A. Moutsatsos

On the implementation of a recently proposed dosimetric formalism to a robotic radiosurgery system.

PURPOSE The aim of this work is to implement a recently proposed dosimetric formalism for nonstandard fields to the calibration and small field output factor measurement of a robotic stereotactic radiosurgery system. METHODS Reference dosimetry measurements were performed in the nonstandard, 60 mm diameter machine specific reference (msr) field using a Farmer ion chamber, five other cylindrical chambers with cavity lengths ranging from 16.25 down to 2.7 mm, and alanine dosimeters. Output factor measurements were performed for the 5, 7.5, 10, and 15 mm field sizes using microchambers, diode detectors, alanine dosimeters, TLD microcubes, and EBT Gafchromic films. Measurement correction factors as described in the proposed formalism were calculated for the ion chamber and diode detector output factor measurements based on published Monte Carlo data. Corresponding volume averaging correction factors were calculated for the alanine output factor measurements using 3D dose distributions, measured with polymer gel dosimeters. RESULTS Farmer chamber and alanine reference dosimetry results were found in close agreement, yielding a correction factor of k(Q(msr),Q)(f(msr),f(ref)) = 0.999 +/- 0.016 for the chamber readings. These results were also found to be in agreement within experimental uncertainties with corresponding results obtained using the shorter cavity length ionization chambers. The mean measured dose values of the latter, however, were found to be consistently greater than that of the Farmer chamber. This finding, combined with an observed inverse relationship between the mean measured dose and chamber cavity length that follows the trend predicted by theoretical volume averaging calculations in the msr field, implies that the Farmer k(Q(msr),Q)(f(msr),f(ref)) correction is greater than unity. Regarding the output factor results, deviations as large as 33% were observed between the different dosimeters used. These deviations were substantially decreased when appropriate correction factors were applied to the measured microchamber, diode, and alanine values. After correction, all diode and microchamber measured output factors agreed within 1.6% with the corresponding alanine measurements, and within 3.1% with the TLD measurements. The weighted mean output factors were 0.681 +/- 0.001, 0.824 +/- 0.001, 0.875 +/- 0.001, and 0.954 +/- 0.001 for the 5, 7.5, 10, and 15 mm beams, respectively. CONCLUSIONS The comparison of Farmer chamber measurements versus alanine reference dosimetry validates the use of the former for dosimetry in the msr field of this treatment delivery system. The corresponding results of this work obtained using chambers with different cavity lengths, combined with previous literature findings, suggest that a k(Q(msr),Q)(f(msr),f(ref)) Farmer chamber dose response correction factor of 1.01 may improve calibration measurement accuracy when using the proposed dosimetric formalism. The k(Q(msr),Q)(f(msr),f(ref)) correction factor is within 0.5% from unity for ion chambers with cavity lengths less than 10 mm. Substantial improvements in small field output factor measurement accuracy can be obtained when using microchambers and diodes by applying appropriately calculated correction factors to the detector measurements according to the proposed dosimetric formalism, and their routine use is therefore recommended.

Dosimetric accuracy of a deterministic radiation transport based 192Ir brachytherapy treatment planning system. Part I: single sources and bounded homogeneous geometries.

PURPOSE The aim of this work is to validate a deterministic radiation transport based treatment planning system (TPS) for single 192Ir brachytherapy source dosimetry in homogeneous water geometries. METHODS TPS results were obtained using the deterministic radiation transport option of a BRACHYVISION v. 8.8 system for three characteristic source designs (VS2000, GMPlus HDR, and GMPlus PDR) with each source either centered in a 15 cm radius spherical water phantom, or positioned at varying distance away from the phantom center. Corresponding MC simulations were performed using the MCNPX code v.2.5.0 and source geometry models prepared using information provided by the manufacturers. RESULTS Comparison in terms of the AAPM TG-43 dosimetric formalism quantities, as well as dose rate distributions per unit air kerma strength with a spatial resolution of 0.1 cm, yielded close agreement between TPS and MC results for the sources centered in the phantom. Besides some regions close to the source longitudinal axes where discrepancies could be characterized as systematic, overall agreement for all three sources studied is comparable to the statistical (type A) uncertainty of MC simulations (1% at the majority of points in the geometry increasing to 2%-3% at points lying both away from the source center and close to the source longitudinal axis). A corresponding good agreement was also found between TPS and MC results for the sources positioned away from the phantom center. CONCLUSIONS Results of this work attest the capability of the TPS to accurately account for the scatter conditions regardless of the size or shape of a given geometry of dosimetric interest, and the position of a source within it. This is important since, as shown in the literature and summarized also in this work, these factors could introduce a significant dosimetric effect that is currently ignored in clinical treatment planning. It is concluded that the implementation of the deterministic radiation transport option of the BRACHYVISION v. 8.8 system for 192Ir brachytherapy dosimetry in homogeneous water geometries yields results of comparable accuracy to the golden standard of Monte Carlo simulation, in clinically viable calculation times.

On the output factor measurements of the CyberKnife iris collimator small fields: Experimental determination of the kQclin,Qmsrfclin,fmsr correction factors for microchamber and diode detectors

PURPOSE To measure the output factors (OFs) of the small fields formed by the variable aperture collimator system (iris) of a CyberKnife (CK) robotic radiosurgery system, and determine the k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factors for a microchamber and four diode detectors. METHODS OF measurements were performed using a PTW PinPoint 31014 microchamber, four diode detectors (PTW-60017, -60012, -60008, and the SunNuclear EDGE detector), TLD-100 microcubes, alanine dosimeters, EBT films, and polymer gels for the 5 mm, 7.5 mm, 10 mm, 12.5 mm, and 15 mm iris collimators at 650 mm, 800 mm, and 1000 mm source to detector distance (SDD). The alanine OF measurements were corrected for volume averaging effects using the 3D dose distributions registered in polymer gel dosimeters. k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factors for the PinPoint microchamber and the diode dosimeters were calculated through comparison against corresponding polymer gel, EBT, alanine, and TLD results. RESULTS Experimental OF results are presented for the array of dosimetric systems used. The PinPoint microchamber was found to underestimate small field OFs, and a k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factor ranging from 1.127 ± 0.022 (for the 5 mm iris collimator) to 1.004 ± 0.010 (for the 15 mm iris collimator) was determined at the reference SDD of 800 mm. The PinPoint k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factor was also found to increase with decreasing SDD; k(Q(clin),Q(msr) ) (f(clin),f(msr) ) values equal to 1.220 ± 0.028 and 1.077 ± 0.016 were obtained for the 5 mm iris collimator at 650 mm and 1000 mm SDD, respectively. On the contrary, diode detectors were found to overestimate small field OFs and a correction factor equal to 0.973 ± 0.006, 0.954 ± 0.006, 0.937 ± 0.007, and 0.964 ± 0.006 was measured for the PTW-60017, -60012, -60008 and the EDGE diode detectors, respectively, for the 5 mm iris collimator at 800 mm SDD. The corresponding correction factors for the 15 mm iris collimator were found equal to 0.997 ± 0.010, 0.994 ± 0.009, 0.988 ± 0.010, and 0.986 ± 0.010, respectively. No correlation of the diode k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factors with SDD was observed. CONCLUSIONS This work demonstrates an experimental procedure for the determination of the k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factors required to obtain small field OF results of increased accuracy.

Dosimetric accuracy of a deterministic radiation transport based 192Ir brachytherapy treatment planning system: Part III. Comparison to Monte Carlo simulation in voxelized anatomical computational models

PURPOSE To compare TG43-based and Acuros deterministic radiation transport-based calculations of the BrachyVision treatment planning system (TPS) with corresponding Monte Carlo (MC) simulation results in heterogeneous patient geometries, in order to validate Acuros and quantify the accuracy improvement it marks relative to TG43. METHODS Dosimetric comparisons in the form of isodose lines, percentage dose difference maps, and dose volume histogram results were performed for two voxelized mathematical models resembling an esophageal and a breast brachytherapy patient, as well as an actual breast brachytherapy patient model. The mathematical models were converted to digital imaging and communications in medicine (DICOM) image series for input to the TPS. The MCNP5 v.1.40 general-purpose simulation code input files for each model were prepared using information derived from the corresponding DICOM RT exports from the TPS. RESULTS Comparisons of MC and TG43 results in all models showed significant differences, as reported previously in the literature and expected from the inability of the TG43 based algorithm to account for heterogeneities and model specific scatter conditions. A close agreement was observed between MC and Acuros results in all models except for a limited number of points that lay in the penumbra of perfectly shaped structures in the esophageal model, or at distances very close to the catheters in all models. CONCLUSIONS Acuros marks a significant dosimetry improvement relative to TG43. The assessment of the clinical significance of this accuracy improvement requires further work. Mathematical patient equivalent models and models prepared from actual patient CT series are useful complementary tools in the methodology outlined in this series of works for the benchmarking of any advanced dose calculation algorithm beyond TG43.

Gamma knife output factor measurements using VIP polymer gel dosimetry.

PURPOSE Water equivalent polymer gel dosimeters and magnetic resonance imaging were employed to measure the output factors of the two smallest treatment fields available in a Gamma Knife model C radiosurgery unit, those formed employing the 4 and 8 mm final collimator helmets. METHODS Three samples of the VIP normoxic gel formulation were prepared and irradiated so that a single shot of the field whose output factor is to be measured and a single shot of the reference 18 mm field were delivered in each one. Emphasis is given to the development and benchmarking of a refined data processing methodology of reduced uncertainty that fully exploits the 3D dose distributions registered in the dosimeters. RESULTS Polymer gel results for the output factor of the 8 mm collimator helmet are found to be in close agreement with the corresponding value recommended by the vendor (0.955 +/- 0.007 versus 0.956, respectively). For the 4 mm collimator helmet, however, polymer gel results suggest an output factor 3% lower than the value recommended by the vendor (0.841 +/- 0.009 versus 0.870, respectively). CONCLUSIONS A comparison with corresponding measurements published in the literature indicates that output factor results of this work are in agreement with those obtained using dosimetric systems which, besides fine spatial resolution and lack of angular and dose rate dependence of the dosimeters response, share with polymer gels the favorable characteristic of minimal radiation field perturbation.

On the output factor measurements of the CyberKnife iris collimator small fields: Experimental determination of the correction factors for microchamber and diode detectors

PURPOSE To measure the output factors (OFs) of the small fields formed by the variable aperture collimator system (iris) of a CyberKnife (CK) robotic radiosurgery system, and determine the k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factors for a microchamber and four diode detectors. METHODS OF measurements were performed using a PTW PinPoint 31014 microchamber, four diode detectors (PTW-60017, -60012, -60008, and the SunNuclear EDGE detector), TLD-100 microcubes, alanine dosimeters, EBT films, and polymer gels for the 5 mm, 7.5 mm, 10 mm, 12.5 mm, and 15 mm iris collimators at 650 mm, 800 mm, and 1000 mm source to detector distance (SDD). The alanine OF measurements were corrected for volume averaging effects using the 3D dose distributions registered in polymer gel dosimeters. k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factors for the PinPoint microchamber and the diode dosimeters were calculated through comparison against corresponding polymer gel, EBT, alanine, and TLD results. RESULTS Experimental OF results are presented for the array of dosimetric systems used. The PinPoint microchamber was found to underestimate small field OFs, and a k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factor ranging from 1.127 ± 0.022 (for the 5 mm iris collimator) to 1.004 ± 0.010 (for the 15 mm iris collimator) was determined at the reference SDD of 800 mm. The PinPoint k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factor was also found to increase with decreasing SDD; k(Q(clin),Q(msr) ) (f(clin),f(msr) ) values equal to 1.220 ± 0.028 and 1.077 ± 0.016 were obtained for the 5 mm iris collimator at 650 mm and 1000 mm SDD, respectively. On the contrary, diode detectors were found to overestimate small field OFs and a correction factor equal to 0.973 ± 0.006, 0.954 ± 0.006, 0.937 ± 0.007, and 0.964 ± 0.006 was measured for the PTW-60017, -60012, -60008 and the EDGE diode detectors, respectively, for the 5 mm iris collimator at 800 mm SDD. The corresponding correction factors for the 15 mm iris collimator were found equal to 0.997 ± 0.010, 0.994 ± 0.009, 0.988 ± 0.010, and 0.986 ± 0.010, respectively. No correlation of the diode k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factors with SDD was observed. CONCLUSIONS This work demonstrates an experimental procedure for the determination of the k(Q(clin),Q(msr) ) (f(clin),f(msr) ) correction factors required to obtain small field OF results of increased accuracy.

On the use of VIP gel dosimetry in HDR brachytherapy

An experimental procedure is discussed with regard to its potential in 192Ir HDR brachytherapy dosimetry. Two samples of VIP normoxic gel formulation are used; one for gel response calibration and the other for acquiring experimental data. Using the same irradiation method for both calibration and experimental purposes (an 192Ir HDR brachytherapy source) and treating the two samples identically (i.e. the two samples are prepared, irradiated and scanned at the same time and stored together at all times) leads to total dose uncertainties comparable to those of other well established dosimetry methods over a significant dose range (~7Gy-40Gy). In this dose range, the described procedure can be used to either acquire absolute dosimetry results for the characterisation of new 192Ir HDR brachytherapy sources, or to facilitate the planning of relative dosimetry experiments for the verification of calculations by new generation treatment planning systems that are currently phasing in, in complex 3D dose distributions involving inhomogeneities and finite medium geometries.

Characterization of system-related geometric distortions in MR images employed in Gamma Knife radiosurgery applications

This work provides characterization of system-related geometric distortions present in MRIs used in Gamma Knife (GK) stereotactic radiosurgery (SRS) treatment planning. A custom-made phantom, compatible with the Leksell stereotactic frame model G and encompassing 947 control points (CPs), was utilized. MR images were obtained with and without the frame, thus allowing discrimination of frame-induced distortions. In the absence of the frame and following compensation for field inhomogeneities, measured average CP disposition owing to gradient nonlinearities was 0.53 mm. In presence of the frame, contrarily, detected distortion was greatly increased (up to about 5 mm) in the vicinity of the frame base due to eddy currents induced in the closed loop of its aluminum material. Frame-related distortion was obliterated at approximately 90 mm from the frame base. Although the region with the maximum observed distortion may not lie within the GK treatable volume, the presence of the frame results in distortion of the order of 1.5 mm at a 7 cm distance from the center of the Leksell space. Additionally, severe distortions observed outside the treatable volume could possibly impinge on the delivery accuracy mainly by adversely affecting the registration process (e.g. the position of the lower part of the N-shaped fiducials used to define the stereotactic space may be miss-registered). Images acquired with a modified version of the frame developed by replacing its front side with an acrylic bar, thus interrupting the closed aluminum loop and reducing the induced eddy currents, were shown to benefit from relatively reduced distortion. System-related distortion was also identified in patient MR images. Using corresponding CT angiography images as a reference, an offset of 1.1 mm was detected for two vessels lying in close proximity to the frame base, while excellent spatial agreement was observed for a vessel far apart from the frame base.

On the use of polymer gels for assessing the total geometrical accuracy in clinical Gamma Knife radiosurgery applications

The nearly tissue equivalent MRI properties and the unique ability of registering 3D dose distributions of polymer gels were exploited to assess the total geometrical accuracy in clinical Gamma Knife applications, taking into account the combined effect of the units mechanical accuracy, dose delivery precision and the geometrical distortions inherent in MR images used for irradiation planning. Comparison between planned and experimental data suggests that the MR-related distortions due to susceptibility effects dominate the total clinical geometrical accuracy which was found within 1 mm. The dosimetric effect of the observed sub-millimetre uncertainties on single shot GK irradiation plans was assessed using the target percentage coverage criterion, and a considerable target dose underestimation was found.

Gamma Knife relative dosimetry using VIP polymer gel and EBT radiochromic films

The VIP polymer gel–MRI method and EBT Gafchromic films were employed to obtain relative dosimetry results for the Gamma Knife (GK) radiation fields of 4 mm and 18 mm nominal diameter. Results are compared to the corresponding calculations of GammaPlan Treatment Planning System (TPS) in the form of 1D profiles and 2D distributions. Measured and planned relative dosimetry datasets are found in close agreement within experimental uncertainties. A corresponding agreement is shown for Dose Volume Histogram (DVH) results that are available only through the application of the polymer gel method.