Purpose: To use daily cone beam CTs (CBCTs) to develop regularized principal component analysis (PCA) models of anatomical changes in head and neck (H&N) patients, to guide replanning decisions in adaptive radiation therapy (ART). Methods: Known deformations were applied to planning CT (pCT) images of 10 H&N patients to model several different systematic anatomical changes. A Pinnacle plugin was used to interpolate systematic changes over 35 fractions, generating a set of 35 synthetic CTs for each patient. Deformation vector fields (DVFs) were acquired between the pCT and synthetic CTs and random fraction‐to‐fraction changes were superimposed on the DVFs. Standard non‐regularized and regularized patient‐specific PCA models were built using the DVFs. The ability of PCA to extract the known deformations was quantified. PCA models were also generated from clinical CBCTs, for which the deformations and DVFs were not known. It was hypothesized that resulting eigenvectors/eigenfunctions with largest eigenvalues represent the major anatomical deformations during the course of treatment. Results: As demonstrated with quantitative results in the supporting document regularized PCA is more successful than standard PCA at capturing systematic changes early in the treatment. Regularized PCA is able to detect smaller systematic changes against the background of random fraction‐to‐fraction changes. To be successful at guiding ART, regularized PCA should be coupled with models of when anatomical changes occur: early, late or throughout the treatment course. Conclusion: The leading eigenvector/eigenfunction from the both PCA approaches can tentatively be identified as a major systematic change during radiotherapy course when systematic changes are large enough with respect to random fraction‐to‐fraction changes. In all cases the regularized PCA approach appears to be more reliable at capturing systematic changes, enabling dosimetric consequences to be projected once trends are established early in the treatment course. This work is supported in part by a grant from Varian Medical Systems, Palo Alto, CA

Purpose: To investigate the variation of TMR for SRS cones obtained by TMR scanning, calculation from PDDs, and point measurements. The obtained TMRs were also compared to the representative data from the vendor. Methods: TMRs for conical cones of 4, 5, 7.5, 10, 12.5, 15, and 17.5 mm diameter (jaws set to 5×5 cm) were obtained for 6X FFF and 10X FFF energies on a Varian Edge linac. TMR scanning was performed with a Sun Nuclear 3D scanner and Edge detector at 100 cm SDD. TMR point measurements were measured with a Wellhofer tank and Edge detector, at multiple depths from 0.5 to 20 cm and 100 cm SDD. PDDs for converting to TMR were scanned with a Wellhofer system and SFD detector. The formalism of converting PDD to TMR, given in Khan's book (4th Edition, p.161) was applied. Sp values at dmax were obtained by measuring Scp and Sc of the cones (jaws set to 5×5 cm) using the Edge detector, and normalized to the 10×10 cm field. Results: Along the central axis beyond dmax, the RMS and maximum percent difference of TMRs obtained with different methods were as follows: (a) 1.3% (max=3.5%) for the calculated TMRs from PDDs versus direct scanning; (b) 1.2% (max=3.3%) for direct scanning versus point measurement; (c) 1.8% (max=5.1%) for the calculated versus point measurements; (d) 1.0% (max=3.6%) for direct scanning versus vendor data; (e) 1.6% (max=7.2%) for the calculated versus vendor data. Conclusion: The overall accuracy of TMRs calculated from PDDs was comparable with that of direct scanning. However, the uncertainty at depths greater than 20 cm, increased up to 5% when compared to point measurements. This issue must be considered when developing a beam model for small field SRS planning using cones.

Purpose:To evaluate the accuracy of a novel open mask for intracranial SRS and to investigate the capability of monitoring intrafraction motion of the patient within this mask using the optical surface monitoring system (OSMS) on a linac‐based platform.Methods:Efficiency was evaluated by measuring mask fabrication time during CT simulation and setup time in the treatment room. Mask shrinkage was assessed by comparing shim settings during treatment to simulation and also by measuring the distance between anterior and posterior sections of the mask on CBCT and Sim CT. Mask comfort was evaluated qualitatively post‐treatment by surveying patients. Intrafraction motion was examined using 2D‐3D autofusion of orthogonal kV images at mid‐treatment. The intrafraction motion accuracy of the OSMS was analyzed by measuring the surface area of mask opening in Eclipse and pictures from the OSMS cameras.Results:The average preparation time for mask setup during simulation and treatment was 12.7 ± 2.3 min and 1.8 ± 0.9 min, respectively. Shims needed to be increased on average 0.8 mm, 0.8 mm, and 0.6 mm for right, left, and superior positions, respectively. CBCT measurements showed mask shrinkage, 0.8 mm both anteriorly and posteriorly. Greatest discomfort was reported on forehead followed by neck and chin. Intrafraction motion was less than 1 mm/1°. Accuracy of OSMS monitoring depended on the selected ROI. The OSMS monitored motion within 1 mm and 1° when the open surface area measured in Eclipse and captured by the camera > 147 cm2 and 33 cm2, respectively. When such conditions were not met, the accuracy somewhat degraded, an issue which can be mitigated by increasing the superior area or by angling the mask to expose greater surface to the camera.Conclusion:The new mask system rigidly immobilizes with sub‐mm accuracy during treatment according to optical camera (OSMS) measurements.Research supported in part by a grant from QFix (Avondale, PA) and Varian Medical System (Palo Alto, CA)

Purpose: To describe the development of a knowledge‐based treatment planning model for lung cancer patients treated with SBRT, and to evaluate the model performance and applicability to different planning techniques and tumor locations. Methods: 105 lung SBRT plans previously treated at our institution were included in the development of the model using Varian's RapidPlan DVH estimation algorithm. The model was trained with a combination of IMRT, VMAT, and 3D–CRT techniques. Tumor locations encompassed lesions located centrally vs peripherally (43:62), upper vs lower (62:43), and anterior vs posterior lobes (60:45). The model performance was validated with 25 cases independent of the training set, for both IMRT and VMAT. Model generated plans were created with only one optimization and no planner intervention. The original, general model was also divided into four separate models according to tumor location. The model was also applied using different beam templates to further improve workflow. Dose differences to targets and organs‐at‐risk were evaluated. Results: IMRT and VMAT RapidPlan generated plans were comparable to clinical plans with respect to target coverage and several OARs. Spinal cord dose was lowered in the model‐based plans by 1Gy compared to the clinical plans, p=0.008. Splitting the model according to tumor location resulted in insignificant differences in DVH estimation. The peripheral model decreased esophagus dose to the central lesions by 0.5Gy compared to the original model, p=0.025, and the posterior model increased dose to the spinal cord by 1Gy compared to the anterior model, p=0.001. All template beam plans met OAR criteria, with 1Gy increases noted in maximum heart dose for the 9‐field plans, p=0.04. Conclusion: A RapidPlan knowledge‐based model for lung SBRT produces comparable results to clinical plans, with increased consistency and greater efficiency. The model encompasses both IMRT and VMAT techniques, differing tumor locations, and beam arrangements. Research supported in part by a grant from Varian Medical Systems, Palo Alto CA.

Purpose: To investigate radiotherapy outcomes by incorporating 4DCT-based physiological and tumor elasticity functions for lung cancer patients. Methods: 4DCT images were acquired from 28 lung SBRT patients before radiation treatment. Deformable image registration (DIR) was performed from the end-inhale to the end-exhale using a B-Spline-based algorithm (Elastix, an open source software package). The resultant displacement vector fields (DVFs) were used to calculate a relative Jacobian function (RV) for each patient. The computed functions in the lung and tumor regions represent lung ventilation and tumor elasticity properties, respectively. The 28 patients were divided into two groups: 16 with two-year tumor local control (LC) and 12 with local failure (LF). The ventilation and elasticity related RV functions were calculated for each of these patients. Results: The LF patients have larger RV values than the LC patients. The mean RV value in the lung region was 1.15 (±0.67) for the LF patients, higher than 1.06 (±0.59) for the LC patients. In the tumor region, the elasticity-related RV values are 1.2 (±0.97) and 0.86 (±0.64) for the LF and LC patients, respectively. Among the 16 LC patients, 3 have the mean RV values greater than 1.0 in the tumors. These tumors were located near the diaphragm, where the displacements are relatively large.. RV functions calculated in the tumor were better correlated with treatment outcomes than those calculated in the lung. Conclusion: The ventilation and elasticity-related RV functions in the lung and tumor regions were calculated from 4DCT image and the resultant values showed differences between the LC and LF patients. Further investigation of the impact of the displacements on the computed RV is warranted. Results suggest that the RV images might be useful for evaluation of treatment outcome for lung cancer patients.

Purpose:Accurate deformable image registration (DIR) between CT and CBCT in H&N is challenging. In this study, we propose a practical hybrid method that uses not only the pixel intensities but also organ physical properties, structure volume of interest (VOI), and interactive local registrations.Methods:Five oropharyngeal cancer patients were selected retrospectively. For each patient, the planning CT was registered to the last fraction CBCT, where the anatomy difference was largest. A three step registration strategy was tested; Step1) DIR using pixel intensity only, Step2) DIR with additional use of structure VOI and rigidity penalty, and Step3) interactive local correction. For Step1, a public‐domain open‐source DIR algorithm was used (cubic B‐spline, mutual information, steepest gradient optimization, and 4‐level multi‐resolution). For Step2, rigidity penalty was applied on bony anatomies and brain, and a structure VOI was used to handle the body truncation such as the shoulder cut‐off on CBCT. Finally, in Step3, the registrations were reviewed on our in‐house developed software and the erroneous areas were corrected via a local registration using level‐set motion algorithm.Results:After Step1, there were considerable amount of registration errors in soft tissues and unrealistic stretching in the posterior to the neck and near the shoulder due to body truncation. The brain was also found deformed to a measurable extent near the superior border of CBCT. Such errors could be effectively removed by using a structure VOI and rigidity penalty. The rest of the local soft tissue error could be corrected using the interactive software tool. The estimated interactive correction time was approximately 5 minutes.Conclusion:The DIR using only the image pixel intensity was vulnerable to noise and body truncation. A corrective action was inevitable to achieve good quality of registrations. We found the proposed three‐step hybrid method efficient and practical for CT/CBCT registrations in H&N.My department receives grant support from Industrial partners: (a) Varian Medical Systems, Palo Alto, CA, and (b) Philips HealthCare, Best, Netherlands

Purpose:Using daily cone beam CTs (CBCTs) to develop principal component analysis (PCA) models of anatomical changes in head and neck (H&N) patients and to assess the possibility of using these prospectively in adaptive radiation therapy (ART).Methods:Planning CT (pCT) images of 4 H&N patients were deformed to model several different systematic changes in patient anatomy during the course of the radiation therapy (RT). A Pinnacle plugin was used to linearly interpolate the systematic change in patient for the 35 fraction RT course and to generate a set of 35 synthetic CBCTs. Each synthetic CBCT represents the systematic change in patient anatomy for each fraction. Deformation vector fields (DVFs) were acquired between the pCT and synthetic CBCTs with random fraction‐to‐fraction changes were superimposed on the DVFs. A patient‐specific PCA model was built using these DVFs containing systematic plus random changes. It was hypothesized that resulting eigenDVFs (EDVFs) with largest eigenvalues represent the major anatomical deformations during the course of treatment.Results:For all 4 patients, the PCA model provided different results depending on the type and size of systematic change in patient's body. PCA was more successful in capturing the systematic changes early in the treatment course when these were of a larger scale with respect to the random fraction‐to‐fraction changes in patient's anatomy. For smaller scale systematic changes, random changes in patient could completely “hide” the systematic change.Conclusion:The leading EDVF from the patientspecific PCA models could tentatively be identified as a major systematic change during treatment if the systematic change is large enough with respect to random fraction‐to‐fraction changes. Otherwise, leading EDVF could not represent systematic changes reliably. This work is expected to facilitate development of population‐based PCA models that can be used to prospectively identify significant anatomical changes early in treatment.This work is supported in part by a grant from Varian Medical Systems, Palo Alto, CA.

Purpose: To estimate the accumulated dose to targets and organs at risk (OAR) for head and neck (H' N) radiotherapy using 3 deformable image registration (DIR) algorithms. Methods: Five H' N patients, who had daily CBCTs taken during the course of treatment, were retrospectively studied. All plans had 5 mm CTV‐to‐PTV expansions. To overcome the small field of view (FOV) limitations and HU uncertainties of CBCTs, CT images were deformably registered using a parameter‐optimized B‐spline DIR algorithm (Elastix, elastix.isi.uu.nl) and resampled onto each CBCT with a 4 cm uniform FOV expansion. The dose of the day was calculated on these resampled CT images. Calculated daily dose matrices were warped and accumulated to the planning CT using 3 DIR algorithms; SmartAdapt (Eclipse/Varian), Velocity (Velocity Medical Solutions), and Elastix. Dosimetric indices for targets and OARs were determined from the DVHs and compared with corresponding planned quantities. Results: The cumulative dose deviation was less than 2%, on average, for PTVs from the corresponding plan dose, for all algorithms/patients. However, the parotids show as much as a 37% deviation from the intended dose, possibly due to significant patient weight loss during the first 3 weeks of treatment (15.3 lbs in this case). The mean(±SD) cumulative dose deviations of the 5 patients estimated using the 3 algorithms (SmartAdapt, Velocity, and Elastix) were (0.8±0.9%, 0.5±0.9%, 0.6±1.3%) for PTVs, (1.6±1.9%, 1.4±2.0%, 1.7±1.9%) for GTVs, (10.4±12.1%, 10.7±10.6%, 6.5±10.1%) for parotid glands, and (4.5±4.6%, 3.4±5.7%, 3.9±5.7%) for mucosa, respectively. The differences among the three DIR algorithms in the estimated cumulative mean doses (1SD (in Gy)) were: 0.1 for PTVs, 0.1 for GTVs, 1.9 for parotid glands, and 0.4 for mucosa. Conclusion: Results of this study are suggestive that more frequent plan adaptation for organs, such as the parotid glands, might be beneficial during the course of H' N RT. This study was supported in part by a research grant from Varian Medical Systems, Palo Alto, CA

Purpose: To use receiver operating characteristic (ROC) analysis to quantify the Varian Portal Dosimetry (VPD) application's ability to detect delivery errors in IMRT fields. Methods: EPID and VPD were calibrated/commissioned using vendor-recommended procedures. Five clinical plans comprising 56 modulated fields were analyzed using VPD. Treatment sites were: pelvis, prostate, brain, orbit, and base of tongue. Delivery was on a Varian Trilogy linear accelerator at 6MV using a Millenium120 multi-leaf collimator. Image pairs (VPD-predicted and measured) were exported in dicom format. Each detection test imported an image pair into Matlab, optionally inserted a simulated error (rectangular region with intensity raised or lowered) into the measured image, performed 3%/3mm gamma analysis, and saved the gamma distribution. For a given error, 56 negative tests (without error) were performed, one per 56 image pairs. Also, 560 positive tests (with error) with randomly selected image pairs and randomly selected in-field error location. Images were classified as errored (or error-free) if percent pixels with γ<κ was < (or ≥) τ. (Conventionally, κ=1 and τ=90%.) A ROC curve was generated from the 616 tests by varying τ. For a range of κ and τ, true/false positive/negative rates were calculated. This procedure was repeated for inserted errors of different sizes. VPD was considered to reliably detect an error if images were correctly classified as errored or error-free at least 95% of the time, for some κ+τ combination. Results: 20mm2 errors with intensity altered by ≥20% could be reliably detected, as could 10mm2 errors with intensity was altered by ≥50%. Errors with smaller size or intensity change could not be reliably detected. Conclusion: Varian Portal Dosimetry using 3%/3mm gamma analysis is capable of reliably detecting only those fluence errors that exceed the stated sizes. Images containing smaller errors can pass mathematical analysis, though may be detected by visual inspection. This work was not funded by Varian Oncology Systems. Some authors have other work partly funded by Varian Oncology Systems.

Purpose: To to determine if tumor control probability (TCP) and normal tissue control probability (NTCP) values computed on the treatment planning image are representative of TCP/NTCP distributions resulting from probable positioning variations encountered during external-beam radiotherapy. Methods: We compare TCP/NTCP as typically computed on the planning PTV/OARs with distributions of those parameters computed for CTV/OARs via treatment delivery simulations which include the effect of patient organ deformations for a group of 19 prostate IMRT pseudocases. Planning objectives specified 78 Gy to PTV1=prostate CTV+5 mm margin, 66 Gy to PTV2=seminal vesicles+8 mm margin, and multiple bladder/rectum OAR objectives to achieve typical clinical OAR sparing. TCP were computed using the Poisson Model while NTCPs used the Lyman-Kutcher-Bruman model. For each patient, 1000 30-fraction virtual treatment courses were simulated with each fractional pseudo- time-oftreatment anatomy sampled from a principle component analysis patient deformation model. Dose for each virtual treatment-course was determined via deformable summation of dose from the individual fractions. CTVTCP/ OAR-NTCP values were computed for each treatment course, statistically analyzed, and compared with the planning PTV-TCP/OARNTCP values. Results: Mean TCP from the simulations differed by <1% from planned TCP for 18/19 patients; 1/19 differed by 1.7%. Mean bladder NTCP differed from the planned NTCP by >5% for 12/19 patients and >10% for 4/19 patients. Similarly, mean rectum NTCP differed by >5% for 12/19 patients, >10% for 4/19 patients. Both mean bladder and mean rectum NTCP differed by >5% for 10/19 patients and by >10% for 2/19 patients. For several patients, planned NTCP was less than the minimum or more than the maximum from the treatment course simulations. Conclusion: Treatment course simulations yield TCP values that are similar to planned values, while OAR NTCPs differ significantly, indicating the need for probabilistic methods or PRVs for OAR risk assessment. Presenting author receives support from Philips Medical Systems.

Purpose: To evaluate the quality of a commercially available MRI‐CT image registration algorithm and then develop a method to improve the performance of this algorithm for MRI‐guided prostate radiotherapy. Methods: Prostate contours were delineated on ten pairs of MRI and CT images using Eclipse. Each pair of MRI and CT images was registered with an intensity‐based B‐spline algorithm implemented in Velocity. A rectangular prism that contains the prostate volume was partitioned into a tetrahedral mesh which was aligned to the CT image. A finite element method (FEM) was developed on the mesh with the boundary constraints assigned from the Velocity generated displacement vector field (DVF). The resultant FEM displacements were used to adjust the Velocity DVF within the prism. Point correspondences between the CT and MR images identified within the prism could be used as additional boundary constraints to enforce the model deformation. The FEM deformation field is smooth in the interior of the prism, and equal to the Velocity displacements at the boundary of the prism. To evaluate the Velocity and FEM registration results, three criteria were used: prostate volume conservation and center consistence under contour mapping, and unbalanced energy of their deformation maps. Results: With the DVFs generated by the Velocity and FEM simulations, the prostate contours were warped from MRI to CT images. With the Velocity DVFs, the prostate volumes changed 10.2% on average, in contrast to 1.8% induced by the FEM DVFs. The average of the center deviations was 0.36 and 0.27 cm, and the unbalance energy was 2.65 and 0.38 mJ/cc3 for the Velocity and FEM registrations, respectively. Conclusion: The adaptive FEM method developed can be used to reduce the error of the MIbased registration algorithm implemented in Velocity in the prostate region, and consequently may help improve the quality of MRI‐guided radiation therapy.

Stereotactic Body Radiation Therapy (SBRT) treatment of patients presenting with multiple primary or oligometastic lung lesions can be complicated with the presence of two or more synchronous lesions. From the dosimetry standpoint there is the challenge of meeting target dose objectives and critical organ constraints when these peripheral lesions are in close proximity to each other, due to interlesion entry and exit dose effects. In terms of delivery, efficient clinical workflow is important. For instance, methods to deliver treatment accurately without image-based immobilization of each lesion in a multiple-isocenter setup must be considered. The objective of this work is to demonstrate improvements in dosimetry and treatment efficiency with plans using a single isocenter for SBRT of multiple lung lesions. Dosimetric criteria in this study were designed to meet the RTOG 0915 (48Gy in 4 fractions) protocol. Treatment planning was done for 6 patients (3 left and 3 right-sided) retrospectively, with two lesions each using the Eclipse AAA algorithm (Varian Medical Systems, Palo Alto CA) and 6X 1000 MU/Min photons. ITVs were obtained from 4DCT scans and PTVs were formed from a uniform, 5 mm expansion of the ITVs. A single isocenter was determined with lesions required to be well within a 15 cm jaw setting to circumvent machine limitations and to minimize low dose spread into normal lung (from scatter and intraleaf leakage). Planning was done with 3D techniques such that 95% of the individual PTVs received the prescribed dose. MLC fields sharing a common gantry angle were merged as one field. Iteration of the plans was made based on DVH criteria. Characteristics of the 6 patients were: 3 male, 3 female with a median age of 74.2 years (range, 62-83 years). The combined PTVs ranged from 8.7 to 124 cc (median 54 cc). Using RTOG coverage objectives as a scoring criteria, conformality index criteria were met (average = 1.22 +/- 0.045). R50% criteria were not met; 2 cases did not meet the D2cm criteria. V20 (%) lung criteria (8.4 +/- 3.7, range = 2.6%,13.4%) were met with two minor deviations. In general, the plans met the coverage and constraint guidelines by limiting high and low dose spill to regions outside the combined PTV volume and thereby critical structures. This study shows that treating two lesions using a single isocenter can be clinically efficient. Further evaluation is needed to validate that a single isocenter IGRT can accurately localize the treatment targets for safe treatment delivery.Scientific Abstract 3778; TableConformality of Dose Coverage and Constraint Objectives to the Combined PTVCombined PTV (cc)Conformality IndexD99% PTV (Gy)R50%Max Dose Outside PTV+2cm (Gy)Lung V20 Gy (%)Lung V5 Gy (%)8.71.447.416.732.25.837.3571.246.4740.110.121.646.91.246.67.532.86.51315.81.22.65.7242.626.271.31.211.7630.911.725.71241.113.46.542.413.429.7 Open table in a new tab

Purpose: To evaluate output factors and dose calculation accuracy on a novel SRS linear accelerator, the Edge (Varian), for treatments of small, elongated targets using flattening filter free (FFF) beam. Methods: Total scatter/output factors (OF’s) for 24 elongated, small, high definition multi-leaf collimator (HDMLC)-defined fields were measured on the Edge machine using 6X FFF beam. 3 detectors were used in water tank: CC01 ion chamber (active volume 10cc), stereotactic photon diode (SFD) (active diameter 0.6mm, active thickness 0.06mm), Edge detector (active volume 0.0019cc). The 24 MLC apertures have widths ranging from 5 to 20mm and length/width ratio from 0.25 to 5. Readings were cross calibrated with CC04 at field size 3×3 cm. A beam model was developed using commissioning measurements for treatment planning in Eclipse (AAA, version 11). One representative patient case (IMRT, target volume 0.2cc, 4×4×14mm) was calculated using AAA 11 and delivered on the Edge. Results: Due to volume averaging effects, CC01 readings were 11.2±0.9% lower than SFD readings for 5mm field sizes. The Edge diode showed a uniform over-response of 2.6±0.7% compared to SFD. Calculation using AAA v11 showed the best agreement with SFD measurements (2.4±1.7% lower than SFD). The largest difference between AAA v11 and SFD occurs at 5mm field sizes. For the patient plan, dose delivered on Edge was measured to be 2.2% higher than AAA v11 calculation. Conclusion: Cross-calibrated SFD output measurements presented the best agreement with commissioned AAA v11 beam model. Field sizes smaller than 1cm posed challenges to both the detectors and the calculation algorithm. For the representative patient with small elongated target, AAA v11 and measurements agreed within ~2% on the Edge linac. Although encouraging, a more comprehensive study is required to validate the overall algorithmic accuracy.

Purpose: Dose-response models that can reliably predict radiation pneumonitis (RP) to guide radiation therapy (RT) for lung cancer presently do not exist. A model is proposed that incorporates non-local radiationinduced bystander effect (RIBE). Methods: A single sigmoid response function, derived from published data for whole lung irradiation, relates RP probability to cumulative lung damage, regardless of fractionation scheme. Lung damage is assumed to be caused by direct local radiation damage, quantified via the linear-quadratic (LQ) model, and RIBE. Based on published data, RIBE is assumed to be activated when per-fraction dose rises above ∼0.6 Gy, but is constant with dose above that threshold. Integral RIBE damage is assumed proportional to lung volume irradiated above ∼0.6 Gy per fraction. Key model parameters include LQ α and β, and two RIBE parameters: the single-fraction probability δ of damage, and a proportionality parameter κ that relates the potential for RIBE damage to irradiated lung volume. All parameters are tentatively fitted from published data, the RIBE parameters from published RP rates for conventionally fractionated RT (CFRT) and stereotactic body RT (SBRT). Results: The model predicts dose-response curves that are consistent with clinical experience. It provides a tentative explanation for why V20 (33 fractions), V13 (20 fractions) and V5 (<10 fractions) are observed to be correlated with RP. It also provides a plausible explanation for the success of SBRT — RIBE damage increases with the number of fractions, so penalizes CFRT relative to SBRT. Conclusion: The proposed model is relatively simple, extrapolates from published data, plausibly explains several clinical observations, and produces dose-response curves that are consistent with clinical experience. While capable of elaboration, its ability to explain doseresponse experience with different fractionation schemes using a small number of assumptions and parameters is an advantage.

PURPOSEThe presence of Calypso Beacon-transponders in patients can cause artifacts during MRI imaging studies. This could be a problem for post-treatment follow up of cancer patients using MRI studies to evaluate metastasis and for functional imaging studies. This work assesses (1) the volume immediately surrounding the transponders that will not be visualized by the MRI due to the beacons, and (2) the dependence of the non-visualized volume on beacon orientation, and scanning techniques.METHODSTwo phantoms were used in this study (1) water filled box, (2) and a 2300 cc block of pork meat. Calypso beacons were implanted in the phantoms both in parallel and perpendicular orientations with respect to the MR scanner magnetic field. MR image series of the phantom were obtained with on a 1.0T high field open MR-SIM with multiple pulse sequences, for example, T1-weighted fast field echo and T2-weighted turbo spin echo.RESULTSOn average, a no-signal region with 2 cm radius and 3 cm length was measured. Image artifacts are more significant when beacons are placed parallel to scanner magnetic field; the no-signal area around the beacon was about 0.5 cm larger in orthogonal orientation. The no-signal region surrounding the beacons slightly varies in dimension for the different pulse sequences.CONCLUSIONThe use of Calypso beacons can prohibit the use of MRI studies in post-treatment assessments, especially in the immediate region surrounding the implanted beacon. A characterization of the MR scanner by identifying the no-signal regions due to implanted beacons is essential. This may render the use of Calypso beacons useful for some cases and give the treating physician a chance to identify those patients prior to beacon implantation.

Purpose/Objective(s)In anticipation of a nationwide incident learning system for Radiation Oncology we have continued our work developing an in-house incident learning system. Using this system, users submit anonymous or named reports. We applied failure mode and effects analysis (FMEA) as described in AAPM report TG-100 to 1033 incident reports generated across 5 clinics from 2001 to 2013.Materials/MethodsFMEA applied to the failure modes generates a Risk Probability Number (RPN) based on the severity, probability of occurrence and probability the event will go undetected. Incidents in our system are sorted into two categories. The first is for those that occurred due to failure to follow established policies and procedures and do not have a dose consequence (e.g., incorrect field naming). The second class is for incidents that could potentially have a dose consequence (e.g., couch not inserted during VMAT treatment planning). Additional data is collected on where the incident occurred, who reported it and what discovery tools were used.ResultsOf the 1033 incidents reported, 55% were reported by physicists, 19% anonymously, 19% by therapists and 7% by dosimetrists. Six hundred twenty-eight were related to clinical workflow and 461 had potential dose consequences. None of the incidents resulted in a reportable medical event. Of the reports with potential dose consequences 13 had RPNs greater than 200. These fell into two categories; incidents related to patients having an implanted electronic medical device and shifts at the treatment unit. Based on our FMEA results, we employed Fault Tree Analysis (FTA) to perform a systematic review of our implanted medical device and treatment policies. This resulted in modifications to existing policy, automated communication with the hospital's electrophysiology clinic and training which enabled us to reduce the RPN from 230 to 60 for reports related to implanted medical devices. Modification to existing policy including a record of shifts for all SRS and SBRT patients and staff training enabled us to reduce the RPN from 220 to 40 for reports related to shifts.ConclusionsAs nationwide incident learning systems become more imminent it is important that each clinic take a critical look at their processes, policies and procedures. Developing one's own incident learning system and applying FMEA and FTA methods are two ways to improve quality and safety in one's clinic. By having reported events reviewed by our Quality Assurance Committee (QAC) on a monthly basis, we are able to triage the reports, enabling us to address the most critical issues affecting patient care. RPN's for these most critical incidents were reduced by automation, updated checklists, updated policies and procedures and staff education. Purpose/Objective(s)In anticipation of a nationwide incident learning system for Radiation Oncology we have continued our work developing an in-house incident learning system. Using this system, users submit anonymous or named reports. We applied failure mode and effects analysis (FMEA) as described in AAPM report TG-100 to 1033 incident reports generated across 5 clinics from 2001 to 2013. In anticipation of a nationwide incident learning system for Radiation Oncology we have continued our work developing an in-house incident learning system. Using this system, users submit anonymous or named reports. We applied failure mode and effects analysis (FMEA) as described in AAPM report TG-100 to 1033 incident reports generated across 5 clinics from 2001 to 2013. Materials/MethodsFMEA applied to the failure modes generates a Risk Probability Number (RPN) based on the severity, probability of occurrence and probability the event will go undetected. Incidents in our system are sorted into two categories. The first is for those that occurred due to failure to follow established policies and procedures and do not have a dose consequence (e.g., incorrect field naming). The second class is for incidents that could potentially have a dose consequence (e.g., couch not inserted during VMAT treatment planning). Additional data is collected on where the incident occurred, who reported it and what discovery tools were used. FMEA applied to the failure modes generates a Risk Probability Number (RPN) based on the severity, probability of occurrence and probability the event will go undetected. Incidents in our system are sorted into two categories. The first is for those that occurred due to failure to follow established policies and procedures and do not have a dose consequence (e.g., incorrect field naming). The second class is for incidents that could potentially have a dose consequence (e.g., couch not inserted during VMAT treatment planning). Additional data is collected on where the incident occurred, who reported it and what discovery tools were used. ResultsOf the 1033 incidents reported, 55% were reported by physicists, 19% anonymously, 19% by therapists and 7% by dosimetrists. Six hundred twenty-eight were related to clinical workflow and 461 had potential dose consequences. None of the incidents resulted in a reportable medical event. Of the reports with potential dose consequences 13 had RPNs greater than 200. These fell into two categories; incidents related to patients having an implanted electronic medical device and shifts at the treatment unit. Based on our FMEA results, we employed Fault Tree Analysis (FTA) to perform a systematic review of our implanted medical device and treatment policies. This resulted in modifications to existing policy, automated communication with the hospital's electrophysiology clinic and training which enabled us to reduce the RPN from 230 to 60 for reports related to implanted medical devices. Modification to existing policy including a record of shifts for all SRS and SBRT patients and staff training enabled us to reduce the RPN from 220 to 40 for reports related to shifts. Of the 1033 incidents reported, 55% were reported by physicists, 19% anonymously, 19% by therapists and 7% by dosimetrists. Six hundred twenty-eight were related to clinical workflow and 461 had potential dose consequences. None of the incidents resulted in a reportable medical event. Of the reports with potential dose consequences 13 had RPNs greater than 200. These fell into two categories; incidents related to patients having an implanted electronic medical device and shifts at the treatment unit. Based on our FMEA results, we employed Fault Tree Analysis (FTA) to perform a systematic review of our implanted medical device and treatment policies. This resulted in modifications to existing policy, automated communication with the hospital's electrophysiology clinic and training which enabled us to reduce the RPN from 230 to 60 for reports related to implanted medical devices. Modification to existing policy including a record of shifts for all SRS and SBRT patients and staff training enabled us to reduce the RPN from 220 to 40 for reports related to shifts. ConclusionsAs nationwide incident learning systems become more imminent it is important that each clinic take a critical look at their processes, policies and procedures. Developing one's own incident learning system and applying FMEA and FTA methods are two ways to improve quality and safety in one's clinic. By having reported events reviewed by our Quality Assurance Committee (QAC) on a monthly basis, we are able to triage the reports, enabling us to address the most critical issues affecting patient care. RPN's for these most critical incidents were reduced by automation, updated checklists, updated policies and procedures and staff education. As nationwide incident learning systems become more imminent it is important that each clinic take a critical look at their processes, policies and procedures. Developing one's own incident learning system and applying FMEA and FTA methods are two ways to improve quality and safety in one's clinic. By having reported events reviewed by our Quality Assurance Committee (QAC) on a monthly basis, we are able to triage the reports, enabling us to address the most critical issues affecting patient care. RPN's for these most critical incidents were reduced by automation, updated checklists, updated policies and procedures and staff education.

Purpose:One primary limitation of using CBCT images for H' N adaptive radiotherapy (ART) is the limited field of view (FOV) range. We propose a method to extrapolate the CBCT by using a deformed planning CT for the dose of the day calculations. The aim was to estimate the geometric uncertainty of our extrapolation method.Methods:Ten H' N patients, each with a planning CT (CT1) and a subsequent CT (CT2) taken, were selected. Furthermore, a small FOV CBCT (CT2short) was synthetically created by cropping CT2 to the size of a CBCT image. Then, an extrapolated CBCT (CBCTextrp) was generated by deformably registering CT1 to CT2short and resampling with a wider FOV (42mm more from the CT2short borders), where CT1 is deformed through translation, rigid, affine, and b‐spline transformations in order. The geometric error is measured as the distance map ||DVF|| produced by a deformable registration between CBCTextrp and CT2. Mean errors were calculated as a function of the distance away from the CBCT borders. The quality of all the registrations was visually verified.Results:Results were collected based on the average numbers from 10 patients. The extrapolation error increased linearly as a function of the distance (at a rate of 0.7mm per 1 cm) away from the CBCT borders in the S/I direction. The errors (μ±σ) at the superior and inferior boarders were 0.8 ± 0.5mm and 3.0 ± 1.5mm respectively, and increased to 2.7 ± 2.2mm and 5.9 ± 1.9mm at 4.2cm away. The mean error within CBCT borders was 1.16 ± 0.54mm . The overall errors within 4.2cm error expansion were 2.0 ± 1.2mm (sup) and 4.5 ± 1.6mm (inf).Conclusion:The overall error in inf direction is larger due to more large unpredictable deformations in the chest. The error introduced by extrapolation is plan dependent. The mean error in the expanded region can be large, and must be considered during implementation.This work is supported in part by Varian Medical Systems, Palo Alto, CA.

Purpose:To evaluate the setup accuracies of image‐guided intracranial radiosurgery across several different linear accelerator platforms.Methods:A CT scan with a slice thickness of 1.0 mm was acquired of a Rando head phantom (The Phantom Laboratory) in a U‐frame mask (BrainLAB AG). The phantom had three embedded BBs, simulating a central, left, and anterior lesion. The phantom was setup with each BB placed at the radiation isocenter under image guidance. Four different setup procedures were investigated: (1) NTX_ExacTrac: 6 degree‐of‐freedom (6D) correction on a Novalis Tx (BrainLAB AG) with ExacTrac localization (BrainLAB AG); (2) NTX_CBCT: 4D correction on the Novalis Tx with cone‐beam computed tomography (CBCT); (3) TrueBeam_CBCT: 4D correction on a TrueBeam (Varian) with CBCT; (4) Edge_CBCT: 6D correction on an Edge (Varian) with CBCT. The experiment was repeated 5 times with different initial setup error at each BB location on each platform, and the mean (μ) and one standard deviation (σ) of the residual error was compared. The congruence between radiation and imaging isocenters on each platform was evaluated by acquiring Winston Lutz (WL) images of a WL jig followed by imaging using ExacTrac or CBCT. The difference in coordinates of the jig relative to radiation and imaging isocenters was then recorded.Results:Averaged over all three BB locations, the residual vector setup errors (μ±σ) of the phantom in mm were 0.6±0.2, 1.0±0.5, 0.2±0.1, and 0.3±0.1 on NTX_ExacTrac, NTX_CBCT, TrueBeam_CBCT, and Edge_CBCT, with their ranges in mm being 0.4∼1.1, 0.4∼1.9, 0.1∼0.5, and 0.2∼0.6, respectively. And imaging isocenter was found stable relative to radiation isocenter, with the congruence to radiation isocenter in mm being 0.6±0.1, 0.7±0.1, 0.3±0.1, 0.2±0.1, respectively, on the four systems in the same order.Conclusion:Millimeter accuracy can be achieved with image‐guided radiosurgery for intracranial lesions based on this set of experiments.

Purpose:Accurate calibration of radiobiological parameters is crucial to predicting radiation treatment response. Modeling differences may have a significant impact on calibrated parameters. In this study, we have integrated two existing models with kinetic differential equations to formulate a new tumor regression model for calibrating radiobiological parameters for individual patients.Methods:A system of differential equations that characterizes the birth‐and‐death process of tumor cells in radiation treatment was analytically solved. The solution of this system was used to construct an iterative model (Z‐model). The model consists of three parameters: tumor doubling time Td, half‐life of dying cells Tr and cell survival fraction SFD under dose D. The Jacobian determinant of this model was proposed as a constraint to optimize the three parameters for six head and neck cancer patients. The derived parameters were compared with those generated from the two existing models, Chvetsov model (C‐model) and Lim model (L‐model). The C‐model and L‐model were optimized with the parameter Td fixed.Results:With the Jacobian‐constrained Z‐model, the mean of the optimized cell survival fractions is 0.43±0.08, and the half‐life of dying cells averaged over the six patients is 17.5±3.2 days. The parameters Tr and SFD optimized with the Z‐model differ by 1.2% and 20.3% from those optimized with the Td‐fixed C‐model, and by 32.1% and 112.3% from those optimized with the Td‐fixed L‐model, respectively.Conclusion:The Z‐model was analytically constructed from the cellpopulation differential equations to describe changes in the number of different tumor cells during the course of fractionated radiation treatment. The Jacobian constraints were proposed to optimize the three radiobiological parameters. The developed modeling and optimization methods may help develop high‐quality treatment regimens for individual patients.