Christian Simonsson1,2,3, Shan Cai1,2, Jens Tellman1, Markus Karlsson1,4, Gunnar Cedersund2,3, Simone Ignatova5, Patrik Nasr6, Mattias Ekstedt2,6, Stergios Kechagias2,6, Nils Dahlström2,7, and Peter Lundberg1,2
1Department of Radiation Physics, Radiology, Department of Medical and Health Sciences, Linköping University, Linköping, Sweden, 2Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden, 3Department of Biomedical Engineering, Linköping University, Linköping, Sweden, 4AMRA Medical AB, Linköping, Sweden, 5Department of Clinical Pathology and Clinical Genetics, Department of Biomedical and Clinical Sciences, Linköping University, Linköping, Sweden, 6Department of Gastroenterology and Hepatology, Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden, 7Department of Radiology and Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden
Synopsis
Keywords: Liver, Liver
Motivation: Metabolic Associated Fatty Liver Disease (MAFLD) will increase the burden on future healthcare. Reliable screening alternatives such as MR-techniques e.g., nP-Dixon imaging and proton MR-spectroscopy are widely used. However, the number of studies comparing both MR-techniques with biopsies in clinically relevant cohorts is limited.
Goal(s): Compare MR Dixon imaging and proton MR-spectroscopy fat fraction with histological evaluation of steatosis
Approach: We have compared three different PDFF measurements; MRS, 6- and 2-point Dixon PDFF, in a prospective study, which included 84 clinically relevant patients with diffuse liver disease.
Results: All three techniques could accurately characterize histology based steatosis scores in this clinically relevant cohort.
Impact: In
this prospective study, with a clinically relevant cohort, a difference between
MRS and Dixon determined PDFF was measured. Both methods could distinguish
between different histological levels of steatosis, and both methods could therefore
characterize the patient group accurately.
Introduction
With a growing obese population, the prevalence of Metabolic
Associated Fatty Liver Disease (MAFLD) is increasing (1), which will lead to an increased economic burden
on future healthcare. The main characteristic of MAFLD is hepatic fat accumulation,
also denoted steatosis. Steatosis is not inherently dangerous, but can progress
into severe conditions such as NASH, cirrhosis and HCC. To identify these
patients early, non-invasive screening methods are important, including
techniques such as MRI- and MRS-derived proton density fat fraction (PDFF).
However, the number of studies
comparing both techniques with biopsies in clinically relevant cohorts is
limited. To this cause, we have
collected prospective data from a clinically relevant cohort comprised of 84
patients, all with suspected diffuse liver disease. The patients underwent an
extensive MR-investigation which included MRS, and two different Dixon
sequences. In addition, the patient also underwent ultrasound guided liver needle
biopsy typically within an hour after the MR-examination.Methods
A
total of 84 patients with suspected diffuse liver disease which were scheduled
to undergo liver biopsy were recruited from the Linköping University hospital
between the years 2016 to 2022. When arriving at the hospital the patients
underwent an extensive MR-examination, including several fat fraction measurements;
1H-MRS (PRESS sequence, voxel size 30x30x30 mm3),
six-point (6PD) mDixon Quant, and T1-weighted 2-point (2PD) mDixon, and the
examinations were performed by experienced radiological nurses. Within an hour after
the MR-examination the patients underwent an ultrasound guided biopsy. The
biopsy was sent for detailed histological evaluation by an experienced
histologist (SI).
Post-processing
of the MRS spectra and PDFF quantification was performed using LCmodel (Provencher,
Canada) by an experienced researcher (JT).
Post-processing and placement of
regions of interests (ROIs) in the 6PD images was done using MATLAB (2022a;
MathWorks, Natick, MA) by an experienced radiologist (ND), and placed in either
the 1H-MRS voxel, or in the liver segment from where the biopsy was
reportedly collected. 2PD images were quantified by fat-referenced (2)
MRI using AMRA Researcher (AMRA Medical AB, Linköping Sweden) and ROIs were
placed throughout the entire liver.Results
The
following measures of PDFF were used: 1H-MRS, measurements for the
two 6PD ROIs (MRS voxel, 6PDMRS-voxel, or biopsy segment, 6PDbiopsy)
and the 2PD which measured the average PDFF over several ROIs. Bland-Altman plots
were used to quantify the differences between the methods (Fig. 2). A larger difference
was observed when comparing the MRS-determined PDFF with each separate Dixon measurement
(Fig. 2A-C), compared to the comparison between the Dixon-measurements (Fig.
2D-F). The 3%, or 5% cut-off values for clinical diagnosis of steatosis (3) are also shown, and no significant bias could
be observed between the measurements, at these normal concentrations.
Linear
regression analysis was performed in order to explore the correlations (Fig. 3).
The highest correlation was found between the paired 6PD measurements (R2
= 0.90, Fig. 3D). These measurements also showed high correlation with the
global 2PD measurement (6PDMRS-voxel, R2 = 0.88 and 6PDbiopsy,
R2 = 0.85, Fig. 3E-F). For the MRS-measurements a similar
correlation between 6PDMRS-voxel and 2PD global measurement was
observed, R2 = 0.71 and R2 = 0.70 respectively (Fig.
3A-B).
The
ability to distinguish between different scores of histological steatosis was
also evaluated (Fig.4). The MRS-measurements were in most cases placed in Couinaud
segments 7-8, whereas the biopsies in most cases were acquired in segments 3,5
and 6. Nevertheless, all measurements could distinguish between different groups
of steatosis scores; zero, low and moderate (0 to 2). Significant p-values ≤0.001
were observed (Wilcoxon rank sum test) when comparing different grades
of steatosis, independent of detection method. Discussion
The
results showed that there were larger PDFF differences between measurement
techniques, as compared to expected disease related to the fat storage heterogeneity
between sites for needle biopsy, and ROI-placements. For example, even though
the MRS-measurement and the 6PDMRS-voxel were positioned in an
identical location, the differences between these measurements were larger compared
with to the comparison between Dixon measurements. Correlation analysis
resulted in similar conclusions. Nevertheless, all techniques could distinguish
between different levels of steatosis. Conclusion
In
conclusion, a small difference between MRS and Dixon derived PDFF values was
observed at a group level. However, all three techniques could accurately distinguish
between histology-based steatosis scores
in this clinically relevant cohort. Acknowledgements
No acknowledgement found.References
1. Riazi K, Azhari H, Charette JH,
Underwood FE, King JA, Afshar EE, et al. The prevalence and incidence of NAFLD
worldwide: a systematic review and meta-analysis. Lancet Gastroenterol Hepatol.
2022 Sep 1;7(9):851–61.
2. Borga M, Ahlgren A, Romu T, Widholm P,
Dahlqvist Leinhard O, West J. Reproducibility and repeatability of MRI-based
body composition analysis. Magn Reson Med. 2020;84(6):3146–56.
3. Nasr P, Forsgren MF, Ignatova S,
Dahlström N, Cedersund G, Leinhard OD, et al. Using a 3% Proton Density Fat
Fraction as a Cut-off Value Increases Sensitivity of Detection of Hepatic
Steatosis, Based on Results from Histopathology Analysis. Gastroenterology.
2017;153(1):53–5.