Lena Vaclavu1, Dilek Betül Arslan2, Lydiane Hirschler1, Carles Falcon3,4, Esin Ozturk-Isik2, Juan Domingo Gispert3,4, Paula Montesinos5, Kim van de Ven6, and Matthias JP van Osch1
1Department of Radiology, C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands, 2Biomedical Engineering Institute, Boğaziçi University, Istanbul, Turkey, 3BarcelonaBeta Brain Research Center (BBRC), Pasqual Maragall Foundation, Barcelona, Spain, 4CIBER-BBN, Madrid, Spain, 5Philips Healthcare Iberia, Madrid, Spain, 6Philips Healthcare, Best, Netherlands
Synopsis
Physiologic fluctuations can lead to the
appearance of artefactual angiographic signal in the perfusion image of
time-encoded pCASL previously dubbed “shine-through effect”. The aim of this
study was to investigate the potential mechanisms inducing this artefact by characterizing
the noise in time-encoded pCASL signal from individual Hadamard columns and
rows as well as without labeling. We observed higher temporal standard
deviation in the arteries compared to gray matter, and found that the
shine-through effect was present even without labeling, suggesting that it is caused
by the selective background suppression pulse applied close to acquisition time
and the associated fresh inflow.
Introduction
Time-encoded arterial spin labeling (te-pCASL) is a multi-time point
perfusion technique in which temporal information is encoded according to, for
example, a Hadamard matrix1. The decoding process ensures that for
each post-label delay (PLD), signals from all other PLDs cancel out. In
practice however, physiologic fluctuations that are present in vessels, lead to
angiographic signal in the perfusion image, previously dubbed “shine-through
effect”2 (Figure 1). This artefact may affect clinical
interpretation, since it can be mistaken for a transit delay artefact
indicative of large vessel pathology and delayed arrival of label3.
The exact cause of the shine-through effect is unknown, but theoretically could
be caused by, 1) decoding errors of the high intensity angiographic signal, 2)
slight mismatches between the Hadamard-encoded bolus due to cardiac pulsations,
3) imperfect interleaving of control and label for certain sub-boli, or 4)
artefacts due to the necessity for selective background suppression (BGS)4,5
pulses applied during the labeling6.
The aim of this study was to investigate the origin of the
shine-through effect in te-pCASL.Methods
Imaging: All images were acquired at 3T (Ingenia CX/Achieva,
Philips, NL) using a 32-channel head-coil. Thirty volunteers were initially
scanned for both standard pCASL and te-pCASL using a 2D GE-EPI sequence with SENSE=2.5,
FOV=220x220x120 mm3, resolution=2.3x2.3x6 mm3 and flip
angle=90°. Standard pCASL|te-pCASL parameters were: 30|8 averages, TR=4600|4127 ms,
TE=12|10 ms, label duration=1800|1800/600/400/300/200/150/150 ms, PLD=2000|200 ms, duration=5:07|5:52 minutes. For BGS after labeling, two global hyperbolic
secant inversion pulses were used. For BGS during labeling, two selective FOCI
inversion pulses were used. Optimal BGS-times were determined automatically by
the scanner console.
Additional experiments: To study the effect of individual blocks, the
full Hadamard scheme was acquired as normal, except that radiofrequency (RF)
pulses were only enabled for a single block at a time (Figure 2). Global
and selective BGS pulses were compared while the RF of the pCASL labeling was disabled
to isolate the effect of the BGS-pulses. Finally, a single-BGS pulse was
applied at varying times (2800/1800/1300/300ms) before the readout.
Analysis: The contribution
of individual Hadamard lines to noise (temporal standard deviation [tStdv]) was
investigated in the fully acquired Hadamard matrix in gray matter (GM) and
arterial regions-of-interest. GM was defined by T1 tissue segmentation, and an arterial
region-of-interest was defined by the average signal of the three shortest PLDs which consist mainly
of angiographic signal. For the additional experiments, tStdv was calculated
over the ASL (subtracted) signal when labeling was enabled, or over the raw unsubtracted
images when labeling was disabled. . Results
The shine-through
artefact was observed in the subtracted ASL images (Figure 1a), but
because it can be either positive or negative, the tStdv of the ASL signal is
more informative to visualize and quantify the effect (Figure 1b). In 30
volunteers, (mean±std age: 55±16y, range: 28-75y, 50% male), compared to
standard pCASL, GM tStdv was the same for all Hadamard lines (Figure 3a).
However, arterial tStdv was significantly higher for all Hadamard lines except for
the first line consisting of only label condition (Figure 3b). Since the
images with increased tStdv were those in which label and control conditions switch
according to the Hadamard scheme, we hypothesized that this higher tStdv could
be caused by mismatches between Hadamard lines (e.g. due to small differences
in the distance travelled by the encoded bolus into the vascular tree). However,
when labeling individual blocks only, we found that the artefact was also
visible in the disabled blocks (Figure 4), and to our surprise, also in
the experiment without any labeling at all (Figure 5a). Finally, lower
tStdv was observed with global versus selective BGS-pulses (without any
labeling, Figure 5b), supporting a potential mechanism of fresh inflow
of untouched spins from below the edge of the BGS-pulse. The timing of the single-BGS
pulses had a small influence on tStdv depending on its proximity in time, to
the readout, with selective pulses leading to higher tStdv (Figure 5c). Discussion
Our results indicate that the shine-through
effect is associated with increased signal fluctuations in arterial vessels. We
observed that the shine-through effect is not dependent on the Hadamard
labeling procedure, since it was also visible in absence of RF labeling. One
major difference between standard and te-pCASL is that BGS is applied globally
during the PLD in standard pCASL, and selectively during the labeling in
te-pCASL (because the PLD is too short for optimal BGS). This creates a
difference in the suppression of spins below the labeling plane such that fresh inflowing blood will result in enhancement
of arterial fluctuations when selective BGS is used. This hypothesis was
supported by our single-BGS experiments showing a difference between the selective
and global pulse placed between 1300 and 1800 ms before the readout. An
explanation for this could be that global BGS inverts all spins within the
range of the transmit coil, while selective BGS inverts only the spins above
the middle of the labeling plane, creating a transition zone of
inverted-uninverted spins flowing into the large vessels and leading to
increased tStdv.Conclusion
We
observed increased signal fluctuations in arteries in te-pACSL compared to standard
pCASL, which was not related to the Hadamard-encoding but due to selective
background suppression.Acknowledgements
This research was supported by the EU under the Horizon2020 program (project: CDS-QUAMRI), and by Philips Healthcare.References
- Günther M. Highly efficient accelerated
acquisition of perfusion inflow series by Cycled Arterial Spin Labeling.
Proceedings of the 15th Annual Meeting of ISMRM. 2007, p. 380 of
ISMRM. Vol. 15. Berlin, Germany, 2007
- Suzuki Y, Helle M, Koken P, et al. Simultaneous
acquisition of perfusion image and dynamic MR angiography using time‐encoded
pseudo‐continuous ASL. Magn. Reson. Med. 2018;79(5):2676-2684.
- Di Napoli A, Cheng SF, Gregson J, et al.
Arterial Spin Labeling MRI in Carotid Stenosis: Arterial Transit Artifacts May
Predict Symptoms. Radiology. 2020;297:652-660.
- Mani S, Pauly J, Conolly S, et al. Background
suppression with multiple inversion recovery nulling: applications to
projective angiography. Magn. Reson. Med. 1997;37: 898–905.
- Maleki N, Dai W & Alsop DC. Optimization
of background suppression for arterial spin labeling perfusion imaging. Magma 2012;25:
127–133.
- Dai W, Robson PM, Shankaranarayanan A, et al.
Benefits of Interleaved Continuous Labeling and Background Suppression.
Proceedings of the 18th Annual Meeting of ISMRM. 2010, p. 1748