J. Kavanaugh

Daily QA of linear accelerators using only EPID and OBI

PURPOSE As treatment delivery becomes more complex, there is a pressing need for robust quality assurance (QA) tools to improve efficiency and comprehensiveness while simultaneously maintaining high accuracy and sensitivity. This work aims to present the hardware and software tools developed for comprehensive QA of linear accelerator (LINAC) using only electronic portal imaging devices (EPIDs) and kV flat panel detectors. METHODS A daily QA phantom, which includes two orthogonally positioned phantoms for QA of MV-beams and kV onboard imaging (OBI) is suspended from the gantry accessory holder to test both geometric and dosimetric components of a LINAC and an OBI. The MV component consists of a 0.5 cm water-equivalent plastic sheet incorporating 11 circular steel plugs for transmission measurements through multiple thicknesses and one resolution plug for MV-image quality testing. The kV-phantom consists of a Leeds phantom (TOR-18 FG phantom supplied by Varian) for testing low and high contrast resolutions. In the developed process, the existing LINAC tools were used to automate daily acquisition of MV and kV images and software tools were developed for simultaneous analysis of these images. A method was developed to derive and evaluate traditional QA parameters from these images [output, flatness, symmetry, uniformity, TPR20/10, and positional accuracy of the jaws and multileaf collimators (MLCs)]. The EPID-based daily QA tools were validated by performing measurements on a detuned 6 MV beam to test its effectiveness in detecting errors in output, symmetry, energy, and MLC positions. The developed QA process was clinically commissioned, implemented, and evaluated on a Varian TrueBeam LINAC (Varian Medical System, Palo Alto, CA) over a period of three months. RESULTS Machine output constancy measured with an EPID (as compared against a calibrated ion-chamber) is shown to be within ±0.5%. Beam symmetry and flatness deviations measured using an EPID and a 2D ion-chamber array agree within ±0.5% and ±1.2% for crossline and inline profiles, respectively. MLC position errors of 0.5 mm can be detected using a picket fence test. The field size and phantom positioning accuracy can be determined within 0.5 mm. The entire daily QA process takes ∼15 min to perform tests for 5 photon beams, MLC tests, and imaging checks. CONCLUSIONS The exclusive use of EPID-based QA tools, including a QA phantom and simultaneous analysis software tools, has been demonstrated as a viable, efficient, and comprehensive process for daily evaluation of LINAC performance.

TH‐A‐116‐07: Validation of Patient Contours for Head and Neck Treatments Using Population‐Based Metrics

PURPOSE To develop a statistical and anatomical population-based model that can be used to validate the accuracy and integrity of head and neck normal tissue structures of individual patients for use in preplanning and/or online adaptive radiation therapy. METHODS Normal tissue contours from 29 patients treated for head and neck cancers were used in development of the model. For each patient, DICOM plan and structure files were exported from the treatment planning system to an in-house developed software program which calculated anatomic metrics for volume, shape, and intra-structure distances for all structures. A statistical analysis of these metrics produced population specific rules that were used within the software program to evaluate the accuracy of head and neck contours for subsequent patients. The contour assessment program included only metrics for which the standard deviation was less than a heuristically determined limit of 15% of the mean for that metric. To verify the softwares utility, 42 common contouring errors were intentionally introduced within 9 specific structures for 9 different patients. These errors included incorrect laterality, position, size and shape, inclusion of small isolated pixels, deleted segments, and empty structures. The evaluation of all 9 head and neck structure sets was blinded to the nature and number of the generated errors. RESULTS The contour accuracy and integrity program correctly identified 40 of 42 generated errors. Small modifications to the structures shape and volume were the most difficult to correctly identify; however the program correctly identified all positional and laterality errors, deleted/isolated segments, small pixels, and deleted contours. CONCLUSION Rules developed from a statistical analysis of anatomic population-based metrics can provide much of the necessary information to correctly and efficiently evaluate the accuracy and integrity of a unique patient contour structure set for IMRT preplanning or for an online adaptive radiation therapy protocol.

TU-FG-201-01: 18-Month Clinical Experience of a Linac Daily Quality Assurance (QA) Solution Using Only EPID and OBI

PURPOSE To describe the clinical use of a Linear Accelerator (Linac) DailyQA system with only EPID and OBI. To assess the reliability over an 18-month period and improve the robustness of this system based on QA failure analysis. METHODS A DailyQA solution utilizing an in-house designed phantom, combined EPID and OBI image acquisitions, and a web-based data analysis and reporting system was commissioned and used in our clinic to measure geometric, dosimetry and imaging components of a Varian Truebeam Linac. During an 18-month period (335 working days), the Daily QA results, including the output constancy, beam flatness and symmetry, uniformity, TPR20/10, MV and KV imaging quality, were collected and analyzed. For output constancy measurement, an independent monthly QA system with an ionization chamber (IC) and annual/incidental TG51 measurements with ADCL IC were performed and cross-compared to Daily QA system. Thorough analyses were performed on the recorded QA failures to evaluate the machine performance, optimize the data analysis algorithm, adjust the tolerance setting and improve the training procedure to prevent future failures. RESULTS A clinical workflow including beam delivery, data analysis, QA report generation and physics approval was established and optimized to suit daily clinical operation. The output tests over the 335 working day period cross-correlated with the monthly QA system within 1.3% and TG51 results within 1%. QA passed with one attempt on 236 days out of 335 days. Based on the QA failures analysis, the Gamma criteria is revised from (1%, 1mm) to (2%, 1mm) considering both QA accuracy and efficiency. Data analysis algorithm is improved to handle multiple entries for a repeating test. CONCLUSION We described our 18-month clinical experience on a novel DailyQA system using only EPID and OBI. The long term data presented demonstrated the system is suitable and reliable for Linac daily QA.

TU‐C‐17A‐04: BEST IN PHYSICS (THERAPY) – A Supervised Framework for Automatic Contour Assessment for Radiotherapy Planning of Head‐ Neck Cancer

PURPOSE Precise contour delineation of tumor targets and critical structures from CT simulations is essential for accurate radiotherapy (RT) treatment planning. However, manual and automatic delineation processes can be error prone due to limitations in imaging techniques and individual anatomic variability. Tedious and laborious manual verification is hence needed. This study develops a general framework for automatically assessing RT contours for head-neck cancer patients using geometric attribute distribution models (GADMs). METHODS Geometric attributes (centroid and volume) were computed from physician-approved RT contours of 29 head-neck patients. Considering anatomical correlation between neighboring structures, the GADM for each attribute was trained to characterize intra- and interpatient structure variations using principal component analysis. Each trained GADM was scalable and deformable, but constrained by the principal attribute variations of the training contours. A new hierarchical model adaptation algorithm was utilized to assess the RT contour correctness for a given patient. Receiver operating characteristic (ROC) curves were employed to evaluate and tune system parameters for the training models. RESULTS Experiments utilizing training and non-training data sets with simulated contouring errors were conducted to validate the framework performance. Promising assessment results of contour normality/abnormality for the training contour-based data were achieved with excellent accuracy (0.99), precision (0.99), recall (0.83), and F-score (0.97), while corresponding values of 0.84, 0.96, 0.83, and 0.9 were achieved for the non-training data. Furthermore, the areas under the ROC curves were above 0.9, validating the accuracy of this test. CONCLUSION The proposed framework can reliably identify contour normality/abnormality based upon intra- and inter-structure constraints derived from clinically-approved contours. It also allows physicians to analytically determine the system parameters to fit various clinic requirements (e.g. as-low-as-possible false positives). It has great potential for improving RT work flow. More geometric attributes and training sets will be investigated to improve framework performance in the future.

SU‐E‐T‐305: Limitations of Using DICOM Data for BrachyVision Treatment Plan Evaluation

PURPOSE To evaluate the accuracy of a real-time automated method of performing dosimetric quality assurance using Eclipse DICOM files for patients receiving HDR-brachytherapy and IMRT. METHODS GYN patients are treated with concurrent high-dose rate brachtherapy and IMRT. The dosimetric parameters were obtained through an in-house QA program developed using Matlab. The DICOM files containing DVH data for organsat-risk (OAR) were analyzed Dosimetric data for 7 patients (total 42 fractions) were collected for bladder, rectum and sigmoid. The accuracy of the dosimetric parameters was estimated by comparing the parameters obtained from the DICOM based QA program and those in BrachyVision. RESULTS The maximal dose values (Dmax) for the OARs obtained using the DICOM-based program are significantly smaller than those valued reported in BrachyVision by 36.2%-48.3%. The mean dose has a deviation from 1% - 2.4%. The dose for the volume of 2cc (D2cc) has a difference up to 7.6% for structures with the volume larger than 200 cc. The average difference of D2cc is 0.5% for structures less than 200 cc. We found that Eclipse BrachyVision only exports DVH data down to a volume equivalent to 1% of the maximum volume for a given structure. Therefore, the reported maximal dose values obtained from DICOM RT dose file do not accurately reflect the maximum dose in a treatment plan. This will also slightly affect the mean dose calculation and D2cc when the structure volume is larger than 200cc. CONCLUSIONS The automatic QA tool based on DICOM files provides a quick retrieval of dose to organs-at-risk and coverage of targets. However, maximal dose to structures is not accurate due to the truncationof the DVH information contained in DICOM files.