The multi-leaf collimator (MLC) dosimetric leaf gap (DLG) offset and transmission are measured parameters which influence the dosimetric accuracy of radiation therapy treatment plans. An international consortium developed an efficient method of automatically measuring the two-dimensional MLC DLG and transmission using the electronic portal imaging device (EPID) on multiple Varian TrueBeam linacs incorporating Millennium and high definition (HD) MLCs. Quality control was implemented as part of each test. Results were compiled including comparisons to baseline measurements and between machines. Validations were accomplished using ion chambers (IC) and a 2D IC array. Sensitivity was investigated by introducing deliberate leaf position offsets, repeatability was assessed, and performance was analyzed using statistical process control methods. The EPID measured DLG (EDLG) and transmission were consistently smaller than the IC measured DLG and transmission for all machines tested. EDLG variations across two dimensions averaged more than 0.24 mm for both Millennium and HD MLCs, demonstrating that leaf-to-leaf DLG variations should be assessed for both MLC types. The two-dimensional correlation coefficient between the IC array and EPID measurements was at least 0.937. A nearly linear relationship between changes in the EPID measured leaf gap and actual leaf positions was measured, with R2 > 0.999. EDLG results were analyzed as a difference relative to a baseline EDLG measurement for each machine, a metric we labeled the leaf offset constancy (LOC). The maximum change in LOC out of the 20 repeatability tests was −0.094 mm. The LOC QC process was found to be capable (Cpk > 1) for 4 machines using a ± 0.15 mm specification limit. LOC results showed that leaves may deviate from their reference positions at a level that approaches dosimetric significance for small stereotactic and highly modulated treatment fields, with average excursions measuring up to 0.21 mm for a Millennium MLC and 0.16 mm for a HD MLC. The effects of initialization, gantry angle, and common repairs are also reported.

Purpose:The MLC dosimetric leaf gap (DLG) and transmission are measured parameters which impact the dosimetric accuracy of IMRT and VMAT plans. This investigation aims to develop an efficient and accurate routine constancy check of the physical DLG in two dimensions.Methods:The manufacturer's recommended DLG measurement method was modified by using 5 fields instead of 11 and by utilizing the Electronic Portal Imaging Device (EPID). Validations were accomplished using an ion chamber (IC) in solid water and a 2D IC array. EPID data was collected for 6 months on multiple TrueBeam linacs using both Millennium and HD MLCs at 5 different clinics in an international consortium. Matlab code was written to automatically analyze the images and calculate the 2D results. Sensitivity was investigated by introducing deliberate leaf position errors. MLC calibration and initialization history was recorded to allow quantification of their impact. Results were analyzed using statistical process control (SPC).Results:The EPID method took approximately 5 minutes. Due to detector response, the EPID measured DLG and transmission differed from the IC values but were reproducible and consistent with changes measured using the ICs. For the Millennium MLC, the EPID measured DLG and transmission were both consistently lower than IC results. The EPID method was implemented as leaf offset and transmission constancy tests (LOC and TC). Based on 6 months of measurements, the initial leaf‐specific action thresholds for changes from baseline were set to 0.1 mm. Upper and lower control limits for variation were developed for each machine.Conclusion:Leaf offset and transmission constancy tests were implemented on Varian HD and Millennium MLCs using an EPID and found to be efficient and accurate. The test is effective for monitoring MLC performance using dynamic delivery and performing process control on the DLG in 2D, thus enhancing dosimetric accuracy.This work was supported by a grant from Varian Medical Systems.

Purpose: To assess the quantitative performance and reproducibility of a new generation linear accelerator using images and trajectory log files across institutions. Methods: A test suite was created to include tests recommended by TG142 (e.g, picket fence at cardinal static gantry angles and during VMAT) and TG179 (e.g. image quality). The test suite, distributed to a consortium of 7 institutions, consisted of DICOM‐RT files and a phantom with BBs at known locations. During each delivery, EPID images were acquired along with trajectory log files. Baseline for each irradiation was set with a flood field without the table or phantom. The phantom was then placed in position and the remaining tests were performed. An analysis program, created in Matlab, assessed the accuracy of leaf, jaw, and collimator positions utilizing the EPID images. Trajectory log files were analyzed as well to assess dynamic parameters such as the reproducibility of gantry motion during arc delivery. Results: Fifteen irradiations were performed on 5 accelerators. Leaf position reproducibility was 0.095 mm for a standard MLC and 0.110 mm for an HDMLC, with maximum standard deviations of 0.019, 0.053, and 0.002 mm for static, IMRT, and arc fields over all linacs. Trajectory logs were consistent with measurements. The maximum gantry deviation was 0.247 ± 0.0160 degrees. Using two different materials, the contrast‐to‐noise ratio was 1.43 ±0.740 and 7.41 ±0.24 for kV and MV images with kV CNR varying by more than a factor of 2 between different machines. Conclusion: EPID and trajectory logs demonstrated thresholds for detection of leaf position errors that were an order of magnitude less than TG142 requirements for different delivery types across institutions. Trajectory log files provided more detailed information regarding stability of gantry position. When tracked over time, these data can be used to reassess the frequency of different test types. This work is supported by Varian Medical Systems.

Melanoma and renal cell carcinoma (RCC) are traditionally classified as radioresistant histologies. However, conflicting single institution data exist using both gamma knife and linear accelerator (LINAC) stereotactic radiosurgery (SRS) approaches. Here we report our institution's experience treating melanoma and RCC with LINAC SRS. Fifty-seven patients were treated with LINAC SRS for melanoma or RCC brain metastases from 2000-2011. 37 patients had at least one MRI imaging follow-up (melanoma n=17, RCC n=20) and were evaluated in the subsequent analysis. Cone-based SRS was used from 2000-2006 for n=24 and micro-multi leaf collimators (mMLCs) from 2007-2011 for n=13. 6 patients received whole-brain radiation prior to SRS. The median SRS dose was 18 Gy (range 15-20 Gy) prescribed to a median isodose line of 85% (range 56-99%). Median number of lesions treated was 2 (range 1-4), median diameter was 13.4 mm (range 4.7-48.9 mm), and median lesion volume was 0.94 cm3 (range 0.01-12.4 cm3). Endpoints were freedom from local failure (FFLF), freedom from distant intracranial failure (FFDICF), and overall survival (OS). Log-rank and Cox proportional hazard models were used for analysis. Median age was 62.0 years with median ECOG = 1. Of the 37 patients, 17 presented asymptomatically, with the remainder presenting with neurologic symptoms. There were 18 single brain metastases. Extracranial metastases were present in n=32 and the primary tumor was controlled in n=29. Actuarial OS at 6 months and 1 year was 69.9% and 30.8%, respectively. For patients, the actuarial FFLF and FFDICF at 1-year was 66.2% and 57.9%, respectively. On univariate analysis melanoma trended towards a worse FFDICF (p=0.08). Analysis of individual lesions demonstrated a 1-year FFLF of 65.0% for melanoma and 59.5% for RCC. Single-lesion univariate analysis showed a trend towards improved FFLF for conformality index (CI) >2 (p=0.10) and mMLCs (p=0.07). On multivariable analysis CI >2 was significantly associated with improved FFLF (hazard ratio = 0.16, p = 0.02), while mMLCs trended towards improved FFLF (HR = 0.16, p = 0.09). Salvage therapy included WBRT (n = 6), SRS (n = 5), surgery (n = 2), and surgery + SRS (n = 1). Toxicity was acceptable with n=3 developing biopsy proven radionecrosis. LINAC-based SRS for melanoma and RCC brain metastases was well tolerated. FFLF was lower than other histologies from historical reports, but consistent with data for melanoma and RCC. There is a trend toward worse DICF in patients with melanoma, suggesting that this patient subset may benefit from WBRT in addition to SRS consistent with level 1 evidence suggesting improved FFLF in patients receiving combined therapy.

Purpose: To investigate the impact of monitor units (MUs) per segment and dose rate on the dose delivery accuracy of step‐and‐shoot intensity modulated radiation (SS‐IMRT) fields on a TrueBeam LINAC.Methods: A step‐and‐shoot multi‐leaf collimator (MLC) sequence was created consisting of 10 identical 10×10cm segments with identical MUs. Beam holding between segments was achieved by moving one out‐of‐field MLC leaf pair. Measurements were repeated for various combinations of MU/segment ranging from 2–40 and dose rates of either 300 or 600MU/min. All measurements were made with a Farmer (0.6cc) ionization chamber placed at the isocenter in a SolidWater phantom at 10cm depth. The measurements were performed on two Varian LINACs: Trilogy and TrueBeam. Each sequence was delivered three times and the charge readings for the corresponding segments were averaged. The effects of MU/segment, dose rate, and LINAC type on the relative dose variation (Delta_i) was compared using F‐tests.Results: On the Trilogy, large Delta_i was observed in small MU segments: at 2MU/segment, the maximum Delta_i was 20.0%/26.2% at 300/600MU/min, respectively. Also, the first segment of each sequence consistently over‐shot(Delta_i>0), while the last segment consistently under‐shot(Delta_i<0). On the TrueBeam, at 2MU/segment, Delta_i ranged from −1.6% to +2.0% at 300 MU/min and −3.0% to +3.1% at 600MU/min; no obvious over/under‐shoot trend was observed. F‐tests showed statistically significant difference ((1—beta)=1.0000) in 2 MU segments between the Trilogy and the TrueBeam, at either dose rate. No significant difference was observed for large (>10 MU/segment) MU segments Conclusions: Earlier generations of Varian LINACs exhibited large dose variations for small MU segments in SS‐IMRT delivery. Our results appeared to confirm these findings. The dose delivery accuracy in small MU segments and at high dose rate for SS‐IMRT is significantly improved on TrueBeam compared to Trilogy, likely due to the faster sampling rate (100Hz vs. 20Hz).

Purpose: Surface matching provides quantitative shifts for patient positioning of superficial treatment sites such as breast. We investigated the reliability of 3D surface matching using AlignRT compared to positioning using skin marks followed by MV portal imaging for WBRT. Methods: Five patients receiving two‐field WBRT without respiratory gating were positioned daily on a breastboard (n=2) or custom alphacradle (n=3). For each treatment fraction guided by MV (n=23), the 3D surface captured using AlignRT (v4.5) was compared to the surface generated from CT simulation data. The AlignRT registration algorithm outputs 3D rotations plus translations to optimize matching between surfaces in user‐defined ROIs. The correlation of AlignRT and MV shifts in 4 degree‐of‐freedom (3D translations and table rotation) was studied for two ROIs: the entire surface (‘all’) and the treated breast (‘Breast’). Two surfaces were re‐captured in the treatment position to compare residual registration errors. Parametric statistical tests were considered significant at p<0.05. Results: The 3D distance of the table shifts was 3.8±5.1mm following MV imaging, 5.9±4.1mm for ‘all’, and 5.7±3.3mm for ‘Breast’ registration. while AlignRT distances were significantly different from MV, only ˈallˈ registration correlated with MV shifts (r=0.76). The 3D difference in initial and final positions correlated with table shifts for both ROIs (r>0.9) although residual 3D distances according to AlignRT remained high for ‘all’ (5.0±4.7mm) and ‘Breast’ (4.7±4.2mm). Furthermore, absolute table rotations calculated by AlignRT for consecutively acquired surfaces exhibited fluctuations that were significantly larger for registrations of ‘Breast’ (0.62±0.58 degrees) than ‘all’ (0.27±0.18 degrees). Conclusion: Breast surface matching using AlignRT depends upon the registered ROI. A large ROI showed higher correlation with MV shifts and increased stability when calculating table rotations. Registrations exhibited a large baseline offset for all ROIs, indicating that quantitative table shifts from AlignRT overestimate those determined from MV imaging by expert physicians for WBRT.

To determine the relation between the incidence of RP and the SBRT dose distribution in a Phase I dose escalation study of multiple metastases. Patients with 1-5 sites of metastatic cancer with a life expectancy of >3 months and good performance status received escalating doses of radiation to all known sites of cancer with SBRT. Twenty-eight patients with 50 lesions in the lung were evaluated for this study. All patients underwent 4DCT simulation and FDG PET (if possible) for internal target volume (ITV) delineation on 4DCT images. Planning target volumes (PTV) varied from 5.7cc to 265.6cc and received 24 Gy to 48 Gy in 3 fractions. In addition, one of the patients received a dose fractionation of 10 x 5 Gy and another one 30 x 2 Gy to one of their lung sites, respectively. Optimized SBRT plans with 9-15 coplanar/noncoplanar beam arrangements were designed using a treatment planning system with convolution/superposition algorithm and tissue heterogeneity corrections. The gated SBRT treatments were delivered between July 2005 and December 2008. Dose volume histograms (DVH) were calculated for the total lung excluding the gross tumor and converted to normalized total dose (NTD) at 2 Gy fractions by using the linear quadratic model with α/β ratio of 3 Gy. Normal tissue complication probabilities (NTCP) were evaluated using the Lyman model as per Kwa et al. (1998). Late toxicities were scored using the NCI Common Terminology Criteria for Adverse Events v3.0. A 61% local control was achieved with 39% of patients surviving in this cohort at the time of study. Seven patients out of 28 experienced grade 2 or higher RP. An increasing RP rate with increasing NTDmean was observed to be significant with a mean NTD of 1514±402c Gy for >grade 2 versus 854±508c Gy for < grade 2 (p = 0.004). The Lyman NTCP modeling with the parameters NTD50 = 30.5 Gy and m = 0.30 produced a well-defined NTCP curve with mean values of 6.0±4.0% for >grade2 and 2.0±3.4% for or < 2 (p = 0.086). The dose-effect relation between the NTDmean and RP established for standard fractionated treatments of lung may be extended to the SBRT treatments of multiple lung lesions up to 3x16 Gy as a predictor of lung toxicity. This tool can facilitate the SBRT treatment planning and analysis process in a dose fractionation regimen not well experienced.

To investigate the potential benefits of non-coplanar large segmented direct machine parameter optimization (DMPO) for primary and metastatic lung SBRT. To assess the delivery accuracy of DMPO beam segments to a gated target in a motion phantom. Seven patients with primary or metastatic lung lesions treated with SBRT according to our institutional IRB protocol were included in this study. Patients with oligometastases received doses of 3 x 10-14 Gy and primary NSCLC patients received 50-60 Gy in 3-10 fractions depending on the tumor size and location. 3D custom treatment plans used 10-12 non-coplanar 6MV beam arrangements with manually optimized beam MLC aperture and weight to meet the clinical goals. Inverse planning was done using the identical beam arrangements. Beam apertures were reset and optimized through the DMPO (Pinnacle, Philips) algorithm with one segment per beam and the minimum segment area of half of the maximum PTV cross-section in the BEV. DMPO plans were normalized at the same identical target coverage level as those for 3D plans. The optimization routine used the collapsed-cone convolution dose calculation after 10 successive iterations. 3D dose distributions, dose volume histograms, and normal tissue complication probabilities (NTCP) were calculated for both 3D conformal and 1-segment/DMPO SBRT plans. Finally, a motion phantom with a lung-equivalent insert was fitted with a small tissue-equivalent material to represent a tumor in lung. Gafchromic EBT films were fitted between the sections of this phantom and were irradiated with optimized single-segment beams to evaluate the dose in the lung, in the target and at the edge of the target. One-segment/DMPO planning improved the conformality of SBRT over 3D CRT delivering on average 12% ± 7%, 8% ± 7%, and 2% ± 2% less dose to the lung at 20, 13, and 5Gy levels, respectively. The maximum dose to the heart, esophagus, and cord were comparable between the 1-segment DMPO and 3DCRT plans. Except for one case, there was no increase in the number of monitor units used with 1-segment/DMPO plans. One -segment/DMPO plans yielded a mean reduction of 17.1% and 30% in the normal lung EUD and estimated lung complication probability, respectively, over 3DCRT plans. A dose to distance agreement of 3%/3 mm between calculation and film measurement for a representative plan in a motion phantom with gating was verified at 99% of points within the fields. Single segment beam DMPO can be used to improve SBRT planning for lung lesions to meet the planning goals in an effective manner. This approach allows for easily deliverable and verifiable beam apertures for gated beam delivery. The automation of our method is a good alternative to more traditional methods and offers significant dosimetric benefits.

Stereotactic radiosurgery/radiotherapy (SRS/SRT) delivers radiation with great spatial precision. To achieve sub-mm precision for intracranial SRS, a head ring is rigidly fixated to the skull to create a fixed reference. For some patients, invasiveness of the ring can be highly uncomfortable and not well-tolerated. In addition, placing and removing the ring requires special expertise from the physician, and patient setup time for SRS can often be long. To reduce the invasiveness and setup time, we are developing a system for performing accurate head positioning without use of a head ring. The system uses real-time 3D optical position feedback for gating the treatment beam and guiding a motor-controlled 3D head motion compensation stage. The proposed system consists of a central control computer with Labview software, an optical patient motion tracking system (Polaris, NDI), and a 3D motion compensation stage. The Polaris 4D infrared tracking system is used to monitor four optical reflective markers attached to a bite block at 30 frames/second with an accuracy of approximately 0.25 mm. A styrofoam head cast is custom-built for patient support and attached to a LINAC couch mount (Radionics) that can be micro-adjusted in 3D. The motion feedback of the markers is processed by the control computer, and thus the motion of the lesion can be calculated. If the lesion deviates beyond a preset 3D motion tolerance, a relay switch is activated and the MV beam is turned off. After a stabilization period, if the lesion has re-entered the tolerance, the beam is automatically turned on. Otherwise, if the patient stabilizes to a new position outside the tolerance, an automatic position correction signal is sent to stepper motors to adjust the head position via the couch mount motion stage. The corrected lesion location is confirmed before the treatment beam is turned on. Initial feasibility study of head motion was conducted with healthy volunteers supported with a head cast cradle and instructed to mimic natural sporadic motions. Analysis showed that the simulated PTV stayed within the 1 mm tolerance for extended periods (approximately 50 seconds) between large motion excursions. Simulations using recorded PTV motion as input for gating, suggested a total beam-on time, or duty cycle of 75%. With active head position correction, the duty cycle and spatial accuracy can be further improved. With an optical tracking system and minimal back-of-the-head support, healthy volunteers demonstrated the ability to maintain sub-mm head position with sufficient time windows for treatment. Adaptive motion compensation showed promise in implementing a high-efficient and accurate, frameless SRS/SRT system.

Purpose: To reduce patient setup time and to perform accurate non-invasive frameless head radiosurgery/radiotherapy by use of real-time position feedback for treatment beam gating and head motion stage guiding. Method and Materials: A Polaris 4D tracker (NDI) was used to monitor four optical reflective markers rigidly fixated to a biteblock at 30fps with an RMS accuracy of 0.25mm. Head motion monitoring was performed on healthy volunteers using a styrofoam head cast for support. Simulation and prototyping were investigated using the head motion as the feedback input for beam gating and head motion compensation. Design specifications include using Labview to import the real-time biofeedback information and to monitor whether the center of the PTV is within a 3D motion tolerance of 1.0mm. If the PTV exceeds the tolerance, a relay switch is activated and the MV beam is turned off. After a 5s stabilization period, provided that the PTV has reentered the set tolerance, the beam is automatically turned on. However, if the patient stabilizes to a new position outside the tolerance, an automatic position correction is performed using a stepper motor controlled head stage. Results: With only a head cast cradle, healthy volunteers mimicked natural sudden sporadic motions on the treatment couch. Data showed that the simulated PTV stayed within the 1mm tolerance for extended periods (∼50s) between large motion excursions. Simulations using recorded head motion data as an input for gating, suggested a total beam-on time, or duty cycle of 75%. With active head position correction, the duty cycle and spatial accuracy can be further improved. Conclusion: With an optical tracking system and minimal back-of-the-head support, healthy volunteers demonstrated the ability to maintain submillimeter head position with sufficient time windows for treatment. Adaptive motion compensation can be helpful in implementing a high-efficient and accurate frameless SRS/SRT system.

To evaluate the accuracy of SBRT treatment planning with tissue heterogeneity corrections for primary and metastatic lung tumors using collapsed cone (CC), pencil beam (PB), and a new commercially available Monte Carlo (XVMC) algorithms. Two patients with lung oligometastases (one patient had four isolated lesions) and one patient with a primary non-small cell lung cancer were treated with 3D-conformal SBRT to 30-60 Gy in 3 fractions. Planning target volumes (PTVs), based on 4D-CT simulations, ranged from 8.1-61.1 cc (mean = 23.4 cc). XVMC algorithm was commissioned on the BrainLAB IPlan system, which also had a PB algorithm. Target and normal tissue volumes on axial CT images were transferred from Pinnacle to IPlan via DICOM protocols. Both planning systems used the same CT electron density table for heterogeneity corrections. Plans using non-opposed, non-coplanar 6MV beams were first generated in IPlan with PB calculations. Dose distributions were then recalculated by XVMC with the same dose grid size (overall statistical uncertainty <0.3%). CC calculations were subsequently performed in Pinnacle with the same beam arrangements. Three dimensional dose distributions and dose-volume-histogram statistics were evaluated for all algorithms. XVMC dose calculations showed that PB algorithm consistently overestimated isocenter dose by 2.5%-6.6%. PTV covered by prescription dose (V100) was also overestimated up to 68% (when the lesion was adjacent to the chest wall). Isocenter dose predicted by CC closely agreed with that calculated by XVMC for all cases, except for the chest wall lesion, where CC underestimated it by 4%. V100 calculated by CC was uniformly underestimated by up to 40% compared to XVMC predictions. The largest deviation occurred for a 3cm-lesion, which was entirely surrounded by the lung tissue. CC underestimated the dose covering 95% of PTV (D95) by 2-3% of XVMC results. Percent lung volumes (excluding PTV) receiving 5, 10, and 20 Gy were similar for all calculation algorithms within 3%. Maximum doses to 1 cc of esophagus and spinal cord were also comparable. Monte Carlo predictions have shown that SBRT dose calculations employing PB and CC algorithms are subject to large errors for small lung targets treated with SBRT. Generally PB overestimates the PTV coverage near margins, while CC underestimates it. The magnitude of the discrepancy varies significantly depending on the location of the target. Although the discrepancy is substantial for V100, D95 is still greater than 92% of prescription for all targets. To predict the volumes of normal lung tissue receiving mid to low doses, the three algorithms provide equivalent results.

The purpose of this study was to assess the efficacy and toxicity of intensity-modulated radiation therapy (IMRT) in the treatment of gastric cancer. Seven patients with gastric cancer were treated with IMRT. Six patients (all Stage III) received post-operative chemoradiotherapy with concurrent 5-fluorouracil and leucovorin. One received planned pre-operative radiation, though did not proceed to surgery. All patients were planned to receive 50.4 Gy in 1.8 Gy fractions. IMRT planning was compared with opposed anterior-posterior: posterior-anterior (AP/PA) and 3-field conventional three-dimensional plans. When compared with either AP/PA or 3-field plans, IMRT significantly reduced the volume exceeding the threshold dose of the liver and at least one kidney. Target coverage with IMRT was excellent, with 98+/-1% of the target receiving >or=100% of the dose. Compared with AP/PA and 3-field plans, IMRT plans had a greater percentage of target receiving the prescribed dose, but also a greater volume receiving >110% of the dose. IMRT was well tolerated; no patients developed acute gastrointestinal toxicity greater than grade 2. All seven experienced grade 2 nausea, three had grade 2 diarrhoea and two had grade 2 oesophagitis. Weight loss ranged from 0-12% (mean 6.1% and median 5.8%). IMRT in the treatment of gastric malignancies reduces the mean and above threshold doses to critical normal tissues. In an initial cohort of seven patients, 50.4 Gy delivered by IMRT is well tolerated and safe.

Objectives To analyze the effects of the B-mode ultrasound acquisition and targeting (BAT) system for positioning of patients with prostate cancer receiving intensity-modulated radiotherapy on acute gastrointestinal (GI) and genitourinary (GU) toxicity. Methods The records of 50 consecutive patients treated using the BAT system were reviewed. Additionally, a comparison (no-BAT) group (ie, a group without a BAT study) treated in a similar manner was identified. The no-BAT group consisted of 49 patients treated immediately before the BAT group. For the two groups, the target definitions and dose prescriptions were identical, the treatment plan acceptance criteria were identical, and intensity-modulated radiotherapy was used for all patients. The daily BAT movements were charted in each of the three principal directions. Acute toxicity was scored for all patients according to the Radiation Therapy Oncology Group GI and GU acute toxicity scales. Results The GU toxicity rates for the BAT versus no-BAT groups were grade 0 in 20% versus 14%; grade 1 in 38% versus 47%; grade 2 in 38% versus 39%; and grade 3 in 4% versus 0%, respectively (P = 0.284). The corresponding GI toxicity rates were grade 0 in 42% versus 27%; grade 1 in 28% versus 29%; and grade 2 in 30% versus 45% (P = 0.040). The incidence of GU and GI toxicity did not correlate with the directions or size of the BAT moves. Regression analysis revealed that for acute GI toxicity, the only variable reaching statistical significance was BAT use; no variable, including BAT use, reached statistical significance for acute GU toxicity. Conclusions The use of the BAT system did not change the rate of acute GU toxicity but did reduce the rate of acute GI toxicity.