C Harris

TH‐C‐M100E‐02: Optically Stimulated Luminescence of Aluminum Oxide Detectors for Radiation Therapy Quality Assurance

Purpose: The purpose of this experiment was to: 1) Determine if a commercially available Al 2 O 3 detector system used for monitoring personnel exposure could be adapted for use as a radiation therapydosimetry system; and 2) Evaluate the systems performance as an in‐vivo dosimeter and its ability to measure absolute surface dose, isocenter dose, and normal tissues dose in a phantom as part of patient‐specific IMRTquality assurance.Method and Materials: The dosimeters were evaluated for: 1) Signal decay; 2) Field size dependence; 3) Energy dependence; and 4) Angular dependence using the Landauer, InLight MicroStar system. In‐Vivo dosimetry measurements were taken for 22 patients treated on a Varian 21EX. The Landauer system was also tested for its ability to measure absolute dose from helical tomotherapy treatments. Results: The variation between dosimeters was evaluated and found to be ±1.6%. The dosimeters appeared to over‐respond in the first 10 minutes, however, after 10 minutes the chips were within 1 percent of the steady‐state reading. Unlike other detectors, the Al2O3dosimeters showed no field size, energy, or angular dependence. The agreement between the dosimeters and the calculated doses for the in‐vivo dosimetry patients was 2.2±6.1 cGy or 3.7±2.5%. The dosimeters were also tested for their ability to measure absolute dose inside an IMRT phantom. The agreement between the dosimeters and the calculated doses was 0.1±5.3 cGy or 0.7±6.7%. Conclusion: Al 2 O 3 dosimeters can be a convenient, inexpensive alternative to TLDs, MOSFETS, and Diodes. The agreement between calculated and measured doses for in‐vivo dosimetry and IMRT QA is comparable to TLDs, MOSFETS, and Diodes. The dosimeters can be quickly read and analyzed after 10 minutes (to allow time for signal decay). The dosimeters do not appear to have an energy, field size, or angular dependence. In addition, the detectors can be erased and re‐used.

TU‐D‐AUD C‐09: In‐Vivo Dosimetry Using Aluminum Oxide Detectors

Purpose: The purpose of this study was to determine if aluminum oxide (Al 2 O 3 ) detectors could be used for in‐vivo dosimetery. The first specific aim of this project was to characterize the performance of a commercially available aluminum oxide detector system for in‐vivo dosimetry. The second specific aim of this study was to compare the accuracy of the new aluminum oxide detector to a commercially available MOSFET system in side‐by‐side patient measurements. Method and Materials: The dosimeters were evaluated for: 1) Signal decay; 2) Field size dependence; 3) Energy dependence; 4) Angular dependence; and 5) their reusability using the Landauer, InLight MicroStar system. In‐Vivo dosimetry measurements were taken for 53 patients treated on a Varian 21EX using (Al 2 O 3 ) dosimeters and 67 patients using OneDose Mosfets.Results: The variation between dosimeters was evaluated and found to be ±1.6%. The dosimeters appeared to over‐respond in the first 10 minutes, however, after 10 minutes the chips were within 1 percent of the steady‐state reading. Unlike other detectors, the Al 2 O 3 dosimeters showed no field size, energy, or angular dependence. In testing the Al 2 O 3 reusability, it was found that the half life of the OSL material was 5.14 ± 0.01 hours. The agreement between the dosimeters and the calculated doses for the in‐vivo dosimetry patients was −1.9±5.9%. The OneDose agreement between mosfet and calculated dose was found to be −1.3±8.5%. Conclusion: Al 2 O 3 dosimeters can be a convenient, inexpensive alternative to TLDs,MOSFETS, and Diodes. The agreement between calculated and measured doses for in‐vivo dosimetry QA is comparable to TLDs,MOSFETS, and Diodes. The dosimeters can be quickly read and analyzed after 10 minutes (to allow time for signal decay). The dosimeters do not appear to have an energy, field size, angular, or sensitivity dependence. In addition, under specific conditions, the detectors can be erased and re‐used.

TU‐FF‐A1‐06: The Impact of Dose Rate Variations On Helical Tomotherapy Delivery

Purpose:Intensity modulated radiation therapytreatments using helical tomotherapy involve detector arrays which collect data and stores it in the form of sinograms. This detector data can be used to verify that treatments are being correctly delivered to the patient. The purpose of this work was to examine the dosimetric effect due to dose rate variations. Method and Materials: During a helical tomotherapy treatment, the intensity of a rotating fan beam is modulated using a 64 leaf binary multi‐leaf collimator (MLC). The beam passes through the patient and is incident on the detector array, where detector counts are stored as sinogram data. For this study, a total of 53 fractions were evaluated for two different patients (1 Prostate and 1 Head & Neck). MLC controller files, which run the delivery sequence for each fraction, were extracted from the data archive system, along with the treatmentdeliverysinograms from all fractions. Because machine output has a cyclical pattern, errors were first simulated as sine curves with magnitudes varying from 2 to 10%. Then machine output sequences were extracted from the treatment fractions and modified by using the actual variation. Results: After visual inspection, it was observed that machine output can vary anywhere from 0−3%. When error simulations were inserted, dose differences ranged from 0.6−3%. The results from inserting actual treatment error back into the delivery sequence show that, with even a 1% variation in output, the dose difference is almost 0.5%. Conclusion: The results show that dose differences for actual error insertions, while not very large, could still cause a discrepancy when routine QA is performed. Changes in machine output vary from patient to patient as well as from day to day, which can cause a difference in the planned and delivereddose.

WE‐D‐AUD B‐01: Automatic Detection of Delivery Errors Using Autoassociative Kernel Modeling

Purpose: The purpose of this work was to evaluate exit detector data for patients treated with helical tomotherapy. A novel technique was developed for automatically evaluating exit dosimetry using autoassociative non‐parametric modeling, which has the ability to learn complex detector data relationships. Method and Materials: The tomotherapy detector array collects and stores exit dosimetry data during treatment delivery in the form of sinograms, which contain a record of the radiation that exits the MLC and passes through the patient during each treatment. Autoassociative Kernel Modeling (AKM) is a non‐parametric technique that makes parameter estimates by calculating a weighted average of a set of historical data called memory vectors. The memory vectors are contained in what is called a memory matrix, which are used to make predictions. Errors between predicted and test values were calculated using the sum of squared errors, which were used to identify faulty projections within each sinogram.Results: A total of 121 delivery sequences were evaluated from 5 patients (4 Prostates and 1 Head & Neck). Major errors were detected in at least one fraction over the course of treatment for each patient in the study. Other detected errors, while smaller in magnitude, could still be an indication of machine faults. Readings from the ionization chambers located in the head of the accelerator can help classify the type of error, whether they are major anatomical misalignments, MLC positional errors, or machine output errors. Conclusion: The key to the model is determining at what cutoff threshold to set so that all significant errors are detected while keeping reducing the number of false alarms. The results show that the model has the ability to detect errors in the exit dosimetry data. They also suggest that AKM modeling can be a useful tool in monitoring the reliability of radiation delivery.

WE‐E‐AUD C‐01: Prediction of Weight Loss, Tumor Response, and Set‐Up Errors For Head and Neck Patients

Purpose: The objective of this study was to develop a novel tool using Kernel Classification that can be used to automatically identify patients that have, or will have, setup issues requiring intervention such as re‐simulation and/or re‐planning. Method and Materials: Inter‐Fraction motion was retrospectively analyzed for 43 H&N patients that were treated on a helical tomotherapy system. For each patient, CTimages were acquired and transferred to the tomotherapy database for treatment planning and image‐guided patient setup. Both custom aquaplastic masks and a positioning mouthpiece were used in 10 of the 43 patients. Results: Fifteen patients had greater than 10% weight loss during the course of treatment. Six patients had a visible reduction in GTV volume. Immobilization effectiveness decreased as the tumors regressed in size and/or the patients lost weight. If the tumor regression was occurred then time could be scheduled to periodically check the mask fit and to make a new mask if needed. The kernel classification technique correctly identified all 43 H&N patients as either having normal or problematic setup using their respective shift data sets. Classifications were made using only the shift values from the first 14 treatments. The predictive performance seriously degraded when data from fewer than 14 treatments were used. However, adding more did little to improve the performance. Conclusion: This study demonstrated that the kernel regression classification method was able to correctly identify the cause behind IGRT positioning problems for H&N patients. The study validated that IGRT positioning problems cause abnormal problem‐specific distributions in the shift data without using statistical distribution tests. Since this technique is fully automated, it could potentially be used during IGRT sessions to help the therapists decipher the factors that hinder patient setup early in a patients treatment so that the proper precautions can be in place.

SU‐DD‐A2‐02: Analysis of Film Registration Techniques in Intensity Modulated Radiation Therapy Quality Assurance

Purpose: The purpose of this study was to evaluate manual and fiducial‐based plan‐to‐film registration techniques, identify advantages and disadvantages of manual and fiducial‐based plan‐to‐film registration techniques, develop a new automatic plan‐to‐film registration technique using a genetic algorithm, and compare the performance of the automatic registration to the manual and fiducial‐based techniques. Method and Materials: Ten patient plans (4 Head & Neck and 6 Prostate) were selected for the IMRT plan‐to‐film registration study. For each patient, the IMRTdose measurements were obtained in both axial and coronal planes using radiographic film, resulting in 20 test films. Dosecalibration films were irradiated at the same time as the test cases, and the test cases were processed at the same time. The calculated and measured dose distributions were registered with one another using RIT113 dosimetry software. Manual registration was performed by visual selection of four points in common on the dose and film images. Fiducial‐based registration was performed using the “Template” tool in RIT113 and marks placed on the films prior to irradiation. An automatic plan‐to‐film registration genetic algorithm was created that performs image registration by optimizing the best longitudinal and rotational shifts. Results: Of the techniques, the manual registration technique was inferior because it was highly susceptible to inter‐ and intra‐user variations. The fiducial‐based technique often resulted in incorrect registrations due to incorrect fiducial placement. Of the three techniques evaluated in this study, the genetic algorithm‐based registration provided the best agreement between calculated and measured dose distributions. Conclusion: Fiducials should be used for an initial registration, and the automatic technique should be used to calcuate the spatial offset between the plan and the film. Using these techniques in combination should improve the time required to perform IMRT QA and decrease the incidence of false positives. Conflict of Interest: Research consultant for RIT.

TU‐C‐M100J‐10: Respiration Motion Prediction Using Time‐Delay Kernel Regression Modeling

Purpose: The purpose of this study was to develop a novel technique for dynamically predicting respiration motion and uncertainty up to 1.5 seconds in the future in real‐time using Time‐Delay Kernel Regression (TDKR) modeling. Unlike neural network based prediction techniques, kernel regression models are continuously learning new respiration cycles for each patient without the need for computationally and time intensive re‐training. Method and Materials: Recorded respiration data from a real‐time respiratory gating system was used to develop a model to predict the amplitude of the marker block at a future time. An empirical TDKR model was designed to compensate for the latency that occurs between the acquisition of the image and the point at which the beam actually turns on/off in beam tracking systems. This non‐parametric model incorporates the temporal information present in the input data. The model was tested using respiration data from 4 patients. Results: The rootmean squared error (RMSE) between the model predictions and the measured data was computed for each patient at different latencies, and then the average was taken over all the patients. For predictions 1.5 seconds into the future the average RMSE was 1.4%. For predictions 1 second into the future, the RMSE dropped to 1.2%, and for 0.5 seconds it was only 0.7%. The average uncertainty for the predictions at 0.5, 1, and 1.5 seconds into the future was 2.4%, 3.2%, and 3.4%, respectively. Conclusion: This study proves that a TDKR model can learn the relationships present in respiration data. The reported results showed that the TDKR model has the same, if not better predictive performance, as previously studied parametric models. However, because TDKR is non‐parametric, it has several distinct advantages over these models that make it more suited for respiratory gating applications.

SU‐FF‐P‐04: Evaluation of Treatment Planning Time Savings Using Direct Machine Parameter Optimization

Purpose: The purpose of this study is to evaluate the capability for increasing treatment‐planning throughput by utilizing Direct Machine Parameter Optimization (DMPO). Traditional inverse treatment planning consists of an optimization phase followed by a separate leaf‐sequencing phase. With DMPO based optimization, there is no leaf‐sequencing step. As such, the treatment planning time for each IMRT case can potentially be decreased and treatment‐planning throughput could be increased. Method and Materials: Ten patients (5 prostate and 5 head & neck) were randomly selected for a comparative treatment planning study. IMRTtreatment plans were created using the standard IMRT optimization algorithm with k‐mean clustering based leaf‐sequencing. A second treatment plan was created for each patient using the DMPO optimization algorithm. The only difference between the two plans for each patient was the optimization algorithm. For each case, the time required to optimize the plans, the number of segments required for step‐and‐shoot delivery, and total monitor units were recorded. Results: Both the DMPO and IMRT optimizations yielded similar dose distributions. The mean time to perform inverse planning and leaf sequencing for prostate patients was 8.00 minutes, verses 8.85 minutes for DMPO based optimization. In contrast, the efficiency of DMPO for Head & Neck cases is substantially different. The mean time to perform a DMPO optimization for Head & Neck cases was 12.05 minutes, as compared with 28.63 minutes for IMRT optimization. Conclusion: Over 16 minutes of treatment planning time can be saved per patient by utilizing the DMPO based optimization in Pinnacle for Head & Neck treatment planning. The time advantage can grow to hours saved per patient if multipe treatment plans are created, or if the planners agressively adjust the plan. Depending on the Head & Neck volumes treated, DMPO optimization can increase treatment planning effiency.

SU‐FF‐T‐125: Comparison of Radiographic Film, Radiochromic Film, and CR Plates for Intensity Modulated Radiation Therapy Quality Assurance

Purpose: The primary purpose of this study was to evaluate different types of IMRT QA media and determine which provides the most accurate results and efficient process for performing IMRTquality assurance. The secondary purpose was to evaluate new correction algorithms that have been developed for use in the RIT113 dosimetry system. Method and Materials: Ten random patients were chosen who received treatment with step‐and‐shoot IMRT.IMRTdose measurements were taken in both a coronal and axial plane using radiographic film (EDR), radiochromic film (EBT), and a computed radiography(CR) plate. Because scanning EBT film with a Vidar scanner will result in artifacts caused by light scattering and lead to a nonuniform response, all EBT films were carefully marked to indicate orientation and scanned prior to irradiation. After exposure, a new background correction algorithm was developed to correct for nonuniformity. CR plates also have variations in their response across the width of the plate. In order to correct for this effect, a special calibration procedure was developed. A dose‐responsecalibration curve was generated for each pixel on the plate, which was accomplished by exposing the plate multiple times to a large uniform dose.Results: In all cases, the application of the new background correction algorithm improved the results for EBT film. Using the correction, the percentage of pixels exceeding the gamma threshold dropped from 8.1% to 4.6%. The IMRT analysis with the CR plate was consistently worse than both the EDR and the EBT films. The CR plate results improved with the uniformity correction, but still exhibited a 10–20% angular dependence and a 2 to 8% field size dependence. Conclusion:IMRT QA results with background corrected EBT films were comparable to EDR films in many cases, while CR plates consistently delivered inferior results. Conflict of Interest: Research consultant for RIT.

WE‐D‐224A‐01: Inter‐ and Intra‐User Variations in Film Based IMRT QA

Purpose: To measure the inter‐ and intra‐user variations in the manual alignment of calculated and measured doses in film‐based IMRT QA. Methods and Materials: Twenty (4 coronal and 16 axial) IMRT film based QA test cases were selected, each detailed by the “QA mode” in the IMPAC information management system. Films were shot in phantoms using a Varian 21EX, and do not contain fiducial marks. The treatment plans were created using the Pinnacle treatment planning system. Four of the films had known MLC problems, and were designed to fail the QA analysis. The films and corresponding calculated doses were placed on the internet for download. Participants were instructed to download the files, perform manual registration using the RIT113 software, save the registration films and return the test package to the investigators. Participants were instructed not to change the regions of interest and to indicate if each case would pass or fail their particular institutional criteria. Returned data was then analyzed for inter‐ and intra‐user variations in the manual alignment. Results: As of abstract submission, six respondents had been analyzed. The respondents had a wide range of passes and failures. Five out of the six respondents correctly identified the four films with known problems. One respondent incorrectly identified Patient #4 as a pass, but did note that the film was overly cropped. On average, the respondents indicated that seven of the films would not pass. Without fiducial marks on the film, each user placed the registration point in unique locations. As a result, each user had a unique QA analysis. Errors in the selection of registration points were directly related to false negatives. Conclusions: In order to minimize inter‐ and intra‐user variation, fiducial marks should be used to register the calculated and measured films in IMRT QA.