Kanishka Sharma1, Bashair Alhummiany2, David Shelley2,3, Margaret Saysell2,3, Maria-Alexandra Olaru4, Bernd Kühn4, Julie Bailey3, Kelly Wroe3, Cherry Coupland3, Michael Mansfield3, and Steven Sourbron1
1Department of Imaging, Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, United Kingdom, 2Department of Biomedical Imaging Sciences, University of Leeds, Leeds, United Kingdom, 3Leeds Teaching Hospitals, Leeds, United Kingdom, 4Siemens Healthcare GmbH, Erlangen, Germany
Synopsis
The iBEAt study MRI biomarker
panel has been developed to determine if imaging biomarkers can improve
predictions of renal function decline in diabetic kidney disease. The aim of
this pilot study was to inform the approach to quality control and image
processing of iBEAt data. Repeatability in 5 healthy volunteers (4 scans per
subject) was analysed with a prototype image-processing pipeline. The results indicated
that repeatability of renal T1, T2* and
phase-contrast renal blood flow was comparable to literature, and measurement
precision was sufficient to detect differences between volunteers. ASL had poor
repeatability presumably from B0 inhomogeneities in labelling plane.
Introduction
Multi-parametric renal MRI allows
characterisation of renal tissue structure and function, facilitating
non-invasive interpretation of pathophysiological changes in the kidney. The
iBEAt study MRI biomarker panel has been developed to determine if imaging
biomarkers can improve predictions of renal function decline in diabetic kidney
disease1 (DKD). The aim of this pilot study was to inform the
approach to quality control and image processing of iBEAt data. Repeatability
data in healthy volunteers were analysed with a prototype image processing pipeline,
and results were compared to literature values. Methods
Data acquisition: Five healthy volunteers (HV: 4F/1M; median
age: 39 yrs) were each scanned 4 times on MAGNETOM Prisma 3T MRI (Siemens
Healthcare GmbH, Erlangen, Germany) using the iBEAt MRI protocol1. HVs
arrived fasted (>8hrs) and were provided with standardised meal & fluid
prior to the MRI scan (same hour/weekday per HV). MRI acquisition included IR-prepared
T1-MOLLI, Multi-echo T2*, 2D-cine PC-MRI,
and 3D-ASL (pCASL, prototype sequence*). The MRI acquisition parameters are available
in Gooding et al. (supplementary material)1.
*Work-In-Progress package, the product is currently under
development and is not for sale in the US and in other countries. Its future
availability cannot be ensured.
Data post-processing:
Relaxation times. Kidney cortex and medulla ROIs were selected on
the middle slice of kidneys for each sequence with six circular regions (3 per
kidney; Figure 1) using ImageJ2. Cortex and medulla ROI masks were imported
in MATLAB3 for pixel-wise nonlinear curve fitting with Levenberg‐Marquardt
algorithm using the standard MOLLI4 model for T1
and a mono-exponential decay for T2*.
Renal perfusion. PC-MRI was analysed on syngo.via (Siemens
Healthcare GmbH, Erlangen, Germany) with an oval-shape ROI, threshold adjusted
to segment the renal artery. RBF was averaged over the cardiac cycle and
normalised to body-surface-area (DuBois formula). ASL perfusion maps were
generated inline on the scanner and average parenchymal perfusion was
extraction by semi-automatic segmentation on the perfusion map.
Summary parameters: For
each parameter and for every volunteer, a repeatability error (RE) and relative
repeatability error (RRE) were calculated as RE = 1.96 x SD and RRE = 1.96 x
SD/AVR, where SD is the standard deviation over 4 measurements and AVR is the
average. With this definition, RE and RRE measure the absolute and relative 95%
confidence interval (CI) for a measurement in an individual. RE and RRE were
averaged over all volunteers and the 95% CI on the average was determined as
1.96 x standard error. Boxplots and pairwise t-tests
were computed in R (v.4.0.3)5.Results
Table 1 shows the main results. RRE was lowest for T1
(<5%), followed by T2* (11-14%), PC-RBF (11%), and renal
perfusion using ASL (61%).
Figure 3 compares all parameters between
individual subjects. Between-subject variability was evident with
statistically significant differences among subjects in T1,
PC-MRI, and ASL. Discussion
After correction for different
definitions of the metrics, the repeatability of relaxation times and arterial RBF
(PC-MRI) were comparable with literature6,7, but ASL was
significantly worse (Table 1).
The ASL RBF maps showed large
signal dropout in individual cases (Figure 2), consistent with the effect of B0
inhomogeneity in the labelling plane. This artifact was largely
reproducible within a subject and the main contributor to the observed
differences between subjects. In the current acquisition protocol, the shim
volume did not include the labelling plane, an area sensitive to B0
effects. For future studies, the acquisition protocol will incorporate volume
shim across the labelling plane.
Inspection of T1
outliers (eg. HV1 scan 4) showed that kidney motion can result in T1
errors (Figure 3). The final processing protocol will mitigate this effect by including
motion-correction in breath hold scans and extend the ROIs to whole
medulla/cortex ROIs to further reduce motion sensitivity. Comparison of T1-MOLLI model fit to the
data also revealed a smaller mismatch, potentially because the T1-MOLLI model does not
account for differently spaced inversion times. Future analyses will use a more
general signal model to correct for this effect.
A unique feature of this pilot
study is the use of 4 rather than 2 repeat scans to measure repeatability
errors, providing relatively accurate assessment of the measurement uncertainty
at a single site. The data indicated that subtle differences in T1,
T2* and phase-contrast RBF between HVs may reflect actual
physiological differences instead of measurement uncertainty. Future studies
are needed to better understand these differences and their relationship to factors
including gender, age, obesity, or ethnicity. Conclusion
The data show good repeatability
of the iBEAt protocol for T1,
T2*, and phase contrast
with potential improvement in relaxation times by refining the processing
protocol. ASL repeatability was poor but expected to improve significantly
after minor adjustments in the acquisition. Data suggest that these markers are
sufficiently sensitive to capture subtle structural differences in healthy
renal tissue. Acknowledgements
iBEAt study is part of the
BEAt-DKD project. The BEAt-DKD project has received funding from the Innovative
Medicines Initiative 2 Joint Undertaking under grant agreement No 115974. This
Joint Undertaking receives support from the European Union’s Horizon 2020
research and innovation programme and EFPIA with JDRF. For a full list of
BEAt-DKD partners, see www.beat-dkd.eu.
B.
Alhummiany is supported by a government scholarship from Saudi Arabia.
References
- Gooding
KM, et al. Prognostic Imaging Biomarkers for Diabetic Kidney Disease (iBEAt):
Study protocol, BMC Nephrol. 2020 Jun; 21(242).
- Schneider,
C. A.; Rasband, W. S. & Eliceiri, K. W. (2012), "NIH Image to ImageJ:
25 years of image analysis", Nature methods 9(7): 671-675, PMID 22930834
(on Google Scholar).
- MATLAB
2018a, The MathWorks, Inc., Natick, Massachusetts, United States.
- Messroghli
DR, Radjenovic A, Kozerke S, Higgins DM, Sivananthan MU, Ridgway JP. Modified
Look-Locker inversion recovery (MOLLI) for high-resolution T1 mapping of the
heart. Magn Reson Med. 2004;52(1):141-146.
- R
Core Team (2013). R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria.URL www.R-project.org/.
- de
Boer A, Harteveld AA, Stemkens B, et al. Multiparametric Renal MRI: An
Intrasubject Test-Retest Repeatability Study [published online ahead of print,
2020 Apr 16]. J Magn Reson Imaging. 2020;10.1002/jmri.27167.
- Cox
EF, Buchanan CE, Bradley CR, et al. Multiparametric Renal Magnetic Resonance
Imaging: Validation, Interventions, and Alterations in Chronic Kidney Disease.
Front Physiol. 2017;8:696. Published 2017 Sep 14.