Effectiveness of compensating filters in the presence of tissue inhomogeneities.

CT based 3D treatment planning systems (3DTPS) can be used to design compensating filters that, in addition to missing tissue compensation, can account for tissue inhomogeneities. The use of computer-driven systems provides a practical, convenient, and accurate method of fabricating compensating filters. In this work, we have evaluated a commercially available PAR Scientific DIGIMILL milling machine linked with FOCUS 3DTPS. Compensating filters were fabricated using refined gypsum material with no additives. Thus, filters were of manageable dimensions and were not sensitive to common machining errors. Compensating filters were evaluated using a homogeneous step phantom and step phantoms containing various internal inhomogeneities (air, cork, and bone). The accuracy of two planning algorithms used to design filters was experimentally evaluated. The superposition algorithm was found to produce better agreement with measurements than the Clarkson algorithm. Phantom measurements have demonstrated that compensating filters were able to produce a uniform dose distribution along the compensation plane in the presence of tissue inhomogeneity. However, the dose variation was greatly amplified in planes located beyond the inhomogeneity when a single compensated beam was used. The use of parallel-opposed compensated beams eliminated this problem. Both lateral and depth-dose uniformity was achieved throughout the target volume.

Matching tomographic IMRT fields with static photon fields.

The matching of abutting radiation fields presents a challenging problem in radiation therapy. Due to sharp penumbra of linear accelerator beams, small (1-2 mm) errors in field positioning can lead to large (>30%) hot or cold spots in the abutment region. With head and neck immobilization devices (thermoplastic mask/aquaplast) an average setup error of 3 mm has been reported. Therefore hot or cold spots approaching 50% of the prescription dose may occur along the matchline. Although abutting radiation fields have been investigated for static fields, there is no reported study regarding matching of tomographic IMRT and static fields. Compared to static fields, the matching of tomographic IMRT fields with static fields is more complicated. Since IMRT and static fields are planned on separate treatment planning computers, the dose in the abutment region is not specified. In addition, commonly used techniques for matching fields, such as feathering of junctions, are not practical. We have developed a method that substantially reduces dose inhomogeneity in the abutment region. In this method, a "buffer zone" around the matchline was created and was included as part of the target for both IMRT and static field plans. In both fields, a small dose gradient (< or =3%/mm) in the buffer zone was created. In the IMRT plan, the buffer zone was divided into three sections with dose varying from 83% to 25% of prescription dose. The static field dose profile was modified using either a specially designed physical (hard) or a dynamic (soft) wedge. When these modified fields were matched, the combined dose in the abutment region varied by < or =10% in the presence of setup errors spanning 4 mm (+/-2 mm) when the hard wedge was used and 10 mm (+/-5 mm) with the soft wedge.

Elimination of field size dependence of enhanced dynamic wedge factors.

Enhanced dynamic wedge factors (EDWF) are characterized by a strong field size dependence. In contrast to physical wedge factors, the EDWF decrease as the field size is increased: for 6 MV 60 degrees wedge, the EDWF decreases by 50% when the field size is increased from 4 x 4 cm2 to 20 x 20 cm2. A method that eliminates the field size dependence of EDWF was developed and investigated in this work. In this method, the wedged field shape is determined by a multileaf collimator. The initial position of the moving Y jaw is determined by the field size and the stationary Y jaw is kept fixed at 10 cm for field sizes < or = 20 cm in the wedged direction. For all other fields, the stationary Y jaw setting is determined by the field size. The modified method results in EDWF that are independent of field size, with no change in the wedge dose distribution when compared with the conventional use of EDW.

A modified technique for RF-LCF interstitial hyperthermia.

To provide uniform heating of a tumour, it is necessary to establish sufficient volumetric control of power deposition. The interstitial Radio-Frequency Localized Current Field (RF-LCF) technique may provide such control when segmented electrodes are used. The length of segments is equal to 1-1.5 cm. Each segment is connected to a separate power source. However, this technique requires an additional implant for interstitial radiotherapy, because the lumen of segmented electrodes is filled with wires necessary to connect each segment to a separate power source. In this work, a modified method of implant that allows delivery of sequential and concomitant controlled thermoradiotherapy was investigated. In this method, each segmented electrode is surrounded by four continuous electrodes. Continuous electrodes pass through vertices of 1.5 x 1.5 cm square and a segmented electrode passes through the centre of the square. The distance between segmented and continuous electrodes is 1.06 cm. The electric field induced between an electrically interacting segment and continuous electrodes is concentrated primarily between this segment and its projection on continuous electrodes. Therefore, control of temperature distribution achieved with a modified implant is similar to that achieved with an implant containing only segmented electrodes. For temperature control during treatment, plastic catheters are inserted at a 0.5 cm distance from each segmented electrode. Temperature is monitored using multisensor temperature probes. The continuous electrodes are also used for placement of radioactive sources. The lateral distance between radioactive sources is equal to 1.5 cm. Besides allowing a sequential and concomitant thermoradiotherapy, the modified method is simpler to implement because it uses several fold less amount of segmented electrodes and power sources.

Improvement of tomographic intensity modulated radiotherapy dose distributions using periodic shifting of arc abutment regions.

Based on the study of treatment arc positioning versus target length, a method that allowed periodic shift of arc abutment regions through the course of intensity modulated radiotherapy (IMRT) was developed. In this method, two treatment plans were developed for the same tumor. The first plan contained the original target (Planning Target Volume as defined by radiation oncologist) and the second one contained a modified target. The modification of the original target consisted of simply increasing its length, adding a small extension to it, or creating a distant pseudo target. These modifications cause arc abutment regions in the second plan to be shifted relative to their positions in the first plan. Different methods of target modification were investigated because in some cases (for instance, when a critical structure might overlap with the target extension) a simple extension of the target would cause an unacceptable irradiation of the sensitive structures. The dose prescribed to the modified portion of the target varied from 10% to 100% of the original target dose. It was found that a clinically significant shift (> or =5 mm) in abutment region locations occurred when the dose prescribed to the extended portion of the target was > or =95% of the original target dose. On the other hand, the pseudo target required only approximately 10% to 20% of the original target dose to produce the same shift in arc positions. Results of the film dosimetry showed that when a single plan was used for the treatment delivery, the dose nonuniformity was 17% and 25% of the prescribed dose with 0.5 and 1 mm errors in couch indexing, respectively. The dose nonuniformity was reduced by at least half when two plans were used for IMRT delivery.

Mix-mode technique of feeding arrays of dipole microwave antennae.

A 4-element array of coherently driven dipole microwave antennae produces a hot spot in the central region of the implanted volume and cold spots in the peripheral regions of the implant. Conversely, an incoherently driven array of antennae predominantly heats peripheral (along the antennae) regions of the implant. These two modes of feeding the antennae are complementary in a sense that the cold zones obtained with the coherently driven antennae coincide with the hot zones obtained with the incoherently driven antennae and vice versa. The SAR distributions resulting from mixing these modes of feeding (mixed-mode technique) were studied theoretically and experimentally. A theoretical model that allowed calculation of SAR distributions of a 4-element array of microwave antennae fed coherently, incoherently or using a mixed-mode technique was developed in this work. The goal of the theoretical study was to determine the proper mix of the coherent and incoherent modes of feeding the antennae such that the adequately (enclosed within a 50% isoSAR surface) heated volume was maximized. In experimental studies, the antennae were driven in a cyclical manner with a duty cycle equal to the weight of the coherent mode in the mix. The duty cycle was defined as the ratio of the time the antennae were driven coherently to the total duration of the cycle. To facilitate the periodical change from the coherent to incoherent feeding, a special electromechanical switch was developed. This switch allowed a wide range of variation of the duty cycle and cycle period. Theoretical and experimental studies have demonstrated that, if the relative weight of the coherent feeding in the mix-mode technique was 30% (duty cycle = 0.3), the adequately heated volume was significantly larger and the SAR distribution was more uniform than those obtained with either the coherent or incoherent mode of feeding.

A modified method of planning and delivery for dynamic multileaf collimator intensity-modulated radiation therapy.

PURPOSE: To develop a modified planning and delivery technique that reduces dose nonuniformity for tomographic delivery of intensity-modulated radiation therapy (IMRT). METHODS AND MATERIALS: The NOMOS-CORVUS system delivers IMRT in a tomographic paradigm. This type of delivery is prone to create multiple dose nonuniformity regions at the arc abutment regions. The modified technique was based on the cyclical behavior of arc positions as a function of a target length. With the modified technique, two plans are developed for the same patient, one with the original target and the second with a slightly increased target length and the abutment regions shifted by approximately 5 mm compared to the first plan. Each plan is designed to deliver half of the target prescription dose delivered on alternate days, resulting in periodic shifts of abutment regions. This method was experimentally tested in phantoms with and without intentionally introduced errors in couch indexing. RESULTS: With the modified technique, the degree of dose nonuniformity was reduced. For example, with 1 mm error in couch indexing, the degree of dose nonuniformity changed from approximately 25% to approximately 12%. CONCLUSION: Use of the modified technique reduces dose nonuniformity due to periodic shifts of abutment regions during treatment delivery.

Dosimetry of very-small (5-10 mm) and small (12.5-40 mm) diameter cones and dose verification for radiosurgery with 6-MV X-ray beams.

The dosimetry and dose verification for 6-MV X-rays were performed for radiosurgery cones of 5- to 40-mm diameter. The total scatter factors decrease slowly from 0.936 (40-mm cone) to 0.893 (10-mm cone; a variation of 5%), but they fall to 0.83 (7.5-mm cone) and 0.67 (5-mm cone). The dmax increases from about 12.9 (5-mm cone) to 16.3 mm (40-mm cone). The full width half maximum (FWHMs) of the beam profiles, measured at 5 cm depth, agree with the cone diameters within 1 mm. The 10-90% beam penumbra/FWHM ratio is 0.23 +/- 0.03 (> or = 20-mm cones); for the smaller-diameter cones this ratio increases reaching 0.84 (5-mm cone). New tissue maximum ratios (TMRs) are reported for the 5-, 7.5-, 32.5-, and 37.5-mm-diameter cones. TMRs for the other diameter cones are consistent with published data. The measured doses in two verification studies using the 12 cones with diameters > or = 12.5 mm with a single 360 degrees arc agreed to 2% with the planned doses, and to about 10% for the three smaller cones. In a simulated treatment neglecting tissue heterogeneties (skull bone), the measured doses for two five arc studies (22.5-mm cone) were within 4% of the calculated dose to isocenter.

Output factors for irregularly shaped electron fields.

It is necessary to know the output factors (dose per monitor unit at depth of maximum) for irregularly shaped electron beam fields to accurately deliver the prescribed dose to the target. Measuring the output factors for individually shaped electron beam fields for each patient is inconvenient. Using the measured output factors for two square fields, one can obtain the output factor for an irregular shaped electron portal with area intermediate between the areas of the two square fields, by obtaining the equivalent square area (as with photons) of the irregularly shaped field, and then interpolating between the output factors of the two square field areas to obtain the output factor for the irregularly shaped field. This empirical method offers a simple, practical solution. The accuracy of the method is about 1% to 2%, depending on the shape and size of the irregularly shaped electron field.

Theoretical limits of SAR distributions of a four-element square array of dipole-type antennas.

Four-element dipole microwave antenna arrays with square insertion patterns are commonly used clinically for interstitial hyperthermia. One major disadvantage with this type of antenna array is the presence of a large dead length at the tips because the current gradually decreases from maximum at the junctions to zero at the tips. This dead length is usually 1.5-2 cm along the central axis of a 2 x 2 cm array of regular dipole antennas. Many attempts to improve the performance of dipole antenna arrays have been made by designing antennas with increased current at the tips. While some dipole antennas of new design show negligible dead lengths at close proximity in the single-antenna configuration, phantom experiments have demonstrated that these antennas exhibit at least 1.1 cm dead space along the central axis of the four-antenna array. Therefore, there seems to be a limit to which the array dead length can be reduced by the improvements in the dipole-type antenna design. The goal of this work is to find the theoretical minimum of the array dead length. This was done by assuming a uniform current distribution along the entire antenna. The specific absorption rate (SAR) patterns were calculated for an array with an insertion depth of 7 cm (resonant length for 915 MHz) and a variable spacing between antennas (1-3 cm). It was found that there is a dead length of 6 mm along the central axis of the 2 x 2-cm array with the uniform current distribution, which can be considered as the theoretical limit of the dead length for this array.(ABSTRACT TRUNCATED AT 250 WORDS)

Phantom evaluation of heating of superficial tissues located above interstitial microwave antennas.

PURPOSE: A technique that improves heating of superficial tissues above an implant of microwave interstitial antennas is presented. METHODS AND MATERIALS: Adequate heating of tumor margins is achieved by extending an implant of microwave antennas beyond the tumor boundary by 1-2 cm. When the tumor infiltrates the superficial tissues including the skin, the implant cannot even reach the superficial margin of the tumor since it requires tissue to support the catheters. This may yield cold spots in the tissues above the implant. Measurements in a phantom with varying thickness of the superficial layer above the implant demonstrated inadequate Specific Absorption Rates of energy distribution in this layer. A method that improves these distributions in the superficial layers was developed and tested in this work. This method requires placing a deionized water bolus on the phantom (patient) surface. Additional microwave antennas are placed on top of the bolus above and parallel to the implanted antennas. The Specific Absorption Rates distributions were evaluated for the thicknesses of superficial layer ranging from 1.5 mm to 16 mm and two bolus thicknesses (5 and 10 mm). RESULTS: The adequate Specific Absorption Rates distributions were achieved for all tested thicknesses of the superficial layer (1.5, 4, 8, 12, and 16 mm). The use of the 5 mm bolus versus 10 mm bolus is discussed. The use of additional antennas did not significantly increase stray radiation. CONCLUSION: This method has the potential to optimize heating of superficial tissues located above a microwave antenna implant.

Evaluation of microwave interstitial antennas in the phantom with varying cross-section.

Dipole-regular microwave interstitial antennas are characterized with a "dead" space located along the tip segment of the antenna. The length of the "dead" space is on the order of 2 cm or larger, depending on the antenna's insertion depth. If the insertion depth is smaller than 4 cm, then coupling of the antennas to tissue becomes a problem. Catheters that facilitate the placement of antennas into tumor frequently protrude beyond the tissue. This provides the opportunity of exposing part of the antenna tips (with low radiation output) beyond the tissue. Decoupling of this part of the antennas from the tissue reduced the dead space and improved microwave power transfer to the tissue. This concept was investigated using a muscle equivalent phantom consisting of five segments with thicknesses varying from 3 cm to 8 cm. The transfer of microwave power to the phantom and SAR distributions along the central axis of a rectangular array of four antennas were evaluated by measuring rates of temperature rise. The protrusion lengths that improved the array performance were found for each segment of the phantom.

Equilibrium temperature distributions in uniform phantoms for superficial microwave applicators: implications for temperature-based standards of applicator adequacy.

Equilibrium temperature distributions are computed using measured SAR distributions for five different superficial microwave (915 MHz) applicators. We assume a model with uniform conduction and blood flow. A Green's function approach is used to calculate equilibrium solutions which identically obey boundary conditions at the surface of the phantom and at infinite depth. The equilibrium solutions are categorized by surface temperature (TS), maximum allowed temperature (TM), and by a parameter (referred to as the diffusion length, lambda) which characterizes the contributions of thermal conduction relative to blood flow. The computed equilibrium temperature distribution at depths of 2 and 3 cm is strongly dependent on lambda and on TM. It is not strongly dependent on surface temperature for TS below 35 degrees C. In previous work we compared the SAR distribution with local control of 53 superficial tumours with over 1 year of follow-up. As an alternative to an SAR-based description of applicator adequacy we consider a temperature-based standard. Tumours are categorized by the minimum value of lambda that would allow full coverage of the tumour volume by the 42 degrees C contour, assuming a TM of 47.5 degrees C and a TS of 35 degrees C. Eighteen of 27 lesions (67%) were locally controlled for lambda less than 1 cm. The local control in 26 lesions with lambda greater than or equal to 1 cm was 31% (p = 0.016). The lesions with the best results were those with both good coverage in theory (lambda less than 1 cm) and with all monitored catheter tracks achieving at least one session with 30 min at or above 43 degrees C. We found that the temperature-based standard of applicator adequacy was not independent of an SAR-based standard, and in this cohort of patients either a minimum SAR criterion or a maximum diffusion length criterion would serve equally well as a screen for inappropriate applicators.

Combined hyperthermia and irradiation in the treatment of superficial tumors: results of a prospective randomized trial of hyperthermia fractionation (1/wk vs 2/wk).

From December 1984 to December 1989, 240 superficially located recurrent/metastatic malignant lesions (173 patients) were enrolled in a prospective randomized study of one versus two hyperthermia fractions per week. In the majority of patients, the dose of radiation therapy was less than 4000 cGy over 4 to 5 weeks. Stratification was by tumor size, site, and histology. The goal of the hyperthermia sessions were 42.5 degrees C for 45-60 min minimum intra-tumor measured temperature. Hyperthermia was given after radiation within 30-60 min. External applicators, both microwave (over 90% of treatments) and ultrasound, were used. Overall, complete response rate in 222 evaluable lesions was 56.3% (125/222) with a minimum follow-up of 6 months and a maximum follow-up of 52 months. The complete response rate for once a week versus twice a week hyperthermia group was 54.7% and 57.8%, respectively. The severe complication rate was 18% (41/222). There was no difference between the two treatment arms. Cox regression analyses were performed to study the prognostic significance of patient characteristics, tumor characteristics, and treatment parameters. Detailed analysis and results are presented.

Phantom studies and preliminary clinical experience with the BSD 2000.

The BSD 2000 system is an array of microwave antennas operating in the 60-120 MHz range. It is a four-quadrant regional hyperthermia device with phase control permitting the operator to choose the point of constructive interference. A computer preplanning system is provided. We have compared the computer predictions with measured relative specific absorption rate (SAR) distributions in muscle equivalent phantoms and pig cadaver. The predicted SAR distribution is in qualitative agreement with observation, although differences of approximately 3-5 cm in the location of iso- SAR contours were noted. The longitudinal dependence of relative of distance cephalad to the central plane of the power distribution was measured, and it was found that a length of 15-20 cm was covered by the 50% iso-SAR. SAR measurements in the spinal canal of a pig cadaver showed that, even at resonance frequency, the measured SAR in the cervical spine was small (0.07 of central axis SAR). However, the spinal canal SAR as measured in the central plane of an evenly balanced configuration was 0.6 of the peak SAR. Marked reduction of the spinal canal SAR could be obtained with steering techniques and was in good agreement with the computer preplan. This paper reviews preliminary clinical experience with 17 patients. All but 2 patients were treated with steering techniques to permit partial sparing of normal tissues. The goal of maintaining central tumour temperature at or above 42 degrees C for at least 30 min was maintained in 41 of 67 sessions (61%) and in at least one session for all but 2 patients. Seventy-seven per cent of monitored tumor points attained at least 42 degrees C in patients for whom thermal mapping was performed. Significant cardiovascular stress developed during hyperthermia in two patients. There have been five complete responses and seven partial responses in the 17 patients.

Evaluation of the Sigma 60 applicator for regional hyperthermia in terms of scattering parameters.

Scattering parameters adequately describe the interference between ports of a multiportal electromagnetic device when the device dimensions are comparable with the wavelength of the electromagnetic waves within the device. Since the Sigma 60 applicator is a four-port electromagnetic device, the interference between ports (quadrants) is described by a 4 x 4 scattering matrix. The load and frequency dependence of the scattering parameters were studied. The exact values of the parameters depends on the load within the applicator, but typically have minima at frequencies around 80 MHz and sometimes at 100-110 MHz. The effects of the coupling between quadrants can be substantial. Marked changes in the heating pattern can occur, particularly if the phase of the coupling element and the phase between quadrants both approach 90 degrees. Examples are shown in which the effects of coupling can qualitatively alter the intended SAR pattern. Simple steps which can be taken to minimize this phenomenon are demonstrated. Recommendations for clinical practice are discussed. Scattering parameters obtained with a non-absorbing phantom can be used for the quality assurance evaluation of the device.

Dual-antenna applicator for hyperthermia of tumours at intermediate depth.

A dual-antenna applicator with 21 x 26 cm2 aperture, that is fully loaded and operates at 74 MHz, was developed at the Mallinckrodt Institute of Radiology. By placing two antennas into an applicator capable of propagating TE10 mode, a significant enlargement of heating pattern was achieved without an increase in applicator dimensions. When antennas are placed symmetrically about a parallel to the antenna axis of symmetry, the sensitivity of the applicator input impedance to variations of load impedance reduces. Stable coupling of the RF power to the treatment area may be provided. Twenty patients with eccentrically located tumours were treated using this device.

Theory and design of "shortened" multiantenna microwave applicators with controllable SAR patterns.

A "shortened" multiantenna hyperthermia applicator has been designed and tested at the Mallinckrodt Institute of Radiology at Washington University School of Medicine. By shortening the distance from antenna to aperture, an applicator is obtained that produces an SAR pattern that is essentially the same as produced by a monopole antenna. By placing several properly spaced probe antennas into the same "shortened" applicator, an applicator is obtained that produces a SAR distribution that is essentially a composite of small overlapping SAR patterns produced by weakly interacting incoherently driven antennas. Such a design significantly improves the applicator's lateral heating efficiency and allows the independent control of temperatures in certain tumor areas by changing the input power to the respective antennas.

Characteristics of improved microwave interstitial antennas for local hyperthermia.

The heating potentials of two newly-developed microwave interstitial antennas are reported in this paper. The longitudinal (parallel to the antenna) and transverse (over a plane perpendicular to the antenna) specific absorption rate (SAR) distributions of single and an array of four parallel antennas were measured in a muscle equivalent phantom and their performance characterized at 915 MHz in terms of the following parameters: peak depth (location of the profile peak with respect to the surface), 50% HL (effective heating length over which SAR greater than 50% of the peak normalized SAR), dead length (axial length at the antenna tip with SAR less than 50% of peak normalized SAR), and the variations of the specific absorption rate pattern relative to the depth of insertion. The results are analyzed and discussed in terms of these parameters and other factors important in the clinical use of these antennas for effective interstitial hyperthermia.

SAR patterns of external 915 MHz microwave applicators.

Analysis of the results from recent clinical trials has shown that tumour size is a significant prognostic factor in eventual tumour control in patients treated with thermoradiotherapy. The critical issue appears to be the adequate coverage of hyperthermia target volume with 'therapeutic temperature'. Therefore one must choose appropriate applicators for the treatment of a given tumour. Accurate knowledge of performance characteristics of the applicators used in clinics thus becomes crucial. In an attempt to take the first step for the appropriate selection of applicators in clinics several commonly used applicators were evaluated according to their 75%, 50%, and 25% two-dimensional SAR (specific absorption rate) contours at depths of 1-3 cm. The data were subsequently approximated by rectangles. This type of information, even with its limitations, is extremely important in implementing quality assurance in hyperthermia. In this communication we will present such information, and the implications in current hyperthermia clinical trials will be discussed.