[en] Purpose: To evaluate errors associated with using different deformable image registration (DIR) pathways to deform dose from planning CT (pCT) to cone-beam CT (CBCT). Methods: Deforming dose is controversial because of the lack of quality assurance tools. We previously proposed a novel metric to evaluate dose deformation error (DDE) by warping dose information using two methods, via dose and contour deformation. First, isodose lines of the pCT were converted into structures and then deformed to the CBCT using an image based deformation map (dose/structure/deform). Alternatively, the dose matrix from the pCT was deformed to CBCT using the same deformation map, and then the same isodose lines of the deformed dose were converted into structures (dose/deform/structure). The doses corresponding to each structure were queried from the deformed dose and full-width-half-maximums were used to evaluate the dose dispersion. The difference between the FWHM of each isodose level structure is defined as the DDE. Three head-and-neck cancer patients were identified. For each patient, two DIRs were performed between the pCT and CBCT, either deforming pCT-to-CBCT or CBCT-to-pCT. We evaluated the errors associated by using either of these pathways to deform dose. A commercially available, Demons based DIR was used for this study, and 10 isodose levels (20% to 105%) were used to evaluate the errors in various dose levels. Results: The prescription dose for all patients was 70 Gy. The mean DDE for CT-to-CBCT deformation was 1.0 Gy (range: 0.3–2.0 Gy) and this was increased to 4.3 Gy (range: 1.5–6.4 Gy) for CBCT-to-CT deformation. The mean increase in DDE between the two deformations was 3.3 Gy (range: 1.0–5.4 Gy). Conclusion: The proposed DDF was used to quantitatively estimate dose deformation errors caused by different pathways to perform DIR. Deforming dose using CBCT-to-CT deformation produced greater error than CT-to-CBCT deformation.

[en] Purpose: To develop a novel method to monitor external anatomical changes in head and neck cancer patients in order to help guide adaptive radiotherapy decisions. Methods: The method, developed in MATLAB, reveals internal anatomical changes based on variations observed in external anatomy. Weekly kV-CBCT scans from 11 Head and neck patients were retrospectively analyzed. The pre-processing step first corrects each CBCT for artifacts and removes pixels from the immobilization mask to produce an accurate external contour of the patient’s skin. After registering the CBCTs to the initial planning CT, the external contours from each CBCT (CBCTn) are transferred to the first week — reference — CBCT_1. Contour radii, defined as the distances between an external contour and the central pixel of each CBCT slice, are calculated for each scan at angular increments of 1 degree. The changes in external anatomy are then quantified by the difference in radial distance between the external contours of CBCT1 and CBCTn. The radial difference is finally displayed on a 2D intensity map (angle vs radial distance difference) in order to highlight regions of interests with significant changes. Results: The 2D radial difference maps provided qualitative and quantitative information, such as the location and the magnitude of external contour divergences and the rate at which these deviations occur. With this method, anatomical changes due to tumor volume shrinkage and patient weight loss were clearly identified and could be correlated with the under-dosage of targets or over-dosage of OARs. Conclusion: This novel method provides an efficient tool to visualize 3D external anatomical modification on a single 2D map. It quickly pinpoints the location of differences in anatomy during the course of radiotherapy, which can help determine if a treatment plan needs to be adapted

[en] Purpose: To compare markerless template-based tracking of lung tumors using dual energy (DE) cone-beam computed tomography (CBCT) projections versus single energy (SE) CBCT projections. Methods: A RANDO chest phantom with a simulated tumor in the upper right lung was used to investigate the effectiveness of tumor tracking using DE and SE CBCT projections. Planar kV projections from CBCT acquisitions were captured at 60 kVp (4 mAs) and 120 kVp (1 mAs) using the Varian TrueBeam and non-commercial iTools Capture software. Projections were taken at approximately every 0.53° while the gantry rotated. Due to limitations of the phantom, angles for which the shoulders blocked the tumor were excluded from tracking analysis. DE images were constructed using a weighted logarithmic subtraction that removed bony anatomy while preserving soft tissue structures. The tumors were tracked separately on DE and SE (120 kVp) images using a template-based tracking algorithm. The tracking results were compared to ground truth coordinates designated by a physician. Matches with a distance of greater than 3 mm from ground truth were designated as failing to track. Results: 363 frames were analyzed. The algorithm successfully tracked the tumor on 89.8% (326/363) of DE frames compared to 54.3% (197/363) of SE frames (p<0.0001). Average distance between tracking and ground truth coordinates was 1.27 +/− 0.67 mm for DE versus 1.83+/−0.74 mm for SE (p<0.0001). Conclusion: This study demonstrates the effectiveness of markerless template-based tracking using DE CBCT. DE imaging resulted in better detectability with more accurate localization on average versus SE. Supported by a grant from Varian Medical Systems

[en] Purpose: To evaluate the inherent accuracy of using a surface guided radiotherapy system (SGRT) in the setup and monitoring of patients receiving stereotactic radiosurgery with an open-face SRS immobilization system. Methods: An anthropomorphic head phantom was set up using the Qfix Encompass SRS Immobilization System on a Varian Edge with OSMS and Varian TrueBeam with AlignRT. The phantom was positioned at 0° gantry and couch. A reference image was acquired using the SGRT system and an ROI was created over the mask opening. The couch and gantry were rotated to different combinations focusing on clinically used SRS gantry/couch combinations and those blocking the SGRT cameras. Perceived surface deviation by the SGRT system from the reference image was recorded. A Winston-Lutz test was performed on couch angles tested and used to exclude couch walkout. The deviation magnitude was calculated using translational values and rotational raw values were recorded. Results: The maximum couch walkouts were: 0.4mm (Edge) and 0.5mm (TB). Solely rotating the gantry resulted in a median couch deviation of 0.2mm and range of 0.1–0.3mm for both linacs. Only rotating the couch (0° gantry) resulted in median deviations of 0.6mm and 0.5mm with ranges of 0.3–1.0mm and 0.3–0.7mm for the Edge and TB, respectively. Combining gantry and couch rotations, the median deviations were 0.7mm and 0.9mm with ranges of 0.3–1.1mm and 0.2–1.9mm for the Edge and TB, respectively. Including all combinations, rotation, roll, and pitch median deviations ranged from 0.1–0.3° with pitch demonstrating consistently higher values and a maximum deviation of 1.0° (both linacs). Conclusion: SGRT is a reliable monitoring tool, though taking into account system fluctuations, 1mm is too restrictive a site tolerance to use with the Qfix Encompass mask. Gantry rotation has little effect on system fluctuation even with camera blockage, whereas couch rotation has a larger effect.

[en] Purpose: To estimate the margin necessary to adequately cover the target using markerless motion tracking (MMT) of lung lesions given the uncertainty in tracking and the size of the target. Methods: Simulations were developed in Matlab to determine the effect of tumor size and tracking uncertainty on the margin necessary to achieve adequate coverage of the target. For simplicity, the lung tumor was approximated by a circle on a 2D radiograph. The tumor was varied in size from a diameter of 0.1 − 30 mm in increments of 0.1 mm. From our previous studies using dual energy markerless motion tracking, we estimated tracking uncertainties in x and y to have a standard deviation of 2 mm. A Gaussian was used to simulate the deviation between the tracked location and true target location. For each size tumor, 100,000 deviations were randomly generated, the margin necessary to achieve at least 95% coverage 95% of the time was recorded. Additional simulations were run for varying uncertainties to demonstrate the effect of the tracking accuracy on the margin size. Results: The simulations showed an inverse relationship between tumor size and margin necessary to achieve 95% coverage 95% of the time using the MMT technique. The margin decreased exponentially with target size. An increase in tracking accuracy expectedly showed a decrease in margin size as well. Conclusion: In our clinic a 5 mm expansion of the internal target volume (ITV) is used to define the planning target volume (PTV). These simulations show that for tracking accuracies in x and y better than 2 mm, the margin required is less than 5 mm. This simple simulation can provide physicians with a guideline estimation for the margin necessary for use of MMT clinically based on the accuracy of their tracking and the size of the tumor

[en] Purpose: Table overrides are relatively common in radiotherapy, yet are a potential safety concern. The goal of this study was to survey current departmental policies on treatment couch overrides and table tolerance values used clinically. Methods: A 25 question electronic survey on couch overrides and tolerances was sent to full members of the AAPM. In the first part of the survey, participants were asked: if table overrides were allowed at their institution, who was allowed to perform these overrides, and if imaging was required with overrides. In the second part of the survey, individuals were asked to provide table tolerance data for the following sites: brain/head & neck, lung, breast, abdomen/pelvis and prostate. Each site was further divided into IMRT/VMAT and 3D conformal techniques. Free-text spaces were provided, allowing respondents to enter any table tolerance data they were unable to specify under the treatment sites listed. Results: A total of 361 individuals responded, of which approximately half participated in the couch tolerances portion of the survey. Overall, 86% of respondents’ institutions allow couch tolerance overrides at treatment. Therapists were the most common staff members permitted to perform overrides, followed by physicists, dosimetrists, and physicians, respectively. Of the institutions allowing overrides, 34% reported overriding daily. More than half of the centers require documentation of the override and/or a setup image (acquired after override) to radiographically verify the treatment site. With respect to table tolerances, groups resulting from the free-text responses were at the two extremes; SRS/SBRT were the tightest, while clinical setup/mets/extremities were the most generous. There was no qualitative difference between IMRT/VMAT and 3D conformal table tolerances. Conclusion: This work is intended to stimulate a discussion within the radiotherapy community. This discussion, supplemented by the survey results provides an opportunity for institutions to enhance patient safety during daily treatment delivery

[en] Purpose: Investigate the impact of tissue inhomogeneities on dose distributions produced by low-energy X-rays in intra-operative radiotherapy (IORT). Methods: A 50-kV INTRABEAM X-ray device with superficial (Flat and Surface) applicators was commissioned at our institution. For each applicator, percent depth-dose (PDD), dose-profiles (DP) and output factors (OF) were obtained. Calibrated GaFchromic (EBT3) films were used to measure dose distributions in solid water phantom at various depths (2, 5, 10, and 15 mm). All recommended precautions for film-handling, film-exposure and scanning were observed. The effects of tissue inhomogeneities on dose distributions were examined by placing air-cavities and bone and tissue equivalent materials of different density (ρ), atomic number (Z), and thickness (t = 0–4mm) between applicator and film detector. All inhomogeneities were modeled as a cylindrical cavity (diameter 25 mm). Treatment times were calculated to deliver 1Gy dose at 5mm depth. Film results were verified by repeat measurements with a thin-window parallel plate ion-chamber (PTW 34013A) in a water tank. Results: For a Flat-4cm applicator, the measured dose rate at 5mm depth in solid water was 0.35 Gy/min. Introduction of a cylindrical air-cavity resulted in an increased dose past the inhomogeneity. Compared to tissue equivalent medium, dose enhancement due to 1mm, 2mm, 3mm and 4mm air cavities was 10%, 16%, 24%, and 35% respectively. X-ray attenuation by 2mm thick cortical bone resulted in a significantly large (58%) dose decrease. Conclusion: IORT dose calculations assume homogeneous tissue equivalent medium. However, soft X-rays are easily affected by non-tissue equivalent materials. The results of this study may be used to estimate and correct IORT dose delivered in the presence of tissue inhomogeneities

[en] Intraperitoneal metastases from ovarian and other gynecologic tumors are a significant source of treatment failure. In recent years, investigators have used radiolabeled monoclonal antibodies to treat this disease with encouraging results. We have developed a dose calculational technique which generates isodose distributions from intraperitoneally administered alpha and beta particle emitters. In this study we apply the calculations to tissue biopsy samples to determine the adequacy of dose to ovarian micrometastases. Tissue samples from staging biopsies at the time of surgical debulking are scanned to identify small metastases. The patient population studied comprised those with ovarian disease who based on clinical criteria would be considered good candidates for intraperitoneal radioimmunotherapy. The regions of interest (which include the tumor and surface of the peritoneum) are digitized and tumor volumes are contoured. Dose calculations based on the modeling of intraperitoneally administered antibodies radiolabeled with various isotopes is performed and the minimum dose to tumor and normal tissue is assessed. For example, with tumor uptake of 0.1% injected dose per gram of tissue, the surface tumor dose from alpha emitters is up to 45,000 rads. The dose falls to 6000 rads at approximately 40 microns from the peritoneal surface. The surface dose from 20 mCi 90Y administered in 1500 ml saline is up to 10,000 rads, and at a 2-mm depth, approximately 2000 rads. From our calculation dose distribution from radioimmunotherapy varies as a function of physical characteristics of the isotope, absorption of activity, and amount of disease being treated

[en] The dosimetric effects of bone and air heterogeneities in head and neck IMRT treatments were quantified. An anthropomorphic RANDO phantom was CT-scanned with 16 thermoluminescent dosimeter (TLD) chips placed in and around the target volume. A standard IMRT plan generated with CORVUS was used to irradiate the phantom five times. On average, measured dose was 5.1% higher than calculated dose. Measurements were higher by 7.1% near the heterogeneities and by 2.6% in tissue. The dose difference between measurement and calculation was outside the 95% measurement confidence interval for six TLDs. Using CORVUS' heterogeneity correction algorithm, the average difference between measured and calculated doses decreased by 1.8% near the heterogeneities and by 0.7% in tissue. Furthermore, dose differences lying outside the 95% confidence interval were eliminated for five of the six TLDs. TLD doses recalculated by Pinnacle³'s convolution/superposition algorithm were consistently higher than CORVUS doses, a trend that matched our measured results. These results indicate that the dosimetric effects of air cavities are larger than those of bone heterogeneities, thereby leading to a higher delivered dose compared to CORVUS calculations. More sophisticated algorithms such as convolution/superposition or Monte Carlo should be used for accurate tailoring of IMRT dose in head and neck tumours

[en] Purpose: To identify areas of improvement in our liver stereotactic body radiation therapy (SBRT) program, using failure mode and effect analysis (FMEA). Methods: A multidisciplinary group consisting of one physician, three physicists, one dosimetrist and two therapists was formed. A process map covering 10 major stages of the liver SBRT program from the initial diagnosis to post treatment follow-up was generated. A total of 102 failure modes, together with their causes and effects, were identified. The occurrence (O), severity (S) and lack of detectability (D) were independently scored. The ranking was done using the risk probability number (RPN) defined as the product of average O, S and D numbers for each mode. The scores were normalized to remove inter-observer variability, while preserving individual ranking order. Further, a correlation analysis on the overall agreement on rank order of all failure modes resulted in positive values for successive pairs of evaluators. The failure modes with the highest RPN value were considered for further investigation. Results: The average normalized RPN values for all modes were 39 with a range of 9 to 103. The FMEA analysis resulted in the identification of the top 10 critical failures modes as: Incorrect CT-MR registration, MR scan not performed in treatment position, patient movement between CBCT acquisition and treatment, daily IGRT QA not verified, incorrect or incomplete ITV delineation, OAR contours not verified, inaccurate normal liver effective dose (Veff) calculation, failure of bolus tracking for 4D CT scan, setup instructions not followed for treatment and plan evaluation metrics missed. Conclusion: The application of FMEA to our liver SBRT program led to the identification and possible improvement of areas affecting patient safety.