Assessment of computational tools for MRI RF dosimetry by comparison with measurements on a laboratory phantom.

This paper presents an extended comparison between numerical simulations using the different computational tools employed nowadays in electromagnetic dosimetry and measurements of radiofrequency (RF) electromagnetic field distributions in phantoms with tissue-simulating liquids at 64 MHz, 128 MHz and 300 MHz, adopting a customized experimental setup. The aim is to quantify the overall reliability and accuracy of RF dosimetry approaches at frequencies in use in magnetic resonance imaging transmit coils. Measurements are compared against four common techniques used for electromagnetic simulations, i.e. the finite difference time domain (FDTD), the finite integration technique (FIT), the boundary element method (BEM) and the hybrid finite element method-boundary element method (FEM-BEM) approaches. It is shown that FDTD and FIT produce similar results, which generally are also in good agreement with those of FEM-BEM. On the contrary, BEM seems to perform less well than the other methods and shows numerical convergence problems in presence of metallic objects. Maximum uncertainties of about 30% (coverage factor k = 2) can be attributed to measurements regarding electric and magnetic field amplitudes. Discrepancies between simulations and experiments are found to be in the range from 10% to 30%. These values confirm other previously published results of experimental validations performed on a limited set of data and define the accuracy of our measurement setup.

Numerical study of RF exposure and the resulting temperature rise in the foetus during a magnetic resonance procedure.

Numerical simulations of specific absorption rate (SAR) and temperature changes in a 26-week pregnant woman model within typical birdcage body coils as used in 1.5 T and 3 T MRI scanners are described. Spatial distributions of SAR and the resulting spatial and temporal changes in temperature are determined using a finite difference time domain method and a finite difference bio-heat transfer solver that accounts for discrete vessels. Heat transfer from foetus to placenta via the umbilical vein and arteries as well as that across the foetal skin/amniotic fluid/uterine wall boundaries is modelled. Results suggest that for procedures compliant with IEC normal mode conditions (maternal whole-body averaged SAR(MWB) < or = 2 W kg(-1) (continuous or time-averaged over 6 min)), whole foetal SAR, local foetal SAR(10 g) and average foetal temperature are within international safety limits. For continuous RF exposure at SAR(MWB) = 2 W kg(-1) over periods of 7.5 min or longer, a maximum local foetal temperature >38 degrees C may occur. However, assessment of the risk posed by such maximum temperatures predicted in a static model is difficult because of frequent foetal movement. Results also confirm that when SAR(MWB) = 2 W kg(-1), some local SAR(10g) values in the mother's trunk and extremities exceed recommended limits.

A random phased array device for delivery of high intensity focused ultrasound.

Randomized phased arrays can offer electronic steering of a single focus and simultaneous multiple foci concomitant with low levels of secondary maxima and are potentially useful as sources of high intensity focused ultrasound (HIFU). This work describes laboratory testing of a 1 MHz random phased array consisting of 254 elements on a spherical shell of radius of curvature 130 mm and diameter 170 mm. Acoustic output power and efficiency are measured for a range of input electrical powers, and field distributions for various single- and multiple-focus conditions are evaluated by a novel technique using an infrared camera to provide rapid imaging of temperature changes on the surface of an absorbing target. Experimental results show that the array can steer a single focus laterally to at least +/-15 mm off axis and axially to more than +/-15 mm from the centre of curvature of the array and patterns of four and five simultaneous foci +/-10 mm laterally and axially whilst maintaining low intensity levels in secondary maxima away from the targeted area in good agreement with linear theoretical predictions. Experiments in which pork meat was thermally ablated indicate that contiguous lesions several cm(3) in volume can be produced using the patterns of multiple foci.

Modelling the interaction of electromagnetic fields (10 MHz-10 GHz) with the human body: methods and applications.

Numerical modelling of the interaction between electromagnetic fields (EMFs) and the dielectrically inhomogeneous human body provides a unique way of assessing the resulting spatial distributions of internal electric fields, currents and rate of energy deposition. Knowledge of these parameters is of importance in understanding such interactions and is a prerequisite when assessing EMF exposure or when assessing or optimizing therapeutic or diagnostic medical applications that employ EMFs. In this review, computational methods that provide this information through full time-dependent solutions of Maxwell's equations are summarized briefly. This is followed by an overview of safety- and medical-related applications where modelling has contributed significantly to development and understanding of the techniques involved. In particular, applications in the areas of mobile communications, magnetic resonance imaging, hyperthermal therapy and microwave radiometry are highlighted. Finally, examples of modelling the potentially new medical applications of recent technologies such as ultra-wideband microwaves are discussed.

Comparison of simulated and experimental results from helical antennas within a muscle-equivalent phantom.

Miniature microwave helical antennas for use in thermal therapy applications are usually investigated using muscle-equivalent phantoms. In this paper, an alternative method using an electromagnetic solver based on the finite integration technique is used to simulate a range of 915 MHz helical antennas within a medium with the dielectric properties of muscle. By avoiding the stair-casing effect associated with many solvers, this method enables accurate simulations of non-orthogonal geometric objects such as helical antennas to be achieved. The effects of coil-spacing and insertion depth on the SAR distribution produced by the antennas were characterized and showed good agreement with previously published results obtained using a muscle phantom and a thermographic camera. The simulations confirm that the performance of helical antennas depends on insertion depth. Modification of the coil density demonstrated improvement of the return loss characteristics and changes to the resulting SAR profile.

Prediction of specific absorption rate in mother and fetus associated with MRI examinations during pregnancy.

There is uncertainty regarding the risk posed by magnetic resonance imaging (MRI) examinations to pregnant patients. The most frequently used methods, such as single-shot fast spin echo (ssFSE), often require operation at the specific absorption rate (SAR) limits imposed by safety guidelines. With the introduction of higher-field systems, such limits will be even more significant for fetal imaging. An electromagnetic solver based on the time domain finite integration technique (FIT) was used to predict SAR in an anatomically realistic model of a pregnant patient (28 weeks' gestation) associated with the radiofrequency (RF) fields from birdcage body coils typical of 1.5 T and 3 T MRI systems (i.e., operating at approximately 64 and 127 MHz, respectively). The results suggest that 1) the highest local SAR is in the mother, with the fetus being exposed to a peak of approximately 40-60% of that value at 64 MHz, increasing to approximately 50-70% at 127 MHz; 2) compliance with U.S. Food and Drug Administration (FDA) and International Commission on Non-Ionizing Radiation Protection (ICNIRP) guidelines requires control of SAR values averaged over 1 g or 10 g of tissue, respectively; and 3) compliance with Medical Device Agency (MDA) guidelines requires control of the maximum SAR(10g) within the fetus.

Five-band microwave radiometer system for non-invasive measurement of brain temperature in new-born infants: system calibration and its feasibility.

Recent simulation studies have shown that a technique of multi-frequency microwave radiometry is feasible for non-invasive measurement of deep brain temperatures in the new-born infants. A five-band microwave radiometer system has been developed, and its operation in a normal electromagnetic environment is checked. Five receivers operating with a waveguide antenna and at center frequencies of 1.2, 1.65, 2.3, 3.0 and 3.6 GHz (0.4 GHz bandwidth) are calibrated using a temperature-controlled water-bath. Temperature resolutions obtained for each receiver are 0.183, 0.273, 0.148, 0.108 and 0.118 K, respectively. A temperature retrieval simulation based on these resolutions and the previously proposed algorithm shows that the confidence interval, as produced by thermal noise, is 0.62 K for the retrieved central brain temperature. If the conductivity of brain is estimated wrong by 10 %, this will result in an error of 0.3-0.4 K. The result of this work is encouraging for realization of radiometric measurement of temperature profile in a baby's head.

Power frequency electromagnetic fields and health. Where's the evidence?

Twenty years ago concerns were raised that exposure to power frequency (or extremely low frequency (ELF)) electromagnetic fields (EMFs) may be associated with an increased risk of cancer or other health hazards. Subsequently no associations have been shown between laboratory magnetic field exposures and carcinogenesis in either animal or cellular models. Indeed, studies have demonstrated that magnetic fields are not associated with cancer. However, the puzzle remains that the results of some epidemiological studies may be interpreted as suggesting that living close to high-voltage transmission (HVT) lines appears to increase slightly the risk of childhood leukaemia. Alternatively, these results could result from small biases and errors in individual studies, which might not necessarily be the same in each study. The nature of the epidemiological studies (power-line, wire code, magnetic field or appliance based) appears to determine whether and how the EMFs associated with HVT lines might be a risk factor. It is possible that a simple association with either magnetic or electric field exposure may not be the whole answer, and an alternative mechanism is always a possibility. Although the interpretation of the available evidence by most expert bodies has led them to conclude that exposure to power frequency electric and magnetic fields is not a human health hazard, a working group under the auspices of the US National Institute of Environmental Health Sciences (NIEHS) concluded that there was a possible low risk associated with certain exposures to ELF magnetic fields. NIEHS itself interpreted the finding as insufficient to warrant aggressive regulatory concern but stated that, because virtually everyone is routinely exposed to ELF EMFs, passive regulatory action is warranted, such as a continued emphasis on educating both the public and the regulated community on means aimed at reducing exposures. These analyses, conclusions and advice are not contradicted by subsequent studies, and therefore the conclusion of the World Health Organisation that further research is needed seems valid.

Numerical modeling of temperature distributions within the neonatal head.

Introduction of hypothermia therapy as a neuroprotection therapy after hypoxia-ischemia in newborn infants requires appraisal of cooling methods. In this numerical study thermal simulations were performed to test the hypothesis that cooling of the surface of the cranium by the application of a cooling bonnet significantly reduces deep brain temperature and produces a temperature differential between the deep brain and the body core. A realistic three-dimensional (3-D) computer model of infant head anatomy was used, derived from magnetic resonance data from a newborn infant. Temperature distributions were calculated using the Pennes heatsink model. The cooling bonnet was at a constant temperature of 10 degrees C. When modeling head cooling only, a constant body core temperature of 37 degrees C was imposed. The computed result showed no significant cooling of the deep brain regions, only the very superficial regions of the brain are cooled to temperatures of 33-34 degrees C. Poor efficacy of head cooling was still found after a considerable increase in the modeled thermal conductivities of the skin and skull, or after a decrease in perfusion. The results for the heatsink thermal model of the infant head were confirmed by comparison of results computed for a scaled down adult head, using both the heatsink description and a discrete vessel thermal model with both anatomy and vasculature obtained from MR data. The results indicate that significant reduction in brain temperature will only be achieved if the infant's core temperature is lowered.

SAR and temperature changes in the leg due to an RF decoupling coil at frequencies between 64 and 213 MHz.

Specific absorption rate (SAR) due to a butterfly surface coil in a realistic model of the leg was calculated for frequencies 64 < or = v < or = 213 MHz. The resulting temperature distribution and temperature changes (deltaT) were found using the bioheat transfer equation. To compare results at different frequencies, the minimum B-field within the coil's footprint in the plane parallel to the coil but displaced 50 mm from it was kept constant. To achieve the same minimum B-field as that associated with operation at 64 MHz that was compliant with safety guidelines (peak SAR in 1 cm3 of tissue of 2.3 W/kg), it was predicted that SAR would exceed recommended levels when v > or = 149 MHz. The corresponding maximum deltaT at 64, 128, and 213 MHz were 0.3, 1.3, and 5.0 degrees C, respectively.

A theoretical assessment of the relative performance of spherical phased arrays for ultrasound surgery.

Computer modeling of spherical-section phased arrays for ultrasound surgery (tissue ablation) is described. The influence on performance of the number of circular elements (68 to 1024), their diameter (2.5 to 10 mm), frequency (1 to 2 MHz), and degree of sparseness in the array is investigated for elements distributed randomly or in square, annular, and hexagonal patterns on a spherical shell (radius of curvature, 120 mm). Criteria for evaluating the quality of the intensity distributions obtained when focusing the arrays both on and away from their center of curvature, and in both single focus and simultaneous multiple foci modes, are proposed. Of the arrays studied, the most favorable performance, for both modes, is predicted for 256 5-mm diameter, randomly distributed elements. For the single focus mode, this performed better than regular arrays of 255 to 1024 elements and, for the case of nine simultaneous foci produced on a coplanar 3x3 grid with 4-mm spacing, better than square, hexagonal, or annular distributed arrays with a comparable number of elements. Randomization improved performance by suppressing grating lobes significantly. For single focus mode, a several-fold decrease in the number of elements could be made without degrading the quality of the intensity distribution.

Pre-clinical evaluation of a two-channel microwave hyperthermia system with adaptive phase control in a large animal.

A pre-clinical assessment of the heating capabilities of a two-channel 915 MHz Microfocus-1000 hyperthermia system, with adaptive phase control, was carried out in a series of experiments using a large animal model. The results of the experimental measurements of specific absorption rate (SAR) and tissue temperature show that when muscle tissue of the hind legs of pigs was compressed to 6.5-7 cm, then a pair of parallel opposed, coherently driven, transverse electromagnetic wave applicators could elevate the temperature in deep tissue to therapeutic levels without overheating superficial tissues when the phase difference between applicators was determined by the adaptive phase control algorithm.

Electromagnetic and thermal modeling of SAR and temperature fields in tissue due to an RF decoupling coil.

The finite difference time domain method is used to calculate the specific absorption rate (SAR) due to a butterfly surface coil in a realistic tissue model of the leg. The resulting temperature distribution and temperature changes are found using a finite difference solution to the bioheat transfer equation. Reasonable agreement is found between predicted temperature changes and those measured in vivo provided that the resulting hyperthermia does not induce noticeable changes in perfusion. The method is applicable to radiofrequency dosimetry problems associated with high Bo field magnetic resonance systems and where knowledge of spatial variation in SAR is important in assessing the safety of new magnetic resonance procedures.

Prevention of tinea corporis in collegiate wrestlers.

OBJECTIVE: To examine the role of a comprehensive skin disease prevention protocol in conjunction with the use of a barrier cream to prevent tinea corporis (ringworm) in collegiate wrestlers. DESIGN AND SETTING: We studied a college wrestling team for 16 weeks during 1 season. During the first 8 weeks, no preventive measures were taken. For the remaining 8 weeks, wrestlers were randomized into 2 groups and used either a barrier or a placebo. SUBJECTS: Twenty-two male college wrestlers with a mean age of 20.4 years (range, 18.1 to 23.2), a mean weight of 68.4 kg (range, 55.8 to 130.2), and a mean height of 177.8 cm (range, 168.7 to 186.9). MEASUREMENTS: We performed skin checks daily. All new or exacerbated lesions were clinically diagnosed by the same team physician and recorded. RESULTS: Cases of tinea corporis declined from 10 diagnosed before initiation of the protocol to 1 after the protocol was initiated. One athlete in the placebo group was found to have tinea corporis versus none in the barrier cream group. CONCLUSIONS: Strict adherence to the prevention protocol for skin infections significantly decreased the number of cases of tinea corporis. The use of the barrier cream in conjunction with the prevention protocol did not result in any further statistical reduction in the number of wrestlers who contracted tinea corporis.

Effectiveness of the Gaussian beam model in predicting SAR distributions from the lucite cone applicator.

The Gaussian beam model (GBM) has been shown to be a successful tool in the development of the current sheet applicator. As a result, the effectiveness of the GBM is investigated in single and dual array applications of the lucite cone applicator (LCA). The LCA is a modified water-filled waveguide applicator with an improved effective field size (EFS > 64 cm2, aperture 10 cm x 10 cm). The GB-source parameters were calculated from the emanating E-field of a single LCA. The SAR distribution from a single LCA was measured by E-field scanning and thermographic (TG) imaging, and compared with the GB-predicted SAR distribution. Deviations in the principal planes were found to be less than 5%. TG-measured and GB-predicted SAR distributions from three different dual LCA configurations were compared and evaluated. When water was used as intermedium between LCAs and phantom, a maximum SAR difference of 27% was calculated. In the absence of water as intermedium, this difference increased to 44%. These large deviations were only found in areas where the measured SAR distribution was disturbed due to antenna interactions. The average SAR differences with and without water as intermedium were 7% respectively 11%, indicating that the GBM can provide good qualitative information about the SAR distribution of dual LCA-arrays.

Ultrasonic hydrophone based on short in-fiber bragg gratings.

We investigate the feasibility of using in-fiber Bragg gratings for measuring acoustic fields in the megahertz range. We found that the acoustic coupling from the ultrasonic field to the grating leads to the formation of standing waves in the fiber. Because of these standing waves, the system response is complex and, as we show, the grating does not act as an effective probe. However, significant improvement in its performance can be gained by use of short gratings coupled with an appropriate desensitization of the fiber. A noise-limited pressure resolution of approximately 4.5 x 10(-3) atm/ radicalHz was found.

Design and calibration of electric field probes in the range 10-120 MHz.

In view of potential thermal hazards, there is a need to determine the specific absorption rate (SAR) distributions associated with radiofrequency coils used in magnetic resonance imaging (MRI) (typically 10-120 MHz). Electric field (E-field) distributions in tissue-equivalent phantoms may be determined using a probe comprising a dipole antenna and a detector. The geometry of the dipole dictates the sensitivity of the device, thus two designs are discussed in this paper. Both probes are compact, have a spatial resolution of 2.5 cm3, operate at MR frequencies and have a response independent of the dielectric characteristics of the phantom material. Calibration of these probes requires a system capable of producing a known E-field both in air and in a tissue-like medium at frequencies between 10 and 120 MHz. Transverse electromagnetic wave (TEM) cells answering these specifications are described and the calibration procedure outlined. Accurately calibrated E-field probes can make field measurements in phantoms which can be used to verify predictions from numerical models. These numerical techniques may then be used to predict E-fields, and hence SAR, in patients.

Experimental verification of numerically predicted electric field distributions produced by a radiofrequency coil.

There are safety issues regarding energy deposition within tissues due to radiofrequency fields used in some magnetic resonance (MR) procedures. Procedures should be compliant with guidelines that specify limits to temperature elevation and specific absorption rate (SAR). In general, direct measurement of these quantities in patients is impractical and an alternative approach is to determine SAR from the electric field (E-field) distributions predicted by numerical models. In this initial study the E-field distribution in a tissue-simulating phantom due to a square coil driven at 31 MHz is predicted using a finite-difference time domain (FDTD) solution to Maxwell's equations. An experimental arrangement of the same problem was constructed and the resulting E-field distribution was measured using a calibrated minimally perturbing E-field probe. A comparison between experimentally and theoretically derived data showed that the numerically predicted E-fields were within +/-1 dB of the fields measured with the E-field probe in the phantom material. The results provide confidence in the use of the FDTD algorithm to determine quantitatively accurate E-field distributions arising from square radiofrequency (RF) coils used in MR procedures. The accuracy of numerical models of other coil designs such as bird cages, saddles and surface coils can be investigated in the same manner. Future studies will evaluate the exposure of patients to these RF fields.

Analysis of thermal parameters obtained during phase III trials of hyperthermia as an adjunct to radiotherapy in the treatment of breast carcinoma.

An analysis of 351 HT treatment sessions administered to 101 patients receiving radiotherapy and hyperthermia (RT + HT) who were entered into Phase III concurrent randomized trials for recurrent (BrR) and intact (BrI) breast tumours is presented. A complete response (CR) was recorded in 50 of 84 (59.5%) fields in the case of recurrent breast patients and in 10 of 17 (59%) fields in the case of the intact breast patients. In comparison, 15 of 60 (25%) patients entered into BrR who received RT alone and 8 of 12 (66.7%) patients receiving RT alone entered into BrI trial achieved CR. A set of thermal parameters is defined and evaluated on a treatment by treatment basis. Patient and tumour characteristics influential on CR are identified and thermal parameters which have additional prognostic value are investigated. Multivariate logistic analysis of the non-thermal data showed that maximum depth of tumour, presence or history of disease outside the treated area and RT regimen were most influential on CR. Tumour volume (cm3) (OR = 0.996, 95% CI = 0.993-1.004, p = 0.08) was not a strong prognostic covariate; tumour area and linear dimensions were even less significant (p = 0.41). The cumulative minimum thermal isoeffect dose (equivalent minutes at 43 degrees C) accrued over the 1st, 1st and 2nd, and 1st, 2nd and 3rd treatment sessions was the only thermal parameter to exhibit an association with CR consistently, Other thermal parameters found to contribute to the predictive models were MINTIME > 42 degrees C calculated for the first treatment session and %sensors > 43 degrees C (peak) calculated for the 2nd treatment session.

Development and investigation of ultrasound linear phased arrays for transrectal treatment of prostate.

We used the rectangular radiator method as a numerical solution to the Rayleigh-Sommerfeld diffraction integral for calculating the acoustic fields produced by linear arrays. The appropriate phases and amplitudes of voltages applied to the elements were computed using the pseudo-inverse method. We have developed and acoustically evaluated several constructions of planar linear ultrasound phased arrays for transrectal thermotherapy of prostate diseases. The designs of the linear phased arrays used in this work were the result of compromise in terms of the choice of frequency and number and size of elements, in order to investigate a means of improving array performance which might lead to increased efficacy and safety of the thermal treatment. The results obtained are in agreement with the data of other studies and show that a linear plane phased array may be potentially useful for thermal therapy.