Natalia Dudysheva1,2, Nicolas Boulant1, Alexandre Vignaud1, and Franck Mauconduit1
1CEA, NeuroSpin, Paris-Saclay University, CNRS, Gif-sur-Yvette, France, 2Multiwave Imaging, Marseille, France
Synopsis
In this
work, we propose to extend the capacities of ultra-low SAR protocol used in
coil design. A developed restricted SAR protocol holds the worst-case SAR below
the threshold by constraining the total transmitted RF power. Since SAR
reflects tissue heating, we exploit a conservative thermal model to derive
relaxed SAR constraints depending on acquisition time. This enables to extend
the options available for sequences played under this mode, including fat
suppression, higher image resolution, or decreased TR.
Introduction
Standard
MRI examinations impose specific absorption rate (SAR) limitations to ensure
patient safety. For head studies, the IEC guidelines set the limits of 3.2 W/kg
and 10 W/kg for the whole-head and local (10 grams of tissue) SARs for 6 min and
double these values for 10 seconds1. These conditions should guarantee that
temperature rise does not exceed 2 degrees under basal conditions of 37 °C. Specific MRI studies suggest a severe SAR limitation to safely study patients
with implants2,3 or evaluate coil design in vivo4 without
SAR pre-simulations. In this context, the restricted SAR methodology4
automatically respects SAR limits for all sequence settings and coils by assuming
that the total transmitted power is entirely absorbed by a 10-gram tissue area.
Thus, the maximum average power is 0.1 W for 6 min (0.2 W for 10 sec). This situation
represents an overstated limitation but does not make any assumption about the
SAR distribution. Here, we propose to estimate conservatively the temperature
rise of this 10-gram sphere of tissue and examine the conditions to relieve the
restrictions while preserving the safety aspects based on the maximum
temperature rise limit.Methods
Assuming
all the power absorbed by a sphere of 10g of tissue (R=1.34 cm considering the
brain tissue density) and only heat conduction in an infinite continuous medium,
the temperature rise over time induced by RF energy absorption obeys the
following diffusion equation: $$ρC\frac{\partial T}{\partial t}-κ\Delta T=ρSAR.$$ Here ρ=1000 kg/m3 is the
density, C=4000 J·kg-1·K-1 denotes the specific heat
capacity, and κ=0.5 Wm-1·K-1 corresponds
to the thermal conductivity. These values do not vary strongly with the head tissues. We first consider SAR as a source with a uniform distribution
within the sphere. By introducing α=κ/(ρC), transforming the equation into Fourier space,
solving several integrals, and going back to the spatial domain, the exact
analytical result for the maximum temperature in the sphere center can be
obtained: $$T_{max}(t)=\frac{ρSAR}{2κ}R^2 \left(1-\sqrt{\frac{4αt}{πR^2}}e^{-\frac{R^2}{4αt}} - \left(1-2\frac{αt}{R^2}\right)erf \left(\frac{R}{2\sqrt{αt}}\right) \right).$$ Next,
we compare the result with the Gaussian distribution case, where $$$SAR(x,y,z)=be^{-\frac{x^2+y^2+z^2}{2σ^2}}.$$$ The same methodology provides the maximum temperature
rise described by the following exact expression: $$T_{max}(t)=\frac{ρbσ^2}{κ}\left(1 - \frac{1}{\sqrt{1+\frac{2αt}{σ^2}}} \right) \quad with \quad b=\frac{10V}{4π}\left(σ^3\sqrt{\frac{π}{2}}erf\left(\frac{R}{σ\sqrt{2}}\right) - Rσ^2e^{-\frac{R^2}{2σ^2}}\right)^{-1},$$ where V is the sphere volume, and b corresponds to the normalization constant so that the average SAR over the 10g sphere is equal to a given value (here 10 W/kg). The parameter σ of the Gaussian distribution must be
less than R/2 so that the analytical calculation is consistent with the
assumption of all power absorbed by the sphere. Given these analytical results,
the parameters (flip angles, pulse bandwidths, TR, TA, SAR levels) of the
sequences (e.g. GRE, XFL, and EPI) can be safely adapted depending on the
duration of the acquisition.Results
The
temperature rise over time for different power levels with a uniform SAR distribution
as well as the maximum exam time for a given SAR are reported in Fig. 1. Experiment
duration of fewer than 50 minutes allows increasing previous power limits (0.1
W for 6 min and 0.2 W for 10 s) by 1.5; below 15 minutes, the augmentation
ratio is 2. A comparison of uniform versus Gaussian SAR distributions is
represented in Figure 2. More peaked SAR distributions define stricter SAR constraints.Discussion
We show
that depending on the exam duration, the SAR thresholds based on the restricted
SAR methodology can be relaxed. The temperature results, more relevant than SAR
from a safety standpoint, strongly depend on the shape of the SAR distribution
in the sphere where all RF power is supposed to be absorbed. However, the model
makes extremely conservative assumptions, such as all RF power absorbed in the
10g area, no heat-dissipation mechanisms besides heat conduction, and no
perfusion. As a result, considering a uniform SAR distribution appears a
reasonable counterpart to relax the SAR constraints. Thus, such a drastic and
simplistic model still allows scanning human subjects safely without prior SAR
simulations using adapted restricted SAR protocols. Relaxing the constraints based
on a temperature analysis permits up to twice the resolution or coverage, or additional
fat suppression, which can be an essential feature, for example, in EPI imaging5,6.Conclusion
The
considered analytical calculations provide the local heating dynamics for given
input power, whether it is a single channel or parallel transmission. Neglect
of perfusion makes the model conservative7. We show that reducing
the acquisition time entails a lower final temperature and thus permits increased
SAR limits. Relaxation of the SAR constraints enables higher image resolution,
better coverage, shorter TR, etc., thereby expanding the restricted SAR
methodology applications and increasing opportunities for low-power research.Acknowledgements
The study has received funding from the European Union
Horizon 2020 Research and Innovation program under Grant Agreements No. 736937
(M-One) and No. 885876 (AROMA). Leducq Foundation (Large Equipement ERPT program, NEUROVASC7T
project) also provided financial support for this work.References
- International
Electrotechnical Commission Medical electrical equipment-part 2–33: particular
requirements for the basic safety and essential performance of magnetic
resonance equipment for medical diagnosis - 3rd ed. Geneva: International
Electrotechnical Commission: 2010- 2–33:601.
- Subhendra N.
S., Alterman R. L. et al. Low Power MRI Techniques for
Neurosurgical Planning and Post-surgical Assessment of Deep Brain Stimulators
in Patients with Medically Refractory Parkinson’s Disease or Dystonia. IOSR
Journal of Dental and Medical Sciences, 2016; Volume 15(10) pp. 71-75.
-
Wheaton A.
J., Borthakur A. et al. Method for Reduced SAR T1ρ-Weighted
MRI. Magnetic Resonance in Medicine, 2004; 51:1096–1102.
-
Vignaud A,
Mauconduit F. et al. Fast and unconditionally safe in vivo MR
head protocol for home-made coil prototype assessment at 7T. Journal
of Physics: Conference Series, IOP Publishing, 2018; 1092, pp.012159.
-
Burakiewicz
J, Charles-Edwards G. D., et al. Water–Fat Separation in
Diffusion-Weighted EPI Using an IDEAL Approach with Image Navigator. Magnetic
Resonance in Medicine, 2015; 73:964–972.
-
Sarlls J. E.,
Pierpaoli C., et al. Robust Fat Suppression at 3T in
High-Resolution Diffusion-Weighted Single-Shot Echo-Planar Imaging of Human
Brain. Magn Reson Med, 2011 Dec; 66(6): 1658–1665.
-
Le Ster
C., Mauconduit F., Mirkes C., Vignaud A., Boulant N. Measuring
radiofrequency field-induced temperature variations in brain MRI exams with
motion compensated MR thermometry and field monitoring. Magn Reson
Med, 2021 online.