Z Xu

SU-FF-T-362: PLanUNC as An Open-Source Radiotherapy Planning System for Research and Education

Purpose: PLanUNC is a radiotherapy planning software package that has been under development and clinical use at the University of North Carolina for approximately 20 years. Under a joint grant from the NCRR and NCI (R01 RR 018615), PLanUNC has been documented, commented, and prepared for distribution as a freely available open‐source treatment planning tool for use as an adaptable and common platform for radiotherapy research. Method and Materials: The software and source code have been made available to qualifying users through a web portal located at http://planunc.radonc.unc.edu. Licenses for PLanUNC are available without fee to institutions, departments, and other facilities engaged in research and education involving radiation therapy.Results: Recent research milestones demonstrating the extensibility and increasing utility of PLanUNC include tools for 4D planning, interfaces with ITK segmentation and registration tools, daily correction of patient positioning, and interfaces with a variety of Monte Carlo dose engines. PLanUNC is currently supported for Linux and Windows operating systems, but has been successfully compiled on Alpha, Macintosh, Solaris, and other platforms. Conclusion: Licensed users will have access to PLanUNC source code, user and development documentation, annual training workshops, and limited support from UNC and the PLanUNC research community. PLanUNC is distributed as source code, making it customizable and extensible to meet the needs of a diverse range of research applications.

3D Analysis of Intensity-Modulated Radiation Therapy Quality Assurance Measurement using a 2D Diode Array

Intensity-modulated radiation therapy (IMRT) quality assurance (QA) is often performed using a 2D device and compares measured and computed fluence maps to determine if a field passes or fails certain dose and position criteria. The effects of a measured deviation to the 3D patient spatial dosimetry and dose volume histogram (DVH) are largely unknown because they cannot be analyzed using commercial 2D array IMRT QA systems. We report an in-house treatment planning system (TPS) PLanUNC based 3D IMRT QA analysis approach that has been used in our institution for the past ten years when 2D fluence map IMRT QA failed. In this approach the measured 2D fluence maps are imported back to PLanUNC and used to re-compute 3D patient dosimetry including DVHs. The 2D fluence map IMRT QA criteria is that the measured dose for 95% of the detectors is within 5% of the planned dose, and that the distance-to-agreement be within 4mm (5%/4mm). 22 IMRT plans that had at least one field fail initial QA using MapCHECK 2 are examined using our 3D QA approach. The DVH analysis shows that 19/22 plans that failed initial QA were within 2% of the planned target and critical structure DVHs. 3/22 IMRT plans were found to have DVH difference greater than 2%. The 3D analysis of 2D IMRT QA result shows that when a fluence map QA fails for a single field, often it is clinically insignificant in terms of patient 3D dosimetry

SU‐E‐T‐347: Evaluation of DQA Results Using a Super‐Sampling Dose Calculation in Helical Tomotherapy

Purpose: The aim of this work is to evaluate the impact of a new supersampling dose calculation method on delivery quality assurance (DQA) results for helical tomotherapy patient plans. Methods: Accurays Tomotherapy treatment planning system performs its dose calculation by approximating the continuous beam of a full gantry rotation into 51 discrete beam projections, with one dose calculation per projection (TomoHD version 1.0). In a recent software release, TomoHD version 1.1, Accuray enhanced this technique by employing three dose calculation samples per projection. This ‘super‐sampling’ methodology is meant to improve agreement between measured and calculated dose. For this study, we compare the results of the 24 patient DQA plans calculated in our clinic with the newer version of dose calculation with the previous 24 patient plans which were calculated with the older method. The plans were delivered to a SunNuclear ArcCHECK cylindrical detector array, and data were compared using a γ evaluation, with criteria of 3%/3mm. To quantify the results, the percentage of points with γ 95%. Results: 21 of 24 DQA plans (87%) calculated with the older TomoHD 1.0 algorithm passed our (Pγ 95% criteria, while all 24 DQA Plans (100%) generated with the TomoHD 1.1 super‐sampling dose calculation passed. The average values for (Pγ<l) was 97.9% and 98.9% for the original and super‐sampling calculation, respectively. The standard deviation for the older software was 2.1, versus 1.5 for the newer super‐sampling method. Conclusions: The increased number of samples per projection angle employed in the new TomoHD Version 1.1 software leads to a reduction in the dose discrepancies seen in patient DQA plan results. This can improve the agreement between the calculated dose and delivered dose to patients.

SU‐E‐T‐859: A New Approach to Improve IMRT Delivery Efficiency by Intensity Map Discretization

Purpose: To minimize IMRT treatment monitor units and delivery time without compromising dosimetric quality in an in‐house treatment planning system. Methods: In‐house treatment planning system PLanUNC employs a two‐step process to generate a MLC‐IMRT treatment plan. First, the index‐dose gradient minimization dose optimization process produces a continuous intensity map. Second, MLC segmentation process converts the intensity map into MLC segments. This two‐step process is sensitive to MLC segmentation technique and can lead to excessive number of segments and dosimetric degradation. We developed a new method that discretizes the intensity map during optimization. The new method considered each beam in turn: its intensity map was optimized using the current dose distribution; the optimized intensity map was discretized according to its statistical distribution; the dose distribution was updated after discretization. In this process, the peaks and valleys in the continuous intensity map were converted into discrete plateaus, and the optimization of remaining beams mitigated possible degradation of dose conformity introduced by the discretization of previous beams. MLC segmentation was applied to the discretized fluence maps. This new approach is being validated using prostate and head and neck cases. Results: For the prostate, with no compromise to dose conformity, the number of segments was reduced by 40%, and the MUs reduced by ∼10%. For the head and neck, the number of segments was reduced by 20%, and a small dose escalation was observed in adjacent normal tissues; this may be due to complex constraints used in the optimization process and could be accommodated by re‐optimizing the beam weight using fixed beam apertures in the future. Conclusions: Discretization based on statistical distribution of beamlet intensity could effectively reduce IMRT beam complexity and result in shorter treatment time.

MO‐F‐BRC‐04: Is the External Point/marker a Good Surrogate for the Breast Surface Motion during the Left Breast DIBH Treatment?

Purpose: The ideal way of treating left breast DIBH is through direct breast surface matching. Many DIBH techniques, however, rely on a surrogate—external point/marker motion. We try to quantify the predictability of the breast surface motion using external point/marker. Methods: AlignRT Beam Hold system is applied to perform real‐time surface matching and the external point/marker tracking simultaneously. The skin rendering of breath hold CT scan is served as the reference. During the treatment, the patient surface is monitored and registered to the reference to calculate the corresponding distance (S(t)). Radiation beam is turned on when this distance is within a preselected threshold. The external point/marker tracking is implemented by tracking the vertical amplitude of a point in the center part of the left breast skin. The real‐time distance (P(t)) of the selected point to the corresponding reference point is calculated. A model is built to predict S(t) using P(t). Statistical and computational complexityanalyses are conducted. Results: 5 patients are included in this study. Based on our statistical analysis, S(t) can be modeled as a proportioned P(f). For each patient, the ratio is calculated for the first treatment day and applied on the subsequent days. The difference between the prediction and the true S(t) is calculated. The average standard deviations of the difference over all the treatment days are 1.69–2.85mm for different patient, which corresponds to 3.8%–14.6% error rate for 3mm threshold. The biggest standard deviation is from the patient with the largest breast. The computational complexity of the fast ICP based surface matching algorithm is O(N), and of the point/marker tracking is O(1). Conclusions: The breast surface motion and external point/marker tracking result do have strong correlation. Surface matching is more accurate at the price of higher computational complexity.

SU-E-J-54: Evaluation of a Toolkit for Automatic Deformable Registration and Segmentation of Treatment Images in Clinical Prostate Cancer IGRT Applications

Purpose: High quality IGRT is only practical when robust tools are available for accurate and fast deformable registration and segmentation of the treatmentimages. We have implemented such a toolkit for the male pelvis using a class of statistically trainable deformable models (SDMs) of anatomical structures, called medial representations (m‐reps). The aim of this study is to evaluate its capabilities and performance in clinical prostate cancerIGRT applications. Methods: An image collection for prostate cancer patients treated with IGRT using a CT‐on‐Rails (CTORs) system was studied retrospectively with this toolkit to determine the actual delivered dose throughout the treatment course. The patient‐specific planning models of the male pelvic organs, including the prostate, bladder, rectum, and femoral heads, were constructed via user‐guided autosegmentation of the planning CT. The planning CT was then automatically registered to the CTORs images via a rigid‐body, soft‐tissue‐based registration method. The planning models were then transformed into initialized treatmentmodels to segment the CTORs images. The essential tissue‐voxel correspondence across treatmentimages was established by this model‐based segmentation process. The dose delivered to the same tissue elements can, therefore, be accumulated for adaptive planning and/or outcome assessment. Results: Preliminary results show that the m‐rep‐based planning image segmentations are clinically acceptable to the physicians and can be readily constructed from the planning CT with nominal user guidance. The automatic image registration with the treatment‐dayimage, occasionally followed by manual refinement, also considered acceptable by clinical staff. Treatmentimage segmentation with intrinsic tissue‐voxel correspondence used substantially less time (5 ± 1.3 minutes) in comparison to manual contouring. Conclusions: Our clinic‐oriented toolkit is effective in segmentation of treatmentimages of the male pelvis. Application for dose accumulation and adaptive IGRT is in progress. This work has been conducted in collaboration with Morphormics, Inc. with grant support from NCI R44 CA119571.

SU‐E‐T‐771: Isodose Line Driven Semi‐Inverse Planning of High Dose Rate Brachytherapy for Cervical Cancer

Purpose: Delineation of tumor is indispensible for adequate tumor coverage in inverse planning of cervical cancerbrachytherapy. However, target definition is challenging in CT/CBCT planning images. In this project, we developed a tool to convert the isodose lines from traditional source loading, which produces dose distribution with good tumor coverage, to a surrogate treatment volume needed in the optimization. Through this, we integrate the clinical knowledge of conventional loading and advantage of inverse planning to spare organs at risks. Method and Materials: Five cervical cancer patients treated with tandem and ovoids HDR brachytherapy are studied. The clinical plans are point‐based (600cGy to point A) with Fletcher‐type loading pattern. Retrospectively, an inverse plan was made for comparisons. A software tool was developed to convert the isodose curves of the conventional plan to closed anatomic structures. In order to limit the dose of bladder and rectum less than 70% of the prescription dose, their contours were subtracted from the 70% isodose line converted‐volume and this new structure (ISD70‐bladder‐rectum ) was placed as the target in Oncentra optimization software (Nucletron). Three dosimetric endpoints, volume coverage of ISD70‐bladder‐rectum and 2cc maximal dose of the bladder and rectum, are used in compassions. Results: The isodose line converted‐ structures are compared with the original dose curves and they are accurate. The inverse planning lowers the dose on the bladder with similar coverage of ISD70‐bladder‐rectum as the conventional plan. The bladder 2cc dose of the inverse plan is 510.4± 92.7 cGy, which is significantly lower than that of conventional plan, 560.3± 93.8 cGy (p=0.05). Conclusion: Isodose surfaces from traditional source loading are good surrogate for 3D treatment volume in HDR inverse planning for cervical cancer. This approach combines the prior clinical experience and strength of inverse planning for better critical structure sparing with the similar tumor coverage.

WE‐C‐BRA‐04: Left Breast Deep Inspiration Breath Hold Treatment Based on 3D Surface Matching

Deep inspiration breath hold (DIBH) is a cardiac sparing technique to treat left‐sided breast cancer. By supervising a patient holding her breath at the deep inspiration level, we can deliver tangent radiation beams to effectively treattumors while reducing the dose to the heart. Research has confirmed both the effectiveness of this treatment, and the reduction in cardiac toxicity. Most clinics in the US are treating DIBH through external surrogate. However, for breast irradiation, the ideal way to implement DIBH would be through direct breast surface matching. This lecture will use AlignRT Beam Hold system (VisionRT Ltd., London, UK) as an example to provide an overview of breast surface matching based DIBH. Clinical implementation will be addressed. Each DIBH patient will have a breath hold CT scan. Patient will also have conventional tattoos. The skin rendering of the breath hold CT is generated and transferred to the AlignRT Beam Hold system. Region‐Of‐Interest (ROI) is selected, and the ROI is the reference image to be matched on during the treatment. There are two criteria for the ROI selection. First, the ROI needs to cover the treatment area. Second, it is important to include some computer vision distinguishable features to assure the accuracy of the surface matching result. On the treatment day, the patient is first setup based on the conventional tattoos. Then, she will be guided to take a deep breath and hold. The AlignRT Beam Hold system will obtain real‐time patient surfaceimages and register it to the reference image. The registration result includes the six degree (translational vertical, longitudinal, and lateral, and rotational vertical, longitudinal, and lateral) displacements and the magnitude displacement. The magnitude displacement is the Euclidian norm of the three translational displacements. All seven displacements are called real‐time deltas. While the patient is still holding her breathe, therapist can adjust the table so that all the real‐time deltas are within pre‐selected tolerances. Now, patient will be instructed to breath normally for a moment followed by another breath hold for verification purpose. If all the real‐time deltas are still within tolerances, the treatment can start. The translational real‐time delta tolerance is selected as follows. On the DRR in the beam eyes view (BEV) with the tangent beam shape and the heart contour, we can measure the shortest distance between the beam edge to the heart. Any number smaller than that shortest distance can be the real‐ time delta tolerance. Rotational tolerance is selected empirically. Either audio or video coaching will be used during the treatment to guide patient breathing in and out to bring the breast surface to the planed level. AlignRT Beam Hold again will report the real‐time deltas. Radiation beam will be turned on automatically when real‐time deltas are within the tolerance and off otherwise. The Learning Objectives of this lecture are: 1. Understand the principal of surface matching on left breast DIBH treatment 2. Get familiar with the clinical flow of left breast DIBH treatment

SU‐GG‐T‐209: A Novel Method for Further Analysis of IMRT QA

Purpose:IMRT QA is often performed with a commercial device that compares 2D intensity maps to determine if a field passes certain dose and position criteria. The effect of an intensity map error on 3D dosimetry and DVH is clinically relevant but often not considered in IMRT QA. We herein exam the 3D dosimetric impact of IMRT failures measured with a 2D IMRT QA tool. Methods and Materials: 22 IMRT plans had at least one field fail QA using MapCHECK 2, and these were further analyzed using in‐house software to perform a DVH analysis. The MapCHECK 2 passing criteria was 95% of measured points had to be within 5% of the planned dose and have a distance‐to‐agreement of 4mm. Software was installed in our treatment planning system, PlanUNC, that reads intensity maps measured with the MapCHECK 2 and creates DVHs for 3D dose analysis. The software allows for field segment adjustment of MLC positions to address field edge failures and modification of prescribed MUs to address detector points that failed QA. Results: Of the 22 plans that had a field fail initial QA with MapCHECK 2, it was determined that 16/22 and 3/22 plans were within 1% and 2% of the planned CTV and critical structure DVHs. 3/22 plans were found to have DVH differences greater than 2%, and MU values were therefore altered in the treatment plan. Conclusions: Further analysis following an initial IMRT QA with the MapCHECK 2 is often needed to address the 3D dosimetric effects of the measured plan. Software has been installed into PLUNC that allows for 3D analysis for IMRT QA. Our analysis shows that using the MapCHECK 2 passing criteria of dose within 5% and 4mm DTA is sufficient, since 19/22 plans failing this criteria are within 2% of the planned DVH.

SU‐FF‐T‐216: Comparison of a 2D and 3D Array of Diodes for IMRT QA

Purpose: To compare the IMRT QA pass/fail rates of a 2D diode array system MapCHECK™ and a cylindrical 3D diode array system Delta4™, and to investigate the benefit of DVH‐based IMRT QA. Methods and Materials: Eight treatment plans totaling 62 IMRT fields were measured using both MapCHECK and Delta4. The data were compared to the treatment planning data using Gamma analysis. Passing criteria was defined as 95% of measured points had to have a gamma value ⩽1.0 using a distance to agreement of 4mm, a 5% dose window, and a 10% dose threshold. Structures, including GTV and organs at risk (OAR), and dose volume histograms (DVHs) were exported from the treatment planning system to Delta4 for comparison to measured DVHs. Results: QA with Delta4 used an average of 466 detector points per field. Using gamma analysis, 60/62 (96.8%) IMRT fields passed with an average of 98.9% of detector points within a gamma value ⩽1.0 when measured on the Delta4. QA with MapCHECK used an average of 88 detector points per field. Application of the same gamma analysis resulted in 14/62 (22.6%) IMRT fields passing with an average of 91.3% of detector points with a gamma value ⩽1.0 as measured by MapCHECK. Further analysis of IMRT fields that failed using MapCHECK QA, indicated that the measured data was within 1–2% of the treatment plan. Planned GTV DVHs corresponded with the Delta4 measured GTV DVHs, however measured OAR DVHs differed from their planned DVHs. Conclusions: This study suggests that QA results acquired with Delta4 correspond more accurately to the actual treatment plan as compared to MapCHECK. Incorporating Delta4 into routine QA will decrease the overall QA analysis time. The increased pass rate with Delta4 may result from the increased amount of detectors per treatment field.