Registration of structurally dissimilar images in MRI-based brachytherapy.

A serious challenge in image registration is the accurate alignment of two images in which a certain structure is present in only one of the two. Such topological changes are problematic for conventional non-rigid registration algorithms. We propose to incorporate in a conventional free-form registration framework a geometrical penalty term that minimizes the volume of the missing structure in one image. We demonstrate our method on cervical MR images for brachytherapy. The intrapatient registration problem involves one image in which a therapy applicator is present and one in which it is not. By including the penalty term, a substantial improvement in the surface distance to the gold standard anatomical position and the residual volume of the applicator void are obtained. Registration of neighboring structures, i.e. the rectum and the bladder is generally improved as well, albeit to a lesser degree.

On verification of hyperthermia treatment planning for cervical carcinoma patients.

PURPOSE: The aim of this study was to verify hyperthermia treatment planning calculations by means of measurements performed during hyperthermia treatments. The calculated specific absorption rate (SAR(calc)) was compared with clinically measured SAR values, during 11 treatments in seven cervical carcinoma patients. METHODS: Hyperthermia treatments were performed using the 70 MHz AMC-4 waveguide system. Temperatures were measured using multisensor thermocouple probes. One invasive thermometry catheter in the cervical tumour and two non-invasive catheters in the vagina were used. For optimal tissue contact and fixation of the catheters, a gynaecological tampon was inserted, moisturized with distilled water (4 treatments), or saline (6 treatments) for better thermal contact. During one treatment no tampon was used. At the start of treatment the temperature rise (DeltaT(meas)) after a short power pulse was measured, which is proportional to SAR(meas). The SAR(calc) along the catheter tracks was extracted from the calculated SAR distribution and compared with the DeltaT(meas)-profiles. RESULTS: The correlation between DeltaT(meas) and SAR(calc) was on average R = 0.56 +/- 0.28, but appeared highly dependent on the wetness of the tampon (preferably with saline) and the tissue contact of the catheters. Correlations were strong (R approximately 0.85-0.93) when thermal contact was good, but much weaker (R approximately 0.14-0.48) for cases with poor thermal contact. CONCLUSION: Good correlations between measurements and calculations were found when tissue contact of the catheters was good. The main difficulties for accurate verification were of clinical nature, arising from improper use of the gynaecological tampon. Poor thermal contact between thermocouples and tissue caused measurement artefacts that were difficult to correlate with calculations.

Radiotherapy and hyperthermia in the treatment of patients with locally advanced prostate cancer: preliminary results.

OBJECTIVE: To report an interim clinical evaluation of combined external beam irradiation (EBRT) and interstitial or regional hyperthermia in the treatment of locally advanced prostate cancer. PATIENTS AND METHODS: From 1997 to 2001, 26 patients with T3-4/NX/0M0 prostate carcinoma were treated with a combination of conformal EBRT and hyperthermia. Fourteen patients received five weekly regional hyperthermia treatments within an optimization (phase II) study, using the coaxial transverse electrical magnetic system. Twelve patients received one interstitial hyperthermia treatment within a feasibility study (phase I), using the multi-electrode current source system. Irradiation was delivered using a conformal three-field technique, administering 70 Gy in 2-Gy fractions in 7 weeks. RESULTS: The mean initial prostate-specific antigen level was 26 ng/mL. Three patients had a T4 and 23 a T3 tumour; the tumours were classified as well (four), moderately (16) and poorly (six) differentiated. The mean follow-up was 36 months. In the combined treatments there was no toxicity of more than grade 2. In regional hyperthermia the mean index temperature (T90 and T50, i.e. exceeded by 90% and 50% of the measurements) was 40.2 degrees C and 40.8 degrees C, and for interstitial hyperthermia 39.4 degrees C and 41.8 degrees C, respectively. All patients survived; seven patients had a biochemical relapse (27%), three in the regional and four in the interstitial group. The actuarial probability of freedom from biochemical relapse was 70% at 36 months for all patients together, 79% for regional and 57% for interstitial. No factors were found that could be used to predict relapse. CONCLUSIONS: The clinical outcome in these patients with advanced localized prostate cancer seems to compare favourably with most series using irradiation alone, and the treatment caused no severe complications.

Comparison of intra-luminal versus intra-tumoural temperature measurements in patients with locally advanced prostate cancer treated with the coaxial TEM system: report of a feasibility study.

A study was performed on regional hyperthermia for patients with locally advanced prostate carcinoma. The primary objective was to analyse the thermometry data with an emphasis on the possibility of replacing invasive thermometry by tumour-related intra-luminal thermometry. Fourteen patients were treated with a combination of conformal external beam radiotherapy (70 Gy) and hyperthermia. Hyperthermia was delivered using the Coaxial TEM system, one treatment per week, to a total of five treatments. Thermometry was performed in bladder, urethra, rectum and esophagus. Invasive thermometry in the prostate was carried out during one or two treatments for each patient by placing transperineally a central and a peripheral catheter. Heterogeneous temperature distributions were measured in the prostate. The mean average invasive temperature range was 1.1 degrees C. Due to the temperature heterogeneity and a limited number of thermometry sensors (mean 7, range 2-13), large variability between treatments and patients existed regarding achieved temperatures and dose. The mean invasive T90 was 40.2 +/- 0.6 degrees C and T50 was 40.8 +/- 0.6 degrees C. The mean Cum min T90>40.5 degrees C per treatment was 22 (range 0-50). Importantly, intra-luminal temperatures did not reliably predict invasively measured temperatures. Invasive thermometry, therefore, remains compulsory to calculate a thermal dose for an individual patient. Changes in temperature during treatment, measured by the urethral sensors, corresponded well with changes in temperature measured by the individual invasive sensors. Similar comparison of rectal temperature changes with intra-prostatic temperature changes was not as predictive. The similarity in temperature changes between the urethral and interstial sites, suggests that urethral temperatures are sufficient for treatment optimization. The SAR profile did not correspond with the temperature profile indicating heterogeneous perfusion. Although regional hyperthermia in combination with external beam radiotherapy for locally advanced prostate carcinoma is clinically feasible, the question on the importance of invasive thermometry remains.

Treatment planning for capacitive regional hyperthermia.

Capacitively coupled hyperthermia devices are widely in use, mainly in Asian countries. In this paper, a comprehensive treatment planning system, including a Specific Absorption Rate (SAR) and thermal model for capacitively coupled hyperthermia, is described and demonstrated using a heterogeneous patient model. In order to accurately model a hyperthermia treatment, simulation at high resolution is mandatory. Using the quasi-static approximation, the electromagnetic problem can be solved at high resolution with acceptable computational effort. The validity of the quasi-static approximation is demonstrated by comparing the Maxwell solution of a phantom problem to the quasi-static approximation. Modelling of capacitive hyperthermia of the prostate reveals the difficulty of heating deep-seated tumours in the pelvic area. Comparison of the SAR distribution in the heterogeneous patient model and a patient shaped agar phantom shows a shielding effect of the pelvic bone and the influence of the fat-muscle distribution. It is shown that evaluation of capacitive hyperthermia with agar phantoms leads to overly optimistic conclusions. Therapeutic relevant tumour temperatures can only be obtained by permitting temperature extrema in normal tissue. This concurs with clinical practice, where treatment-limiting hot spots restrict the tumour temperature. It is demonstrated that the use of very cold overlay bolus bags has only a very superficial effect. The presented model can be used for individual treatment planning and optimization, for the evaluation of capacitive applicator modifications and comparison with other devices.

Increasing the systemic temperature during regional hyperthermia: effect of a cooling strategy on tumour temperatures and side-effects.

In the application of regional hyperthermia, optimization of the temperature distribution remains necessary. One of the tools that might be used is a modest increase in the systemic temperature to diminish cooling by blood perfusion. This study investigates (1) if it is feasible to increase the systemic temperature by applying other cooling strategies, without inducing unacceptable systemic stress, and (2) whether a rise in systemic temperature results in improvement of tumour temperatures. Eleven patients with locally advanced cervical carcinoma and 12 patients with locally advanced prostate carcinoma were treated with our Coaxial TEM regional hyperthermia system. In this system, the temperature of the open water bolus can be easily adjusted. Two cooling methods were applied alternately, one with a relatively low water temperature (method A), the other with a higher water bolus temperature in combination with extensive head/chest cooling by a hand shower (method B). Method B resulted in significantly higher systemic temperatures, for both patient groups separately (0.8, respectively, 0.5 degrees C) and for the total patient group (0.7 degrees C). Additionally, all tumour index temperatures were higher. For the combined group (for T50: 0.4 degrees C) and for the cervix group (for T50: 0.7 degrees C), it reached statistical significance. The raise in core temperature led to a significantly higher increase in heart rate. For the group of cervix patients, higher systemic temperatures resulted in more treatment-limiting systemic stress. For the prostate patients, systemic stress was not an important issue. Since the raise in systemic temperature did not influence the overall tolerance of treatment, method B could be applied to this group. However, the increases in tumour temperatures were small, and potential hazards of systemic temperature increase should be considered.

Improvement of absorbing structures used in regional hyperthermia.

Local pain is a major limiting factor in regional hyperthermia treatment with radiative applicators. Absorbing structures, consisting of agar bound saline water, have been used successfully to reduce peripheral hot spots. However, both clinical experience and simulation results indicate a SAR elevation in the tissue under the edges of the absorber block. This paper investigates the effect of modification of shape, position and spatial composition of the absorber blocks on the central attenuating effect and the SAR elevating effect at the edges. A selection from a set of five options is made based on simulations with a phantom and a single ring dipole applicator. The simulations have been performed with the FDTD core of the regional hyperthermia treatment planning system. It is shown that tapering of the absorber edge and introduction of a water layer between the absorber and the skin can reduce the edge effect in the superficial fat layer by approximately 50% with respect to a rectangular absorber. A further reduction of 15% can be obtained by an absorber with an appropriate gradient of its conductivity in the direction of the dominant E-field. The modified absorbers produce a central attenuating effect comparable to the rectangular type. The use of a water layer type and a sigma gradient type absorber is also analysed in a patient anatomy, both in the dipole ring applicator, operating at 70 MHz, as well as in a three ring Cavity Slot (CS) applicator, operating at 150 MHz. The mutual influence of phase-amplitude steering and the application of absorbers is investigated in the CS applicator. It appears that absorbers have a significant influence on the interference pattern in the patient model, possibly causing substantial reduction of the SAR value in the tumour and limiting the possibility of ad hoc application of absorbers. Re-optimization can only partly cancel this effect. Local SAR reduction by phase-amplitude control alone can match or improve the effect obtained with modified absorbers.

CT-resolution regional hyperthermia treatment planning.

Recent development of quasistatic zooming has enabled the computation of mm-resolution SAR distributions within reasonable computation times. These high-resolution SAR distributions proved to be completely dissimilar from the cm-resolution distributions. To study the impact of high-resolution SAR modelling on regional hyperthermia treatment planning (HTP), a conventional, low-resolution treatment plan is compared to an high-resolution plan. This comparison shows that the high-resolution plan yields totally different SAR and temperature distributions when compared to the conventional plan. Both SAR and temperature maxima predicted by the low-resolution plan are not predicted by the high-resolution plan and, even worse, the low-resolution plan fails to predict maxima that are predicted by the high-resolution plan. Furthermore, it appears that small-scale SAR maxima can result in temperature maxima which may cause treatment-limiting hot spots. These small-scale SAR maxima appear to be highly determined by the dielectric geometry of the patient. This demonstrates the need for an accurate, high-resolution description of this dielectric geometry. Moreover, it suggests that it may be very difficult to reduce potential treatment-limiting hot spots in clinical practice. This study demonstrates the need for high-resolution regional hyperthermia treatment planning.

Regional hyperthermia applicator design using FDTD modelling.

Recently published results confirm the positive effect of regional hyperthermia combined with external radiotherapy on pelvic tumours. Several studies have been published on the improvement of RF annular array applicator systems with dipoles and a closed water bolus. This study investigates the performance of a next-generation applicator system for regional hyperthermia with a multi-ring annular array of antennas and an open water bolus. A cavity slot antenna is introduced to enhance the directivity and reduce mutual coupling between the antennas. Several design parameters, i.e. dimensions, number of antennas and operating frequency, have been evaluated using several patient models. Performance indices have been defined to evaluate the effect of parameter variation on the specific absorption rate (SAR) distribution. The performance of the new applicator type is compared with the Coaxial TEM. Operating frequency appears to be the main parameter with a positive influence on the performance. A SAR increase in tumour of 1.7 relative to the Coaxial TEM system can be obtained with a three-ring, six-antenna per ring cavity slot applicator operating at 150 MHz.

Quasistatic zooming of FDTD E-field computations: the impact of down-scaling techniques.

Due to current computer limitations, regional hyperthermia treatment planning (HTP) is practically limited to a resolution of 1 cm, whereas a millimetre resolution is desired. Using the centimetre resolution E-field distribution, computed with, for example, the finite-difference time-domain (FDTD) method and the millimetre resolution patient anatomy it is possible to obtain a millimetre resolution SAR distribution in a volume of interest (VOI) by means of quasistatic zooming. To compute the required low-resolution E-field distribution, a low-resolution dielectric geometry is needed which is constructed by down-scaling the millimetre resolution dielectric geometry. In this study we have investigated which down-scaling technique results in a dielectric geometry that yields the best low-resolution E-field distribution as input for quasistatic zooming. A segmented 2 mm resolution CT data set of a patient has been down-scaled to 1 cm resolution using three different techniques: 'winner-takes-all', 'volumetric averaging' and 'anisotropic volumetric averaging'. The E-field distributions computed for those low-resolution dielectric geometries have been used as input for quasistatic zooming. The resulting zoomed-resolution SAR distributions were compared with a reference: the 2 mm resolution SAR distribution computed with the FDTD method. The E-field distribution for both a simple phantom and the complex partial patient geometry down-scaled using 'anisotropic volumetric averaging' resulted in zoomed-resolution SAR distributions that best approximate the corresponding high-resolution SAR distribution (correlation 97, 96% and absolute averaged difference 6, 14% respectively).

The use of absorbing structures during regional hyperthermia treatment.

Local pain is the main factor that limits regional hyperthermia treatment. Using the SAR model of the regional hyperthermia treatment planning system, the capability of absorbing blocks to reduce peripheral hot spots was investigated. The effect of rectangular absorbers of various size and salinity on an elliptical phantom in the Coaxial TEM was evaluated. The computed results were compared with SAR values measured in the phantom. Absorbers of 9 x 9 x 4 cm3 and a salinity of 18 gram l(-1) provide a SAR reduction in the muscle equivalent material, centrally under the absorber of at least 50% at a depth of up to 3 cm. The effect on the central (i.e. tumour) region is less than 20%. Larger absorbers have a more global effect and cause more attenuation in the central region. The attenuating effect depends strongly on the thickness of the fat layer between muscle and absorber. More than 2 cm fat limits the effective use of absorbers. Absorbers can induce a significant increase of SAR in muscle and fat near their edges. This effect also depends on absorber size and salinity and the thickness of the fat layer. The effect of an absorber was also evaluated with a patient anatomy, yielding results in agreement with the phantom experiments.

Development of a regional hyperthermia treatment planning system.

A flexible and fast regional hyperthermia treatment planning system for the Coaxial TEM System has been devised and is presented. Using Hounsfield Unit based thresholding and manually outlining of the tumour, a 40 cm CT data set (slice thickness 5 mm) is segmented and down scaled to a resolution of 1 cm, requiring only 30 min. The SAR model is based on the finite-difference time-domain (FDTD) method. The number of time steps to achieve numerical stability has been determined and was found to be 7000. Various optimizations of the SAR model have been applied, resulting in a relatively short computation time of 3.7 h (memory requirements 121 MB) on a Pentium III, 450 MHz standard personal computer, running GNU/Linux. The model has been validated using absolute value(Ez) measurements in a standard phantom inserted in the Coaxial TEM Applicator under different conditions and a good agreement was found. Hyperthermia treatment planning in combination with the homemade visualization tools have provided much insight in the regional hyperthermia treatment with the Coaxial TEM Applicator.

Quasistatic zooming for regional hyperthermia treatment planning.

Due to current computer limitations, specific absorption rate (SAR) distributions in regional hyperthermia treatment planning (HTP) are limited to centimetre resolution. However, since patient anatomy is highly structured on a millimetre scale, millimetre-resolution SAR modelling is required. A method called quasistatic zooming has been developed to obtain a high-resolution SAR distribution within a volume of interest (VOI): using the low-resolution E-field distribution and the high-resolution patient anatomy, the high-resolution SAR distribution is computed within a small zoom volume Q (small compared with the wavelength in water (lambda(w))). Repeating this procedure yields the zoomed-resolution SAR distribution in an arbitrary VOI. To validate this method for a VOI that is not small compared with lambda(w), high-resolution finite-difference time-domain (FDTD) modelling is needed. Since this is impractical for a clinical applicator, a computer model of a small applicator has been created. A partial patient anatomy is inserted into the applicator and both high- and low-resolution SAR distributions are computed for this geometry. For the same geometry, zoomed-resolution SAR distributions are computed with different sizes of Q. To compare the low- and zoomed-resolution SAR distributions with the high-resolution one, the correlation and averaged absolute difference are computed. These numbers are improved considerably using zooming (correlation 58% to 92%; averaged absolute difference 43% to 20%). These results appear to be independent of the size of Q, up to 0.3 lambda(w). Quasistatic zooming is a valuable tool in high-resolution regional HTP.

The significance of accurate dielectric tissue data for hyperthermia treatment planning.

For hyperthermia treatment planning, dielectric properties of several tissue types are required. Since it is difficult to perform patient specific dielectric imaging, default values based on literature data are used. However, these show a large spread (approximately 50%). Consequently, it is important to know what limit this spread imposes on the accuracy of the SAR and subsequently on the temperature distributions. Hyperthermia treatment plans performed with different values for the dielectric properties were compared. This showed that a spread of 50% resulted in the average absolute difference of approximately 20% in both SAR and temperature distributions (heat sink approach) for regional hyperthermia. For interstitial hyperthermia, a spread of 25% resulted in the averaged absolute difference of approximately 10% in the SAR distributions and 5% in the temperature distributions (heat sink approach). Considering other problems that hamper hyperthermia treatment planning, it can be concluded that default values for the dielectric properties suffice.

High-resolution SAR modelling for regional hyperthermia: testing quasistatic zooming at 10 MHz.

Present-day regional hyperthermia treatment planning systems are limited to centimetre resolution. To obtain CT-resolution SAR distributions, a method called quasistatic zooming has been developed: using the centimetre-resolution E-field distribution and the CT-resolution tomogram, the CT-resolution SAR distribution is obtained. For a low frequency of 10 MHz this method has been validated sucessfully using CT-resolution SAR computations. It appears that these CT-resolution SAR distributions are completely different from centimetre-resolution SAR distributions, indicating the necessity for high-resolution SAR modelling. Using the presented zooming technique, reliable CT-resolution SAR modelling is now possible with relatively short computation times. So far, the zooming method has only been validated for low frequencies, but clinically relevant frequencies appear to be possible.

Determination and validation of the actual 3D temperature distribution during interstitial hyperthermia of prostate carcinoma.

To determine the thermal dose of a hyperthermia treatment, knowledge of the three-dimensional (3D) temperature distribution is mandatory. The aim of this paper is to validate an interstitial hyperthermia treatment planning system with which the full 3D temperature distribution can be obtained in individual patients. Within a phase I study, 12 patients with prostate cancer were treated with interstitial hyperthermia using our multi electrode current source interstitial hyperthermia treatment (MECS IHT) system. The temperature distribution was measured from within the heating devices and by additional thermometry. The perfusion level was estimated and the heating implant reconstructed. The steady-state temperature distribution was calculated using our interstitial hyperthermia treatment planning system. The simulated temperature distribution was validated by individually comparing the measured and simulated thermo-sensors, both for the thermometry integrated with the heating applicators and the additional thermometry. The entire procedure was also performed on a no-flow agar-agar phantom. It was shown that the calculated temperature distribution of an individual patient during MECS interstitial hyperthermia is very heterogeneous. The validation indicates that the calculated temperature elevations match the measurements within approximately 1 degrees C. Possible improvements are more precise reconstruction, incorporation of discrete vasculature and using a temperature-dependent, heterogeneous perfusion distribution. Further technical improvements of the MECS-IHT system may also result in better temperature calculations.

Regional hyperthermia of pelvic tumours using the Utrecht 'Coaxial TEM' system: a feasibility study.

Between August 1989 and July 1992 a total of 22 patients (64 treatments) with inoperable or recurrent deep seated pelvic tumours were treated with regional hyperthermia and radiotherapy. The 70 Mhz Coaxial TEM applicator with its characteristic open waterbolus was used as heating device. The main objective of this pilot study was to evaluate the feasibility, toxicity and temperature data. The results showed that the major treatment limiting factors were insufficient power and systemic stress. Local pain was observed in only 10% of all treatments. Most of the treatments resulted in elevated systemic temperatures with the overall mean maximum oesophagus temperature reaching 38.9 +/- 0.7 degrees C, however, in only 6% of these treatments this was found to be treatment limiting. From the measured data the following intratumoral temperatures were calculated: T90 = 39.9 +/- 1.0 degrees C; T50 = 40.7 +/- 1.0 degrees C; T10 = 41.4 +/- 1.0 degrees C. In addition, the overall mean average normal tissue temperatures were determined: Trectum = 40.8 +/- 0.7 degrees C; Tvagina = 41.3 +/- 0.9 degrees C; Turethra = 40.8 +/- 0.9 degrees C. The temperatures in normal tissue were frequently higher than in tumour, indicating that a large volume was heated. The open waterbolus allows strong cooling, but the strategy was changed during the study: higher systemic temperatures were allowed to improve the pelvic temperatures. This pilot study proved that the open waterbolus is clinically a success, because it offers patient comfort and SAR-steering by patient repositioning, and that regional hyperthermia with the Coaxial TEM is feasible.

Quality assurance in various radiative hyperthermia systems applying a phantom with LED matrix.

The Amsterdam phantom with LED-matrix is applied as an instrument in testing the performance of four types of radiative deep-body hyperthermia systems, which are in clinical use in Germany and The Netherlands. The devices tested were Essen's BSD-1000, Berlin's BSD-2000, Utrecht's Coaxial TEM applicator and Amsterdam's Four-waveguide-array. Photographs were taken of the matrix of dipoles loaded with light-emitting diodes (LED) to visualize the distribution of the RF power deposition or specific absorption rate (SAR) in the aperture midplane. The utility of the phantom with LED matrix for various types of radiative hyperthermia systems is demonstrated. Within this preliminary study, the influence of important parameters on the SAR-pattern in the aperture midplane was demonstrated. After corrections on the phase relation of the applicators a central focus in the SAR distribution could be realized in all systems and could also be moved in any direction. The patterns of the central focus changed in its absolute values and its proportions depending on the relative relations of phase and amplitude of the lateral applicators with respect to the top and bottom applicator. Frequency dependency was recognized for the central focus of the BSD-1000 as well as for the irradiation pattern of a single applicator for the BSD-2000. In the Coaxial TEM applicator it was demonstrated that the dimension of the open water bolus influenced the absolute value of SAR in the aperture midplane.

Temperature and SAR measurements in deep-body hyperthermia with thermocouple thermometry.

Multisensor (7-14) thermocouple thermometry is used at our department for temperature measurement with our 'Coaxial TEM' regional hyperthermia system. A special design of the thermometry system with high resolution (0.005 degrees C) and fast data-acquisition (all channels within 320 ms) together with a pulsed power technique allows assessment of specific absorption rate (SAR) information in patients along catheter tracks. A disadvantage of thermocouple thermometry, EM interference, is almost entirely eliminated by application of absorbing ferrite beads around the probe leads. We investigated the effect of remaining disturbance on the temperature decay after power-off, both experimentally in phantoms and in the clinic, and with numerical simulations. Probe and tissue characteristics influence the response time tau dist of the decay of the disturbance. In our clinical practice a normal pulse sequence is 50 s power-on, 10 s power-off: a response time longer than the power-off time results in a deflection of the temperature course at the start. Based on analysis of temperature decays correction of temperature is possible. A double-pulse technique is introduced to provide an initial correction of temperature, and fast information about accuracy. Sometimes disturbance with a relatively long response time occurs, probably due to a bad contact between probe, catheter and/or tissue. Thermocouple thermometry proved to be suitable to measure the SAR along a catheter track. This is used to optimize the SAR distribution by patient positioning before treatment. A clinical example illustrates this.