; HBP, hepatobiliary phase. Correlation analysis between ALBI score and MRI parameters |r|>0.7, |r|=0.4–0.7, |r|=0.2–0.4 respectively indicate strong, moderate, and weak correlation |r|<0.2 indicates no correlation (Table 4). Table 4Correlation of MRI parameters with the ALBI scoringMRI parametersrP valueRelative enhancement rate of hepatic parenchyma0.472<0.0001Liver-muscle contrast ratio0.471<0.0001Liver-spleen contrast ratio0.532<0.0001Liver-portal vein contrast ratio0.503<0.0001Portal vein-spleen contrast ratio0.1160.253Liver-muscle ratio growth rate0.451<0.0001Liver-spleen ratio growth rate0.522<0.0001Relative signal intensity of biliary tract0.368<0.0001ALBI, albumin-bilirubin; MRI, magnetic resonance imaging. Correlation of MRI parameters with the ALBI scoring ALBI, albumin-bilirubin; MRI, magnetic resonance imaging. Correlation between liver cancer grade and MRI parameters The cutoff of the relative SI of biliary tract, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-portal vein contrast ratio, portal vein-spleen contrast ratio, liver-muscle ratio growth rate, relative enhancement rate of hepatic parenchyma, and liver-spleen ratio growth rate, in descending order, are 4.3656, 2.624, 1.7536, 1.4763, 1.2192, 0.9116, 0.7874, and 0.3937. The area under the curve (AUC) [95% confidence interval (CI)] of the liver-portal vein contrast ratio, liver-spleen contrast ratio, liver-muscle contrast ratio, relative enhancement rate of hepatic parenchyma, liver-spleen ratio growth rate, liver-muscle ratio growth rate, relative SI of biliary tract, and portal vein-spleen contrast ratio, in descending order, are 0.806 (0.715–0.897), 0.799 (0.707–0.891), 0.772 (0.675–0.870), 0.765 (0.663–0.868), 0.762 (0.664–0.860), 0.730 (0.624–0.835), 0.678 (0.567–0.790), and 0.532 (0.418–0.647). All effective MRI parameters showed a moderate correlation with ALBI scores (|r|: 0.368–0.532, P<0.001), indicating that MRI SI parameters can reflect the changes of liver function, but there are other factors affecting liver function that cannot be captured by these parameters. Portal vein-spleen contrast ratio showed no significant group difference between Group A (ALBI grade 1) and Group B (ALBI grade 2–3) (P=0.589), had the lowest AUC (0.532) with no diagnostic value, and no significant correlation with ALBI score (r=0.116, P=0.253). Therefore, portal vein-spleen contrast ratio is a non-informative parameter for liver function assessment in this study population and is not recommended for clinical application. Discussion Theoretical basis for evaluating liver reserve function through liver parenchymal signal-to-noise ratio: cirrhosis is a diffuse liver parenchymal disease characterized by progressive fibrosis, bridging of the portal area, and regeneration of liver parenchymal nodules (18). Therefore, it is reasonable to hypothesize that the number of functional hepatocytes is reduced in the case of cirrhosis. The value of Gd-EOB-DTPA specifically ingested by the liver is related to the activity of hepatocytes. When hepatocytes are damaged, the uptake capacity decreases, and the SI decreases accordingly. Many studies have shown that the degree of liver parenchymal uptake and biliary excretion on Gd-EOB-DTPA-enhanced MRI is helpful for evaluating liver function during the progression of cirrhosis (3,4). Therefore, liver reserve function can be assessed by measuring the SI value of the liver in the hepatobiliary-specific phase. Using the SI ratio to assess liver function is the simplest and most convenient method in Gd-EOB-DTPA-enhanced MRI imaging (1). Our data showed that with the aggravation of liver function impairment, the SI of the hepatic parenchyma gradually decreased. Previous studies (19-21) have also shown that the severity of cirrhosis can significantly affect the absorption of gadolinium, thereby affecting the degree of liver enhancement, which is consistent with our study result. Direct measurement of SI can be used to evaluate liver function; however, MRI SI is a relative value and is easily affected by scanning parameters and so on (22). When parallel imaging techniques are used, the signal-to-noise ratio becomes unreliable (15). Moreover, the SI of contrast-enhanced MRI has a nonlinear relationship with the concentration of the contrast agent. Therefore, the paravertebral muscles and spleen are often used as the reference for the SI of liver parenchyma after Gd-EOB-DTPA injection. Accordingly, this study adopted parameters such as relative enhancement rate of hepatic parenchyma, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-portal vein contrast ratio, liver-muscle ratio growth rate, liver-spleen ratio growth rate, relative SI of biliary tract, and portal vein-spleen contrast ratio as indicators for evaluating liver function, exploring the feasibility of assessing liver function by Gd-EOB-DTPA-enhanced MRI and conducting comparative study on each parameter, in order to provide a basis for clinical application. The reason why there is no difference in portal vein-spleen contrast ratio: in previous studies, the value of portal vein-spleen contrast ratio in assessing liver function in cirrhosis has been rarely explored. In our study, all Gd-EOB-DTPA MRI-based indicators were associated with ALBI scores, except for portal vein-spleen contrast ratio. Our research results indicate that, portal vein-spleen contrast ratio cannot reflect the liver function of patients with cirrhosis, which is consistent with the results reported by Yang et al. (1). From normal liver to ALBI grade 3 cirrhosis, the portal vein SI showed a continuous mild increase, but there was no difference between groups. The mean value of spleen signal might be equal between groups. Consequently, there may be no differences in portal vein-spleen contrast ratio among groups. Liver-muscle ratio growth rate, liver-muscle contrast ratio In a simple regression analysis that excluded individual liver volume (LV), Haimerl et al. (14) identified the growth rate of the liver-spleen ratio as having the strongest correlation among the SI based liver function paraments (relative enhancement rate of hepatic parenchyma, liver-spleen contrast ratio, liver-muscle contrast ratio, liver-spleen ratio growth rate, liver-muscle ratio growth rate). Kamimura et al. (15) found that the growth rate of the liver-muscle ratio had a stronger correlation with indocyanine green retention test at 15 minutes (ICG-R15) than the absolute liver-muscle ratio. It was shown that by adding plain scanning imaging data, the inter-subject variability of liver SI before contrast agent injection could be reduced, which facilitated an accurate assessment of Gd-EOB-DTPA uptake by the liver parenchyma. The results of the present study have shown comparable sensitivity between the growth rate of the liver-muscle ratio and the absolute liver-muscle ratio in discriminating ALBI grades. This finding could be attributable to the relatively small inter-individual variation within our cohort. Furthermore, our statistical analysis revealed that the SI of the erector spinae muscle increased during the HBP in some cases, whereas it decreased in others. We speculate that this inconsistency may be influenced by technical parameters such as radiofrequency amplifier gain and receiver coil sensitivity, which could also explain the relatively low sensitivity observed for both the growth rate and the absolute value of the liver-muscle ratio. Liver-muscle Ratio Growth Rate, Relative Enhancement Rate of Hepatic Parenchyma Motosugi et al. (23) reported that the liver-muscle ratio increase rate exhibited a higher correlation with the fibrosis score than the relative enhancement (relative enhancement rate of hepatic parenchyma) of the liver parenchyma. In contrast, Kamimura et al. (15) found that ICG R15 showed a stronger correlation with relative enhancement rate of hepatic parenchyma than with the liver-muscle ratio increase rate. Our results have demonstrated that the correlation between relative enhancement rate of hepatic parenchyma and the ALBI grade was slightly higher than that of the liver-muscle ratio increase rate, which is consistent with the findings of Kamimura et al. This discrepancy might be attributed to the SI of the paravertebral muscles, which can vary due to differences in fat infiltration among individuals. Consequently, the results may depend on individual variations in fat infiltration within the paravertebral muscles. However, it should be noted that relative enhancement rate of hepatic parenchyma can also be influenced by image registration errors between pre-contrast and 20-minute post-contrast images following Gd-EOB-DTPA administration, as well as variations in pre-scan parameters such as transmitter gain, receiver gain, and other factors (24). Comparison between Liver-spleen Contrast Ratio and Liver-portal Vein Contrast Ratio Takatsu Y et al. examined the relationship between liver-portal vein contrast ratio and liver-spleen contrast ratio (25). Their results indicated a strong correlation between liver-portal vein contrast ratio and liver-spleen contrast ratio, with liver-portal vein contrast ratio values being lower than liver-spleen contrast ratio values in all groups. The authors suggested that the underlying cause might be portal vein SI, which reflects the blood pool more effectively than the spleen. Yang et al. (1) reported findings contrary to those of Takatsu et al., indicating that the liver-portal vein contrast ratio value was more effective than the liver-spleen contrast ratio in distinguishing the severity of liver cirrhosis patients. Our study also demonstrates that both liver-portal vein contrast ratio and liver-spleen contrast ratio exhibit strong correlations with the ALBI grade, and the AUC of liver-portal vein contrast ratio is larger than that of liver-spleen contrast ratio. These discrepancies in these results may be attributed to the following factors. (I) Different etiologies may lead to variations in the uptake and excretion of Gd-EOB-DTPA. Our patient predominantly consisted of those with hepatitis B-induced cirrhosis, whereas their study mainly included patients with chronic liver disease; (II) the use of different MRI equipment and imaging sequences. The biliary excretion of Gd-EOB-DTPA provides positive intraductal contrast on T1WI (13,26). The use of Gd-EOB-DTPA can potentially enhance MRI cholangiography (27). The findings from Takao et al. (13) indicated that the presence of chronic liver disease significantly affects the bile duct SI on Gd-EOB-DTPA-enhanced MRI, and that this SI reflects the underlying liver function. In our study, a decreasing trend in the maximum common bile duct SI was observed with increasing ALBI grades, which also confirms that chronic liver disease can affect the bile duct SI on Gd-EOB-DTPA-enhanced MRI imaging. However, the relative bile duct SI showed only a weak correlation with the ALBI grades and demonstrated poor efficacy in distinguishing between Group A and Group B (ROC curve area: 0.678, 0.567–0.790). The reasons for this may be as follows. Firstly, this is likely because the 20-minute HBP was used in our study, whereas the bile duct SI peaks at 30 minutes after Gd-EOB-DTPA administration (13). Secondly, the common bile duct is often narrow, which could lead to inaccuracies in SI measurements. The advantages of liver-portal vein contrast ratio The present results indicate that, among all MRI signal indices, liver-portal vein contrast ratio demonstrated the best performance in distinguishing Group A (ALBI grade 1) from Group B (ALBI grade 2–3), with the largest AUC (0.806, 0.715–0.897) and the highest sensitivity (85.2%). We hypothesize that the good performance of liver-portal vein contrast ratio may be related to the following factors to the mechanistic explanation for the favorable performance of liver-portal vein contrast ratio. This might be attributable to the central location of the portal vein within the body, which also positions it near the isocenter of both the static magnetic field and the receiver coil. This location facilitates the reception of homogeneous signals and potentially reduces susceptibility to artifacts (24). Furthermore, the SI of the portal vein has shown a good correlation with blood contrast agent concentration (25). The changing trend of the portal vein SI strengthens the association between liver-portal vein contrast ratio and liver function. The study by Takatsu et al. (25) also confirmed the value of liver-portal vein contrast ratio for assessing liver function. They proposed that Q-liver-portal vein contrast ratio could serve as a substitute for Q-liver-spleen contrast ratio to simply evaluate the degree of hepatic contrast enhancement, particularly in scenarios involving splenectomy or the presence of Gamna-Gandy bodies. They also suggested that MRI imaging allows for a more intuitive, semi-quantitative assessment of the contrast between the liver parenchyma and the adjacent large portal vein. Additionally, liver-spleen contrast ratio cannot be calculated in the setting of splenectomy and may be less reliable in cirrhotic patients with Gamna-Gandy bodies. The study population was focused on hepatitis B-related early and middle-stage liver disease patients with liver function impairment. Cirrhosis can damage liver cells, increase the spleen volume, and lead to portal hypertension (1), so we mainly assessed the relevant indicators of the liver, spleen, and portal vein. Liver function impairment caused by hepatic parenchymal damage directly reduces the uptake capacity of hepatocytes for Gd-EOB-DTPA, which in turn results in decreased SI of the hepatic parenchyma in the HBP. Meanwhile, the secondary splenic volume increase and portal hypertension induce alterations in the SI of the spleen and portal vein, rendering the ratio parameters based on the liver, spleen, and portal vein (liver-portal vein contrast ratio, liver-spleen contrast ratio) more sensitive to reflecting changes in liver function. These results collectively suggest that liver-portal vein contrast ratio may represent a more useful alternative imaging biomarker for evaluating liver function compared to other MRI signal indices such as liver-spleen contrast ratio. The study has several limitations. First, we employed a clinical scoring system, the ALBI score, for comparison. Although histology has long been regarded as the gold standard for assessing liver function, clinical staging systems are more frequently used in practice and are associated with fewer complications. Moreover, sampling error in histological assessment can lead to underestimate the severity of disease (18). Second, we did not use the ICG clearance test to validate liver function. Previous studies have demonstrated that ICG-R15 is a predictor of sufficient liver enhancement on HBP images (28,29). In our patient series, very few patients underwent the ICG clearance test; therefore, our analysis was limited to comparing Gd-EOB-DTPA uptake values with ALBI grades. Further studies are needed to validate our results. Third, our study included a relatively small number of patients with severe liver dysfunction. However, enrolling a large cohort of such patients is challenging, as individuals with severe liver impairment are generally not considered candidates for Gd-EOB-DTPA-enhanced MRI or therapeutic interventions for liver tumors (15). Fourth, the influence of hepatic steatosis or iron deposition on contrast indices was not evaluated. The impact of steatosis or iron deposition on liver contrast enhancement indices should be investigated in larger future studies. Fifth, manual ROI delineation and SI measurement were adopted in our study. Although discrepancies in delineation were resolved through the independent evaluation of a third senior radiologist, manual delineation has inherent limitations: the position and size of ROIs are slightly influenced by the subjective clinical experience of observers, precluding the achievement of absolutely standardized definition across different cases and potentially leading to minor fluctuations in the measurement results of individual parameters. In the future, a customized automated ROI segmentation model for hepatitis B-related liver diseases may be developed to realize the standardization and unification of ROI delineation, thereby further improving the reproducibility and clinical translatability of measurement results. Sixth, the study was a single-center retrospective study, and the research population was only limited to hepatitis B-related early and middle-stage liver disease patients, which may lead to selection bias and limit the generalizability of the results. Seventh, the ROC cutoff values derived from this study only have exploratory significance, and lack of external validation in other centers and populations. Conclusions In summary, the relative enhancement rate of hepatic parenchyma, liver-muscle contrast ratio, liver-spleen contrast ratio, liver-muscle ratio growth rate, and liver-spleen ratio growth rate obtained from the HBP of Gd-EOB-DTPA-enhanced MRI can be used for the assessment and grading of liver function. The results of this study indicate that the liver-portal vein contrast ratio was the best indicator for distinguishing between ALBI grade 1 and grades 2–3. Therefore, liver-portal vein contrast ratio has the potential to serve as an MRI-based biomarker for liver function evaluation. Additionally, the relative SI of the biliary system demonstrated poor efficacy in assessing liver function, and the relative SI of biliary tract did not reflect liver function status in patients with cirrhosis. Supplementary The article’s supplementary files as Acknowledgments None. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Shanghai University of Medicine and Health Sciences, and individual consent for this retrospective analysis was waived. Footnotes Reporting Checklist: The authors have completed the STARD reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2463/rc Funding: None. Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2463/coif). The authors have no conflicts of interest to declare. Data Sharing Statement Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2463/dss References Associated Data Supplementary Materials The article’s supplementary files as Data Availability Statement Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2463/dss