Multi-echo Dixon and breath-hold T2-corrected multi-echo single-voxel MRS for quantifying hepatic iron overload in rabbits

Abstract

Background

Three-dimensional (3D) multi-echo-Dixon (ME-Dixon) and breath-hold T2-corrected multi-echo single-voxel MR spectroscopy (HISTO) can simultaneously quantify liver fat and liver iron. However, their diagnostic efficacy and application scope for quantitative iron in co-existing fatty liver have not been adequately evaluated.

Purpose

To evaluate the accuracy of ME-Dixon and HISTO for quantitative analysis of hepatic iron in rabbits with iron deposition and fatty liver using liver–iron concentration (LIC) as a reference standard.

Material and Methods

ME-Dixon, HISTO, and conventional two-dimensional multi-echo gradient echo (GRE) sequences were performed on 42 rabbits. The following parameters were calculated: R2* from ME-Dixon and GRE; proton density fat fraction (PDFF) from the ME-Dixon, HISTO (normal TE range), and HISTO-H (extended TE range); and R2_water from HISTO and HISTO-H. The LIC and liver–fat concentration (LFC) were measured through chemical analysis, and their relationship with the MRI parameters were assessed. Receiver operating characteristic (ROC) curves and the area under the curve (AUC) were used to evaluate the diagnostic efficiency.

Results

LIC was significantly correlated with R2_HISTO-H, R2*_Dixon, and R2*_GRE (r = 0.858, 0.910, 0.931, respectively; P < 0.001) and weakly with R2_HISTO (r = 0.424; P = 0.008). A strong correlation was also observed between the LFC and PDFF obtained from HISTO, HISTO-H, and ME-Dixon (r = 0.776, 0.811, 0.888, respectively; P < 0.001). ME-Dixon showed the best performance with moderate iron overload (AUC = 0.983).

Conclusion

3D ME-Dixon is useful for quantifying the LIC, especially with co-existing fatty liver. Its diagnostic performance is also superior to that of the HISTO sequence.

Keywords

Magnetic resonance imaging fatty liver iron liver

Introduction

Hemochromatosis, transfusion-dependent anemia, hemolysis, chronic liver disease, and other such conditions can cause iron overload in the body. The liver is the main iron-storage organ. Therefore, quantitative assessment of the liver–iron concentration (LIC) is of great significance in the treatment of iron-overload patients (1,2). Such assessments have become even more important with the increase in the incidence of fatty liver disease in recent years (3). Hepatic iron deposition and fatty liver represent the early stages of chronic diffuse liver disease that can lead to liver fibrosis and liver cirrhosis (4,5). Iron overload and fatty liver deposits often exist concurrently in patients with chronic liver disease and exert synergistic promoting effects (6,7). Therefore, simultaneous quantification of both liver iron and liver fat and timely interventions are important to protect hepatocytes from further damage.

Magnetic resonance imaging (MRI) is a useful tool for non-invasive and effective detection of the liver–iron overload (8,9). Ferriscan-R2 and GRE-R2* mappings are some of the conventional techniques used for the abovementioned purpose. However, the quantitative accuracy of these techniques reduces when the liver has both iron and fat deposition (10). In recent years, various novel MRI sequences have been developed such as three-dimensional (3D) multi-echo-Dixon (ME-Dixon) and T2-corrected multi-echo single-voxel MR spectroscopy (HISTO) sequences that allow simultaneous quantification of iron and fat deposits in the liver without interferences (11,12). ME-Dixon is a variant of 3D multi-gradient-echo sequences that afford short imaging times and simultaneous fat fraction and transverse-relaxation fittings. The simultaneously estimated relaxation rate, R2*, can also be used to quantify the iron overload (13 –15). Inline calculations of the proton density fat fraction (PDFF) (fat) and R2* (iron) values for the entire liver improves the workflow without the use of other software or postprocessing methods. HISTO is a single-voxel high-speed T₂-corrected multiple-echo ¹H-MRS sequence with a stimulated echo acquisition mode (STEAM) that can quantify lipid percentages by correcting for the R2 relaxation of water and fat (16,17). The R2 value of the water component, obtained with the HISTO sequence after data fitting, can also reflect the LIC.

However, few studies have used ME-Dixon and HISTO sequences to quantify the high liver–iron overload. Moreover, the efficiency and scope of these sequences for quantitative detection of hepatic iron overload at 3 T also need to be explored further (1,15,16). In addition, the whole liver fat and iron have rarely been used as a reference standard and their chemical analysis has not been adequately investigated (11,13 –15,18). The aim of the present study was to investigate the feasibility and scope of using ME-Dixon and HISTO sequences at 3 T to quantify liver iron using rabbit models with or without fatty liver. The study also evaluates the pros and cons and diagnostic performance of these two sequences.

Material and Methods

Animal model

This prospective study was approved by the Animal Research Committee of our Guangxi Medical University First Affiliated Hospital (approval number 2020(KY-E-106)). The care of laboratory animals and all animal experiments were conducted as per the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (19). A total of 42 adult New Zealand white rabbits were obtained from the animal experimentation center of our institution. The rabbits were aged 3–4 months and they weighed 2.0–3.0 kg. They were randomly divided into a control group (Group A, n = 8) and two experimental groups (Group B, n = 17, and Group C, n = 17). Rabbits in Groups B and C were injected weekly with 15 mg/kg body weight of iron dextran (iron content 150 mg/mL) into the thigh muscle for 16 weeks to increase their iron overload. The control group was not injected with any iron. Group C rabbits were also fed with a high-fat and high-cholesterol diet (83% normal feed, 2% cholesterol, 10% lard oil, and 5% sucrose). Groups A and B were fed a normal diet, and all rabbits were provided access to food and water ad libitum.

MRI data acquisition

Two rabbits from each group were randomly selected for MR scanning each week. Before scanning, 3% pentobarbital sodium was injected into the ear vein for anesthesia (1.0 mL/kg). Then, the rabbit limbs were paralyzed to stop their movement. Bandages were placed around their abdomens to reduce respiratory motion artifacts. Next, the rabbits were placed on the examination table for MRI scanning. A 3 T MRI system (MAGNETOM Prisma; Siemens Healthcare, Erlangen, Germany) with an eight-channel coil of research device for rabbit (Zhongzhi Medical Healthcare, Suzhou, China) was used. The MRI protocol included prototype ME-Dixon and two HISTO sequences with different TE ranges (HISTO protocol for moderate iron load and HISTO-H protocol for high iron load) and a conventional two-dimensional (2D) GRE-based R2* mapping sequence.

ME-Dixon was performed with the following parameters: TE1/spacing/TE6 = 1.26/1.34/7.96 ms; repetition time (TR) = 9.25 ms; slice thickness = 3.5 mm; a low flip angle (4°) to reduce the T₁ effects; matrix size = 160 × 120; field of view (FOV) = 120 × 100 mm²; 40 slices with a thickness of 3.5 mm; and bandwidth = 1200 Hz/pixel. Parallel imaging (CAIPIRINHA) with a total acceleration factor of 4 was used to reduce the acquisition time, which was 16 s. The HISTO sequence was executed twice with different TE ranges: 12/24/36/48/72 ms (HISTO) and 12/15/18/21/24 ms (HISTO-H). All the other parameters remained the same: TR = 3000 ms; bandwidth = 1200 Hz; average = 1; voxel size = 3–5 cm³; and acquisition time = 15 s. The spectroscopy voxel was positioned on three-plane localizing images, avoiding the main vessels and liver edges with edge lengths of 1.4–1.7 cm. The conventional GRE sequence for R2* was as follows: TR = 200 ms; TE1/spacing/TE12 = 0.96/1.42/16.48 ms; bandwidth = 1950 Hz/pixel; FOV = 120 × 100 mm²; matrix size = 128 × 128; and 10 slices with a thickness of 5.0 mm. Parallel imaging (GRAPPA) with an acceleration factor of 2 was used.

Image processing

The ME-Dixon inline implementation automatically calculated PDFF (PDFF_Dixon), R2* (R2*_Dixon), and fit error (GOF_Dixon) maps as a measure of signal model consistency. Data fitting was performed by using the multistep adaptive method proposed in a previous study (20), accounting for the spectral complexity of fat, with an additional magnitude-noise first-moment modeling (21). The HISTO inline processing integrated water and fat signals separately. Next, an exponential fit of the resulting data points was performed to extrapolate to TE = 0. The fat fraction (PDFF_HISTO) was calculated as the ratio of the fat signal over the total signal (water + fat). HISTO_PDFF, as well as the R2 of water (R2_HISTO) and Rsquare (Rsqr_HISTO) from the log-linear exponential fit, were reported as a DICOM page. An R2* map from the 2D multi-echo GRE data (R2*_GRE) was calculated using a simple exponential fit.

Two radiologists with >5 years of experience in abdominal MRI independently analyzed these images and measured the parameters. For the ME-Dixon images, the criterion for acceptable data quality was defined as Dixon_GoF < 5%. On the GRE_R2* maps, 3–5 regions of interest (ROIs) of approximately 1.0 cm² were drawn over different hepatic lobes using the slice with the most extensive liver coverage, while avoiding visible vessels, bile ducts, and artifacts. The ROIs were copied to the corresponding slices of the PDFF_Dixon and R2*_Dixon maps. The final mean ROI value of each parameter was calculated by averaging the values measured by the abovementioned two physicians. For HISTO, only fitting results with Rsqr_HISTO > 0.95 were considered. The inline report was used to record both PDFF_HISTO and R2_HISTO values.

Pathologic and chemical liver analyses

After MR scanning, the rabbits were euthanized by intravenously injecting 100 mg/kg of pentobarbital. First, whole fresh livers were obtained, and the gallbladder, large vessels, and bile ducts were removed. The liver volume was measured as the volume of water displaced when placing the liver in a container. Each liver lobe was cut into two or three parts and washed 8–10 times with double distilled water. Then, 0.5–1 cm³ of the tissues were selected from each liver lobe. These tissues were immersed and fixed in 10% formalin, embedded in paraffin, and stained with Prussian blue and hematoxylin and eosin (H&E). Half of the remaining liver tissue was placed in an oven at 60 °C for approximately five days, until a constant weight was achieved. The dried liver tissue was sent to the Analysis and Testing Research Center (Guangxi, PR China) for LIC quantification using an atomic absorption spectrophotometer (iCap7000, Thermo Fisher Scientific). The remaining halves of the fresh liver tissues were sent to the center to measure the LFC as the wet weight (ww). The LICs were classified according to the LIC grades reported by Olivieri et al. (22): normal grade = <1.8 mg/g; mild grade = 1.8–3.2 mg/g; moderate grade = 3.2–7.0 mg/g; heavy grade = 7.0–15.0 mg/g; and dry weight (dw).

Statistical analysis

Statistical analyses were performed using SPSS software version 22.0 and MedCalc version 13.1.2.0. The consistency of the results reported by the two physicians was tested using the intraclass correlation coefficient (ICC). The data were expressed as the mean ± standard deviation. Data of all three groups were compared with normal distributions and uniform variances using the one-way analysis of variance (ANOVA). Pearson and Spearman analyses were used to assess the correlation among each MRI index and pathologic and chemical results, respectively. Finally, ROC curves were used to evaluate the diagnostic efficiencies and thresholds of the MRI sequences for iron content. A P value <0.05 was considered statistically significant.

Results

General condition of the rabbits

During the experiments, the control group (Group A) and high-iron-deposition group (Group B) rabbits grew well and had normal weights and appetites. In contrast, Group C rabbits had decreased appetites and lost body weight after two months on the high-fat diet (Table 1).

Table 1.

Metrics of rabbits in each group (n = 42).

Variable	Group			P
Variable	Group A (n = 8)	Group B (n = 17)	Group C (n = 17)	P
Age (weeks)	9.0 ± 4.9	9.5 ± 5.34	9.5 ± 5.34	0.971
Weight (kg)	2.92 ± 0.38	3.06 ± 0.37	2.62 ± 0.31	<0.05
Liver volume (mL)	82.8 ± 12.9	106.9 ± 20.6	142.8 ± 25.4	<0.05^*
LFC (g/100 g)	2.66 ± 0.38	2.80 ± 0.38	11.12 ± 2.79	<0.05
LIC (mg/g dw)	0.54 ± 0.09	5.29 ± 2.55	3.44 ± 1.57	<0.05^*

Values are given as mean ± SD.

Differences between all groups are statistically significant.

dw, dry weight; LFC, liver–fat concentration; LIC, liver–iron concentration; PDFF, proton density fat fraction.

Pathologic and chemical liver analyses

Hepatocyte steatosis and iron particle deposition were not observed in the livers of the control group rabbits (Fig. 1a). The livers of Group B rabbits were dark and had varying degrees of iron deposition in the liver sinusoids and hepatocytes, as revealed by microscopic evaluation (Fig. 1b). The mean liver volume of rabbits in Group C was significantly higher than those in Groups A and B. The liver surfaces of the former were also rough, greasy, and yellowish-brown, and their hepatocytes had varying degrees of edema, steatosis, and iron deposition (Fig. 1c). The average LIC and LFC values of each group are listed in Table 1.

Fig. 1.

Representative images of rabbit liver from different groups assessed with (left to right) T₁ StarVIBE, HISTO-H (extended TE range) at TE = 12 ms, ME-Dixon R2* map, 2D GRE R2* map, Prussian blue histologic staining, and gross pathology. (A1–A5) A normal rabbit from Group A with the following metrics: R2_HISTO-H = 50 s⁻¹; R2*_Dixon = 74 s⁻¹; R2*_GRE = 89 s⁻¹; no abnormalities are seen in the cellular architecture; the liver surface was smooth; the color was bright red; LIC = 0.51 mg/g dw; LFC = 2.3 g/100 g ww. (B1–B5) Group B rabbit with iron overload and the following metrics: R2_HISTO-H = 107 s⁻¹; R2*_Dixon = 330 s⁻¹; R2*_GRE = 356 s⁻¹; iron deposition was observed in hepatic sinusoids and hepatocytes; the liver turned dark brown and the texture was soft; LIC = 5.13 mg/g dw; LFC = 2.7 g/100 g ww. (C1–C5) Group C rabbit liver with iron and lipid deposition with the following metrics: R2_HISTO-H = 88 s⁻¹; R2*_Dixon = 271 s⁻¹; R2*_GRE = 319 s⁻¹; hepatocyte became swollen; iron deposition was observed; the liver is yellowish-brown and has a firm texture; LIC = 4.23 mg/g dw; LFC = 11.9 g/100 g ww. dw, dry weight; GRE, 2D multi-echo gradient echo; HISTO, T2-corrected multi-echo single-voxel; LFC, liver–fat concentration; LIC, liver–iron concentration; ME-Dixon, 3D multi-echo-Dixon; ROC, receiver operating characteristic; ww, wet weight.

MRI measurements

Four rabbits (4/42, Group B) showed poor GOF_Dixon performance with the ME-Dixon sequence because of the severe iron overload (the LIC measurements of the four rabbits were 7.1, 8.7, 9.1, and 10.9 mg/g dw), and they were therefore excluded from the analyses. Thus, a total of 38 rabbits were finally selected (Group A: n = 8, Group B: n = 13, and Group C: n = 17) for further analyses. As a measure of repeatability, the intra-observer agreements of PDFF_Dixon, R2*_Dixon, and R2*_GRE were found to be excellent with the ICCs of 0.954 (95% confidence interval [CI] = 0.907–0.977), 0.985 (95% CI = 0.972–0.992), and 0.953 (95% CI = 0.900–0.977), respectively.

The MRI results of each animal group are summarized in Table 2. The rabbits on a high-fat-diet (Group C) had significantly higher PDFF than that of the normal-diet rabbits (Groups A and B) on both the HISTO and ME-Dixon maps (P < 0.001). The control group (Group A) had significantly lower R2_HISTO-H, R2*_Dixon, and R2*_GRE values than those in the experimental groups (P < 0.001). The R2_HISTO-H and R2*_Dixon values of the groups also differed significantly from each other (Fig. 2).

Fig. 2.

Box plots of (a) PDFF_HISTO-H, (b) R2_HISTO-H, (c) PDFF_Dixon, and (d) R2*_Dixon values between the experimental and control groups. GRE, 2D multi-echo gradient echo; HISTO, T2-corrected multi-echo single-voxel; PDFF, proton density fat fraction.

Table 2.

MRI parameters for all rabbit groups (n = 38).

	Group A (n = 8)	Group B (n = 13)	Group C (n = 17)	F	P value
PDFF_HISTO (%)	1.6 ± 0.6	1.8 ± 0.9	10.2 ± 3.4	61.59	<0.001
PDFF_HISTO-H (%)	1.9 ± 0.7	2.1 ± 1.2	7.6 ± 1.5	91.70	<0.001
PDFF_Dixon (%)	1.5 ± 0.7	1.6 ± 0.9	8.1 ± 1.7	124.99	<0.001
R2_HISTO (1/s)	38.5 ± 4.2	48.2 ± 6.9	46.1 ± 8.4	4.67	0.016
R2_HISTO-H (1/s)	43.7 ± 3.6	93.0 ± 11.5	74.0 ± 12.5	50.03	<0.001^*
R2*_Dixon (1/s)	70.4 ± 19.0	283.9 ± 78.8	204.6 ± 75.3	23.56	<0.001^*
R2*_GRE (1/s)	75.5 ± 18.7	318.3 ± 70.8	273.1 ± 104.6	23.06	<0.001

Values are given as mean ± SD.

Differences among all groups are statistically significant.

GRE, 2D multi-echo gradient echo; HISTO, T2-corrected multi-echo single-voxel; PDFF, proton density fat fraction.

Relationship between liver MRI measurements and chemical analyses

Fig. 3 shows the scatter plots of the MRI relaxation rates and LIC measurements for all rabbits. R2_HISTO-H, R2*_Dixon, and R2*_GRE showed strong and positive correlations with LIC (r = 0.858, P < 0.001; r = 0.910, P < 0.001; r = 0.931, P < 0.001, respectively), while R2_HISTO showed a weakly positive correlation with LIC (r = 0.424, P = 0.008). The LIC measurement (mg/g dw) determined from the correlation slope and intercept for R2*_Dixon was 0.017 × R2*_Dixon + 0.400, and that for R2*_GRE was 0.014 × R2*_GRE + 0.570 (R2* values in 1/s). Correlation analysis between the MRI PDFF and LFC measurements was performed for all rabbits. The results showed that PDFF_HISTO, PDFF_HISTO-H, and PDFF_Dixon positively correlated with the LFC (r = 0.776, P < 0.001; r = 0.811, P < 0.001; r = 0.888, P < 0.001, respectively).

Fig. 3.

Scatter plots of the R2_HISTO, R2_HISTO-H, R2*_Dixon, and R2*_GRE maps between LICs with the linear fit (straight line) and a 95% confidence interval (curved lines). (a) Correlation between LICs and the R2_HISTO map (r = 0.424, P = 0.008); (b) LICs and the R2_HISTO-H map (r = 0.858, P < 0.001); (c) LICs and the R2*_Dixon map (r = 0.910, P < 0.001); and (d) LICs and the R2*_GRE map (r = 0.931, P < 0.001). GRE, 2D multi-echo gradient echo; HISTO, T2-corrected multi-echo single-voxel; LIC, liver–iron concentration.

Diagnostic performance of MRI sequence parameters

The optimal cutoff values, area under the curve (AUC) values, and corresponding diagnostic performance for the R2_HISTO-H, R2*_Dixon, and R2*_GRE values are summarized in Table 3. Corresponding to the LIC thresholds of 1.8 and 3.2 mg/g dw, the AUCs of HISTO, HISTO-H, ME-Dixon, and 2D GRE were 0.861, 0.978, 0.971, and 0.994, and 0.652, 0.906, 0.983, and 0.957, respectively (Fig. 4). As the LIC increased, the diagnostic performance of HISTO decreased. No significant differences were found between the AUCs of the R2*_Dixon and R2*_GRE values at both LIC thresholds (P = 0.149 and P = 0.319, respectively).

Fig. 4.

(a) ROC curve to discriminate between the normal LIC (LIC < 1.8 mg/g) and liver–iron overload (LIC ≥ 1.8 mg/g). ROC curve showed the AUC of R2_HISTO, R2_HISTO-H, R2*_Dixon, and R2*_GRE was 0.816, 0.978, 0.971 and 0.994, respectively (P < 0.001). (b) ROC curve to discriminate between the normal and mild liver–iron overloads (LIC < 3.2 mg/g) vs. moderate liver–iron overload (LIC ≥ 3.2 mg/g). ROC curve showed the AUC of R2_HISTO was 0.652 (P = 0.117) and the AUC of R2_HISTO-H, R2*_Dixon, and R2*_GRE was 0.906, 0.983, and 0.957, respectively (P < 0.001). AUC, area under the ROC curve; GRE, 2D multi-echo gradient echo; HISTO, T2-corrected multi-echo single-voxel; LIC, liver–iron concentration; ME-Dixon, 3D multi-echo-Dixon; ROC, receiver operating characteristic.

Table 3.

Diagnostic performance of HISTO, ME-Dixon, and conventional 2D GRE sequences for the assessment of liver–iron concentration at 1.8 and 3.2 mg/g dry weight using ROC curve analyses.

	Statistic	R2_HISTO	R2_HISTO-H	R2*_Dixon	R2*_GRE
LIC > 1.8 mg/g	AUC	0.816 (0.680–0.951)	0.978 (0.931–1.000)	0.971 (0.927–1.000)	0.994 (0.977–1.000)
	P value	<0.001	<0.001	<0.001	<0.001
	Cutoff value (1/s)	45.60	58.27	147.67	217.75
	Sensitivity	0.731	1.000	0.923	0.923
	Specificity	1.000	0.971	0.917	1.000
LIC > 3.2 mg/g	AUC	0.652 (0.466–0.838)	0.906 (0.808–1.000)	0.983 (0.952–1.000)	0.957 (0.902–1.000)
	P value	0.117	<0.001	<0.001	<0.001
	Cutoff value (1/s)	47.95	80.90	208.75	260.12
	Sensitivity	0.625	0.875	1.000	1.000
	Specificity	0.727	0.864	0.909	0.818

Values in parentheses are 95% confidence intervals.

AUC, area under the ROC curve; GRE, 2D multi-echo gradient echo; HISTO, T2-corrected multi-echo single-voxel; LIC, liver–iron concentration; ME-Dixon, 3D multi-echo-Dixon; ROC, receiver operating characteristic.

Discussion

This study showed that the ME-Dixon and HISTO-H sequences can quantify the liver–iron overload and fatty liver with a good diagnostic accuracy. The AUC calculations revealed that the LIC in a rabbit model can be diagnosed with a high accuracy using ME-Dixon, GRE, and HISTO-H sequences. The HISTO sequence exhibited a lower diagnostic performance than that of the other sequences, and it became poorer as the LIC increased.

In this study, we used two HISTO sequences with long and short TEs to assess the iron overload. Since MR signals are known to decrease when TEs increase, the relaxation times become very short and MR signals could be close to the noise of longer TEs when the LIC is high. We also observed this phenomenon in our study, especially in rabbits with high LIC. Our study showed that the R2 values of the short-TE HISTO sequences had higher correlations with the LIC than did the longer-TE results (r = 0.858, P < 0.001 vs. r = 0.424, P = 0.008). This finding could be explained by the presence of a minimal signal close to the noise of the longer TEs in some rabbits with high LIC. It also affected the measurement of PDFF. Therefore, the short-TE HISTO sequence should be used when there is iron deposition in the liver. Notably, the HISTO sequence can only measure the local liver–iron deposition. Additionally, factors such as the repeatable and stable ROI location selections are affected by the respiratory motion (23).

ME-Dixon is an improved 3D Dixon sequence that can simultaneously evaluate fat fractions and iron depositions in the entire liver (24). Our results showed high correlations between the R2*_Dixon values and LICs, as well as between the PDFF_Dixon values and LFCs. The correlation slopes and intercepts for the R2*_Dixon and R2*_GRE maps versus the LIC measurements were consistent with the literature values (25).

In clinical medicine, monitoring and treating the liver–iron overload depends on the LIC assessment. MRI can measure the LIC, help stage the disease, and provide clinical information (22,26). Our study results showed that the ME-Dixon and 2D GRE sequences quantified different levels of LIC with a similar high diagnostic performance. In our study, the conventional 2D GRE sequence was obtained with a somewhat shorter TE than that of the 3D ME-Dixon sequence (TE = 0.96 ms vs. 1.26 ms, respectively), which is advantageous for capturing higher LIC signals. Conversely, the 3D ME-Dixon sequence might have had higher signal-to-noise ratio (SNRs) and lower sensitivities for susceptibility-related signal dephasing. The ME-Dixon sequence affords high spatial resolution, wide coverage, and simultaneous determination of the fat and iron content.

However, the present study has some limitations. First, some of the rabbits with fatty liver developed liver fibrosis. The degree of liver fibrosis and the potential influence of fibrosis on the MR results could not be assessed. Heterogeneity and small size of the sample will affect the LIC and LFC measurement error. Second, the pathophysiologic characteristics of iron overload via dextran-loading in rabbits might differ from that in human beings. Third, no comparative study was performed on the effects of different equipment and field strength. Therefore, further clinical studies are needed to test and verify the results.

In conclusion, ME-Dixon and HISTO-H sequences are promising and useful tools to quantify the LIC, especially in patients with co-existing fatty liver. Compared with that of the HISTO sequence, multi-echo Dixon—a 3D sequence—affords a high image resolution and can simultaneously quantify fat and iron deposition with the whole-liver coverage.

Footnotes

Acknowledgements

The authors thank Editage for English language editing. They also thank Tang Cheng and Lei Yiwu, technologists in their department, for their work performing measurements for this study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received the following financial support for the research, authorship and/or publication of this article: This study has received funding from the Natural Science Foundation of Guangxi (grant number: AB20238016).

ORCID iD

Fanyu Zhao

Supplemental material

Supplemental material for this article is available online.

References

Papanikolaou

Pantopoulos

. Systemic iron homeostasis and erythropoiesis. IUBMB Life 2017;69:399–413.

Angelucci

Brittenham

McLaren

, et al. Hepatic iron concentration and total body iron stores in thalassemia major. N Engl J Med 2000;343:327–331.

Fan

Kim

Wong

. New trends on obesity and NAFLD in Asia. J Hepatol 2017;67:862–873.

Maliken

Nelson

Klintworth

, et al. Hepatic reticuloendothelial system cell iron deposition is associated with increased apoptosis in nonalcoholic fatty liver disease. Hepatology 2013;57:1806–1813.

Marmur

Beshara

Eggertsen

, et al. Hepcidin levels correlate to liver iron content, but not steatohepatitis, in non-alcoholic fatty liver disease. BMC Gastroenterol 2018;18:78.

Xue’e

Shishen

Wei

, et al. The combination and comprehensive influence of iron in the diet and the influence of selected risk factors on the risk of NAFLD in the Chinese population. Sci Rep 2019;9:4069.

Powell

Ali

Clouston

, et al. Steatosis is a cofactor in liver injury in hemochromatosis. Gastroenterology 2005;129:1937–1943.

Zhuo

, et al. The application value of ultra-short echo time MRI in the quantification of liver iron overload in a rat model. Quant Imaging Med Surg 2019;9:180–187.

Peng

Huang

Long

, et al. Liver iron quantification by 3 Tesla MRI: calibration on a rabbit model. J Magn Reson Imaging 2013;38:1585–1590.

10.

Henninger

Alustiza

Garbowski

, et al. Practical guide to quantification of hepatic iron with MRI. Eur Radiol 2020;30:383–393.

11.

Galimberti

Trombini

Bernasconi

, et al. Simultaneous liver iron and fat measures by magnetic resonance imaging in patients with hyperferritinemia. Scand J Gastroenterol 2015;50:429–438.

12.

Sharma

Altbach

Galons

, et al. Measurement of liver fat fraction and iron with MRI and MR spectroscopy techniques. Diagn Interv Radiol 2014;20:17–26.

13.

Fukuzawa

Hayashi

Takahashi

, et al. Evaluation of six-point modified Dixon and magnetic resonance spectroscopy for fat quantification: a fat-water-iron phantom study. Radiol Phys Technol 2017;10:349–358.

14.

Henninger

Plaikner

Zoller

, et al. Performance of different Dixon-based methods for MR liver iron assessment in comparison to a biopsy-validated R2* relaxometry method. Eur Radiol 2021;31:2252–2262.

15.

Yang

Huang

, et al. 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;9:1014–1024.

16.

Zhan

Olsen

Zhang

, et al. Detection of hepatic steatosis and iron content at 3 Tesla: comparison of two-point Dixon, quantitative multi-echo Dixon, and MR spectroscopy. Abdom Radiol (NY) 2019;44:3040–3048.

17.

Pineda

Sharma

, et al. Measurement of hepatic lipid: high-speed T2-corrected multiecho acquisition at 1H MR spectroscopy–a rapid and accurate technique. Radiology 2009;252:568–576.

18.

Satkunasingham

Besa

Bane

, et al. Liver fat quantification: comparison of dual-echo and triple-echo chemical shift MRI to MR spectroscopy. Eur J Radiol 2015;84:1452–1458.

19.

National Research Council (US) Committee for the Update of the Guide for the Care and Use of Laboratory Animals. Guide for the care and use of laboratory animals. 8th ed. Washington, DC: National Academies Press (US), 2011.

20.

Zhong

Nickel

Kannengiesser

, et al. Liver fat quantification using a multi-step adaptive fitting approach with multi-echo GRE imaging. Magn Reson Med 2014;72:1353–1365.

21.

Jhaveri

Kannengiesser

SAR

Ward

, et al. Prospective evaluation of an R2* method for assessing liver iron concentration (LIC) against FerriScan: derivation of the calibration curve and characterization of the nature and source of uncertainty in the relationship. J Magn Reson Imaging 2019;49:1467–1474.

22.

Olivieri

Brittenham

. Iron-chelating therapy and the treatment of thalassemia [published correction appears in Blood 1997 Apr 1;89(7):2621]. Blood 1997;89:739–761.

23.

Sharma

Martin

Pineda

, et al. Quantitative analysis of T2-correction in single-voxel magnetic resonance spectroscopy of hepatic lipid fraction. J Magn Reson Imaging 2009;29:629–635.

24.

França

Alberich-Bayarri

Martí-Bonmatí

, et al. Accurate simultaneous quantification of liver steatosis and iron overload in diffuse liver diseases with MRI. Abdom Radiol (NY) 2017;42:1434–1443.

25.

d’Assignies

Paisant

Bardou-Jacquet

, et al. Non-invasive measurement of liver iron concentration using 3-tesla magnetic resonance imaging: validation against biopsy. Eur Radiol 2018;28:2022–2030.

26.

St Pierre

El-Beshlawy

Elalfy

, et al. Multicenter validation of spin-density projection-assisted R2-MRI for the noninvasive measurement of liver iron concentration. Magn Reson Med 2014;71:2215–2223.