4070

Combined Effect of Iron Overload and Steatosis on Liver R2* Using Morphological Modeling and MRI Signal Synthesis via Monte Carlo Simulations

Utsav Shrestha¹, Juan Pablo Esparza¹, Sanjaya Satapathy², Jason Vanatta³, and Aaryani Tipirneni-Sajja^1,4
¹Biomedical Engineering, The University of Memphis, Memphis, TN, United States, ²Department of Medicine, North Shore University Hospital/Northwell Health, Manhasset, NY, United States, ³College of Medicine, University of Tennessee Health Science Center, Memphis, TN, United States, ⁴St. Jude Children’s Research Hospital, Memphis, TN, United States

Synopsis

Chemical-shift based multi-spectral fat-water models accounting for single or dual R2* correction are used to assess hepatic iron concentration (HIC) and fat fraction (FF). In this study, we developed a Monte-Carlo based approach for simulating steatosis and iron overload models mimicking the in-vivo characteristics and synthesizing MRI signal to compare the accuracy of the R2* models to estimate FF in the presence of iron. Our results show that R2* estimated by single R2* model is highly governed by iron whereas dual R2* model predicts R2*-FF relationship close to the in-vivo calibration. Nevertheless, FF predicted by both the models were close to the true FF.

Introduction

Chemical shift based multi-spectral fat-water MRI signal models with single or dual R2*(1/T2*) correction have been proposed for simultaneous assessment of hepatic iron content (HIC) and steatosis^1,2. For simplicity, most multi-spectral models that incorporate the R2* correction assume that the R2* for water (R2*W) and fat (R2*F) are the same (single R2*). However, this assumption may not hold true in co-existing conditions of iron overload and steatosis as iron could affect the MRI signal of water and fat molecules differently due to their difference in molecular sizes³. The objective of this study is to compare the accuracy of R2* and FF estimation of single and dual R2* models in the presence of both iron and fat by building combined liver iron overload and steatosis models using morphological descriptors from histology^4,5 and synthesizing MRI signal using Monte-Carlo simulations.

Methods

Fat droplets (FDs) were segmented from 68 human liver histology samples. Morphometric (radius, circularity) and space organizational (NN distance and regional anisotropy) characteristics of FDs were extracted from the segmented images and were statistically modeled using gamma distribution function (GDF). Relationships between the extracted features and GDF parameters were computed using regression analysis. Similar relationships were obtained for iron from a study conducted by Ghugre et al⁴. Virtual liver volume with dimension 120X120X120 µm³ was simulated with different FFs and HICs based on those relationships.
MRI signal was generated using the Monte-Carlo approach described in previous studies^4,5, by accounting for both iron-induced and fat-induced susceptibilities. The MRI signals from water and fat were computed and superimposed to obtain the final combined MRI signal. The signals were synthesized for FFs ranging from 0-50% and HIC = [0.4, 0.6, 0.8, 1, 5] mg of iron/g of dry tissue weight with echo times (TE): 1, 1.5, 2, …, 14.5ms at 1.5T.
R2* and FF were calculated using fat water toolbox¹ (FWT) (single R2*) and auto regressive moving average² (ARMA) (dual R2*) models. R2*-FF as well as R2*-HIC relationship were compared with their corresponding in-vivo calibration and the relationship between R2*, FF and HIC was computed using multi-regression analysis.

Results & Discussion

Figure 1 shows (a) 3D and (b) its corresponding 2D slice of a representative virtual liver model with FF of 10% and HIC of 5mg/g which mimics steatosis and iron overload with in-vivo sizes and organization of iron spheres and FDs. R2* predicted by FWT was 60s^-1 and 240s^-1 for HIC<=1mg/g and HIC=5mg/g respectively for all FFs simulated. However, R2*W predicted by ARMA changed with FF for the simulated HICs. Hence, rest of the R2* results are based on R2*W predicted by ARMA. As shown in Figure 2(a), predicted R2*W values showed positive correlation with FFs, with slopes similar to the Karlsson calibration for HIC < 1.2mg/g⁶. Figure 2(b) shows decrease in the slope of R2*W-FF relationship at HIC of 5mg/g suggesting that at higher HIC most of the signal is dominated by iron. R2*F predicted by the ARMA model shows weak-moderate correlation with FF (R² < 0.6) as demonstrated in Figure 3. Figure 4 compares the estimated R2*W-HIC relationship with the Wood⁷ and Karlsson calibration⁶. The estimated relationship was closer to the Wood calibration. The predicted relationship between R2*, FF and HIC for this study is:
$$R_2^*=37.43*HIC+0.60*FF+20.13$$
The predicted FF values were in excellent agreement with true FFs, with slopes ~1 for both ARMA and FWT (Figure 5).
Hence, our simulations performed by creating iron overload and steatosis models and synthesizing MRI signal demonstrate that FWT single R2* has less sensitivity to the presence of FDs and its R2* estimation is dominated by iron particles whereas ARMA dual R2* estimation shows a positive correlation with FF. Even though the R2* predicted by FWT did not change with FF, it was able to predict FF with high accuracy which might be because fat accumulation in liver causes only a slight increase in liver R2*⁸. For higher HICs, iron signal will be much greater than the fat signal which can cause FWT to predict FF incorrectly. Moreover, this work demonstrates the feasibility of performing Monte-Carlo modeling to study the association between R2*, FF and HIC and determine the optimal R2* fat-water model for in-vivo studies.

Conclusion

This study shows that under the condition of steatosis and iron overload, the R2* predicted by single R2* model is dominated by iron signal whereas dual R2* can predict R2* values based on FF, which are close to the published in-vivo calibrations. Future work will perform a systematic investigation to study the association between R2*, HIC and FF for a wide range of HICs and FFs.

Acknowledgements

We acknowledge the support of National Institute of Health (NIH), #1R21EB031298.

References

1. Hamilton G, Yokoo T, Bydder M, et al. In vivo characterization of the liver fat ¹H MR spectrum. NMR Biomed. 2011;24(7):784-790.

2. Tipirneni-Sajja A, Krafft AJ, Loeffler RB, et al. Autoregressive moving average modeling for hepatic iron quantification in the presence of fat. J Magn Reson Imaging. 2019;50(5):1620-1632.

3. Horng DE, Hernando D, Reeder SB. Quantitation of liver fat in the presence of iron overload. J Magn Reson Imaging. 2017;45(2):428-439.

4. Ghugre NR, & Wood JC. Relaxivity‐iron calibration in hepatic iron overload: probing underlying biophysical mechanisms using a Monte Carlo model. Magn Reson Med. 2011;65(3): 837-847.

5. Shrestha U, van der Merwe M, Kumar N, et al. Morphological characterization of hepatic steatosis and Monte Carlo modeling of MRI signal for accurate quantification of fat fraction and relaxivity. NMR Biomed. 2021;34(6):e4489.

6. Karlsson M, Ekstedt M, Dahlström N, et al. Liver R2* is affected by both iron and fat: A dual biopsy‐validated study of chronic liver disease. J Magn Reson Imaging. 2019;50(1):325-33

7. Wood JC, Enriquez C, Ghugre N, et al. MRI R2 and R2* mapping accurately estimates hepatic iron concentration in transfusion-dependent thalassemia and sickle cell disease patients. Blood. 2005;106(4):1460-5.

8. Mamidipalli A, Hamilton G, Manning P, et al. Cross-sectional correlation between hepatic R2* and proton density fat fraction (PDFF) in children with hepatic steatosis. J Magn Reson Imaging. 2017;47:418–424.

Figures

Figure 1. Model predicted fat and iron morphology with 10% FF and 5mg/g HIC in a liver volume of 120×120×120μm3. Distribution of FDs (white spheres) and iron spheres (blue) in 3D liver geometry (a) and the corresponding 2D 4-μm-thick cross-sections (b).

Figure 2. Predicted water R2* (R2*W) in the presence of iron obtained by using ARMA dual R2* fat-water model for FFs ranging from 0% to 30% for HIC <= 1mg/g (a) and 0% to 50% for HIC = 5mg/g (b). Solid black line denotes the in-vivo R2*-FF calibration in the presence of HIC <= 1.2mg/g. The R2*-FF relationship in (a) showed close agreement with the in-vivo calibration.

Figure 3. Estimated fat R2* (R2*F) obtained by using ARMA dual R2* fat-water model for FF ranging from 0% to 30% for HIC <= 1mg/g (a) and 0% to 50% for HIC = 5mg/g (b). In the presence of iron, R2*F showed low correlation with FF for HIC <= 1 mg/g and moderate correlation with FF for HIC = 5 mg/g.

Figure 4. Estimated water R2* (R2*W) obtained by using ARMA dual R2* model for FFs ranging from 0% to 50% and HIC = [0.4, 0.6, 0.8, 1, 5]mg/g. Solid black and red line denotes the in-vivo R2*-FF calibration by Wood et al. and Karlsson et al. respectively. The R2*-HIC relationship showed close agreement with the Wood calibration.

Figure 5. Linear regression analysis between predicted FF and simulated FF for FF ranging from 0% to 30% for HIC <= 1mg/g (a) and 0% to 50% for HIC = 5mg/g (b). The dotted line indicates the line of unity. For both the R2* models, the predicted FF values showed an excellent agreement with simulated FFs, with slopes close to unity and intercepts close to zero.

Proc. Intl. Soc. Mag. Reson. Med. 30 (2022)

4070

DOI: https://doi.org/10.58530/2022/4070