4472 ieee transactions on microwave theory and...

4472 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 54, NO. 12, DECEMBER 2006

Evaluations of Specific Absorption Rateand Temperature Increase Within Pregnant

Female Models in Magnetic ResonanceImaging Birdcage Coils

Dagang Wu, Student Member, IEEE, Saad Shamsi, Ji Chen, Member, IEEE, and Wolfgang Kainz, Member, IEEE

Abstract—This paper presents a detailed numerical study ofspecific absorption rate (SAR) and temperature increase cal-culations within pregnant female models exposed to magneticresonance imaging (MRI). Nine pregnant female models, rep-resenting different pregnant stages, were used for this study.SAR and temperature increase within and around fetuses atdifferent pregnancy stages were calculated for two MRI operatingmodes (normal mode and first-level controlled mode) at 64 and128 MHz. Local fetus energy deposition exceeds the InternationalElectrotechnical Commission limit of 10 W/kg in the first-levelcontrolled mode at 64 MHz. Fetus temperature exceeds or ap-proaches 38 C for both frequencies in the first-level controlledmode at later stages of pregnancy. The core temperature limitsfor both modes and both frequencies are not exceeded. The resultsshow higher maximum SAR and higher temperature at 64 MHzand during later pregnancy stages with a significant increasestarting with the fifth month of pregnancy. Based on the resultsof this study, radiologists can minimize local fetus heating, espe-cially late in pregnancy, by using normal mode sequences, whichminimize the whole body SAR in the mother.

Index Terms—Electromagnetic heating, finite-differencemethod, magnetic resonance imaging (MRI), pregnant woman,safety standards.

I. INTRODUCTION

MAGNETIC resonance imaging (MRI) exams are widelyused in the medical field to obtain topographic images

of patients’ internal structures. This procedure can produce im-ages with millimeter resolution. Physicians use these images todiagnose defects in ligaments, muscles, brain tissue, and othersoft tissues. During the imaging, patients are subjected to sev-eral kinds of electromagnetic fields produced by MRI coils [1].While these electromagnetic fields are required for the purposeof imaging, they also deposit electromagnetic energy withinpatients. Due to such energy deposition, temperature increasewithin patients is inevitable [2], [3].

Manuscript received April 20, 2006; revised July 21, 2006. This workwas supported in part by the National Science Foundation under GrantBES-0332957.

D. Wu, S. Shamsi, and J. Chen are with the Electrical and Computer Engi-neering Department, University of Houston, Houston, TX 77204 USA (e-mail:[email protected]).

W. Kainz is with the Center for Devices and Radiological Health, U.S.Food and Drug Administration, Rockville, MD 20852 USA (e-mail: [email protected]).

Color versions of Figs. 1–11 are available online at http://ieeexplore.ieee.org.Digital Object Identifier 10.1109/TMTT.2006.884655

To ensure patient safety, the International ElectrotechnicalCommission (IEC) has developed a standard to limit the max-imum energy deposition within human subjects undergoing anMRI [4]. This standard sets the limits for the specific absorp-tion rate (SAR) for whole-body SAR, partial-body SAR, andlocal SAR and limits for the increase of the body core temper-ature and the localized temperature. Since MRI procedures canbe operated at two different modes (normal mode and first-levelcontrolled mode), limits for both modes are specified in [4]. Thenormal mode is suitable for all individuals, while the first-levelcontrolled mode is more suitable for patients who can toleratea higher dosage of energy deposition. For example, the max-imum 10-g SAR and the core temperature increase are limitedto 10 W/kg and 0.5 C during normal mode operation, whereasthe limits are 10 W/kg and 1 C for first-level controlled modeoperation. These limits were developed and tested using manyadult male and female models and have been used to guide prac-tical MRI exams [5]–[9].

While these limits are proven to be accurate for malesand females, their application to pregnant women or fetusesneeds to be further investigated since the fetus is surroundedby high-conductive liquid [10]. Due to the high-conductivefluid, high doses of electromagnetic energy deposition may beexpected, which could then lead to a high temperature increasearound the fetus. To address these concerns, some research onfetus MRI safety have been conducted, but only on animals.However, these investigations are not very conclusive. Theresults can vary significantly depending on factors such asanimal species, strength of the magnetic field, and duration ofthe scan. Some studies indicated a considerable rate of infantmortality or abnormalities in animal fetuses [11], [12], whileothers established that there were no harmful effects of MRIon animal fetuses [13]–[15]. To evaluate the energy depositionwithin pregnant women undergoing an MRI, pregnant femalemodels were recently developed and numerical evaluationswere performed [16], [17]. However, only preliminary investi-gations on energy deposition around fetuses were performed,and no temperature increase near fetuses was calculated.

The purpose of this study is to perform a comprehensive nu-merical evaluation of the energy deposition and the tempera-ture increase within pregnant female models exposed to MRIelectromagnetic radiation generated by RF birdcage coils at 64and 128 MHz. The simulations were carried out using the fi-nite-difference time-domain (FDTD) method. Nine models, one

0018-9480/$20.00 © 2006 IEEE

WU et al.: EVALUATIONS OF SAR AND TEMPERATURE INCREASE WITHIN PREGNANT FEMALE MODELS IN MRI BIRDCAGE COILS 4473

for each month of pregnancy, were used in the study to deter-mine if a certain stage of pregnancy was more susceptible to ab-sorbing more electromagnetic radiation. A computer model ofa physical birdcage coil, rather than pre-extracted current distri-butions, was used in this study [6], [7], [16]–[20]. This enablesus to obtain a more realistic study on the interactions betweenthe coils and the pregnant female models.

The remainder of this paper is organized as follows. InSection II, we describe the methodologies and the models(including both the birdcage coil model and the nine pregnantfemale models) used in the study. Section III presents theresults from both electromagnetic and thermal simulations.Detailed discussions are also given in Section III. Based on theresults and discussions, a conclusion is given in Section IV.

II. METHODOLOGY AND MODELS

A. Simulation Methods

The simulation methods used in this study are describedin [17]. The FDTD scheme was applied to both Maxwell’sequations and bio-heat equation to obtain the SAR distributionsand the temperature increase within our models [22]–[25].SAR, used as a fundamental metric for RF heating, is definedas [5]–[7]

(1)

In this expression, is the conductivity (in siemens per meter),is the mass density (kg m ), and is the rms value of elec-

tric fields (volts per meter) within the biological subjects. Thestrength of the electric field at any location within the biologicaltissue can be directly obtained from the results of electromag-netic FDTD simulations [22], [23].

To evaluate the temperature increase within biological tis-sues, the finite-difference technique is applied to the bio-heatequation [22]. With electric fields obtained from the previouselectromagnetic simulations, temperature distribution ( ) in-side biological tissues can then be evaluated via

(2)

where is the specific heat capacity (J kg C), is the thermalconductivity (W m C), is the blood perfusion coefficient(W m C), is the metabolic heat production rate (W m ),and is the blood temperature. Including the blood perfusionin the modeling generally will reduce the temperature varia-tion due to energy deposition. This is a more accurate modelfor our applications. At the external surfaces between the bio-logical tissue and the surrounding air, the following convectiveboundary condition is applied:

(3)

where is the convection coefficient (W m C), is theambient temperature, and is the normal direction to the sur-face. Using the convection coefficient as a boundary conditionhas limited accuracy. Detailed discussion on using convection,radiation, and evaporation are given in [25].

Fig. 1. Nine pregnant female models used in this investigation from [17]. Theupper row shows the models month-1 to month-4 using the same body model.The lower row shows the models month-5 to month-9 using the scaled pregnantfemale model. For all nine models, the uterus, placenta, and fetus were scaledaccording to the different sizes for the different gestational age.

The simulation procedure starts with using the electromag-netic FDTD method to evaluate the electric and magnetic fieldswithin the pregnant female models due to the RF field of thebirdcage coil. Once the electric field distributions are obtained,we can predict the temperature increase using (2) and (3). Itshould be noted that a basal temperature distribution within thebiological tissue should be evaluated first. This is evaluated bysetting the value of the electric field equal to zero in (2). In thefollowing studies, the temperature of the blood is set at 37C, and the ambient temperature is assumed to be 24 C. A

convection coefficient of 10.5 W m C is used in all sim-ulations [25].

B. Pregnant Female Models

The nine pregnant female models used in this study wereco-developed by the University of Houston, Houston, TX,and the Food Drug Administration (FDA), Rockville, MD[30]. Data that describe the shape of the body surface of apregnant woman in the 34th gestational week (body 1) wereobtained from FarField Technology Limited, Christchurch,New Zealand. A second computer-aided design (CAD) modelof a nonpregnant female body surface (body 2), was obtainedfrom 21st Century Solutions Ltd., Gibraltar, U.K. Fetus,bladder, uterus, placenta, and bones are based on MRI data ofa woman in the 35th gestational week. The MRI data has aresolution of 5 mm and was provided by Stanford University,Stanford, CA. We converted the data sets to stereolithography(STL) format and combined them into one CAD model. Forthe models month-1 to month-4 we used body 2, and for themodels month-5 to month-9 we used body 1. Assuming thebelly of a pregnant woman begins noticeable growing betweenthe third and fourth month, we scaled the belly from body 1 forthe pregnant female models month-5 to month-7. Body 2, usedfor the models month-1 to month-4, and body 1, used for themodels month-8 and month-9, remained unchanged. Uterus,placenta, and fetus were scaled according to the different sizesfor the different gestational age. These nine models are shownin Fig. 1. The dielectric properties of each tissue, obtainedfrom the materials database at 64 and 128 MHz are tabulated


TABLE IDIELECTRIC PROPERTIES OF TISSUES

TABLE IITHERMAL PROPERTIES OF TISSUES

in Table I [31]. Blood perfusion rates are 277 mL/kg/min forthe uterus and 1760 mL/kg/min for the placenta [26]–[28].Based on an approximate proportional relation between bloodperfusion and and [25], we calculated the values and

for the uterus and the placenta accordingly (see Table II).The thermal properties for the fetus are a weighted average ofmuscle and fat (55% muscle, 45% fat). The thermal propertiesfor the body are also a weighted average of muscle and fat (70%muscle, 30% fat). We found that varying the placenta perfusionrate from 1760 mL/kg/min to 200 mL/kg/min does not givea significant difference (less than 0.2 C) in the temperaturevariation for the fetus region.

C. Birdcage Coil Model

The MRI coil used in this study is the commonly used bird-cage coil. The coil was designed using the birdcage builder soft-ware package developed by Pennsylvania State University, Her-shey, PA [33]. Like most birdcage coils, this coil is composed oftwo end-rings connected by 16 equally spaced rungs, as shownin Fig. 2. The rungs and rings of the birdcage coil are constructedusing a perfect electric conductor (PEC). Each rung is split, anda capacitor is placed inside the gap. The parasitic capacitancebetween gaps must also be considered in the design since it alsoaffects the coil resonant frequency. To reduce RF interference,the coil is surrounded by an RF shield, which is also modeled asa PEC. This shield affects the coil tuning because eddy currentsinduced in the shield affect the resonant frequency [34]. Whenthe female models are added to the structure, the coil needs to bere-tuned in order to achieve the exact MRI resonant frequency.The detailed dimensions for the coil are shown in Fig. 2.

For our coil design, we used the quadrature excitation. Inthe quadrature excitation, two voltage sources are used to ex-

Fig. 2. Constructed birdcage coil used in this study from [17]. (a) Side view.(b) Top view. (c) Three-dimensional view. (d) Coil with shielding.

Fig. 3. Positioning of the pregnant female models within the MRI RF coil from[17].

cite the 16-leg birdcage coil. These two sources need to be onone end-ring, 90 spatially separated, with 90 out-of-phase inthe excitation waveform. Theoretically, having such an arrange-ment would quadruple the effective power of the coil. The re-sulting magnetic field is highly homogenous, leading to a largeroperating region [1].

The next step is to tune the capacitors to achieve either 64-or 128-MHz coil resonance. To determine the correct capac-itor values, we used Gaussian pulse excitation and recordedtime-domain signals at various locations within the coil. The ca-pacitor value was altered until the highest response is detectedat 64 or 128 MHz. After this tuning, we can obtain the capac-itor value for an empty coil without any model placed within it(unloaded coil).

D. Simulation Setup and Final Tuning

Fig. 3 shows the positioning of the model in the coil. Eachmodel had to be placed within the coil consistently, while con-sidering the changes in the size of the models. To simulate theeffect of a woman supine on a bed while being scanned, the backof the model is always placed 22.7 cm away from the inner edge


Fig. 4. Normalized maximum 10-g averaged SAR (64-MHz normal mode) asa function of pregnant stage. From [17].

of the coil [17]. The navel of the model is always placed on the– -plane, as shown in Fig. 3. This positioning ensured that the

models would be in the same location for each simulation.Once the coil is loaded, the resonant frequency can shift due

to the coupling between the coil and models. The tuning pro-cedure needs to be repeated in order to bring the resonant fre-quency back to 64 or 128 MHz. This procedure was repeatedfor each of the nine pregnant female models because each modelhad a different effect on the coil loading. The detuning of the coildue to the pregnant woman for both frequencies was 1%–3%.Several iterations were necessary before the correct capacitorvalue was found.

After setting up the simulation for each month of the preg-nant model and final capacitor tuning, the simulations were exe-cuted in series. Computation time for each month typically takesaround one day, depending on the exact mesh used in simula-tions. After the simulations were complete, a post processingprocedure was used to extract the distributions of temperatureincrease and energy deposition.

III. RESULTS AND DISCUSSIONS

Here, the calculated maximum 10-g cubical averaged SARand the maximum temperature increase within various tissuesare given. These values are calculated for 64 and 128 MHz.All values (SARs and temperature) are normalized to the IEC60601-2-33 whole-body averaged SAR limit [4], which are 2and 4 W/kg for normal mode and first-level controlled mode,respectively. The local 10-g averaged SAR is 10 W/kg for bothmodes. The core body temperature increase limits are 0.5 C forthe normal mode and 1 C for the first-level controlled mode.The maximum local temperature limits are defined as 38 C forthe head and 39 C for the torso.

A. SAR and Local Temperature Increase for the 64-MHzBirdcage Coil

For the 64-MHz normal mode, the 10-g averaged SARs foreach tissue at different stages of pregnancy shown in Fig. 4 arebelow the limit of 10 W/kg for all tissues, except the body.

Fig. 5. Maximum temperature within different tissues as a function of pregnantstage (64-MHz normal mode).

Fig. 6. Normalized maximum 10-g averaged SAR (64-MHz first-level con-trolled mode) as a function of pregnant stage. From [17].

The maximum temperatures within the body and the amnioticfluid (amn. fluid)) reaches 38 C starting at the fifth month ofpregnancy, as shown in Fig. 5. The maximum fetus heating, in-cluding the fetus brain, does not exceed 38 C. The figures onlyshow the maximum SAR and the local temperature increase.Therefore, slight fluctuations on both were observed for the dif-ferent months. Especially for the critical tissues, fetus, placenta,uterus, and amniotic fluid, the maximum SAR and the temper-ature increase are, in general, higher in later pregnancy stages.

For the first level controlled mode, the maximum 10-g av-eraged SAR and the maximum temperature for each tissue atdifferent pregnant stages are shown in Figs. 6 and 7. Due toincreased RF power, the maximum energy depositions withinthe fetus exceed the 10-W/kg limit beyond the fourth monthof pregnancy. The maximum temperature for the body is abovethe limit of 39 C starting with the sixth month. The maximumfetus and amniotic fluid temperature exceeds 38 C after the fifthmonth.


Fig. 7. Maximum temperature within different tissues as a function of pregnantstage (64-MHz first-level controlled mode).

Fig. 8. Normalized maximum 10-g averaged SAR as a function of pregnantstage (128-MHz normal mode).

B. SAR and Local Temperature Increase for the 128-MHzBirdcage Coil

For normal mode operation at 128 MHz, the maximum 10-gaveraged SAR and the maximum temperature increase in eachtissue are plotted in Figs. 8 and 9. We found that SAR for thebody tissue exceeds the limit for all months by approximately60%, whereas for all other tissues, the 10-g averaged SAR iswell below the limit. The maximum temperatures for all tissuesare below 38 C. However, the same trend as for 64 MHz canbe observed: later stages of pregnancy are more susceptible tofetus heating.

The maximum 10-g averaged SAR for each tissue under thefirst-level controlled mode is shown in Fig. 10. Again, due to theincreased RF power, the energy deposition within body, bone,uterus, and amniotic fluid exceed the limit at some stages ofpregnancy. However, the energy depositions within the fetus arebelow the limit for all stages of pregnancy, but reach the limitin the ninth month. The maximum temperature increase of eachtissue is shown in Fig. 11. Although a higher temperature in-crease within the fetus is observed for later stages of pregnancy,

Fig. 9. Maximum temperature within different tissues as a function of pregnantstage (128-MHz normal mode).

Fig. 10. Normalized maximum 10-g averaged SAR as a function of pregnantstage (128-MHz first-level controlled mode).

Fig. 11. Maximum temperature within different tissues as a function of preg-nant stage (128-MHz first-level controlled mode).

the maximum temperature within the fetus are all below the limitof 38 C. Similar to the results for normal mode, the maximum


TABLE IIICORE TEMPERATURE INCREASE WITHIN FETUS (UNIT IN DEGREES CELSIUS)

TABLE IVCORE TEMPERATURE INCREASE WITHIN THE ENTIRE

PREGNANT FEMALE MODEL (UNIT IN DEGREES CELSIUS)

temperatures within body and liquid exceed the maximum tem-perature within the fetus.

From these results, we find that, for 64 MHz, the maximum10-g averaged SARs within the fetus are below the IEC limitunder normal mode operation, whereas these values exceed thelimit during the first-level controlled-mode scans. As expected,the maximum temperatures at the first-level controlled modewithin all tissues are also larger than those at normal mode op-eration. For 128 MHz, the maximum 10-g averaged SAR andtemperature increase inside the fetus models are below the limitsfor both operating modes. A different SAR increase pattern wasobserved for 64 and 128 MHz. This is mainly caused by the factthe maximum value of SAR is related to both the MRI operatingfrequency, as well as the conductivity of the pregnant models.

The core temperature increase for fetus is defined as the av-erage temperature within the entire fetus region, while the coretemperature increase for the entire pregnant model is definedas the average temperature increase within the entire femalemodel. Tables III and IV gives the fetus and pregnant womancore temperature increase for both operating modes and for bothfrequencies. As indicated in the tables, the core temperature in-crease within the fetus actually decreases as the pregnant stagesincrease. This appears to be contradictory to the previous re-sults. However, further analysis shows that due to the increaseof the fetus, the average temperature increase can indeed de-crease. Overall, the average temperature within the entire preg-nant female model is directly related to the total energy deposi-tion with the body. As expected, the core temperature increase

for the first level controlled mode is almost twice as that of thenormal mode, but in both modes and for both frequencies, theIEC limits for core temperature are not exceeded. We see thatthe distributions of energy deposition and temperature increaseare quite different at the two frequencies. Our simulation resultsseem to suggest higher maximum SAR and higher temperatureat 64 MHz. This can be explained by the reduced penetrationdepth at 128 MHz.

IV. CONCLUSIONS

We have found for both frequencies an increase in SAR andtemperature at later stages of pregnancy. The SAR within thebody exceeds the IEC limit for both frequencies and both op-erating modes with a maximum of 3.8 times higher than therecommended limit of 10 W/kg. The 10-g averaged SAR forthe fetus is only exceeded for the first-level controlled mode at64 MHz. The results suggest that the relation of the local SARto the whole body averaged SAR needs to be further evaluatedusing anatomically correct models for all types of human beingsundergoing an MRI. The results also indicate that, for pregnantwomen, instead of the whole body averaged SAR, the local 10-gaveraged SAR might be the limiting factor. Whole body het-erogeneous models of pregnant women for different stages ofpregnancy would be needed to answer this question. Only forthe first-level controlled mode at 64 MHz does the local fetustemperature exceeds 38 C, while it approaches the limit of38 C also for the first-level controlled mode at 128 MHz forlater stages of pregnancy.

Based on the results of this study, radiologists can minimizelocal fetus heating, especially late in pregnancy, by using normalmode sequences, which minimize the whole-body SAR in themother.

REFERENCES

[1] J. Jin, Electromagnetic Analysis and Design in Magnetic ResonanceImaging. New York: CRC, 1999.

[2] D. J. Schaefer, “Safety aspects of radio-frequency power deposition inmagnetic resonance,” Magn. Reson. Imaging Clin. N. Amer., vol. 6, pp.775–789, 1998.

[3] F. G. Shellock, “Radiofrequency-induced heating during MR proce-dures: A review,” J. Magn. Reson. Imaging, vol. 12, pp. 30–36, 2000.

[4] Medical Equipment–Part 2: Particular Requirements for the Safety ofMagnetic Resonance Equipment for Medical Diagnosis, IEC Standard60601-2-33, 2004.

[5] O. P. Gandhi and X. B. Chen, “Specific absorption rates and inducedcurrent densities for an anatomy-based model of the human for expo-sure to time varying magnetic fields of MRI,” Magn. Res. Med., vol.41, no. 5, pp. 816–823, 1999.

[6] C. M. Collins, W. Liu, J. Wang, R. Gruetter, J. T. Vaughan, K. Ugur-bill, and M. B. Smith, “Temperature and specific absorption rate cal-culations for a human head within volume and surface coils at 64 and300 MHz,” J. Magn. Res., vol. 19, pp. 650–656, 2004.

[7] U. D. Nguyen, S. Brown, I. A. Chang, J. Krycia, and M. S. Mirotznik,“Numerical evaluation of heating of the human head due to magneticresonance imaging,” IEEE Trans. Biomed. Eng., vol. 51, no. 8, pp.1301–1309, Aug. 2004.

[8] L. A. Zaremba, “FDA guidelines for magnetic resonance equip-ment safety,” Center for Devices and Radiol. Health, Food,Drug Admin., Rockville, MD, YEAR [Online]. Available:http://www.aapm.org/meetings/02AM/pdf/8356-8054.pdf

[9] J. Behrens, “Module #5: MRI safety,” Siemens Med. Syst., Greensboro,NC, 2002.

[10] J. P. De Wilde, A. W. Rivers, and D. L. Price, “A review of the cur-rent use of magnetic resonance imaging in pregnancy and safety im-plications for the fetus,” Prog. Biophys. Molecular Biol., vol. 87, pp.335–353, 2005.


[11] K. I. Carnes and R. L. Magin, “Effects of in utero exposure to 4.7 TMR imaging conditions on fetal growth and testicular development inthe mouse,” Magn. Res. Imag., vol. 14, pp. 263–274, 2003.

[12] Y. P. Yip, C. Capriotti, S. G. Norbash, S. L. Talagala, and J. W. Yip,“Effects of MR exposure on axonal outgrowth in the sympathetic ner-vous system of the chick,” Magn. Res. Imag., vol. 4, pp. 742–748, 1994.

[13] K. P. Behr, H. W. Tiffe, K. H. Hinz, H. Luders, M. Friederichs, M.Ryll, and H. Hundeshagen, “Nuclear magnetic resonance (NMR) andthe development of chicken embryos,” Tsche. Tierarztl. Wochenschr.,vol. 98, pp. 149–152, 1991.

[14] Y. P. Yip, C. Capriotti, S. L. Talagala, and J. W. Yip, “Effects of MRexposure at 1.5 T on early embryonic development of the chick,” Magn.Res. Imag., vol. 5, pp. 457–462, 1994.

[15] Y. P. Yip, C. Capriotti, and J. W. Yip, “Effects of MR exposure on cell-proliferation and migration of chick motonuerons,” Magn. Res. Imag.,vol. 5, no. 4, pp. 457–462, 1995.

[16] M. Strydom, K. Caputa, M. A. Stuchly, and P. Gowland, “Numericalmodeling interaction of RF field in MRI with a pregnant female model,”in IEEE/ACES Wireless Commun. Appl. Comput. Electromagn. Conf.,Honolulu, HI, 2005, pp. 389–392.

[17] S. Shamsi, D. Wu, J. Chen, R. Liu, and W. Kainz, “Specific absorptionrate evaluation in pregnant woman models radiated from MRI birdcagecoil,” presented at the IEEE MTT-S Int. Microw. Symp., San Francisco,CA, Jun. 2006.

[18] C. M. Collins, S. Li, and M. B. Smith, “Specific absorption rate andB1 field distributions in a heterogeneous human head model within abirdcage coil,” Magn. Res. Med., vol. 40, pp. 847–856, 1998.

[19] J. Jin and J. Chen, “On the specific absorption rate and field inhomo-geneity of birdcage coils loaded with the human head,” Magn. Res.Med., vol. 38, pp. 953–963, 1997.

[20] T. S. Ibramhim, A. M. Abduljalil, B. A. Baertlein, R. Lee, and P. M.Robitaille, “Analysis of B1 field profiles and specific absorption ratevalues for multi-strut transverse electromagnetic RF coils in high fieldMRI applications,” Phys. Med. Biol., vol. 46, pp. 2545–2555, 2001.

[21] T. S. Ibrahim, “A numerical analysis of radiofrequency power require-ments in magnetic resonance imaging experiments,” IEEE Trans. Mi-crow. Theory Tech., vol. 52, no. 8, pp. 1999–2003, Aug. 2004.

[22] A. Taflove, Computational Electrodynamics: The Finite-DifferenceTime-Domain Method. Norwood, MA: Artech House, 1995.

[23] K. S. Kunz and R. J. Luebbers, The Finite Difference Time DomainMethod for Electromagnetic.s. Boca Raton, FL: CRC, 1993.

[24] H. H. Pennes, “Analysis of tissue and arterial blood temperatures inresting forearm,” J. Appl. Physiol., vol. 1, pp. 93–122, 1948.

[25] P. Bernardi, M. Cavagnaro, S. Pisa, and E. Piuzzi, “Specific absorptionrate and temperature increases in the head of a cellular-phone user,”IEEE Trans. Microw. Theory Tech., vol. 48, no. 7, pp. 1118–1125, Jul.2000.

[26] W. E. Huckabee, J. Metcalfe, H. Prystowsky, and D. H. Barron,“Blood flow and oxygen consumption of the pregnant uterus,” Amer.J. Physiol., vol. 200, pp. 274–278, 1961.

[27] P. A. Gowland, S. T. Francis, K. R. Duncan, A. J. Freeman, B. Issa, R. J.Moore, Bowtell, P. N. Baker, I. R. Johnson, and B. S. Worthington, “Invivo perfusion measurements in the human placenta using echo planarimaging at 0.5 T,” Magn. Reson. Med., vol. 40, no. 3, pp. 467–473,1998.

[28] S. T. Francis, K. R. Duncan, R. J. Moore, P. N. Baker, I. R. Johnson,and P. A. Gowland, “Non-invasive mapping of placental perfusion,”Lancet, vol. 351, pp. 1397–1399, 1998.

[29] A. Ibrahiem, “Analysis of the temperature increase linked to the powerincluded by RF source,” Prog. Electromagn. Res., vol. PIER 52, pp.23–46, 2005.

[30] W. Kainz, T. Kellom, R. Qiang, and J. Chen, “Development of preg-nant woman models for nine gestational ages and calculation of fetusheating during magnetic resonance imaging (MRI),” in BEMS Conf.,Dublin, Ireland, Jun. 2005, pp. 137–139.

[31] C. Gabriel, “Compilation of the dielectric properties of body tissues atRF and microwave frequencies,” Brooks AFB, Brooks AFB, TX, Tech.Rep. AL/OE-TR-1996-0037, 1996.

[32] F. A. Duck, Physical Properties of Tissues: A Comprehensive Refer-ence Book. New York: Academic, 1990.

[33] Birdcage Builder. Pennsylvania State Univ., Philadelphia, PA, 2006[Online]. Available: http://psunmr.hmc.psu.edu/birdcage/index.htm

[34] W. Renhart, O. Biro, P. Wach, and R. Strollberger, “Investigation ofthe resonance behavior of an MR-birdcage applying a 3-D FEM code,”IEEE Trans. Magn., vol. 37, no. 9, pp. 3688–3692, Sep. 2001.

Dagang Wu (S’99) was born in Nanjing, China, onNovember 10, 1978. He received the B.S. and M.S.degrees in electrical engineering from Southeast Uni-versity, Nanjing, China, in 1999 and 2002, respec-tively.

Since September 2002, he has been a ResearchAssistant with the Department of Electrical andComputer Engineering, University of Houston,Houston, TX. His current research interestsinclude computational electromagnetics, bioelec-tromagnetics and numerical logging-while-drilling

(LWD)/measurement-while-drilling (MWD) modeling.

Saad Shamsi, photograph and biography not available at time of publication.

Ji Chen (S’87–M’90) received the Bachelor’s degreefrom the Huazhong University of Science and Tech-nology, Wuhan, Hubei, China, in 1989, the Master’sdegree from McMaster University, Hamilton, ON,Canada, in 1994, and the Ph.D. degree from theUniversity of Illinois at Urbana-Champaign in 1998,all in electrical engineering.

He is currently an Assistant Professor with theDepartment of Electrical and Computer Engineering,University of Houston, Houston, TX. Prior to joiningthe University of Houston, he was a Staff Engineer

with Motorola Personal Communication Research Laboratories, Chicago, IL,from 1998 to 2001.

Dr. Chen was the recipient of the 2000 Motorola Engineering Award.

Wolfgang Kainz (M’02) received the M.S. andPh.D. degrees in electrical engineering from theVienna University of Technology, Vienna, Austria,in 1997 and 2000, respectively.

Following his affiliation with the Austrian Re-search Centers Seibersdorf (ARCS), where he wasinvolved with electromagnetic compatibility ofelectronic implants and exposure setups for bio-ex-periments, he joined The Foundation for Research onInformation Technologies in Society (IT’IS), Zürich,Switzerland, as Associate Director. He co-managed

the IT’IS together with Prof. N. Kuster and was involved with the developmentof in vivo and in vitro exposure setups for bioexperiments. In 2002, he movedto the U.S., where he is currently with the U.S. Food and Drug Administration,Center for Devices and Radiological Health, Rockville, MD. His researchinterest is focused on the safety and effectiveness of medical devices and safetyin electromagnetic fields. This includes computational electrodynamics (FDTDand finite-element method (FEM) simulations) for safety and effectivenessevaluations, MRI safety, performance and safety of wireless technology used inmedical devices, electromagnetic compatibility of medical devices, especiallyelectronic implants, and dosimetric exposure assessments.

Dr. Kainz is chairman of the IEEE Standard Coordination Committee 34,Sub-Committee 2, which develops compliance techniques for wireless devices.

4472 ieee transactions on microwave theory and...

Documents