Clinical assessment of hepatic steatosis and iron content using three-dimensional (3D) multi-echo chemical shift encoded magnetic resonance imaging (MRI): correlation with histopathology

Introduction

Chronic liver diseases represent a significant global health burden, with hepatic steatosis and iron overload being common pathological conditions associated with various etiologies (1). These conditions can lead to progressive liver damage, fibrosis, and ultimately cirrhosis if left undiagnosed or untreated. Therefore, accurate quantification of hepatic steatosis and iron content is of great importance for disease prevention, disease assessment, clinical diagnosis and therapeutic intervention (2). Traditionally, liver biopsy has been considered the gold standard for diagnosing and staging these hepatic disorders (3). However, this invasive procedure is associated with potential complications, limiting its widespread use in clinical practice (4).

Non-invasive imaging techniques have emerged as attractive alternatives for quantifying liver fat and iron content. Although ultrasound (US) and computed tomography (CT) can assess hepatic steatosis non-invasively, both have inherent limitations (5,6). US demonstrates relatively low sensitivity for detecting hepatic fat content below 20% (7). Whereas CT evaluation is constrained by radiation exposure risks, interference from coexisting hepatic conditions, and low sensitivity to early or subtle changes in hepatic fat content (8,9). Magnetic resonance imaging (MRI) offers superior tissue characterization. Various magnetic resonance (MR) techniques, including conventional in-phase and out-of-phase imaging, proton MR spectroscopy, and water-fat suppression methods based on water-lipid frequency difference, provide diverse approaches for assessing hepatic fat and iron deposition (10-12). Among these, multi-echo chemical shift encoded (CSE) methods for fat quantification and R2* mapping for iron assessment have demonstrated significant promise in accurately quantifying both hepatic steatosis and iron overload (10,13). Previous investigations have established correlations between MRI-derived proton density fat fraction (PDFF) and histological steatosis grading, as well as the reliability of R2* measurements for iron quantification (13-15). However, many studies have been limited by small sample sizes or have focused on quantifying either fat or iron alone (16).

Recent advances in MRI technology have led to the development of sophisticated multi-echo Dixon techniques that enable simultaneous quantification of hepatic fat and iron content. This technique is based on a three-dimensional (3D) gradient echo sequence, using a low flip angle (typically 3°) to acquire 6 gradient echoes with precisely designed echo spacing for optimal signal-to-noise ratio (SNR). It uses a 6-peak spectral model of fat for fitting, and simultaneously generates multiple contrast-weighted images and parametric maps (including water-only, fat-only, in-phase, out-of-phase images, and PDFF and R2* maps) in a single scan, thereby enabling comprehensive assessment of hepatic composition in a single acquisition (17,18).

This study aims to evaluate the clinical utility of a 3D multi-echo Dixon MRI sequence for simultaneous quantification of hepatic steatosis and detection of iron overload in patients with liver disease. We sought to validate the accuracy and reliability of this quantitative MRI technique against histopathological analysis in a large cohort of 177 patients, representing one of the most extensive studies of its kind. The multi-echo Dixon approach offers high spatial resolution and enables comprehensive assessment of liver composition in a single, time-efficient acquisition. By conducting direct comparison with histopathology in a substantial patient cohort using advanced 3.0T MRI technology, this study aims to establish quantitative MRI as a reliable, non-invasive alternative to liver biopsy, with potential to significantly impact clinical practice and improve patient care. We present this article in accordance with the STARD reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1795/rc).

Methods

Study design

This retrospective study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine (No. 2021-1062-137). As this was a retrospective study with no interventions performed and no disclosure of patients’ personal identifying information, the requirement for written informed consent was waived by the ethics committee. We analyzed data from patients with chronic liver disease who underwent liver biopsy with hematoxylin and eosin (HE) staining and Prussian blue staining at Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine between January 2021 and January 2025. Inclusion criteria were: (I) percutaneous liver biopsy yielding adequate tissue for reliable steatosis grading, and (II) MRI examination performed within 180 days of the biopsy. Exclusion criteria included: (I) MRI images with significant motion or susceptibility artifacts, (II) an interval exceeding 180 days between MRI and biopsy, (III) biopsy-confirmed liver malignancy. The patient selection process is illustrated in Figure 1. A final cohort of 177 patients met all criteria and were included in the analysis.

Figure 1 Flow chart of population selection. After exclusions based on missing assessments, MRI issues, and long intervals between MRI and biopsy, the final cohort included 177 patients, categorized by steatosis grade and iron content. MRI, magnetic resonance imaging.

MR imaging examination

All patients underwent MRI examinations using a 3.0T uMR790 scanner (United Imaging Healthcare, Shanghai, China) equipped with a 32-channel coil. All examinations were performed using a standardized protocol with patients in the supine position. For PDFF and R2* mapping, a 3D multi-echo Dixon sequence utilizing the proprietary Fat Analysis & Calculation Technique (FACT) was employed. The imaging parameters were as follows: repetition time (TR) of 12.06 ms; echo times (TE) of 1.71, 3.22, 4.73, 6.24, 7.75, and 9.26 ms; flip angle of 3°; field of view (FOV) of 420×420 mm²; acquisition matrix of 240×240; slice thickness of 10 mm; bandwidth of 900 Hz/pixel; parallel acquisition acceleration factor of 2; and acquisition time of 15 seconds. This sequence produces multi-contrast-weighted images and quantitative maps (encompassing fat fraction, R2*, water-only, fat-only, in-phase, and out-of-phase images) within one acquisition. To ensure accurate quantification of the PDFF and R2*, the FACT reconstruction algorithm incorporates several correction techniques, widely validated in the literature, to mitigate the effects of known confounding factors: (I) multi-peak fat spectral modeling: the fat signal is modeled as a superposition of multiple resonance frequencies (6 peaks in this study), rather than a single frequency. This model corrects for the underestimation of PDFF and overestimation of R2* caused by the complex nature of the fat spectrum (19,20). (II) T2* decay correction and R2 estimation: By acquiring multiple echoes, the algorithm simultaneously fits the water, fat, and T2* decay. The R2 value is estimated based on the composite signal decay from the entire voxel, not solely from the water signal or a specific fat peak, ensuring measurement accuracy in the presence of coexisting fat and iron overload (19,21). This is achieved using a complex signal model that incorporates a multi-peak fat spectrum, a common field map, and a single R2* decay term. (III) T1 bias minimization: The sequence utilizes a low flip angle design to minimize signal bias arising from differences in T1 relaxation times between water and fat (20). (IV) Noise bias and eddy current correction: The reconstruction algorithm includes corrections for noise bias (using magnitude discrimination methods) and for phase errors induced by eddy currents from gradient switching, thereby enhancing the reliability of quantification, particularly in regions with low fat fraction (19,20). (V) Region-enhanced algorithm: This algorithm improves the robustness of parameter estimation in areas with magnetic field inhomogeneity or low SNR through pixel-wise fitting and regional consistency optimization.

Image analysis

The acquired image data were transferred to a dedicated post-processing workstation (uWS, United Imaging Healthcare, Shanghai, China) for quantitative analysis. Two radiologists, one with 8 years of experience (reader 1) and another with 3 years of experience (reader 2) in diagnostic abdominal imaging, independently reviewed and analyzed the images. Both readers were blinded to the clinical information and biopsy results of patients to prevent bias. For each patient, multiple regions of interest (ROIs) were carefully selected on both the PDFF and R2* maps. The ROIs were positioned in segment eight of the right hepatic lobe, corresponding to the anatomical location of the biopsy site, with a standardized area of 1.5–2.0 cm². The two readers first referred to the R2* image, precisely avoiding the areas of major blood vessels and bile ducts, while excluding liver lesions and regions affected by motion artifacts. Subsequently, they drew the ROI at the same anatomical location on the PDFF image to ensure the spatial consistency between the PDFF and R2* measurement areas. For each patient, at least three independent ROI measurements were performed on both PDFF and R2* maps. Finally, the average value of the three measurements was used as the PDFF value and R2* value for each patient. Inter-reader differences were assessed, and final PDFF and R2* values were obtained by averaging the measurements from both readers for subsequent statistical analysis.

Histopathological assessment

Percutaneous liver biopsies were performed using 18-gauge needles under US guidance by experienced hepatologists. Biopsy specimens were processed using standard techniques and stained with HE and Prussian blue stains. Histopathological evaluation was performed by a single experienced hepatopathologist with over 15 years of experience. Hepatic steatosis was graded according to the percentage of hepatocytes with macrovesicular fat: grade 0 (S0, <5%), grade 1 (S1, 5–33%), grade 2 (S2, 34–66%), and grade 3 (S3, >66%) (22). Hepatic iron deposition was assessed using Prussian blue staining and classified as a binary variable: iron-negative (Prussian blue staining negative) or iron-positive (Prussian blue staining positive) (23).

Statistical analysis

All analyses were performed using software SPSS22.0 (IBM, Chicago, USA). Continuous data are presented as mean ± standard deviation or median (interquartile range), and categorical data as frequencies (percentages). To evaluate inter-reader agreement between the two radiologists in measuring PDFF and R2* values, the intraclass correlation coefficient (ICC) with 95% confidence intervals (CIs), linear regression and Bland-Altman analysis were used. Data normality was assessed using the Shapiro-Wilk test and visual inspection of histograms. Since PDFF and R2* values did not demonstrate normal distribution and homogeneity of variance, non-parametric statistical methods were employed. Spearman rank correlation analysis was used to assess the correlation between MRI-derived PDFF values and histological steatosis grades, and between R2* values and Prussian blue staining results. The Kruskal-Wallis test was used to evaluate differences in PDFF values across steatosis grade and differences in R2* values between Prussian blue staining groups. To evaluate the diagnostic performance of PDFF values for detecting different grades of steatosis, receiver operating characteristic (ROC) curves were constructed for the following clinically relevant dichotomizations: S0 vs. S1 or above; S1 and below vs. S2 and above; and S2 and below vs. S3. Similarly, ROC analysis was conducted to evaluate the diagnostic performance of R2* values in discriminating between iron-positive and iron-negative cases. The area under the curve (AUC) with 95% CIs was calculated to evaluate diagnostic accuracy. Optimal threshold values were determined using the Youden index. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each threshold. Statistical significance was defined as P<0.05.

Results

Patients

A total of 177 patients (mean age 47.70±12.69 years; 93 men and 84 women) were included in the study. All underwent MRI scanning, with a mean time interval of 19.5 days between MRI and biopsy. Patient selection is illustrated in Figure 1, and baseline characteristics are summarized in Table 1.

Inter-reader reliability

Strong correlations and high consistency were demonstrated between the measurements from two readers for both PDFF and R2* values, with correlation coefficients of determination (R2) of 0.997 (P<0.001) and 0.982 (P<0.001), respectively. Bland-Altman analysis revealed mean differences of −0.050 (95% CI: −0.957 to 0.856) for PDFF and 0.107 (95% CI: −6.050 to 6.264) for R2*. The ICC was 0.998 (95% CI: 0.998–0.999) for PDFF and 0.991 (95% CI: 0.988–0.993) for R2*. Figure 2 shows the results of linear regression (A, PDFF; C, R2*) and Bland-Altman analysis (B, PDFF; D, R2*) for measurements obtained using the 3D FACT sequence.

Figure 2 Enhanced linear regression and Bland-Altman analysis illustrating the consistency and agreement of PDFF and R2* values derived from two independent measurements. (A) Linear regression plot demonstrating the correlation between PDFF values from the first and second measurements (R²=0.997). (B) Bland-Altman plot showing the agreement between the two PDFF measurements, with the mean difference and 95% limits of agreement. (C) Linear regression plot demonstrating the correlation between R2* values from the first and second measurements (R²=0.982). (D) Bland-Altman plot showing the agreement between the two R2* measurements, with the mean difference and 95% limits of agreement. PDFF, proton density fat fraction; SD, standard deviation.

Assessment of hepatic steatosis using PDFF

PDFF values showed excellent correlation with histopathological hepatic steatosis grades (r=0.856, P<0.001), with significant differences observed across all steatosis grades (P<0.001). The box plot of PDFF values between various degrees of hepatic steatosis is shown in Figure 3.

Figure 3 The relationship between PDFF and pathological hepatic steatosis grading. Statistically significant differences are indicated by *** (P<0.001) between all groups. PDFF increases progressively with the severity of steatosis. PDFF, proton density fat fraction.

ROC curve analysis was performed to assess the diagnostic performance of PDFF for different steatosis grades. For detecting ≥ grade 1 steatosis, PDFF achieved an AUC of 0.932 (95% CI: 0.897–0.968), with an optimal threshold being 3.58%. This threshold yielded a sensitivity of 82.8%, specificity of 92.7%, PPV of 96.2%, NPV of 70.8%, and accuracy of 85.9%. For distinguishing ≥ grade 2 steatosis, PDFF achieved an AUC of 0.974 (95% CI: 0.955–0.992) at an optimal cut-off of 8.55%, with sensitivity of 95.1%, specificity of 87.9%, PPV of 80.6%, NPV of 97.1%, and accuracy of 90.4%. For identifying grade 3 steatosis, PDFF exhibited exceptional diagnostic performance with an AUC of 0.992 (95% CI: 0.981–1.000). At the optimal threshold of 25.06%, with a sensitivity of 100%, specificity of 98.2%, PPV of 66.7%, NPV of 100%, and accuracy of 98.4%. Detailed results are presented in Table 2 and Figure 4, indicating excellent diagnostic accuracy of PDFF across all steatosis grades, with particularly outstanding performance for severe steatosis. As shown in Figure 5, representative cases demonstrate the imaging and pathological features of different hepatic steatosis grades (S1–S3).

Figure 4 ROC curve analysis of PDFF for liver biopsy-based steatosis grading. The ROC curves show the diagnostic performance of PDFF with AUC values of 0.932, 0.974, and 0.992 for differentiating between 0 vs. ≥1, ≤1 vs. ≥2, and ≤2 vs. 3 grade steatosis, respectively. AUC, area under the curve; PDFF, proton density fat fraction; ROC, receiver operating characteristic.

Figure 5 Representative PDFF maps, HE staining images, R2* maps, and Prussian blue staining images from individual patients. (A) PDFF =5.6%, corresponding to hepatic steatosis grade S1 on HE staining, R2* =90.0 s⁻¹, with positive iron staining on Prussian blue; (B) PDFF =17.9%, corresponding to steatosis grade S2 on HE staining, R2* =94.0 s⁻¹, and Prussian blue staining positive for iron; (C) PDFF =37.4%, corresponding to steatosis grade S3 on HE staining, R2* =84.0 s⁻¹, with positive iron deposition shown by Prussian blue staining. HE, hematoxylin and eosin; PDFF, proton density fat fraction.

Assessment of hepatic iron content using R2*

R2* values exhibited a strong positive correlation with hepatic iron deposition (r=0.617, P<0.001), with significant intergroup differences as shown in Figure 6. ROC curve analysis was performed to evaluate the diagnostic performance of R2* for detecting hepatic iron deposition based on Prussian blue staining results. The AUC was 0.880 (95% CI: 0.822–0.937), demonstrating good diagnostic accuracy. Using an optimal cut-off value of 69.75 s⁻¹, the sensitivity was 66.1%, specificity was 98.3%, PPV was 94.6%, NPV was 86.2%, and overall accuracy was 88.1%. Figure 5 respectively presents the variations in R2* values and the conditions of iron deposition in representative cases.

Figure 6 The relationship between R2* and Prussian blue staining group. A statistically significant difference is observed between the two groups, as indicated by the asterisks (***, P<0.001). This finding suggests a strong association between Prussian blue staining and elevated R2* values.

Correlation between PDFF and R2*

Cross-tabulation analysis revealed a significant association between histopathologic steatosis and iron deposition grades (Pearson Chi-squared, P=0.005). From an imaging perspective, the Spearman correlation analysis indicated a positive correlation between PDFF and R2* values, which was statistically significant (r=0.589, P<0.001). The relationship between PDFF and R2* is illustrated in the scatter plot with linear fitting (Figure 7).

Figure 7 The linear fitting of PDFF and R2*. Scatter plot showing the correlation between PDFF (%) and R2* (s⁻¹). The solid line represents the linear regression fit. PDFF, proton density fat fraction.

Discussion

The present study evaluated the diagnostic performance of 3D Multi-Echo Chemical Shift Encoded MRI for simultaneous quantification of hepatic steatosis and detecting iron content in 177 patients with chronic liver disease. Our findings demonstrate excellent correlation between MRI-derived metrics and histopathological assessment, supporting the clinical utility of this non-invasive imaging technique as a reliable alternative to liver biopsy for comprehensive hepatic evaluation.

The PDFF measurements in our study showed excellent correlation with histological steatosis grades, which is consistent with previous investigations but extends these findings to a larger patient cohort (24). The exceptional diagnostic performance across all steatosis grades, with AUC values ranging from 0.932 to 0.992, demonstrates the robust capability of 3D multi-echo Dixon sequences for accurate fat quantification. Particularly noteworthy is the outstanding performance for detecting severe steatosis (grade 3), achieving 100% sensitivity and 98.2% specificity at the optimal threshold of 25.06%. This level of accuracy surpasses many previously reported studies and establishes PDFF as a highly reliable biomarker for hepatic fat content assessment (13,25). The superior performance observed in our study can be attributed to several technical advantages of the 3D multi-echo Dixon approach. The utilization of multiple TEs enables comprehensive correction for T2* decay effects, while the low flip angle minimizes T1 bias, ensuring more accurate PDFF quantification (17,18). Additionally, the FACT incorporates advanced signal attenuation correction methods and region-enhanced algorithms, which contribute to improved measurement precision and reproducibility (26,27). These technical refinements likely account for the excellent correlation with histopathology observed in our cohort.

Our results demonstrate that R2* measurements provide reliable assessment of hepatic iron deposition, with strong correlation to Prussian blue staining results and good diagnostic accuracy. The optimal threshold of 69.75 s⁻¹ achieved high specificity with reasonable sensitivity, suggesting that elevated R2* values are highly indicative of iron deposition. However, the interpretation of these diagnostic metrics requires careful consideration of the reference standard’s limitations. Studies have shown that the 3D CSE MRI method has high reproducibility in quantitatively measuring liver iron concentration (LIC) under different magnetic field strengths (1.5T and 3T) (28,29). Given that MRI has been proven to be highly reliable, the relatively low sensitivity observed in our study is likely due to the unsatisfactory histological methods rather than the inherent flaws of the imaging technology. The binary classification (positive/negative) based on Prussian blue staining cannot quantitatively capture the changes in iron deposition and may classify patients with very slight and clinically insignificant iron content fluctuations as “positive”, or conversely, miss the early accumulation detected by MRI. Moreover, the definitions of relevant R2*/LIC thresholds in clinical practice have varied in the past. The correlation and standardization of these thresholds are crucial for accurate clinical interpretation (30). Therefore, our binary organizational reference standard may underestimate the true quantitative accuracy of the R2* assessment. In the future, we will utilize the characteristics of the 3D multi-echo Dixon sequence of the equipment to optimize the scanning parameters, and at the same time, adopt quantitative biochemical iron content measurement methods to establish an “R2*-LIC” conversion equation applicable to this sequence. This will enable us to more effectively verify the high repeatability advantage of MRI technology.

The significant positive correlation between PDFF and R2* values reflects the complex pathophysiological relationship between hepatic fat accumulation and iron deposition. This association has been observed in various chronic liver diseases, including metabolic dysfunction-associated steatotic liver disease (MASLD) and chronic hepatitis, where oxidative stress and inflammatory processes contribute to both steatosis development and iron accumulation (31-33). Cross-tabulation analysis confirmed an association between histopathological steatosis grade and iron deposition, further confirming the clinical significance of comprehensive liver assessment. However, it is worth noting that in patients with PDFF lower than 5%, the R2* value showed a certain degree of dispersion, indicating the presence of simple iron overload, thus demonstrating that these two lesions can either coexist or exist independently. In patients with both fatty degeneration and iron overload, this may represent a subgroup with more severe liver damage and a higher risk of disease progression, emphasizing the importance of identifying these two conditions for achieving the best patient management (34).

The ability of 3D Multi-Echo Chemical Shift Encoded MRI to simultaneously assess both hepatic fat and iron content in a single, time-efficient acquisition represents a significant advancement in hepatic imaging. The 15-second acquisition time enables comprehensive liver assessment with minimal patient discomfort and high throughput, making it particularly suitable for clinical implementation. This efficiency is crucial for patients with chronic liver disease who require regular monitoring and may have difficulty tolerating prolonged imaging procedures (35). The excellent inter-reader reliability observed in our study demonstrates the reproducibility of quantitative measurements, which is essential for clinical application and longitudinal monitoring. The minimal inter-reader variability supports the potential for standardized implementation across different institutions and readers with varying levels of experience, facilitating widespread clinical adoption (36).

Limitation

Several limitations of our study warrant consideration. First, the retrospective design may introduce selection bias, as patients undergoing liver biopsy represent a specific subset of individuals with suspected liver disease. Second, the binary classification of iron status based on Prussian blue staining may not fully capture the spectrum of iron deposition severity, potentially limiting the correlation analysis. Third, while the 180-day interval between MRI and biopsy was chosen to ensure temporal relevance, liver composition can change over time, particularly in patients with active liver disease or those receiving specific treatments. A shorter interval or concurrent imaging and biopsy would provide more precise correlation data.

Conclusions

Our evaluation of 3D Multi-Echo Chemical Shift Encoded MRI in 177 patients demonstrates excellent performance for simultaneously quantifying hepatic steatosis and detecting iron content. Strong correlations with histopathology, superior inter-reader reliability, and time-efficient acquisition establish this as a valuable non-invasive alternative to liver biopsy. The ability to assess both hepatic fat and iron in a single examination provides comprehensive diagnostic information for disease management. These findings support implementing quantitative MRI as a standard hepatic evaluation tool, potentially improving patient care while reducing invasive procedures.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STARD reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1795/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1795/dss

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-2025-1795/coif). Z.Q. is employed by Shanghai United Imaging Healthcare Co., Ltd. The other authors have no conflicts of interest to declare.

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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine (No. 2021-1062-137). As this was a retrospective study with no interventions performed and no disclosure of patients’ personal identifying information, the requirement for written informed consent was waived by the ethics committee.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Koenig AB, Tan A, Abdelaal H, Monge F, Younossi ZM, Goodman ZD. Review article: Hepatic steatosis and its associations with acute and chronic liver diseases. Aliment Pharmacol Ther 2024;60:167-200. [Crossref] [PubMed]
2. Younossi ZM. Non-alcoholic fatty liver disease - A global public health perspective. J Hepatol 2019;70:531-44. [Crossref] [PubMed]
3. El-Badry AM, Breitenstein S, Jochum W, Washington K, Paradis V, Rubbia-Brandt L, Puhan MA, Slankamenac K, Graf R, Clavien PA. Assessment of hepatic steatosis by expert pathologists: the end of a gold standard. Ann Surg 2009;250:691-7. [Crossref] [PubMed]
4. Vuppalanchi R, Unalp A, Van Natta ML, Cummings OW, Sandrasegaran KE, Hameed T, Tonascia J, Chalasani N. Effects of liver biopsy sample length and number of readings on sampling variability in nonalcoholic Fatty liver disease. Clin Gastroenterol Hepatol 2009;7:481-6. [Crossref] [PubMed]
5. Han A, Zhang YN, Boehringer AS, Montes V, Andre MP, Erdman JW Jr, Loomba R, Valasek MA, Sirlin CB, O'Brien WD Jr. Assessment of Hepatic Steatosis in Nonalcoholic Fatty Liver Disease by Using Quantitative US. Radiology 2020;295:106-13. [Crossref] [PubMed]
6. Starekova J, Hernando D, Pickhardt PJ, Reeder SB. Quantification of Liver Fat Content with CT and MRI: State of the Art. Radiology 2021;301:250-62. [Crossref] [PubMed]
7. Ito K. US-derived Fat Fraction Screening of Hepatic Steatosis. Radiology 2022;304:83-4. [Crossref] [PubMed]
8. Pickhardt PJ, Park SH, Hahn L, Lee SG, Bae KT, Yu ES. Specificity of unenhanced CT for non-invasive diagnosis of hepatic steatosis: implications for the investigation of the natural history of incidental steatosis. Eur Radiol 2012;22:1075-82. [Crossref] [PubMed]
9. Cropp RJ, Seslija P, Tso D, Thakur Y. Scanner and kVp dependence of measured CT numbers in the ACR CT phantom. J Appl Clin Med Phys 2013;14:4417. [Crossref] [PubMed]
10. Tang A, Tan J, Sun M, Hamilton G, Bydder M, Wolfson T, Gamst AC, Middleton M, Brunt EM, Loomba R, Lavine JE, Schwimmer JB, Sirlin CB. Nonalcoholic fatty liver disease: MR imaging of liver proton density fat fraction to assess hepatic steatosis. Radiology 2013;267:422-31. [Crossref] [PubMed]
11. Yokoo T, Serai SD, Pirasteh A, Bashir MR, Hamilton G, Hernando D, Hu HH, Hetterich H, Kühn JP, Kukuk GM, Loomba R, Middleton MS, Obuchowski NA, Song JS, Tang A, Wu X, Reeder SB, Sirlin CBRSNA-QIBA PDFF Biomarker Committee. Linearity, Bias, and Precision of Hepatic Proton Density Fat Fraction Measurements by Using MR Imaging: A Meta-Analysis. Radiology 2018;286:486-98. [Crossref] [PubMed]
12. Henninger B, Plaikner M, Zoller H, Viveiros A, Kannengiesser S, Jaschke W, Kremser C. Performance of different Dixon-based methods for MR liver iron assessment in comparison to a biopsy-validated R2* relaxometry method. Eur Radiol 2021;31:2252-62. [Crossref] [PubMed]
13. Gu J, Liu S, Du S, Zhang Q, Xiao J, Dong Q, Xin Y. Diagnostic value of MRI-PDFF for hepatic steatosis in patients with non-alcoholic fatty liver disease: a meta-analysis. Eur Radiol 2019;29:3564-73. [Crossref] [PubMed]
14. Loomba R, Neuschwander-Tetri BA, Sanyal A, Chalasani N, Diehl AM, Terrault N, Kowdley K, Dasarathy S, Kleiner D, Behling C, Lavine J, Van Natta M, Middleton M, Tonascia J, Sirlin CNASH Clinical Research Network. Multicenter Validation of Association Between Decline in MRI-PDFF and Histologic Response in NASH. Hepatology 2020;72:1219-29. [Crossref] [PubMed]
15. Xu L, Duanmu Y, Blake GM, Zhang C, Zhang Y, Brown K, Wang X, Wang P, Zhou X, Zhang M, Wang C, Guo Z, Guglielmi G, Cheng X. Validation of goose liver fat measurement by QCT and CSE-MRI with biochemical extraction and pathology as reference. Eur Radiol 2018;28:2003-12. [Crossref] [PubMed]
16. Hernando D, Cook RJ, Qazi N, Longhurst CA, Diamond CA, Reeder SB. Complex confounder-corrected R2* mapping for liver iron quantification with MRI. Eur Radiol 2021;31:264-75. [Crossref] [PubMed]
17. Li G, Xu Z, Li X, Zuo X, Chang S, Wu D, Dai Y. Adding marrow R2* to proton density fat fraction improves the discrimination of osteopenia and osteoporosis in postmenopausal women assessed with 3D FACT sequence. Menopause 2021;28:800-6. [Crossref] [PubMed]
18. Fellner C, Nickel MD, Kannengiesser S, Verloh N, Stroszczynski C, Haimerl M, Luerken L. Water-Fat Separated T1 Mapping in the Liver and Correlation to Hepatic Fat Fraction. Diagnostics (Basel) 2023;13:201. [Crossref] [PubMed]
19. Yu H, Shimakawa A, McKenzie CA, Brodsky E, Brittain JH, Reeder SB. Multiecho water-fat separation and simultaneous R2* estimation with multifrequency fat spectrum modeling. Magn Reson Med 2008;60:1122-34. [Crossref] [PubMed]
20. Reeder SB, Cruite I, Hamilton G, Sirlin CB. Quantitative Assessment of Liver Fat with Magnetic Resonance Imaging and Spectroscopy. J Magn Reson Imaging 2011;34:729-49. [Crossref] [PubMed]
21. Kühn JP, Meffert P, Heske C, Kromrey ML, Schmidt CO, Mensel B, Völzke H, Lerch MM, Hernando D, Mayerle J, Reeder SB. Prevalence of Fatty Liver Disease and Hepatic Iron Overload in a Northeastern German Population by Using Quantitative MR Imaging. Radiology 2017;284:706-16. [Crossref] [PubMed]
22. Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999;94:2467-74. [Crossref] [PubMed]
23. SCHEUER PJ. WILLIAMS R, MUIR AR. Hepatic pathology in relatives of patients with haemochromatosis. J Pathol Bacteriol 1962;84:53-64.
24. Bachtiar V, Kelly MD, Wilman HR, Jacobs J, Newbould R, Kelly CJ, Gyngell ML, Groves KE, McKay A, Herlihy AH, Fernandes CC, Halberstadt M, Maguire M, Jayaratne N, Linden S, Neubauer S, Banerjee R. Repeatability and reproducibility of multiparametric magnetic resonance imaging of the liver. PLoS One 2019;14:e0214921. [Crossref] [PubMed]
25. Tang A, Desai A, Hamilton G, Wolfson T, Gamst A, Lam J, Clark L, Hooker J, Chavez T, Ang BD, Middleton MS, Peterson M, Loomba R, Sirlin CB. Accuracy of MR imaging-estimated proton density fat fraction for classification of dichotomized histologic steatosis grades in nonalcoholic fatty liver disease. Radiology 2015;274:416-25. [Crossref] [PubMed]
26. Peng H, Zou C, Cheng C, Tie C, Qiao Y, Wan Q, Lv J, He Q, Liang D, Liu X, Liu W, Zheng H. Fat-water separation based on Transition REgion Extraction (TREE). Magn Reson Med 2019;82:436-48. [Crossref] [PubMed]
27. Cheng C, Zou C, Liang C, Liu X, Zheng H. Fat-water separation using a region-growing algorithm with self-feeding phasor estimation. Magn Reson Med 2017;77:2390-401. [Crossref] [PubMed]
28. Hernando D, Zhao R, Yuan Q, Aliyari Ghasabeh M, Ruschke S, Miao X, Karampinos DC, Mao L, Harris DT, Mattison RJ, Jeng MR, Pedrosa I, Kamel IR, Vasanawala S, Yokoo T, Reeder SB. Multicenter Reproducibility of Liver Iron Quantification with 1.5-T and 3.0-T MRI. Radiology 2023;306:e213256. [Crossref] [PubMed]
29. Gandon Y, Olivié D, Guyader D, Aubé C, Oberti F, Sebille V, Deugnier Y. Non-invasive assessment of hepatic iron stores by MRI. Lancet 2004;363:357-62. [Crossref] [PubMed]
30. Müller L, Hernando D, Nazeef M, Reeder SB. Liver iron concentration thresholds: Where do they really come from? Hemasphere 2025;9:e70122. [Crossref] [PubMed]
31. Wang L, Qu HJ, Yang WJ, Ding JP, Shi JP, Zhang QL. Metabolic study of iron deposition based on magnetic resonance in patients with nonalcoholic fatty liver disease. Zhonghua Gan Zang Bing Za Zhi 2023;31:1204-8.
32. Bashir MR, Wolfson T, Gamst AC, Fowler KJ, Ohliger M, Shah SN, et al. Hepatic R2* is more strongly associated with proton density fat fraction than histologic liver iron scores in patients with nonalcoholic fatty liver disease. J Magn Reson Imaging 2019;49:1456-66. [Crossref] [PubMed]
33. Lv H, Zhang Y, Wang X, Liu H, Zhao H. Correlation Between Hepatic Iron and Fat in Hemodialysis Patients: A Quantitative MRI Analysis. Kidney Med 2025;7:101035. [Crossref] [PubMed]
34. Mamidipalli A, Hamilton G, Manning P, Hong CW, Park CC, Wolfson T, Hooker J, Heba E, Schlein A, Gamst A, Durelle J, Paiz M, Middleton MS, Schwimmer JB, Sirlin CB. Cross-sectional correlation between hepatic R2* and proton density fat fraction (PDFF) in children with hepatic steatosis. J Magn Reson Imaging 2018;47:418-24. [Crossref] [PubMed]
35. Hu F, Yang R, Huang Z, Wang M, Yuan F, Xia C, Wei Y, Song B. 3D Multi-Echo Dixon technique for simultaneous assessment of liver steatosis and iron overload in patients with chronic liver diseases: a feasibility study. Quant Imaging Med Surg 2019;9:1014-24. [Crossref] [PubMed]
36. Koo TK, Li MY. A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med 2016;15:155-63. [Crossref] [PubMed]