JOURNAL OF MAGNETIC RESONANCE IMAGING 19:650 ­ 656 (2004)

Technical Note

Temperature and SAR Calculations for a Human

Head Within Volume and Surface Coils at 64 and

300 MHz

Christopher M. Collins, PhD,1* Wanzhan Liu, MS,1 Jinghua Wang, PhD,1

Rolf Gruetter, PhD,2 J. Thomas Vaughan, PhD,2 Kamil Ugurbil, PhD,2 and

Michael B. Smith, PhD1

age with radio frequency (RF) energy during MRI, the

Purpose: To examine relationships between specific energy ab-

United States' Food and Drug Administration (FDA) and

sorption rate (SAR) and temperature distributions in the human

the International Electrotechnical Commission have

head during radio frequency energy deposition in MRI.

devised limits on both specific energy absorption rate

Materials and Methods: A multi-tissue numerical model of

(SAR) and temperature (T) increases in the human body

the head was developed that considered thermal conduc-

(1,2). Limits on SAR are given for the average SAR over

tivity, heat capacity, perfusion, heat of metabolism, electri-

the whole body, average in the head, and for the maxi-

cal properties, and density. Calculations of SAR and the

mum volume average over any 1 g (FDA) or 10 g (IEC)

resulting temperature increase were performed for different

region of tissue in the head, body, or extremities (max-

coils at different frequencies.

imum local SAR). It is possible to estimate the average

Results: Because of tissue-dependant perfusion rates and

SAR in the head or entire body during an experiment by

thermal conduction, there is not a good overall spatial cor-

dividing the time-average input RF power by the mass

relation between SAR and temperature increase. When a

of the appropriate body section, but determination of

volume coil is driven to induce a head average SAR level of

either 3.0 or 3.2 W/kg, it is unlikely that a significant

the local SAR levels or temperature changes in the hu-

temperature increase in the brain will occur due to its high

man subject is not so easily accomplished. Due to the

rate of perfusion, although limits on SAR in any 1 g of tissue

complicated induced current patterns in the heteroge-

in the head may be exceeded.

neous human head, calculations or measurements of

Conclusion: Attempts to ensure RF safety in MRI often rely

the maximum local SAR in simple geometries are not

on assumptions about local temperature from local SAR

likely to be indicative of the maximum local SAR levels

levels. The relationship between local SAR and local tem-

in the head (3). Furthermore, existing limits on the

perature is not, however, straightforward. In cases where

maximum local SAR in any 1 g region of tissue are likely

high SAR levels are required due to pulse sequence de-

to be exceeded before the limits on average SAR are

mands, calculations of temperature may be preferable to

exceeded (3,4), especially when using local coils, such

calculations of SAR because of the more direct relationship

as surface coils.

between temperature and safety.

When assessing the risk of the RF power in an MR

Key Words: calculations; temperature; MRI; SAR; safety

examination, the most important factor is the potential

J. Magn. Reson. Imaging 2004;19:650 ­ 656.

increase in core or local body temperature (1,2). Due to

© 2004 Wiley-Liss, Inc.

the practical limitations of assessing local temperature

increases, however, limits of power deposition are also

expressed in SAR limits (1,2). While there are well-

TO REDUCE THE RISKS of inducing thermoregulatory

studied correlations between T, time, and tissue dam-

imbalance and causing local hyperthermic tissue dam-

age or thermoregulatory distress (5­7), the precise re-

lationship  between  SAR  and  temperature  is

multifactorial. The effect of whole-body SAR on core

1

Center for NMR Research, Department of Radiology, The Pennsylvania

body temperature is a function of the ability of the body

State University College of Medicine, Hershey, Pennsylvania.

to lose heat by radiation, convection, and evaporation of

2

Center for MR Research, Department of Radiology, School of Medicine,

University of Minnesota, Minneapolis, Minnesota.

sweat and is affected by, among other things, body size

Contract grant sponsor: NIH; Contract grant number: R01 EB000454.

and type, amount of clothing worn, ventilation, humid-

Continuation of work presented at ISMRM, Honolulu, 2002. p 704.

ity, ambient temperature, and individual response to

*Address reprint requests to: C.M.C., Center for NMR Research, NMR/

heat stress. The effect of local SAR on local temperature

MRI Building, Department of Radiology H066, The Pennsylvania State

University College of Medicine, 500 University Drive, Hershey, PA

depends on, among other things, the effects of thermal

17033. E-mail: cmcollins@psu.edu

conduction, perfusion by blood, and the complex het-

Received September 5, 2003; Accepted January 29, 2004.

erogeneous geometry of the human body. If the local or

DOI 10.1002/jmri.20041

core temperature changes enough, there can be local or

Published online in Wiley InterScience (www.interscience.wiley.com).

© 2004 Wiley-Liss, Inc.

650