Manil Chouhan1, Rajiv Ramasawmy2, Alan Bainbridge3, Adrienne Campbell-Washburn2, Jack Wells2, Shonit Punwani1, Rajeshwar Mookerjee4, Simon Walker-Samuel2, Mark Lythgoe2, and Stuart Taylor1
1UCL Centre for Medical Imaging, University College London, London, United Kingdom, 2UCL Centre for Advanced Biomedical Imaging, University College London, London, United Kingdom, 3Department of Medical Physics, University College London Hospitals NHS Trust, London, United Kingdom, 4UCL Institute for Liver and Digestive Health, University College London, London, United Kingdom
Synopsis
Non-invasive preclinical liver perfusion measurements
could be used to develop biomarkers and assess new treatments for liver disease
and primary/secondary malignant liver lesions. ASL can provide regional hepatic perfusion
maps, and in this study we compare FAIR ASL tissue perfusion measurements with
caval subtraction phase-contrast MRI, a validated method for measuring total
liver blood flow, to demonstrate ASL overestimation but encouraging agreement
between both methods.Purpose
Non-invasive preclinical liver perfusion measurements
could be used to develop biomarkers and assess new treatments for liver disease(1) and primary/secondary
malignant liver lesions(2). Arterial spin labelling (ASL) has been applied
in other organs(3,4)
but has been challenging to implement in the liver, mainly because of its dual hepatic
arterial (HA) and portal venous (PV) vascular supply, and susceptibility to
respiratory motion.
Previously we have demonstrated that a Look-Locker
Flow-Sensitive Alternating Inversion Recovery (FAIR) ASL technique is suitable
for pre-clinical liver perfusion measurements(5) and compared these with phase-contrast (PC) MRI
measurements of the portal vein (PV) in rats(6). However,
the PV only contributes approximately 75% of liver perfusion and as such the
comparative PCMRI measurement excludes the contribution of the HA. Caval subtraction PCMRI is a novel, validated
method for assessment of total liver blood flow (TLBF) using measurements of
bulk flow in the inferior vena cava(7). In this study we combine ASL with caval
subtraction PCMRI at 9.4T to assess agreement between methods within subjects.
Methods
Scans were performed on a 9.4T Agilent 20 cm
horizontal-bore system (Agilent Technologies, USA), using a 72 mm birdcage coil
(RAPID Technologies, Germany). Healthy Sprague-Dawley rats (n=9) were
anaesthetised with isoflurane, and their respiratory rate, ECG and temperature
monitored (SA Instruments, USA).
Two-dimensional
cine PCMRI
Axial and angled coronal gradient echo images were used
to plan caval PCMRI studies. Cardiac and respiratory-triggered 2D cine PCMRI
was performed (TR/TE=10/1.2ms, α=10°, 2 mm slice thickness, 192x192 matrix, FOV
40x40mm2, 10-15 cardiac cycle phases, Venc=33 and 66 cm/s for proximal
and distal IVC flows). ROIs were positioned manually on each vessel for each
frame of the cardiac cycle and flow quantification was performed using in-house
developed Matlab code. TLBF was
estimated by subtracting proximal IVC flow (above renal but below hepatic
venous inlets) from distal IVC flow (above hepatic venous inlets, but below the
IVC-right atrial junction). Estimated
TLBF measurements were normalised to explanted liver weight.
ASL acquistion
An axial slice that enabled good visualisation
of hepatic parenchyma was selected from respiratory-gated anatomical
images. A FAIR Look-Locker ASL sequence
was used using an end-expiration triggered segmented acquisition with a spoiled
gradient-echo readout (FOV 60x60 mm2, 128x128 matrix, 2 mm slice
thickness, TE=1.18 ms, TI=110 ms, TRRF=2.3 ms, αLL=8˚, TRI=13
seconds, 50 inversion recovery readouts, 15 minute acquisition time). Images were processed using in-house
developed Matlab code and ASL perfusion maps were calculated using the Belle
model(3), as described previously(5). Three
identically sized circular ROIs were placed on the hepatic parenchyma avoiding major
vascular structures and extra-hepatic tissues. ASL perfusion estimates were based on averages
obtained from the three ROIs (figure 1).
Results
Mean
ASL bulk liver perfusion (351.1±26.4 ml/min/100g) was less than mean caval
subtraction PCRMI estimated TLBF (459.9±40.5 ml/min/100g). The mean difference (bias) between ASL and
PCMRI measurements of bulk liver perfusion was 144.8 ml/min/100g, suggesting ASL
underestimation. The Bland-Altman 95%
Limits of Agreement (BA 95% LoA) were ±166.9 ml/min/100g, with significant
positive correlations between measurements (r=0.7162, p=0.0150, figure 2). The coefficient of variation was similar for
both ASL (25.1%) and caval subtraction PCMRI TLBF (26.4%).
Discussion
Previously,
we have demonstrated the feasibility of measuring localised liver perfusion using
FAIR ASL(5) and reported expected overestimation of FAIR
ASL bulk liver perfusion relative to weight normalised PCMRI bulk PV flow(6). While
ASL perfusion maps are a mixture of both the arterial and portal venous
contribution, in this study we have demonstrated a tendency for FAIR ASL bulk
liver perfusion to underestimate perfusion relative to caval subtraction PCMRI
estimated TLBF. Nonetheless, allowing
for the systematic bias, the BA 95% LoAs and strong correlation are both
encouraging for the use of FAIR ASL for the assessment of regional hepatic
perfusion. The underestimation of FAIR
ASL may be due to imperfect labelling due to the large coil loading(8). Our use
of caval subtraction PCMRI to validate FAIR ASL is a useful application of this
method and will be a valuable tool for the assessment of pre-clinical models of
hepatic disease.
Conclusion
FAIR ASL tends to overestimate hepatic perfusion, but
demonstrates encouraging agreement with caval subtraction PCMRI.
Acknowledgements
This work was supported by a Wellcome Trust Clinical
Research Training Fellowship (grant WT092186), a Wellcome Trust Senior
Research Fellowship (grant WT100247MA), an MRC Capacity Building
Studentship and the British Heart Foundation, King’s College London and UCL
Comprehensive Cancer Imaging Centre CR-UK & EPSRC, in association with the
DoH (England).References
1. Van Beers BE, Leconte I, Materne R,
Smith AM, Jamart J, Horsmans Y. Hepatic perfusion parameters in chronic liver
disease: dynamic CT measurements correlated with disease severity. AJR Am J
Roentgenol 2001;176(3):667-673.
2. Jackson
A, Haroon H, Zhu XP, Li KL, Thacker NA, Jayson G. Breath-hold perfusion and
permeability mapping of hepatic malignancies using magnetic resonance imaging
and a first-pass leakage profile model. NMR in biomedicine 2002;15(2):164-173.
3. Belle
V, Kahler E, Waller C, et al. In vivo quantitative mapping of cardiac perfusion
in rats using a noninvasive MR spin-labeling method. Journal of magnetic
resonance imaging : JMRI 1998;8(6):1240-1245.
4. Golay
X, Hendrikse J, Lim TC. Perfusion imaging using arterial spin labeling. Topics
in magnetic resonance imaging : TMRI 2004;15(1):10-27.
5. Ramasawmy
R, Campbell-Washburn AE, Wells JA, et al. Hepatic arterial spin labelling MRI:
an initial evaluation in mice. NMR in biomedicine 2015;28(2):272-280.
6. Chouhan
M, Ramasawmy R, Campbell-Washburn A, et al. Measurement of bulk liver
perfusion: initial assessment of agreement between ASL and phase-contrast MRI
at 9.4T. Proc Intl Soc Mag Reson Med. Volume 21; 2013. p. 2190.
7. Chouhan
M, Mookerjee R, Bainbridge A, et al. Caval subtraction 2D phase-contrast MRI to
measure total liver and hepatic arterial blood flow: preclinical validation and
initial clinical translation. Radiology 2015 (accepted, in submission).
8. Wells JA, Siow B, Lythgoe MF, Thomas DL. The importance
of RF bandwidth for effective tagging in pulsed arterial spin labeling MRI at
9.4T. NMR in biomedicine 2012;25(10):1139-1143.