AAPM Spring Clinical Meeting – Abstracts

PO‐BPC‐Exhibit Hall‐01

Improvements in MR Quality Assurance Workflow and Outcomes with a Web‐Based Quality Control Database

X Yang*, K Little, X Jiang, D Hintenlang

The Ohio State University, Columbus, OH

Purpose: To describe a custom‐built, web‐based MR Quality Control QC database, and to make a preliminary assessment of its impact on the Quality Assurance QA process in a large U.S. hospital.

Methods: The MR QC database was built with Microsoft Access 2010 and published on a Microsoft Sharepoint website owned and maintained by the authors’ institution. Authorized users can access the database remotely with mainstream web browsers. QC technologists were granted access to add, review, and print daily and weekly QC records. Qualified medical physicists (QMPs) were granted additional access to edit, review, and approve existing QC records, and to change tolerance limits. A macro was utilized to conduct an automatic weekly review of QC status, and to email the results to a QMP. This web‐based QC database was implemented on five clinical MRIs at the authors’ institution. Weekly ACR QC findings within five months before and after implementation were compared.

Results: Retrospective review of the conventional QC records before implementation revealed 26 QC issues. After the adoption of the new web database, the number of QC issues reduced by 38% to 16. Since only a small fraction of these issues were reported at the time of occurrence, the web‐based database permitted the QMPs to more quickly identify a QC issue (before/after: 7.6 ± 7.5/1 ± 0 weeks, P = 0.0002). Among the 26 QC issues occurred before the implementation, 8 issues (31%) would have been handled differently had they been identified earlier. The time to corrective actions was also found to be slightly shorter with the web‐based system, although the difference is not statistically significant (before/after: 3.2 ± 2.7/1.7 ± 2.1 weeks, P = 0.3).

Conclusion: The web‐based QC database provides a positive impact on our MR QA process. It enables early detection and facilitates resolution of potential issues that may affect the quality of clinical MRI studies.

PO‐BPC‐Exhibit Hall‐02

Four‐Dimensional Digital Tomosynthesis Based On Visual Respiratory Guidance

D Kim¹, S Kim², T Suh^1,*

(1) The Catholic University of Korea, Seoul, Republic of Korea, (2) Virginia Commonwealth University, Glen Allen, VA

Purpose: The aim of this research was to introduce and evaluate a respiratory‐guided slow gantry rotation 4D digital tomosynthesis (DTS).

Methods: For each of ten volunteers, two breathing patterns were obtained for 3 minutes, one under free breathing condition and the other with visual respiratory‐guidance using an in‐house developed respiratory monitoring system based on pressure sensing. Visual guidance was performed using a 4 s cycle sine wave with an amplitude corresponding to the average of end‐inhalation peaks and end‐exhalation valleys from the free‐breathing pattern. The scan range was 40° for each simulation, and the frame rate (FR) and gantry rotation speed (GRS) were determined so that one projection per phase should be included. Both acquisition time (AT) and the number of total projections to be acquired (NPA) were calculated. Applying the obtained respiration pattern and the corresponding sequence, virtual projections were acquired under a typical geometry of Varian on‐board imager for two virtual phantoms, modified Shepp‐Logan (mSL) and XCAT (extended Cardiac‐Torso). For the XCAT, two different orientations were considered, anterior‐posterior (i.e., coronal) and left‐right (i.e., sagittal). Projections were sorted to ten phases and image reconstruction was made using a modified filtered back‐projection. Reconstructed images were compared with the planned breathing data (i.e., ideal situation) by SSIM (Structural Similarity) and NRMSE (Normalized Root‐mean‐square Error).

Results: For each case, simulation with guidance (SwG) showed motion‐related artifact reduction compared to that under free‐breathing (SuFB). SwG required less NPA but provided slightly higher SSIM and lower NRMSE values in all phantom images than SuFB did. In addition, the distribution of projections per phase was more regular in SwG.

Conclusion: Through the proposed respiratory‐guided 4D DTS, it is possible to reduce imaging dose while improving image quality.

This research was supported by the Mid‐career Researcher Program (NRF‐2018R1A2B2005343) through the National Research Foundation (NRF) funded by the Ministry of Science and ICT (MSIT) of Korea.

PO‐BPC‐Exhibit Hall‐03

Is Measurement of CTDIvol with a Helical Scan an Acceptable Alternative to the Standard Single‐Slice Methodology?

S Leon^1,*, R Kobistek², I Barreto³, B Schwarz⁴, Z Zhang⁵, E Olguin⁶

(1) University of Florida, Gainesville, FL, (2) Nat'l Physics Consultants, Ltd, Mentor, OH, (3) University of Florida, Gainesville, FL, (4) University of Florida, Gainesville, FL, (5) University of Florida, Gainesville, FL, (6) University of Florida, Gainesville, FL

Purpose: Most clinical CT protocols use helical scanning; however, the standard methodology for CTDI (vol) measurement replaces the helical protocol with an equivalent axial scan, which is not easily accomplished on many scanners. Altering protocols may result in unavoidable changes to bowtie filters and collimation, which impede measurement accuracy. The purpose of this study was to determine whether CTDI (vol) can be accurately measured with a helical scan.

Methods: CTDI (vol) was measured for helical protocols on 15 CT scanners from all major manufacturers. Adult and pediatric head and abdomen protocols were tested. CTDI (vol) was measured using the axial methodology, then again using the helical protocol. For the helical measurements, the scan length was equal to the active area of the pencil chamber, as seen on the topogram, and the CTDI (vol) equation was altered by setting both nxT and table increment to 100 mm. CTDI (vol) measurements using each method were compared to each other and to the CTDI (vol) displayed by the scanner.

Results: Excluding systems in which it was impossible to match collimation between the axial and helical protocols, the difference between the axial and helical methods averaged 0.4 mGy (range −2.9–3.9). Among all systems, the difference between the axial measurements and the displayed CTDI (vol) averaged −0.9 mGy (range −7.5–5.1). The difference between the helical measurements and the displayed CTDI (vol) averaged −1.6 mGy (range −9.1–2.0). The helical measurements agreed with the displayed CTDI (vol) better than the axial measurements for 70% of protocols tested. Two axial measurements differed from the displayed CTDI (vol) by >20%; no helical measurements did.

Conclusion: There was excellent agreement between the two measurement methods and to the displayed CTDI (vol). The helical measurement method can be accomplished more easily than the axial method on many scanners and is a reasonable testing method for QC purposes.

PO‐BPC‐Exhibit Hall‐04

Surface Imaging Accuracy for Stereotactic Radiosurgery Setup Treatment Delivery

J McCulloch*, V Bry, D Saenz, P Myers, N Kirby, K Rasmussen

UTHSCSA, San Antonio, TX

Purpose: To determine the accuracy of using the C‐RAD Catalyst HD surface imaging system for Stereotactic Radiosurgery treatment setup.

Methods: An anthropomorphic head phantom was imaged on a GE LightSpeed CT scanner in a head‐first supine SRS treatment position. Plans were created with couch angles at 270°, 225°, 180°, 135°, and 90° in the BrainLab treatment planning software with a simulated lesion in the right frontal lobe and the left occipital lobe. White tape was placed on the phantom's exterior in order to make it more visible to the Catalyst HD's cameras. The phantom was set up at the 180° using Brainlab ExacTrac. The coordinates were captured with the C‐Rad Catalyst HD system. The phantom was then set up at each of the different couch positions with ExacTrac x‐ray imaging to verify internal anatomy accuracy. At each position the Brainlab shifts to isocenter were compared with the C‐Rad shifts to isocenter in each dimension.

Results: External surface imaging and internal x‐ray imaging had agreement within 0.62, 0.84, and 0.60 mm in the vertical, lateral, and longitudinal dimensions, respectively, across all couch angles. The total range of values for each was 0.02–0.62 mm, 0.01–0.84 mm, and 0.01–0.60 mm, respectively. The maximal vector shifts were calculated to be ranging from 0.11 to 1.05 mm. Rotation deviation was negligible for all angles (<0.2°).

Conclusion: For the phantom analyzed, C‐RAD Catalyst HD has a discrepancy of less than 1 mm from internal x‐ray imaging in each cardinal direction for the full range of couch angles. These results are promising for its application for SRS treatment and verification.

Research funded in part by a vendor grant.

PO‐BPC‐Exhibit Hall‐05

Characterization of Siemens FORCE CT Bowtie Filters

J Salazar*, K Lewis

Vanderbilt University, Nashville, TN

Purpose: To characterize the four bow tie filters (BTF) (two head and two body) on a dual‐source Siemens FORCE CT scanner through non‐invasive means by using a real‐time dosimeter.

Methods: The Characterization of Bowtie Relative Attenuation (COBRA) method, developed by JM Boone and further elaborated by BR Whiting, was augmented with additional with beam quality measurements. A 0.6 cc ion chamber was positioned near the periphery of the scan field of view for each of the two CT x‐ray sources (tube A and tube B); the detector remained stationary while the tubes were separately activated and rotated around the detector; the tube collimation was open to its maximum so that the x‐ray cone‐beam encompassed the dosimeter throughout its rotation. Measurements were made at 80 and 120 kVp. The acquired dose‐rate waveforms were distance corrected and then fitted to a polynomial equation for smoothing purposes. Additional stationary tube exposure measurements were obtained at three different fan angles in order to better characterize the quality of the x‐ray beam (HVLs of Al) through different amounts of bow tie material. The bow tie filter thickness filter was then determined in mm of aluminum‐equivalent thickness for the four BTFs at various fan angles relative to the center (0o) of the filter.

Results: The shapes of the bow tie filter profiles determined for the two energies had the expected shape and were in close agreement with each other (e.g., the max thickness deviation between the two profiles for the Tube B head filter was 0.06 mm).

Conclusion: We have determined the shape and relative thickness of four BTFs (two head and two body) in a Siemens FORCE CT scanner into aluminum‐equivalent thicknesses. Length‐wise measurements of HVLs along the fan beam was an effective method to account for varying beam hardening that occurs.

PO‐BPC‐Exhibit Hall‐06

Use of a Conditional Generative Adversarial Network to Synthesize MR of the Prostate

J Kielbasa*, A Kubli, R Mathews, T Willoughby, O Zeidan, A Shah, P Kelly, S Meeks

UF Health Cancer Center at Orlando Health, Orlando, FL

Purpose: To train a conditional generative adversarial network (cGAN) to generate synthetic MR (synMR) images of the prostate from CT data and evaluate the quality of the synMR images produced using various training datasets.

Methods: Seventy‐seven prostate patients with both MR and CT exams were used in this study. Seventy‐three patients were randomly selected as training sets, and the remaining four patients were used as testing sets to evaluate cGAN performance. The cGAN was trained using registered axial MR and CT slices. In order to characterize the relationship between training data and synMR quality, the following three training sets were used: (1) entire scans of 25 patients, (2) entire scans of 73 patients, (3) 73 scans limited to slices ±1 cm around the prostate in the craniocaudal direction. This quality was evaluated by calculating the mean absolute error around the prostate, MAE (prostate ± 1 cm), between synMR and true MR images in each testing case.

Results: In 2/4 cases, the calculated MAE¬prostate ±1 cm was smallest using 73 scans localized to the prostate. Although no consistent improvement was established between the two 73‐patient training sets, qualitative improvements were observed in synMR images generated from the larger datasets when compared to the 25‐patient training set. In one case, the discrepancy in patient position between CT and true MR lead to relatively inaccurate synMR (with respect to the true MR).

Conclusion: Machine learning can be used to generate synMR from CT. Choice of training dataset has a significant effect on synMR quality. A more sophisticated training set may yield more accurate results. Future work will train using multi‐channel cGAN input data with marked anatomical structures and a more robust metric for image quality. This can lead to significant improvements in overall synMR data.

PO‐BPC‐Exhibit Hall‐07

A Team‐centered Approach to Implementing New Technologies in a Multi‐Institutional Cancer Center Setting

C Abing*

HSHS St. Vincent Hospital, Green Bay, WI

Purpose: Emerging technology introduction can be a challenge in a single institution, but becomes significantly more difficult when dealing with multiple centers. Our team is spread out across four cancer center locations and has implemented a standardized approach to the implementation of new technologies in a streamlined manner at any of the facilities.

Methods: Our combined four cancer centers provide external beam, IMRT, SRS, SBRT, HDR, and LDR treatments. We have developed a standard policy for the introduction of new technologies into the department across the system. The policy consists of a multi‐disciplinary team which oversees the following aspects of implementation: safety review (NRC, FDA, and vendor notifications), staff communication, education, vendor training, commissioning, policy creation, in‐house training, dry runs, go‐live, follow, and quality review after implementation.

Results: Over the course of a year and a half, we have successfully implemented the following programs using this process: cranial MLC‐based SRS, lung SBRT, tandem and ovoid applicator, universal cylinder applicator, APBI applicator, and LDR prostate seed implant. One key component to producing a high quality program is routine follow‐up and evaluation of both the overall new process and dosimetric results from the patients treated.

Conclusion: Creating a robust, standardized process for implementation of emerging technologies can be successfully performed if the team is committed to providing a robust quality approach to the process.

PO‐BPC‐Exhibit Hall‐08

Correlation of Dosimetric Parameters with Clinical Outcomes in the Treatment of Unresectable Hepatic Metastases with Yttrium‐90 Microspheres

W Wei, V Sehgal, M Al‐Ghazi*

UC Irvine, Orange, CA

Purpose: Yttrium‐90 (Y‐90) microspheres are used for the treatment of unresectable hepatic metastases. This study examines whether dosimetry related factors serve as prognostic indicators of clinical outcome, including overall survival (OS) and progression‐free survival (PFS).

Methods: Thirty‐two patients with hepatic metastases treated with Y‐90 were included in this study. Treatment dose was prescribed and the initial activity required to deliver such dose was calculated. OS and PFS probabilities were determined using Kaplan‐Meier actuarial analysis. Median Y‐90 activity administered, pulmonary shunt, and resulting lung activity were calculated. Patients were stratified into high and low activity groups with respect to each parameter's median value. Statistical differences of OS and PFS between the two groups were compared using log‐rank test.

Results: All tumors were metastatic with the most common primary site being colorectal adenocarcinoma (68.8%). Median follow‐up duration was 10.5 months. Median Y‐90 activity delivered was (40.8 ± 22.3 mCi). Fourteen patients received activity higher than the median value (46.3 ± 23.4 mCi). Eighteen patients received lower activity (30.1 ± 7.5 mCi; P < 0.001). Higher median PFS was observed with higher activity compared to lower activity (28.2 ± 3.9 vs. 9.3 ± 1.4 months, P < 0.001). The trend showed higher activity improved median OS (28.4 ± 3.7 vs. 10.8 ± 3.6 months, P = 0.068). Median OS probability was 80.8 % for high activity versus 57.1 % for low activity at the median follow‐up time.

Conclusion: Our results suggest that the initial activity delivered to patients could potentially serve as a prognostic indicator of PFS. More patients and longer follow‐up time need to be included in a future analysis to determine whether total activity could also be used as OS indicator. Other dosimetric parameters such as lung activity and pulmonary shunt did not correlate with OS and PFS.

PO‐BPC‐Exhibit Hall‐09

Patient‐Specific QA On Halcyon 2.0

E Laugeman*, A Heermann, J Hilliard, M Roberson, M Watts, R Morris, S Goddu, I Zoberi, H Kim, S Mutic, G Hugo, B Cai

Washington University School of Medicine, St. Louis, MO

Purpose: Determine local process‐based patient specific quality assurance action limits using Halcyon 2.0 and multiple measurement devices to 1 evaluate agreement with universal action limits recommended by TG‐218 and 2 provide reference to other clinics with similar equipment.

Methods: Ten plans were developed for Halcyon 2.0 via the TG119 test suite with both IMRT and VMAT techniques as well as 28 clinical plans from a variety of sites including: brain, head and neck, lung, breast, abdomen, and pelvis. Six additional breast VMAT plans were created using the two‐isocenter technique, which includes an automatic longitudinal shift. A new feature available with Halcyon 2.0, the dynamically flattened beam, was used for an AP/PA spine and four field boxes, as well as included in ten field‐in‐field breast plans. All plans (56) were measured with an ion chamber, portal dosimetry, MatriXX within MULTICube phantom, ArcCheck, and Delta4. Action limits were calculated and compared to TG‐218 recommendations.

Results: Fifty‐three out of the 56 plans are within 3% for ion chamber measurements with an average expected to measured ratio of 1.007 (0.7%) for all plans. The average gamma passing rates with a 3% dose difference and 2 mm distance‐to‐agreement (global normalization, 10% threshold) criteria for all plans are: PD – 99.9%, MatriXX – 91.6%, ArcCheck – 98.7%, and Delta4 – 99.2%. Therefore, the local action limits will be set at: PD – 99%, MatriXX – 72%, ArcCheck – 95%, and Delta4 – 96%.

Conclusion: Halcyon plan deliveries using IMRT, VMAT, two‐isocenter technique, and dynamically flattened beams successfully passed patient‐specific QA with multiple measurement devices. Calculated action limits bettered the recommendations of TG‐218 (90% with 3%/2 mm) using portal dosimetry, ArcCheck, and Delta4. MatriXX action limits are lower due to the measurement geometry and inability to use an inclinometer to apply angular corrections.

Financial support was provided by Varian Medical Systems.

PO‐BPC‐Exhibit Hall‐10

Evaluation of a Dense Planar Diode Array for True Composite SRS Measurements

S Ahmed*, G Zhang, E Moros, V Feygelman, H Lee

Moffit Cancer Center and Research Insitution, Tampa, FL

Purpose: This study is aiming to evaluate a 2D detector array (SRS MapCHECK [SMC] from Sun Nuclear Corp.) for SRS plan verification.

Methods: The SMC was placed in a StereoPHAN phantom and irradiated with a conventional 6MV or a 6MVFFF beam from a Varian TrueBeam accelerator. A comprehensive suite of evaluation measurements was performed, among them the 180° rotation, angular dependence (against an ion chamber, IC), field size dependence (against a scintillator), and end‐to‐end (against radiochromic film) tests.

Results: The mean differences in the flip test detector readings were within 0.01 ± 0.2%. The SMC sensitivity angular dependence in the axial plane, with both energies, was under 2% for all gantry angles, except with the beam central axis within ±5° from the array plane. Most combinations of gantry and couch angles resulted in <2% deviation from the IC. Discrepancies of 2%–3% were observed with the beams incident at ±30° to the detector array and the couch rotations >50°. The output factors agreed with the scintillator within 2%, with the only exception of the 6X 5 mm field (3.2% error). For the SRS plans, the average γ‐analysis passing rates were 98.03% ± 0.8% and 97.95% ± 1.2% for 3%/1 mm and 2%/2 mm criteria, respectively.

Conclusion: SRS MapCHECK detector spacing (2.5 mm) makes it theoretically capable of faithfully representing any gradients encountered in modern x‐ray radiotherapy. Our data show that after application of built‐in correction factors, its calibration is sufficiently accurate to make it a feasible device for the composite SRS measurements, provided that the couch/gantry angle combinations do not result in direct irradiation of the device electronics. SMC exhibits common sensitivity variations when the beam is nearly parallel to the detector plane, which does not meaningfully affect the composite dose.

This work was supported in part by a grant from Sun Nuclear Corporation. SA is a graduate student supported by SNC grant, VF is the PI on the project.

PO‐BPC‐Exhibit Hall‐11

Automated Verification Plan Preparation 2D‐3D Gamma Analysis for Proton Patient‐Specific Quality Assurance

D Hernandez Morales^1,*, J Shan², W Liu¹, K Augustine¹, M Bues¹, M Davis¹, M Fatyga¹, J Johnson³, D Mundy³, J Shen¹, J Younkin¹, J Stoker¹

(1) Mayo Clinic Arizona, Phoenix, AZ, (2) Arizona State University, Scottsdale, AZ, (3) Mayo Clinic, Rochester, MN

Purpose: Physician‐approved treatment plans undergo patient‐specific quality assurance (PSQA) prior to beginning of treatment. For pencil beam scanning proton therapy, quality assurance is complex and time consuming, involving multiple measurements per field. We evaluated the PSQA process to identify routine steps that could be automated for a comprehensive and efficient workflow.

Methods: We used the treatment planning system's (TPS) capability to support C# scripts to develop an Eclipse Application Programming Interface (API) script to automate the preparation of the verification‐phantom plan. The API script evaluated the gradient in the target volume of each verification field based on established criteria to identify adequate depth‐dose profiles and depths for PSQA measurements. A local area network (LAN) connection between our measurement equipment and shared database was established to facilitate equipment control, measurement data transfer and storage. To improve measurement data analysis, a Python script was developed to automatically perform a 2D‐3D γ‐index analysis between the measurement plane and the corresponding TPS in‐water volume for each acquired measurement. We evaluated a subset of patient plans representing the various disease sites treated at our clinic with the previous and automated methods to quantify changes in efficiency.

Results: The device connection via LAN granted immediate access to the plan and measurement information for analysis using an online software suite. Automated verification plan preparation reduced the task time by more than 50%, decreasing the time from 5 to 20 min per field to 1–3 min per field. The γ‐index analysis time reduction is more pronounced, being reduced by an order of magnitude for all disease sites. With these automations we observed an average overall PSQA time savings of 57% per patient plan.

Conclusion: Automating routine PSQA workflow elements improve time efficiency, reduce user fatigue, and focus efforts on evaluation of key quality metrics.

Dr. Wei Liu and Jie Shan were supported by the National Cancer Institute (NCI) Career Development Award K25CA168984, Arizona Biomedical Research Investigator Award, the Lawrence W. and Marilyn W. Matteson Fund for Cancer Research, and the Kemper Marley Foundation.

PO‐BPC‐Exhibit Hall‐12

Associations Between Dose‐Volume, Quality of Life, and Clinical Toxicity Correlations After Lung SBRT

S Devpura*, I Chetty, S Brown, S Rusu, E Mayyas, R Araj, J Kim, J Liu, C Liu, Z Sun, D Snell, S Vance, M Ajlouni, S Siddiqui, B Movsas

Henry Ford Health System, Detroit, MI

Purpose: To measure the effect of radiation dose and volume on patient‐reported quality of life (QOL) outcomes and clinical toxicity of lung cancer patients up to 36 months after SBRT.

Methods: Non‐small cell lung cancer (NSCLC) patients, n = 122, received SBRT (12 Gy x 4). Symptoms including cough, dyspnea, fatigue, and pneumonitis were measured at baseline (before treatment), after treatment and 3, 6, 12, 18, 24, and 36 months post‐treatment. Toxicity, graded from 0 to 5, followed the Charlson comorbidity and toxicity index. Quality of life was determined using the previously validated Functional Assessment of Cancer Therapy‐Trial Outcome Index (FACT‐TOI) Lung questionnaire which incorporated three subscale endpoints: lung subscale (LSC), physical well‐being (PWB), and functional well‐being (FWB). Dosimetric parameters included the mean lung radiation dose (MLD) and the volume of normal lung receiving at least 5, 10, 13, or 20 Gy (V5, V10, V13, and V20), esophagus receiving at least 5 Gy, maximum and mean dose (E_V5, E_Dmax, and E_Dmean) obtained from the treatment plan. Pearson correlation and t‐test analyses were used to measure correlations between lung metrics with QOL and toxicities.

Results: SBRT produced minimal toxicities. QOL (TOI, LSC, PWB, or FWB) at 24 and 36 months posttreatment were significantly correlated with V5, V10, V13, V20, and MLD for stage I&II (P < 0.05). Moreover, correlations were found for dyspnea with V10 and V13 for stage I&II, for esophagitis with E_V5 for stage III, and for dyspnea with V20 for stage IV. Overall, QOL improved for patients with stage IV lung cancer at 18, 24, and 36 months.

Conclusion: Lung SBRT treatment for patients with NSCLC, using a 12 Gy x 4 dose regimen, was well tolerated. Unique QOL data (not previously reported) and clinical toxicities at up to 36 months follow‐up showed correlations with lung dose and subvolumes for different stages.

This work was supported in part by a research grant from Varian Medical Systems, Palo Alto, CA.

PO‐BPC‐Exhibit Hall‐13

Predicting Individual Target V12 Gy for Single Isocenter Multi‐Target (SIMT) Stereotactic Radiosurgery (SRS): A Comparison of Three Empirical Methods

T Li^1,*, H Zhang¹, W Shi², H Liu²

(1) University of Pennsylvania, Philadelphia, PA, (2) Thomas Jefferson University, Philadelphia, PA

Purpose: V12 Gy is the key indicator for brain necrosis following SRS. This study compares three empirical methods for predicting V12 Gy from target volume and prescription information prior to planning.

Methods: Sixteen patients with a total of 112 targets were retrospectively studied. Plans were generated using Brainlab Elements™ automatic SIMT system. Three methods were compared for predicting V12 Gy in cc using only the target volume (cc) and prescription (Gy) information. (1) Linear regression based on individual target volume weighted by prescription. (2) Neuro‐network–based regression using ten hidden layers and Bayesian Regularization backpropagation training algorithm. The training was performed in MATLAB® with 85% of samples randomized to training set and 15% as test set. To assess the randomness of training/testing partitioning, the training was repeated 100 times, and the Root‐mean‐square‐error (RMSE) of test data prediction was summarized. (3) Geometric expansion (1–6 mm) of the target's equivalent sphere by a fixed distance to approximate the location of 12 Gy isodose volume outside the target. The volume of this expanded target was then compared to true V12 Gy to determine residual prediction error.

Results: (1) Multivariate analysis showed that V12 Gy is highly correlated to the product of target volume (TV) and its prescription dose (Rx). Linear regression between V12 Gy and achieved an RMSE of 1.09 cc. (2) Neuro‐network‐based regression was performed 100 times and showed RMSE of 1.14 ± 0.04 cc (range 1.026–1.336). (3) Residual analysis showed that a fixed expansion of 4 mm achieved smallest and most stable prediction with RMSE at 1.12 cc.

Conclusion: Despite more complicated algorithm and 3D geometric information that were included in methods 2 and 3, simple linear regression achieved the lowest RMSE. It is feasible to predict individual target’ V12 Gy during SIMT SRS using simple linear regression based on target volume weighted by prescription with an RMSE of ˜1.1 cc.

PO‐BPC‐Exhibit Hall‐14

Efficiency and Accuracy Improvements for Patient Plan Quality Assurance with a Passive Scattering Proton Therapy System

B Tonner*, C Curley, G Georgiev, D Murff, P Marvin, R Tokarz

Ackerman Cancer Center, Jacksonville, FL

Purpose: All proton therapy plans for treatment at our institution using the Mevion S250 passive scattering therapy system are verified for correct monitor units by patient‐specific dosimetric measurement. We describe software methods that are designed to streamline the workload, improve accuracy, and reduce the possibility of errors in the overall process of plan quality assurance and monitor unit determination.

Methods: The RayStation version 6 treatment planning system (TPS) incorporates a complete implementation of the Python scripting language. The scripting interface allows for directly writing custom spreadsheets that are used in each of the QA measurements. The measurements are performed using the IBA MatrixX detector and “myQA” software. To aid in the accuracy of the measurements, a predictive starting monitor unit is extracted from the TPS. A large set (over 500) of measurements were made, covering all of the 24 beam options, to determine the correlation of the plan “meterset” with the actual, measured, clinically valid monitor unit.

Results: The relationship between the treatment plan meterset and the clinically validated treatment monitor unit has been shown to be linear with R‐factor's in excess of 0.999. The measured linear function is incorporated into the software that generates the patient QA spreadsheet, to give a starting point for the QA measurement that will be close to the final value, improving the accuracy of the measurement. A limitation of the commercial software is that regions‐of‐interest (ROI's) used in analysis can only be rectangular in shape. To overcome this limitation, and provide additional analysis information, a software tool was written that can compare the measured and TPS generated dose distributions with an arbitrary, user defined, ROI.

Conclusion: Using custom produced software can improve both the efficiency and accuracy of patient‐specific quality assurance of double‐scattering (aperture and compensator) treatment plans.

PO‐BPC‐Exhibit Hall‐15

MRIdian Linac Daily QA

P Yadav*, K Mittauer, K Vredevoogd, J Bayouth

School of Medicine and Public Health UW‐Madison, Madison, WI

Purpose: MRgRT is an emerging technology and requires establishment of QA procedures that support the novel types of therapy being provided with the MRIdian Linac. The purpose of this study is to develop and validate a series of daily radiation dose and imaging QA tests to be performed on the MRIdian Linac system to establish and maintain an effective and safe MRgRT. These highly sensitive tests evaluate the precision of MR and RT isocenter alignment, spatial fidelity imaging and patient handling systems, and the performance of the linear accelerator.

Methods: On daily basis safety interlocks, laser alignment with known offset from isocenter, imaging, and treatment isocenter coincidence, radiation dose output, couch motion, and small‐field MLC positioning as commonly required for IMRT are tested. The ViewRay daily QA phantom is used for all tests except safety interlock tests. At MR and RT isocenter coincidence, the ViewRay daily QA phantom is irradiated with an extremely small sensitive field of 0.4 × 0.83 cm2. This test requires submillimeter agreement MLC positioning, beam steering, sub‐degree gantry angle accuracy, to provide a radiation dose within ±10% of the expected dose. Radiation dose and dose rate constancy of the linear accelerator are measured using standard field size of 9.96 × 9.96 cm2, with a specification of ±3%.

Results: Over 4 months period the radiation output was within ±3%. Our monthly output results were compared with daily output and were within 0.5%–2%. Ratio of daily dose and reference dose for small field size shows that MLC position is within set spec of 2 mm with random variation which might be due to setup errors.

Conclusion: Tests performed on daily basis help us to keep track of key component such as linac, couch, MLC, MR imaging system which are critical for accuracy of MRgRT.

PO‐BPC‐Exhibit Hall‐16

Evaluation of Backscatter Factors (Bw) and Percentage Depth Dose (Pdd) Data in TG‐61 Protocol and BJR Supplement 25 for Superficial X‐Ray Radiotherapy

Oliver Chan*, Michael Lee

Pamela Youde Nethersole Eastern Hospital, Hong Kong

Purpose: In TG‐61 protocol and BJR Supplement 25, only the first HVL was used to tabulate the correction factors, for example, backscatter factors Bw, without considering other factors that could affect the beam quality. The purpose of this work was to i) measure the percentage depth dose Pdd of a machine with combination of tube potentials and HVL different from those in BJR Sup.25, and ii) evaluate the appropriateness of using published Bw for open‐ended cones to derive cutout factor for regular field defined with lead cutout.

Methods: Pdd of Filter7 (120 kV, 5.15 mmAl), Filter8 (140 kV, 8.03 mmAl), and Filter9 (150 kV, 1.01 mmCu) of the Xstrahl 150 units were measured using NACP chamber and Markus chamber, in water phantom (iba blue phantom) and Solid water slabs (Gammex RMI‐457), respectively. Cutout factor for regular field was measured using Markus chamber in Solid water slabs under lead cutout with opening from 2 to 14 cm diameter. The result was compared to cutout factors calculated with published Bw for open‐ended cones in TG‐61 protocol.

Results: There was a considerable difference between the Pdd data by measurement and those in BJR Supp.25. The difference is less than 3% in general. However, the difference of some data points at shallow depth was up to about 10%. The difference between the cutout factor for regular field defined with lead cutout (opening down to 2 cm in 15 cm cone) by measurement and that derived from open‐ended cones in TG‐61 protocol could be up to 3.5% for different tube potentials and HVL combination.

Conclusion: Since TG‐61 protocol and BJR Supp.25 are still the important references for kilo‐voltage beam data commissioning, the Bw and Pdd data in these documents should be used with awareness of the size of lead cutout, and combination of tube potential (kV) and HVL when calculating treatment time for superficial x‐ray radiotherapy.

PO‐BPC‐Exhibit Hall‐17

A Method for Measuring the Setup Accuracy of An IGRT System for Single‐Isocenter, Multiple Target SRS Deliveries

E Gete*, K Luchka, A Mestrovic, B Gill

BC Cancer, Vancouver Center, Vancouver, BC

Purpose: A specialized Hidden Target Test HTT was developed to measure the targeting accuracy of a patient positioning system used for single‐isocenter multiple target Stereotactic Radiosurgery SRS treatment deliveries.

Methods: The HTT measurements were made using an anthropomorphic head phantom, into which six radio‐opaque markers were inserted simulating treatment targets within the head. The treatment procedure consisted of fixation mask construction, acquisition of reference CT images, treatment planning and delivery. The treatment plan was generated in Eclipse and contained static fields defined by a multi‐leaf collimator (MLC) with sizes ranging from 7.5 × 7.5 mm2 to 15.0 × 15.0 mm2 with each field exposing a target marker. The plan was delivered using a TrueBeam STx linac equipped with a six degree of freedom (6DoF) couch and On‐board Imaging System (OBI) capable of acquiring cone beam CT (CBCT) and 2D kV projection images. Acquired images were registered with a reference planning dataset to correct translational and rotational misalignments of phantom position to isocenter. The positional accuracy of the system was evaluated by comparing the portal images of the multi‐port treatment fields acquired during delivery with the corresponding digitally reconstructed radiographs calculated from the planning CT.

Results: For the CBCT setup, deviations of the six target markers from their expected locations ranged between 0.3 and 0.8 mm, with mean deviation of 0.5 ± 0.2 mm. For the setup with orthogonal 2D x‐ray pairs, deviations ranged between 0.2 and 1.2 mm, with mean deviation of 0.7 ± 0.2 mm.

Conclusion: This work indicates that a positional accuracy of <1.0 mm can be achieved for cranial off‐axis targets when corrections are made using CBCT imaging in combination with the 6DoF couch.

PO‐BPC‐Exhibit Hall‐18

Statistical Evaluation of the Isocenter Dose Measurement Option for ArcCHECK CavityPlug Based Patient Specific IMRT/VMAT QA

T Ren^{1,2, *}, J Bourland^1,2

(1) Wake Forest School of Medicine, Winston Salem, NC, (2) Wake Forest University, Winston‐Salem, NC

Purpose: To determine the role of ionization chamber IC dose measurements for IMRT/VMAT patient specific quality assuranceQA using the ArcCHECK CavityPlug™ system Sun Nuclear Corp, Melbourne, FL.

Methods: Results for a total of 557 IMRT/VMAT QA cases from the past 365 days were analyzed. For IC measurements the ArcCHECK phantom was shifted to place the IC in a low gradient, high dose region found from the expected dose map from the treatment planning system (TPS). IMRT/VMAT QA plans were generated in three commercial TPSs, Philips Pinnacle³, RaySearch Laboratories RayStation and Elekta Monaco. The VMAT plans had more than 1 arc. IMRT/VMAT QA plans were measured on four linear accelerators: Elekta Axesse, Infinity and Versa HD and a Varian SC2100, using the ArcCHECK phantom with inserted small‐volume (0.05 cc) Capintec PR‐05P ion‐chamber (PTW, Freiburg, Germany) calibrated by an ADCL. Measurements and Gamma (Γ) analysis were recorded and performed using the SNC patient software with distance to agreement (DTA) and gamma (γ) analysis. A Γ pass was set to be ≥90% for 3 mm/3%, with threshold (TH) greater than 10% of the maximum dose. An IC measured dose pass was set to be ≤5% difference, compared to the TPS expected dose at the isocenter.

Results: For the two criteria of Γ analysis and dose comparison, none of the cases passed Γ and failed IC dose comparison, 93.18% passed both criteria, 6.10% failed Γ and passed IC dose comparison, and a small fraction, 0.72%, passed neither Γ nor the IC dose comparison.

Conclusion: For patient specific IMRT/VMAT QA in our clinic, an independent dose measurement, obtained by an independent IC and the ArcCHECK system, alone never vetoes the QA pass criteria. An independent IC dose measurement does provide an additional dose metric beyond the entrance and exit dose from the ArcCHECK diodes.

PO‐BPC‐Exhibit Hall‐19

Dosimetric Impact of Diaphragm Motion and Dynamic MLC Interplay in Lower Thoracic Spine Radiosurgery Using VMAT Technique

B. Zhao*, N. Wen, I.J. Chetty, K.C. Snyder, Z. Sun, M.S. Siddiqui, Y. Huang

Henry Ford Health System, Detroit, MI

Purpose: VMAT technique with full gantry rotations has become popular in spine radiosurgery SRS. In most institutions, free‐breathing CTs FBCT are utilized for treatment planning. However, for treatments at the level of diaphragm, the diaphragm motion may have an impact on dose delivery accuracy. The aim of this study was to evaluate the dosimetric impact of breathing motion in lower thoracic spine SRS.

Methods: Four representative patients with FBCT and 4DCT datasets were selected for this preliminary study. For each case, an 18 Gy VMAT plan composing of 2–3 full arcs was generated. The dose distribution was re‐calculated on average CT (AVGCT) and two extreme phases (PH0%CT and PH50%CT, to represent the possible scenarios of FBCT). In addition, each arc was divided into 64–155 sub‐arcs based on total monitor units, dose‐rate, and timestamp relative to the breathing trace. Dose distributions from sub‐arcs were recalculated onto corresponding phased CT images. 4D plans, which were summations of sub‐arcs doses, were generated to simulate the delivered dose. Dose‐volume metrics evaluated were D95%, V(Rx) for PTV, and D10%, D0.035 cc for cord.

Results: No correlation was observed between the dose‐volume metrics calculated on FBCT and other images (PH0%CT, PH50%CT, AVGCT). When comparing doses calculated on FBCT, PH0%CT, and PH50%CT to AVGCT, the largest changes in D95% and V(Rx) were 1.0 Gy and 40.7%, respectively. In contrast, the largest differences between 4D and AVGCT plans were 0.02 Gy and 0.75%. For cord dose, average changes (±standard deviation) in D10% and D0.035 cc were 0.2 ± 0.2 and 0.3 ± 0.3 Gy, respectively.

Conclusion: Due to diaphragm‐induced motion, the use of FBCT for VMAT planning in lower thoracic spine SRS sometimes leads to large deviations between planned and delivered dose. Dose calculation on AVGCT was found to be consistent to that on 4DCT, thus planning on AVGCT is recommended if diaphragm is in arc's path.

PO‐BPC‐Exhibit Hall‐20

Electron Treatment Plan Comparisons Between 3D‐Electron Monte Carlo Dose Calculations

R George*, P Myers, D Saenz, K Rasmussen, N Kirby, S Stathakis

University of Texas Health, San Antonio, TX

Purpose: Electron pencil beam dose calculation algorithm is used for by many treatment planning systems (TPS) for electron therapy treatments. Recently, more TPS have implemented Monte Carlo (MC) dose calculation for electron therapy treatments. The purpose of this work was to retrospectively evaluate possible dosimetric differences between pencil beam and Monte Carlo calculations of previously treated patients.

Methods: Twenty previously treated patients with electrons were included in this study. The electron energies used for the treatment ranged from 6 to 15 MeV. The treatment sites included sinuses, ears, testicles, skin, and intra‐mammary nodes. All patient treatment plans were initially created using the Pinnacle TPS with the 3D Electron dose calculation algorithm. The plans were then re‐calculated using the Monaco TPS with the electron Monte Carlo. The dose grid resolution was kept to 3 mm³ for all calculations. For the Monaco calculations, the number of histories was set to 50,000cm². The doses were compared using DVH, and differences within the isodose lines.

Results: Differences between the two dose calculation algorithms were observed for the majority of the cases. The dose differences were within 1% for cases with minimal tissue inhomogeneities. Larger differences of more than 3% were observed where inhomogeneities were present (sinus, skin folds, skull). The DVH analysis showed higher doses for most of the organs when calculated with Monaco as well as broader and deeper dose distributions. In the cases of photon‐electron beam matching, the dose on the photon side increased by 2%–5%.

Conclusion: It is important to carefully consider the electron dose calculation accuracy when the treatment site has tissue inhomogeneities, as in the case of air‐bone‐tissue interfaces. When photon beams and electron beams are matched, consideration should be given to the hotspot in the photon side and if necessary measurements should be made to verify the dose.

Research is partially supported by ELEKTA, Inc.

PO‐BPC‐Exhibit Hall‐21

Ideal Planning Assistant Application for Guidance in Radiation Treatment Planning

C Kabat*, N Papanikolaou, S Stathakis

University of Texas Health Science Center San Antonio, San Antonio, TX

Purpose: Intensity‐modulated radiation therapy, rotational or helical delivery technologies were developed to focus on maximizing dose to target volume while sparing health tissue. To achieve this, the dose distribution to surrounding healthy tissues is determined by organ type and physician constraints and published normal tissue tolerance criteria. With defined constraints, dosimetrists optimize each patient's plan to meet these objectives. However, for unidentified organs or healthy tissue for which no explicit dosimetric constraints are defined optimization procedure might omit them, allowing high levels of dose to be unnecessarily deposited. To enhance plan development and reduce dose to healthy tissue an application was developed that provides dosimetrists and physicians a means of outlining and working with an ideal patient plan.

Methods: An application was developed that incorporates the planning CT set with structures and calculates an ideal theoretical dose distribution for a patient plan using real machine PDD parameters. Each structure is capable of having multiple objectives that are manipulated in real time to develop an ideal plan based on physician's preferences. A few historical patient plans from the head & neck, thorax, and pelvis regions were outlined using the application and then utilized in the Pinnacle TPS to compare to original plans.

Results: The application demonstrated the ability to create effective plans quickly and provided an outline defining which OARs or health tissue areas allowed for greater sparing without impacting PTV coverage. Historical plan comparisons showed the capability of developing plan outlines that could be used in developing real patient plans that require less time and have preferred DVHs over historical.

Conclusion: The application has demonstrated its practicality when compared to historical plans and provides insight into coverage and acceptable dosimetric criteria. Implementations are being made to include NTCP/TCP values and plan achievability.

PO‐BPC‐Exhibit Hall‐22

Characterization of Optically Stimulated Luminescent Dosimeters in a Simulated Fluoroscopy Beam

N Quails*, C Schaeffer, A Heshmat, N Correa, C Olguin, I Barreto, M Arreola, L Rill

University of Florida, Gainesville, FL

Purpose: A prior characterization of Optically Stimulated Luminescent Dosimeters OSLDs in a simulated fluoroscopy beam only assessed angular response in‐air and at the center of a cylindrical CTDI head phantom. Considering the goal to assess radiation dose to the skin and lens of the eye, OSLDs angled on the surface of a patient's skin must be assessed as to whether they need an angular correction factor, based on the different geometry and scatter conditions.

Methods: OSLDs were placed in strips of two rows perpendicular to the simulated fluoroscopy beam anode‐cathode axis. The strips were placed at ten degree intervals with respect to the plane perpendicular to the x‐ray beam direction on the surface of a cylindrical CTDI head phantom. A 0.6 cc ion chamber was placed next to the strips in the radiation field for a reference measurement. Three sets of strips and the ion chamber were exposed to an 80 kVp beam, 630 mA, and 1000 ms x‐ray beam.The OSLDs were geometrically corrected back to the distance of the ion chamber from the focal spot.

Results: OSLDs after being corrected for beam quality yielded correction factors ranging from 0.87 to 1.32 (σ = ±0.13). The maximum correction factor comes when the OSLDs are at 90° with respect to the beam. The minimum correction factor is where the OSLD is perpendicular to the beam (0°).

Conclusion: Angularity correction factors were generated for the use of OSLDs in interventional fluoroscopy rooms for the purpose of placing on patient skin for skin and lens of the eye dose measurements in‐vivo.

PO‐BPC‐Exhibit Hall‐23

Using Modulated and Reflected Light to Correct Clinical Cherenkov Images for Patient‐Specific Tissue Variations During Whole Breast Radiotherapy

R Hachadorian^1,*, P Bruza², L Jarvis³, D Gladstone⁴, B Pogue⁵

(1) Dartmouth College, Hanover, NH, (2) Thayer School Of Engineering, Dartmouth College, Hanover, NH, (3) Dartmouth‐Hitchcock Med. Ctr., Lebanon, NH, (4) Dartmouth‐Hitchcock Medical Center, Lebanon, NH, (5) Thayer School Of Engineering, Dartmouth College, Hanover, NH

Purpose: Imaging Cherenkov emission during radiotherapy has been developed to establish real‐time, treatment field verification in vivo. Signal linearity between Cherenkov emission and absorbed dose exists, which yields potential for establishing quantitative dose verification. However, intensities vary as much as 45% between patients due to differences in tissue optical properties, alone. Our goal is to correct Cherenkov images for large scale differences in absorption, including those due to variable melanin concentrations, and subcutaneous vasculature.

Methods: Cherenkov emission was imaged for both calibration data and patient treatment data, using an intensified CMOS camera on a Varian Clinac with online background subtraction. Optical property maps were acquired using spatial frequency domain imaging (SFDI) over five spatial frequencies and four NIR wavelengths, then optimized for correction by spectral weighting. Reflectance images were taken using the iCMOS camera, where an Align RT system integrated into clinical workflow was used as a light source. Corrections to patient data were applied using a previously established calibration.

Results: When using SFDI as a means for subsurface vasculature correction, the percent differences associated with the corrected image lateral, inferior and medial vessels (6%, 4%, and 7% respectively) adhered more closely to the intensity gradient associated with the treatment plan, compared to that of the uncorrected image (22%, 14%, and 10% respectively). Using optimization, corrected image intensity absorbed due to areolar pigment improved almost completely, relative to the treatment plan, as compared to the uncorrected image. Establishing an Align RT reflectance‐based calibration for variable skin tones allowed us to normalize patient data.

Conclusion: Using Align RT light as a source for reflectance images allowed us to characterize patient data for large‐scale tissue optical properties and SFDI optical property maps have enabled the development and application of a pixel‐by‐pixel method to correct Cherenkov images for surface and subsurface attenuation.

Brian Pogue (PI) is the president and co‐founder of DoseOptics LLC, a company which designs and manufactures Cherenkov Cameras. This work was not financially supported by DoseOptics.

PO‐BPC‐Exhibit Hall‐24

Pre‐ and Post‐ESD in Clinicopathologic Analysis in Early Gastric Cancer

Y Gao*

Hebei General HospitalShijiazhuang

Purpose: We aimed at investigating clinicopathological and endoscopic characteristics of early gastric cancer and precancerous lesions, comparing differences in diagnosis between endoscopic forceps biopsy EFB and endoscopic submucosal dissection ESD, and identifying risk factors of under‐diagnosis of high‐grade intraepithelial neoplasia HGIN/carcinoma for a better guidance of the ESD decision‐making process.

Methods: From June 2015 to November 2018, the 60 patients with gastric mucosal HGIN, and early gastric carcinoma (EGC) which were treated by ESD were included. We reviewed the relationship between lesion area, expression of white light endoscopy, magnification endoscopy, narrowband imaging endoscopy, and postoperative pathological classification and depth of lesion was analyzed. Differences in those features between EFB and ESD were compared and risk factors for under‐diagnosis by EFB were analyzed.

Results: There were 60 patients with a total of 62 lesions. There was no statistically significant difference between pathological type and lesion part, lesion size. There were statistically significant difference between LGIN and HGIN, severe dysplasia and early gastric cancer in clear boundary, irregular capillaries, and irregular surface microstructure. Although concordant in most (91.94%) cases between EFBs and ESDs, pathological diagnoses in five (8.06%) cases were upgraded in ESDs. Compared to the concordant group, the lesion size is more than 2 cm, and depressed, excavated, and ulcer patterns were significantly more frequent in the upgraded group.

Conclusion: Magnification endoscopy combined with NBI can find early detection of HGIN and early gastric cancer. Preoperative biopsy had guidance significance in diagnosis and treatment, which can reduce the missed diagnosis of early gastric cancer. Lesion size more than 2.0 cm and the depressed pattern at initial EFB were independent risk factors for pathologic upgrade to advanced diseases in ESD.

PO‐BPC‐Exhibit Hall‐25

Commissioning a 2.5 MV Portal Imaging Beam in the Eclipse Treatment Planning System

W Ferris*, W Culberson, Z Labby

University of Wisconsin‐Madison, Madison, WI

Purpose: To test the feasibility of a commercial treatment planning system to model the TrueBeam 2.5 MV imaging beam and verify the accuracy of Acuros XB and the Anisotropic Analytical Algorithm AAA. This beam has not been modeled by any commercial TPS in the literature.

Methods: Standard beam commissioning data were measured for a 2.5 MV beam line. Both algorithms were commissioned in Eclipse using these data. A diode was used to acquire dose profiles at one depth for jaw‐collimated fields and spot sizes were tuned to these profiles. Both beam models were validated using MPPG 5a recommendations. Three heterogeneity cork setups were used, with varying field sizes and cork thicknesses. The ratios of dose above the cork to dose below the cork was compared between measured and computed.

Results: The TPS was able to model the low energy imaging beam. The optimal algorithm spot sizes differed substantially from those in typical models of therapeutic beam energies. The mean energy along the CAX for the 2.5 MV beam was 0.49 MV. For MPPG 5a Tests 5.4–5.7, the gamma pass rate was greater than 95% for a 2%/2 mm criteria for 21 of 23 profiles and PDDs for Acuros and 18 of 23 for AAA. Both algorithms underestimate out‐of‐field dose. For MPPG Test 6.2, the per cent difference of the ratio of dose above cork to dose below cork between measured and Acuros was less than 1.5% for all three setups. The per cent difference between measured and AAA was greater than 3% for all three setups.

Conclusion: Both Acuros and AAA can be used to model a 2.5 MV imaging beam. In general, Acuros passes gamma analysis with slightly higher passing rates than AAA for basic photon tests. Acuros performs significantly better than AAA for heterogeneity calculations.

PO‐BPC‐Exhibit Hall‐27

Improve the Dosimetric Outcome in Bilateral Head Neck Cancer HNC Treatment Using Spot‐ScanningProton Arc SPArc Therapy: A Feasibility Study

G Liu, A Qin, X Li, D Yan, C Stevens, P Kabolizadeh, X Ding*

Willam Beaumont Hospital, Royal Oak, MI

Purpose: To exploit the dosimetric improvement, delivery efficiency, and plan robustness for bilateral head and neck cancer(HNC) treatment utilizing a novel proton therapy technique – the spot‐scanning proton arc (SPArc) therapy.

Methods: Ten bilateral HNC patients were retrospectively evaluated. Both SPArc and 3‐field robust optimized Intensity Modulated Proton Therapy (RO‐IMPT) plans were generated using the same robust optimization parameters (±3.5% range and 3 mm setup uncertainties). The prescription dose was 7000 cGy [RBE] for CTV_high and 6000 cGy [RBE] for CTV_low. Clinically significant dosimetric parameters and potential clinical benefit for parotid glands were extracted and compared. Root‐mean‐square deviation doses (RMSDs) was used to evaluate the plan robustness. Total treatment delivery time was estimated based on a full gantry rotation with 1 rotation‐per‐min, 2 ms spot switching time, 0.01 minimum spot monitor unit per spot, and energy‐layer‐switching‐time (ELST) from 0.1 to 5 s. Statistical analysis was evaluated using paired‐t test.

Results: The SPArc plan was able to provide equivalent or better robust target coverage while demonstrating significant dosimetric improvements over RO‐IMPT in most of OARs sparing. More specifically SPArc reduced the mean dose of ipsilateral parotid (P < 0.001), contralateral parotid (P < 0.001), and oral cavity (P < 0.001) by 27.1%, 28.1%, and 32.6% respectively compared to RO‐IMPT. The D1% of brain stem and cord were also reduced by 21.3% (P = 0.003) and 6.2% (P = 0.276) using SPArc, respectively. SPArc reduced the dose uncertainties in cord and ispilateral parotid 119.6 cGy [RBE] versus 91.3 cGy [RBE] (P = 0.017), 268.2 cGy [RBE] versus 243.6 cGy [RBE] (P = 0.027), respectively. Based on ELST of 0.1 s, SPArc was comparable in average total estimated delivery time (284.0 s vs 304.5 s P = 0.244). SPArc will decrease the mean probability of salivary flow dysfunction by 8.6% (P < 0.001).

Conclusion: SPArc could significantly reduce the dose to OARs (parotids and oral cavity etc.) while providing a similar or better robust target coverage compared with RO‐IMPT. SPArc could potentially achieve comparable treatment delivery time with ELST less than 1 s.

PO‐BPC‐Exhibit Hall‐28

Geometric and Dosimetric Differences Due to Image Registration Workflows for Frameless GammaKnife Icon Patients

E Hubley*, K Mooney, W Shi, Y Yu, H Liu

Thomas Jefferson Univ. Hospital, Philadelphia, PA

Purpose: We investigate the geometric and dosimetric differences between two registration workflows for the Leksell GammaKnife^® Icon™: registering the planning MRI image directly to the on‐board CBCT for stereotactic definition per the vendor recommendation, and an alternate workflow where the MRI is registered to a diagnostic CT, which is then in turn registered to the stereotactic CBCT. We aim to identify situations which are more susceptible to these differences.

Methods: For five patients, eight 4 mm diameter and one 0.14 mm diameter spherical lesions were contoured on the 3D T1 MRI scan. For each patient, the MRI was registered to the diagnostic CT, which was in turn registered to the CBCT. A single‐shot plan was created; shot location and prescription isodose were selected to maximize conformity index while maintaining 100% PTV coverage. The MRI was then fused directly to the CBCT, and the plan was copied to the new registration. All registrations were done using GammaPlan®'s automatic registration process using a volume of interest encompassing the entire CBCT volume (whole skull). Geometric differences between the target center locations and the resulting changes in PTV coverage were recorded.

Results: The mean 3D displacement of target centers between registrations was 0.4 ± 0.2 mm (max = 0.9 mm). The mean decrease in target coverage was 3.0 ± 2.4% (max = 10%). The targets exhibiting largest displacements and most coverage loss were the two most posteriorly located. The patient with the largest displacements (0.6 ± 0.2 mm) and most coverage loss (5.0 ± 2.9%) had limited shoulder clearance which prevented the CBCT from capturing the base of skull region, potentially reducing the registration accuracy.

Conclusion: Geometric and dosimetric differences exist between the two registration workflows. Preliminary data indicate patients with posteriorly located targets or limited CBCT FOV may be more susceptible to CBCT‐to‐MR registration uncertainties.

PO‐BPC‐Exhibit Hall‐29

Robustness Evaluation of a Neural Network‐Based Photon Beam Profile Deconvolution Method

K Mund*, G Yan

University of Florida, Gainesville, FL

Purpose: The authors have previously shown that neural network NN can successfully perform photon beam profile deconvolution–the elimination of volume average effect VAE of scanning ionization chambers IC. The purpose of this work was to evaluate the robustness of the method when applied to ICs of various sizes and beams of different modalities FF & FFF.

Methods: The proposed NN has an input, hidden, and output layer. It inputs data extracted from cross beam profiles using a sliding window, and outputs deconvolved data at the center of the window. Cross beam profiles of fields ranging from 2 × 2 to 10 × 10 cm² were measured with CC04, CC13, Farmer chamber, and an EDGE diode detector for 6 MV FF and FFF beams. The profiles measured with each chamber were divided into training, validation, and testing sets to train and test a 3‐layer feed forward NN. The diode‐measured data were used as the reference. The sliding window length and the number of hidden neurons were optimized such that the same NN structure could be applied for all tested chambers, fields, depths, and beam modalities. The NN's performances were quantified by evaluating mean square error (MSE) and penumbra width difference (PWD) between the deconvolved and EDGE‐measured profiles.

Results: Excellent agreement between the deconvolved and reference profiles was achieved using a sliding window width of 15 and five hidden neurons for all the tested ICs and both beam modalities. The average PWD decreased from 2.70 ± 0.47, 2.66 ± 0.41, and 3.99 ± 0.42 mm to 0.09 ± 0.39, 0.03 ± 0.35, and 0.04 ± 0.38 mm for the CC04, CC13, and Farmer chambers, respectively.

Conclusion: We found that the NN‐based deconvolution method can be effectively applied to ICs of various sizes and beams of different modalities. Separate NNs are needed for different ICs but, for a specific IC, one NN works for both beam modalities.

PO‐BPC‐Exhibit Hall‐30

Dosimetric Comparison of Biologically Guided Radiotherapy and X‐Ray‐Guided Stereotactic Ablative Radiotherapy for Oligometastatic Prostate Cancer

WT Hrinivich^1,*, R Phillips¹, AJ Da Silva², N Radwan¹, MA Gorin³, SP Rowe⁴, KJ Pienta³, MG Pomper⁴, J Wong¹, P Tran¹, K Wang¹

(1) Department of Radiation Oncology, Molecular Radiation Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, (2) RefleXion Medical Inc., Hayward, CA, (3) Department of Urology, Johns Hopkins University School of Medicine, Baltimore, MD, (4) Department of Radiology, Johns Hopkins University School of Medicine, Baltimore, MD

Purpose: Prostate‐specific membrane antigen PSMA‐targeted PET tracers have improved early detection of oligometastatic OM prostate cancer. Evidence suggests that stereotactic ablative radiotherapy SABR for OM prostate cancer may improve clinical outcomes, but localization of multiple small lesions for treatment is challenging with conventional X‐ray‐based imaging systems, limiting deliverable doses. Biologically guided radiotherapy BgRT is being developed to utilize strong contrast of PET signal for real‐time intra‐fraction tracking of multiple targets. This study investigates PSMA‐directed BgRT using a cohort from our Phase II randomized trial of SABR to men with recurrent hormone sensitive OM prostate cancer.

Methods: Fifteen patients from our trial treated with SABR and imaged with PSMA PET‐CT were used to compare clinical SABR plans versus BgRT plans generated from a prototype treatment planning system (TPS). The TPS models a 6MV linear accelerator mounted on a ring gantry that includes a PET detector and high‐speed binary multi‐leaf collimator (MLC). The PET signal is used as fiducial to indicate target location. Photon fluence is thus optimized as a function of PET activity within a user‐defined “firing zone”. Clinical SABR and BgRT plans were compared in terms of maximum PTV dose (Dmax), mean dose to proximal organs at risk (DOAR), conformity index, gradient index, and the correlation of Dmax with maximum PET activity.

Results: Clinical SABR and BgRT plans resulted in mean ± standard deviation Dmax of 128 ± 11% and 150 ± 13% (P < 0.001), DOAR of 11 ± 8% and 10 ± 8% (P = 0.02), conformity indices of 0.74 ± 0.08 and 0.72 ± 0.08 (P = 0.4), and gradient indices of 4.47 ± 0.63 and 5.40 ± 0.83 (P = 0.003), respectively. Dmax and maximum PET activity had correlation coefficients of −0.15 (P = 0.6) and 0.59 (P = 0.02) for clinical SABR and BgRT.

Conclusion: BgRT plans resulted in increased PTV dose and decreased OAR dose while incorporating underlying PET activity to enable real‐time biological guidance, demonstrating the feasibility of PSMA‐directed BgRT for OM prostate cancer.

This work was supported in part by RefleXion Medical.

PO‐BPC‐Exhibit Hall‐31

Increase in Superficial Dose in Whole‐Breast Irradiation with Halcyon Linac Compared to Traditional C‐Arm Linacs: In‐Vivo Dosimetry and Planning Study

F O'Grady PhD*, A Barsky MD*, S Anamalayil MS*, G M Freedman MD*, C Kennedy PhD*, L Dong PhD*, J M Metz MD*, N K Taunk*, T Li*

University of Pennsylvania, Philadelphia, PA

Purpose: Superficial dose is an important parameter in breast cancer radiotherapy. When treated with conventional linacs, bolus is commonly applied to improve target coverage near the surface while also managing the risk of severe skin reactions and negative cosmesis. With the introduction of the new Halcyon™ linac with exclusively 6X flattening filter free FFF photon beams, the impact on superficial dose and the need for bolus in breast irradiation must be studied.

Methods: In‐vivo measurements of superficial dose were made with optically stimulated luminescence dosimeters on 11 breast cancer patients treated with the Halcyon™ 6X FFF linac. Additionally, measurements were made of 6X FFF and 6X beams delivered to an anthropomorphic phantom. A planning study was carried out in which 16 patients treated on the Halcyon were re‐planned with a conventional linac to determine the difference in superficial dose predicted by the treatment planning system (TPS). Measures were taken to optimize the accuracy of the TPS surface dose.

Results: The use of 6X FFF beams led to higher superficial dose compared with 6X beams. The in‐vivo measurements show an average superficial dose of 84.2 ± 0.5% which is an increase in approximately 13 % compared with published measurements for a 6X linac with flattening filter. Comparison of surface dose for 6X and 6X FFF beams in the phantom measurements show an increase from 70 % ± 1.3% to 84.7 ± 1.3%, which is consistent with the in‐vivo measurements. The planning comparison shows an increase in skin structure receiving >70% Rx from 59 ± 4.3 % to 78 ± 2.9 % for Halcyon compared with the standard c‐arm linac with flattening filter.

Conclusion: The use of 6X FFF beams were associated with substantially greater superficial dose and may obviate the need for bolus as used with traditional linacs.

JMM is on the advisory board of obtained personal fee from, and received grant funding unrelated to this work from Varian Medical Systems, LD received grant funding unrelated to this work from Varian Medical Systems, and CK received honoraria for speaking engagements from Varian Medical Systems.

PO‐BPC‐Exhibit Hall‐32

Measurement of Output Factors with Various Detectors in the Presence of the ViewRay MRIdian Magnetic Field

B Barraclough*, D Dunkerley, K Vredevoogd, B Paliwal, K Mittauer, P Yadav

School of Medicine, Public Health UW‐Madison, Madison, WI

Purpose: Many detectors have been already been examined for their ability to accurately measure OFs in standard and small fields. However, the introduction of magnetic fields into the treatment field necessitates a re‐examination of these detectors. In this work, the ability of a variety detectors to measure OFs in the presence of the magnetic field of a ViewRay MRIdian Linac was examined.

Methods: OFs were measured at 5 cm depth in either a 1D water phantom or a solid water phantom. The phantom was placed at 75 cm SSD with field sizes ranging from 0.415 × 0.415 cm² to 25.5 × 25.5 cm² and readings normalized to the 9.96 × 9.96 cm² field. The magnetic field strength was 0.35T. The detectors examined were the SunNuclear Edge diode, Standard Imaging Exradin W2 scintillator, and Best Medical MOSFET. These were compared to Gafchromic EBT3 radiochromic film. Additionally, all detectors were compared with Monte Carlo (MC) calculated OFs from ViewRay's TPS.

Results: OFs measured by all detectors were within 2% of the film OFs. The largest discrepancies were for the diode and MOSFET at the smallest field size measured by film, 0.83 × 0.83 cm². They were 1.9% and 1.4% larger than the radiochromic film, respectively. The W2 at this field size was within 0.6%. Its greatest difference from film was in the 3.32 × 3.32 cm² field, where it was 1.0% lower. The MC calculation's largest deviation was 1.0% higher than film at the 0.83x0.83 cm2 field size. Comparing each detector against the MC calculation, the film, diode, and W2 were all within 1.6%. The MOSFET had a 2.2% difference at the 12.45 × 12.45 cm² field.

Conclusion: The detectors examined accurately measured output factors for ViewRay's MRIdian Linac. The 0.35T field has a minimal effect on the measurement of OFs. The W2 scintillator measured OFs most similar to the film measurements. Agreement of all detectors with MC calculations was within 2.2%.

PO‐BPC‐Exhibit Hall‐33

An Open‐Source Tool to Visualize Potential Cone Collisions While Planning SRS Cases

A Licon^1,*, A Alexandrian², N Kirby¹

(1) University of Texas HSC SA, San Antonio, TX, (2) Seattle, WA

Purpose: To create an open‐source visualization program that allows one to find potential cone collisions while planning stereotactic radiosurgery cases.

Methods: Measurements of physical components in the treatment room (gantry, cone, table, localization SRS frame, etc.) were incorporated into a set of scripts in MATLAB (MathWorks, Natick, MA) that produce 3D visualization of the components. The localization frame fully contains the patient so it was used to represent patient collisions. To approximate simulated components graphically, a box model was used. The gantry, couch, cone, and SRS localization frame were all represented by boxes with the measured dimensions. The equations used to create the boxes allowed for accurate representation of motion in a treatment room and related all motion to the machine isocenter. It allowed for the target to be moved in the three translation directions, the couch to be changed by 180°, and for the gantry to move a full 360°. A simple graphical user interface (GUI) was made in MATLAB to allow users to pass the target coordinate (vert, lat, long) relative to the localization frame, the initial and terminal gantry and couch angles, and the number of angular points to visualize between the initial and terminal gantry angle.

Results: The GUI provides a fast and simple way to discover collisions in the treatment room before the treatment plan is completed. By inputting the treatment isocenter the GUI shifts the localization box to the appropriate location, symbolized by a diamond as a visual assurance.

Conclusion: This simple GUI can be used by the planner, physician, or physicist to find the best orientation of beams for each patient. By finding collisions before a plan is being simulated in the treatment room, the clinic can save time due to replanning of cases.

PO‐BPC‐Exhibit Hall‐34

Comparison of Collapsed Cone Convolution Superposition and Monte Carlo Dose Calculations for Spine SBRT Dose

H Parenica*, D Saenz, N Papanikolaou, S Stathakis

UT Health San Antonio, San Antonio, TX

Purpose: To quantify and compare the effect of calculating dose using a Collapsed Cone Convolution Superposition CCCS algorithm and a Monte Carlo algorithm for spine patients being treated with SBRT.

Methods: Four spine patients (two thoracic, two lumbar) were included in this study. All patients were prescribed 1600 cGy over one fraction. Plans were optimized in the Pinnacle3 TPS and dose was calculated using a CCCS algorithm. Plans were then exported to the Monaco TPS and the dose was recalculated using a Monte Carlo algorithm. Comparison metrics included Conformity Index (CI), Conformal Index (CoIn), Gradient Index (GI), maximum spinal cord dose, and PTV D98 (near minimum) and D2 (near maximum) values.

Results: For all cases, the Monte Carlo algorithm in Monaco showed a lower overall PTV coverage than what the CCCS algorithm in Pinnacle3 initially calculated. Lower dose regions were not affected. The average CI for all patients was 0.947 and 0.808 for Pinnacle and Monaco, respectively. The average CoIn for all patients was 0.837 and 0.736 for Pinnacle and Monaco, respectively. For all evaluation metrics and for all patients, the Monte Carlo algorithm in Monaco calculated a lower dose than the CCCS in Pinnacle3.

Conclusion: Monte Carlo remains in the most accurate of dose calculation algorithms. SBRT planning and delivery require a higher level of precision and accuracy than conventional radiation therapy. These results indicate the possible need for Monte Carlo calculations as a second check for spine SBRT patients to ensure adequate target coverage.

This research is supported by the CPRIT Research Training Award (RP 170345).

PO‐BPC‐Exhibit Hall‐35

A Comparison Between Traditional and Continuous Delivery Methodologies for the Gamma Knife Perfexion

S Blake^1,*, D Check²

(1) Hux Cancer Center, Terre Haute, IN, (2) Nicholasville, KY

Purpose: By utilizing multiple shots in Gamma Knife radiosurgery, one can improve the dose distribution Cheek et. al 2005. Traditionally, a single shot is delivered and the radiation is suspended while transitioning to the next shot. If a high number of shots are intended to be delivered, then the transitioning time in between shots can add to the overall treatment time. The purpose of this study is to investigate the differences between a traditional delivery method versus a continuous delivery.

Methods: An optimized Gamma Knife plan was developed for a 4.3 cc lesion that utilized 13 8 mm and two 16 mm shots. A traditional delivery dose distribution was calculated. For the continuous shot dose distribution, a custom program was written that determines a minimal travel path, discretizes the path in 1 mm increments, and calculates the resulting dose distribution for the continuous shot movement. The two resulting dose distributions were compared.

Results: For the 15‐shot plan, traditional transitioning time between shots was calculated to be approximately 46 s. The prescription coverage, conformity index, and gradient index for the two delivery methods were approximately identical. The target maximum doses differed by 0.4% of Rx while the target minimum differed by 0.2% of the Rx.

Conclusion: Transitioning to a continuous delivery methodology can decrease treatment time while providing a comparable dose distribution. Preliminary data indicate that the continuous delivery methodology could allow for an increased number of shots. Previous studies have shown the dose distribution can be improved by delivering a high number of shots; therefore future studies will investigate increasing the number of shots.

PO‐BPC‐Exhibit Hall‐36

Development of a Stand‐Alone Comprehensive Collision Prediction System for Non‐Coplanar Radiotherapy

N Islam^1,2,*, J Kilian‐Meneghin^1,2, M Podgorsak^1,2

(1) State University of New York at Buffalo, Buffalo, NY, (2) Roswell Park Cancer Institute, Buffalo, NY

Purpose: The determination of potential collisions between a patient, treatment accessories, and linac mechanical components during treatment planning is important. Studies exploring the use of non‐coplanar 4π radiotherapy involving non‐standard treatment couch and gantry orientations have demonstrated significant dosimetric benefits. In this work, a stand‐alone collision prediction system CPS was developed using MATLAB. This system is based on a straight‐forward linac model, developed in this work, that can be used in conjunction with a treatment planning system to predict collisions in 4π treatments prior to delivery.

Methods: Physical dimensions of the outer mechanical components of a Varian TrueBeam were quantified to construct a geometric model of the linac. A combination of contours from the planning CT and 3‐D depth data acquired with a Kinect camera was used to develop a patient specific model. An algorithm that relies on a dot product between vectors denoting linac hardware and patient anatomy is used to predict collisions. Virtual collision test cases were experimentally verified in the treatment room using an ArcCheck phantom to simulate a patient.

Results: For a set of 52 different collision test cases, the positive predictive value and the negative predictive value of the CPS model were calculated to be 0.83 and 1.00, respectively. Validation experiments demonstrate that the CPS consistently estimates collisions at conservative distances from actual collision locations. The average difference between predicted and measured collision states was 2.14 cm for lateral couch movements. The predicted couch rotational position for a collision between the gantry and a patient analog differed from actual values on average by 3.4°.

Conclusion: This study outlines the development and clinical implementation of a solution to 4π radiotherapy collision management challenges during treatment planning. The CPS will be a valuable tool for treatment planners who seek to implement 4π radiotherapy.

PO‐BPC‐Exhibit Hall‐37

SRS Spine Treatment Planning Study on a Magnetic Resonance Image Guided Linear Accelerator

J Dolan*, J Kim, K Snyder, I Chetty, N Wen

Department of Radiation Oncology, Henry Ford Health System, Detroit, MI

Purpose: MRI guidance provides superior soft‐tissue contrast, making it increasingly appealing for radiation therapy requiring high localization accuracy. Enhanced visualization of tissues (e.g., CSF and spinal cord) during patient positioning enables more exact targeting and dose escalation. This work investigates treatment plan quality and deliverability of spine Stereotactic Radiosurgery SRS cases planned for delivery on an MR‐Linac.

Methods: In this retrospective study, six spine‐SRS patients were re‐planned on a commercial Monte Carlo‐based MR‐Linac TPS. The population included C/T/L vertebral bodies with varying degrees of epidural involvement. The treatment technique utilized 9‐10 posterior step‐and‐shoot IMRT beams. Treatment plans were optimized to achieve coverage, constraints, and quality criteria dictated by RTOG‐0631. Three cases, with varying degrees of modulation, were delivered on a solid water phantom. Deliverability and dosimetric accuracy were assessed via ion chamber and absolute film dosimetry patient‐specific quality assurance (PSQA) measurements.

Results: The MC‐based TPS provided IMRT plans that met all critical‐tissue constraints outlined in RTOG‐0631, while easily covering 90% of the target with the 18‐Gy single‐fraction prescription dose. Critical tissues encountered in the sampled population include the spinal cord, pharynx, esophagus, great vessels, stomach, kidneys, lungs, and liver. In all cases, the spinal cord was the most limiting organ, where the max dose encountered was kept below 10.55‐Gy. Plan quality metrics controlling hot spots and high‐dose spillage (e.g., <115% permitted outside the target) were achieved but provided the greatest planning challenge due to the nature of static beams. Efforts were made to minimize plan complexity (˜7‐segments/beam) and delivery time (˜24‐min). PSQA measurements proved accurate deliverability (average ion chamber agreement −2.34% and 3%/1 mm global gamma analysis 97.11%) and were within institutional tolerances.

Conclusion: This study demonstrated the feasibility of planning RTOG‐0631 compliant spine‐SRS treatments via an MC‐based TPS for delivery on an MR‐Linac. PSQA measurements validated deliverability and dosimetric accuracy.

The submitting institution holds research agreements with Philips Healthcare and ViewRay, Inc.

PO‐BPC‐Exhibit Hall‐38

A Simplified Total Body Irradiation Scatter Dose Calculation

N Astorga*, M Naessig, S Stathakis, D Saenz, K Rasmussen, P Myers, N Papanikolaou, N Kirby

University Of Texas Health, San Antonio, TX

Purpose: To commission a simplified total body irradiation TBI scatter dose calculation.

Methods: Our TBI patients are positioned semi standing for AP‐PA treatments and lying down in a crib for lateral and decubitus treatments at 340 cm source‐to‐axis distance (SAD). An extended SAD dose reference is utilized, which is measured with a stack of solid water at this SAD. Scatter, depth, and off‐axis corrections are made to this reference. The scatter correction is made with dose calculations after density overrides of elliptical contours, with various widths and thicknesses, relative to a contour of the solid water stack. The depth corrections were made with solid water based tissue phantom ratio measurements. Our TBI treatment field is divided into fixed sectors for the use of dose compensators. Off‐axis measurements were performed at these sectors across the field. End‐to‐end measurements were performed in solid water stacks and in an anthropometric phantom. Dosimeter measurements were performed on the phantom surface, to replicate the in‐vivo measurement setup for our patients, and additionally at the phantom midline. Our final dose equation included all correction factors previously mentioned and was computed in Excel from the measured patient width and thicknesses, which yielded the number of beam monitor units necessary for treatment and compensator thickness for each sector.

Results: Internal and external dose verification readings were within a 10 per cent dose deviation as recommended by TG‐29. Our previous dose calculation, which excluded off‐axis and patient scatter correction factors, had an average dose measurement deviation of six per cent. This same number is reduced to two per cent with the new technique.

Conclusion: The lateral scatter correction factor improved dose accuracy. We have implemented a simplified procedure for commissioning of a total body irradiation program with scatter corrections.

PO‐BPC‐Exhibit Hall‐39

Implementation of Risk‐Based Plan Evaluation and NTCP Estimation From Intermediate Dose Levels

C Velten, P Brodin, W Bodner, S Kalnicki, C Guha, M Garg, W Tome

(1) Montefiore Medical Center, Bronx, NY, (2) Albert Einstein College of Medicine, Bronx, NY

Purpose: To perform risk‐based plan evaluation within the Eclipse treatment planning system using normal tissue complication probability NTCP models and derive planning guidance based on intermediate sequential boost levels.

Methods: Utilizing the Varian Eclipse Scripting API we created a module implementing dose volume histogram manipulation including normalization to standard fractionation and reduction in organ‐equivalent doses (OED) and generalized‐equivalent‐uniform‐doses (gEUD). Lyman‐Kutcher‐Burman NTCP models (QUANTEC recommended) were used for rectum (D50 = 76.9 Gy) and small bowel (D50 = 59 Gy) organs at risk (OARs). We randomly selected 27 patients treated for prostate cancer with radiation therapy to 77.4 Gy in 1.8 Gy/fraction with 45 Gy to the pelvis followed by sequential boosts to the prostate and seminal vesicles. Volumetric modulated arc therapy was used for all patients. We calculated OEDs, gEUDs, and NTCP values in all plans and total dose plan sums. Spearman's rank correlation coefficients (ρ) and polynomial fits were calculated for the relationship between NTCP estimated from the total plan dose (NTCPx) and dose metrics from the sequential plans.

Results: Small bowel NTCP correlated most strongly (P < 0.01) with gEUD45 and Dmean45, with ρ = 0.91 and 0.82, respectively. For the sequential boosts (54 Gy/77.4 Gy) significant albeit weaker correlation was found with ρ = 0.56/0.57 and 0.72/0.74, respectively. Values of NTCP45 were within 8.3% ± 2.0% of NTCPx on average. Correlation between NTCPx and gEUD45 was significant (P = 0.03) with ρ = 0.41, while Dmean45 was not significantly associated. Stronger correlation (P < 0.01) was found for gEUD and Dmean on sequential boost plans with ρ = 0.76/0.72 and 0.64/0.66, respectively. Values of gEUD45 leading to 5% NTCPx were 29.5 Gy and 36.7 Gy for small bowel and rectum, respectively.

Conclusion: Our scripting module can be used for NTCP‐based plan evaluation. Using the correlation between gEUD45 and NTCPx we can derive gEUD targets that can be used in optimizing 45 Gy primary prostate plans to achieve a certain acceptable NTCP after sequential boosts to 77.4 Gy.

PO‐BPC‐Exhibit Hall‐40

Dosimetric Impact Associated with the New 25‐Source GammaPodTM Design

P Sabouri^1,*, Y Niu², C Yu², E Nichols¹, M Guerrero¹, B Yi1, S Becker¹

(1) University of Maryland School of Medicine, Baltimore, MD, (2) Xcision Medical Systems, Columbia, MD

Purpose: The GammaPod GPOD is a novel prone‐breast stereotactic radiotherapy device. Highly conformal plans are delivered using conical non‐coplanar arcs from Co‐60 sources with 18 or 25 mm collimation. The original GPOD design used 36 sources located between 18° and 53° below the horizontal. The new design uses a more lateral placement of 25 sources 25S spanning from 18° to 42° below the horizontal. We assess the dosimetric impact, delivery accuracy, and feasibility of the 25S‐GPOD.

Methods: 25S‐GPOD plans were generated using the planning CT scans and contours of 13 patients who received an 8 Gy (RX = 8 Gy) to 95% of the PTV boost on the 36S‐GPOD. Delivery feasibility and accuracy was evaluated using patient‐specific QA with a point‐dose ion chamber measurement in an acrylic phantom, and relative gamma analysis with the SRS‐MAPCHECK device (Sun Nuclear). The statistical significance of the reduction in OAR max‐dose and differences in PTV coverage (D95%, D2%, D95%) were determined using a paired sample t‐test.

Results: On average, the maximum heart dose was reduced by 3% (P‐value = 0.0047). Similarly, the average maximum lung dose was reduced by 2.8% (P‐value = 7e‐6). The 25S‐GPOD plans used the 15‐mm collimator more frequently; while PTV coverage remained unchanged, use of the smaller collimator resulted in more conformal plans and an 18% average (P‐value = 0.009) reduction was observed in the conformity index. However, a systematic 1% increase (P‐value = 6e‐5) between the 25S‐GPOD TPS and measured point‐dose was observed and is associated with more frequent use of the 15‐mm collimator.

Conclusion: While PTV coverage remained unchanged, better heart and lung sparing was achieved using the 25S‐GPOD design. This new design also exhibits a more frequent use of the 15‐mm collimator, which resulted in more conformal plans at the cost of a slight increase between the measured and planned dose difference.

Ying Niu and Cedric Yu are affiliated with Xcision Medical Systems.

PO‐BPC‐Exhibit Hall‐41

CyberKnife‐Based SBRT for High Intermediate and High Risk Prostate Cancer

K Lambson*, K Huang

Christiana Care Health Systems, Wilmington, DE

Purpose: Non‐isocentric CyberKnife stereotactic radiotherapy techniques for treatment to the prostate have been shown to give good local control and prostate‐specific antigen response. This study looks to investigate the viability of using CyberKnife SBRT to treat the prostate and nodes in combination with Cesium‐131 implants for selected favorable high‐intermediate/high‐risk, non‐metastatic prostate cancer. This will provide convenience for patients traveling long distances to receive highly conformal hypofractionated radiation with the potential for similar outcomes as a standard external beam boost fractionation.

Methods: An SBRT test plan is generated and optimized in MultiPlan (Version 5.1) software. The planning target volume includes the prostate, seminal vesicles, and lymph nodes and is treated to a dose of 25 Gy in five fractions to 95% of the volume. Treatment quality is evaluated based on target coverage, normal tissue sparing, homogeneity and conformality index, and treatment efficiency.

Results: A preliminary result showed adequate coverage and comparable results to standard fractionation external beam delivery. Due to the large volume, the conformality for the prostate and nodal targets individually are higher, but acceptable for the whole PTV with a value of 1.28 at 95% coverage. The homogeneity index was calculated to be 1.37 and the delivery time per fraction is 51 minutes.

Conclusion: CyberKnife‐based SBRT for treatment of prostate and nodes to a uniform dose appears to be a practical method of delivery. The conformal dose distribution takes a feasible amount of time to deliver in significantly fewer fractions than LINAC‐based VMAT. Additional plans are currently being optimized to decrease delivery time and increase conformality and dose fall‐off for normal tissue sparing.

PO‐BPC‐Exhibit Hall‐42

Planning and Delivery Performance of the Halcyon™ for Single‐Isocenter Multi‐Target Stereotactic Radiosurgery

P Irmen^1,*, H Liu², W Shi², M Alonso‐Basanta¹, J Zou¹, B Teo¹, J Metz¹, L Dong¹, T Li¹

(1) University of Pennsylvania, Philadelphia, PA, (2) Thomas Jefferson University, Philadelphia, PA

Purpose: To evaluate the planning and delivery performance of the Halcyon™ V2 for the treatment of multiple small cranial targets with a single‐isocenter as compared to the Halcyon™ V1 and noncoplanar Multi‐Aperture Dynamic Conformal Arc DCA.

Methods: Ten patients were retrospectively studied, each with 6–10 targets with volumes from 0.11 to 8.57 cc. VMAT plans were created for the Halcyon™ V1 (1 cm leaf width) and Halcyon™ V2 (dual‐layer modulation with 0.5 cm effective resolution) and compared with the clinical non‐coplanar DCA plan generated with Truebeam HDMLC. All Halcyon™ plans were coplanar due to machine limitation. Conformity Index (CI), gradient index (GI), V12 Gy, V6 Gy, V3 Gy, and brain mean dose were compared. Additionally, the platform efficiency was compared analyzing optimization time, delivery time, and total MU.

Results: Halcyon™ V2 showed improved CI compared to V1 for targets with a diameter <1 cm, and performed similarly in almost all parameters. Compared to the DCA plan, Halcyon™ V2 exhibited a superior CI for targets >1 cm in diameter but an inferior CI for smaller targets. The DCA plan exhibited a superior GI for all targets. V12 Gy, V3 Gy, and brain mean dose were similar. All plans met clinical constraints for critical structures. Halcyon™ VMAT optimization took twice as long as Truebeam VMAT plans. The total required MUs were approximately 7500 and 12000 MU for the DCA and Halcyon™ plans, respectively, with an estimated delivery time of approximately 15 min for each considering dose rate and couch rotation time.

Conclusion: Halcyon™ V2 improved dose conformity for small targets (<1 cm) compared to V1, and achieved better CI for large targets using VMAT compared to DCA. GI for Halcyon™ is inferior to DCA, likely due to the coplanar limitation. Overall, the planning performance of the Halcyon™ V2 system is clinically acceptable for targets >0.7 cm.

PO‐BPC‐Exhibit Hall‐43

Impact of the MLC Leaf‐Tip Model: TPS Dose Calculation Deficiencies and IROC‐H Phantom Failures

B Koger*, R Price, D Wang, D Toomeh, S Geneser, E Ford

University of Washington, Seattle, WA

Purpose: Radiotherapy treatment planning system TPS dose calculation is sensitive to MLC modeling, especially when treating with IMRT or VMAT. AAPM recommendations such as TG‐119 and MPPG‐5a suggest that patient‐specific IMRT QA can be used to detect modeling errors. This study investigates the dosimetric impact of two MLC model parameters leaf‐tip width and leaf‐tip offset in a commercial TPS. In addition, the detectability of introduced errors and the relationship with IROC‐H head‐and‐neck phantom failures is assessed.

Methods: An Agility MLC (Elekta Inc.) was commissioned clinically in RayStation v.6.1. Nine IMRT and VMAT plans were optimized to treat the IROC‐H phantom. Dose distributions were re‐calculated on 27 different beam models, varying the leaf‐tip width (ranging from 2.0 to 6.5 mm) and leaf‐tip offset (ranging from −2.0 to +2.0 mm), and doses were compared to IROC‐H TLD and ArcCheck measurements.

Results: Leaf‐tip width had a modest dosimetric impact with <2% and 5.6% differences in the PTV and spinal cord, respectively. Dose calculations were more sensitive to leaf‐tip offset. Offsets of 1.0 mm caused differences up to 10% in the PTV and 15% in the spinal cord. Offsets of 2.0 mm caused dose deviations up to 50% in the spinal cord. These errors were not apparent during the modeling process, and IMRT QA was unable to reliably detect these dose deviations; ROC AUCs were 0.513–0.756 for detecting 10%–20% dose differences.

Conclusion: Small errors in MLC model parameters can cause large changes in the calculated dose that may be unapparent both in dose curves and through standard IMRT QA methods. This may, in part, explain the high failure rate of IROC‐H phantom tests reported by the community. It is recommended that external validation be performed as part of the commissioning process of IMRT treatment planning.

PO‐BPC‐Exhibit Hall‐44

An Efficient Automatic Dose‐Space Registration Technique for Clinical IMRT/VMAT Quality Assurance with Radiochromic Film Dosimetry

T Ren^1,2*, J Bourland^1,2

(1) Department of Physics, Wake Forest University, Winston Salem, NC, (2) Department of Radiation Oncology, Wake Forest School of Medicine, Winston‐Salem, NC

Purpose: To develop a contrast‐based, marker‐free, automatic dose‐space registration technique for patient specific IMRT QA verification using radiochromic film. Dose‐space registrations of the reference and evaluation dose distributions of clinical IMRT QA plans are typically done either through the recognition of made‐ahead marks on films, CT fiducial marks or through the subjective method involving the tedious process of manual rotation and translation.

Methods: The evaluation dose pixels were resampled to match the pixel spacing of film measurement through bicubic interpolation by automatically detecting the pixel spacing information from both the reference film measurement and the treatment planning calculation. The algorithm employs the dose pixels above 20% of the maximum dose from the evaluation dose map to perform the registration by minimizing the normalized dose value‐based metric residuals. The algorithm was programmed in MATLAB code. As a test case, a clinical lung IMRT plan was measured by Gafchromic™ EBT 3 film inside a solid‐water phantom on an Elekta VersaHD™ linear accelerator. The triple‐channel dosimetry method was deployed to generate the reference dose plane from film. An unexposed film piece was used to correct the scan‐to‐scan artifact of an Epson 10000XL photo scanner. The evaluation dose plane was generated by the RayStation 6 treatment planning system. Gamma index analysis was used for the 2‐D spatial dose comparison after the automatic registration.

Results: The passing rate (γ < 1) for global and local gamma analysis is 100% and 94.8% respectively with 2 mm/2% and 10% threshold dose. The mean global and local gamma index value is 0.11 and 0.40 respectively.

Conclusion: An efficient, automatic dose comparison registration technique for radiochromic film dosimetry has been developed and evaluated, showing robust performance. The method requires no made‐ahead marks or other fiducial markers for either film or phantom use during the routine clinical IMRT/VMAT QA.

PO‐BPC‐Exhibit Hall‐45

Can the Student Outperform Its Master? A Comparison of Two Automated Inverse Planning Engines Using a Quantitative Quality Plan Metric

A Smith, A Granatowicz, C Stoltenberg, S Wang, X Liang, S Zhou, C Enke, A Wahl, D Zheng

(1) University of Nebraska Medical Center, Omaha, NE, (2) Nebraska Medicine, Omaha, NE, (3) University of Florida, Jacksonville, FL

Purpose: Auto‐Planning™ and RapidPlan™ are two commercial planning engines to automate inverse planning and are based on fundamentally different methods: Auto‐Planning uses a generic template of dose goals to iteratively but automatically modify the planning structures and objectives for plan optimization, and RapidPlan generates static objectives from DVH estimations predicted by models configured using a library of “training” plans. This study objectively compared the performance of these two algorithms on IMRT planning for prostate fossa and lymph nodes adopting the plan quality metric (PQM) used in the 2011 AAMD Plan‐Challenge.

Methods: All plans used an identical 9‐coplanar‐IMRT‐beam setup and the 56 Gy/68 Gy simultaneous‐integrated‐boost prescribed by the Plan‐Challenge. Auto‐Planning was used without manual intervention on 20 post‐prostatectomy patients and these 20 plans were subsequently employed as the library to build the RapidPlan model. To compare the two engines, an independent test‐set of ten patients plus the Plan‐Challenge patient were planned by both Auto‐Planning (master) and RapidPlan (student) without manual intervention and evaluated using the PQM that included 14 quantitative plan metrics ranging over target coverage, spillage, and OAR dose. PQM scores were compared between the two engines using the Mann‐Whitney U‐test.

Results: There was no significant difference between the performances of the two engines on the 11 test plans (P = 0.764). On the Plan‐Challenge patient, Auto‐Planning scored 133 and RapidPlan scored 130.3, as compared with the average human‐plan score of 116.9 ± 16.4 (range: 58.2–142.5) among the 125 AAMD Plan‐Challenge participants.

Conclusion: Using an innovative study design, an objective comparison has been made between two major commercial automated inverse planning engines. The two engines performed comparably with each other and both yielded plans on par with average human planners. Using a constant‐performing planner (Auto‐Planning) to train and to compare, RapidPlan was found to yield plans no better than, but as good as, its library plans.

We would like to acknowledge the research license support of Plan IQ by Sun Nuclear and Auto‐Planning by Phillips. We thank Dr. Nelms from ProKnow for some helpful discussion.

PO‐BPC‐Exhibit Hall‐46

Outcome Analysis in Stereotactic Body Radiation Therapy for Spine Metastases: Dose‐Response for Toxicity and Pain Relief

B Ziemer*, M Susko, S Braunstein, L Ma

University of California San Francisco, San Francisco, CA

Purpose: To report clinical outcomes in patients with metastatic cancer following stereotactic body radiotherapy SBRT to vertebral spine metastases with assessment of dose‐volume associations to late toxicity and pain relief.

Methods: A single‐institution retrospective analysis of 123 metastatic vertebral spine lesions in 89 patients treated with 1, 3, or 5 fraction SBRT course between 2011 and 2017 was performed. Median clinical follow‐up was 19.6 months (range: 4.8–73.2) with 63% of patients receiving follow up imaging at 17.4 months (range: 1.1–70.0). Late toxicities, including vertebral compression fracture (VCF), per CTCAE v5.0 and patient reported pain relief (change in brief pain inventory scale) were recorded. Data were used to test a previously published myelopathy probability model. Dosimetric variables were extracted from the plans to use in logistic regression to determine predictors of toxicity.

Results: A total of 16 (13.0%) and 13 (10.6%) patients experienced 2 grade toxicity and VCF, respectively. Sixty‐seven (54.5%) patients reported pain relief with mean pain relief of 5 (range: 1–10) on the 10‐point BPI scale. Overall local control was 86.2%. One case of grade 2 asymptomatic radiographic myelitis was observed. The model indicated >5% myelitis risk based on dosimetric data. Biological effective dose (BED3) for the spinal cord and maximum point target dose within the vertebral elements were associated with pain relief (P = 0.047) and VCF (P = 0.042), respectively. Fracture risk of 6.2%–19.2% from maximum point target dose of 21.6–42.4 Gy was calculated from the data.

Conclusion: SBRT treatment for spine metastases was associated with moderate risk of late toxicity and VCF while achieving high local control. Greater BED and maximum point target dose were associated with patient‐reported pain relief and VCF. The extracted dose‐dependent fracture risk shows agreement with other studies. The data set will be used to test existing and further develop models to ensure safe SBRT practice.

PO‐BPC‐Exhibit Hall‐47

Failure Modes and Effects Analysis for Gamma Knife Icon Frameless SRS/FSRT

R Jakubovic*, A Goenka, G Gill, J Chang

Radiation Medicine, Northwell Health, Lake Success, NY

Purpose: Frameless stereotactic radiosurgery/fractionated radiotherapy SRS/FSRT using the Gamma Knife Icon™ GKI treatment planning system with onboard cone‐beam computed tomography CBCT imaging has introduced a new treatment workflow for frameless gamma knife radiosurgery. To assess the risk associated with frameless GKI treatment we implemented the recommendations of the AAPM Task Group 100 for failure modes and effects analysis FMEA at our institution.

Methods: A process tree for frameless GKI was mapped, identifying three focus areas specific to the frameless workflow (Simulation, Treatment Planning, and Treatment). Potential failure modes were identified for each of the focus areas excluding all processes related to framed Gamma Knife treatment. Failure modes will be assessed using the TG‐100 FMEA grading system specific for radiotherapy outcomes and observations. Briefly, each member of the FMEA team evaluates likelihood of occurrence, severity, and detectability, to generate a risk priority number.

Results: Three subprocesses comprising 45 steps were identified during the preliminary FMEA evaluation. Twenty‐five steps displayed equivalency to the Gamma Knife Perfexion framed stereotactic radiosurgery previously published and were not included in the analysis. Potentially hazardous failure modes during simulation comprised improper mask fabrication, missing setup documentation, and CBCT registration errors. During treatment planning, skull definition based on the MRI segmentation, improper registration to the simulation CBCT and verification of dose were identified as failure modes. With regard to treatment, failure modes associated with patient registration and dose alignment as compared to the planning images (CBCT or MRI), intrafractional patient motion and collisions during treatment (sweat on cap, arms colliding etc.) were prominent.

Conclusion: FMEA analysis is a robust method to identify risk associated with frameless Gamma Knife SRS/FSRT. Implementation of this approach at our institution based on the recommendations of AAPM Task Group 100 will enable a better understanding of the frameless GKI and improve overall workflow.

PO‐BPC‐Exhibit Hall‐48

Characterizing the Response of a Conical, Scintillation‐Based Detector for Machine and Patient‐Specific Quality Assurance Applications in Photon Radiotherapy

L Johnson^1,*, M Price¹, B Nelson², A Yock¹

(1) Vanderbilt Univ. Medical Ctr., Nashville, TN, (2) Logos Systems Int'l, Scotts Valley, CA

Purpose: To extend the application of a conical, scintillation‐based detector for machine, and patient‐specific QA by assessing signal dependence on irradiation conditions and device acquisition parameters.

Methods: Extending the use of a scintillation detector typically used for targeting accuracy to QA applications featuring a broad range of energies, dose rates, and aperture sizes requires further characterization of the device regarding these variables. For this purpose, we used a conventional c‐arm linac to irradiate the device with square fields of varying size (1 × 1 cm–8 × 8 cm), position, angle (45° increments), energy (6–15 MV, FF and FFF), and dose rate (100–600 MU/min). Vendor‐provided software was used to analyze both the entrance and exit scintillation spots observed for each irradiation to determine the effect of these variables on the device response. The dependence of device response on acquisition settings such as frame rate was also analyzed.

Results: Differences in scintillation intensity with varying incident gantry angles were small (0.8%) although larger systematic differences were observed based on longitudinal position of the device (1.1%). The relative signal from irradiations of varying field sizes appear to be correlated with dose output factors acquired during machine commissioning. The device exhibits a bilinear response with respect to dose rate and acquisition frame rate. The scintillation intensity was observed to vary considerably (18.7%) and preliminary data suggest the behavior of the entrance and exit scintillation spot signals may differ in their energy‐dependence.

Conclusion: Our initial characterization of a conical, scintillation‐based detector suggests the response may depend on irradiation and acquisition parameters such as field size, position, angle, energy, dose rate, and frame rate. Understanding the influence of these factors on the device is imperative for its application to a broader range of QA applications such as patient‐specific QA for IMRT, SBRT, and SABR.

This work was supported in part by Logos Systems Int'l.

SA‐A‐Osceola BRC‐00

The Effect Increased Training Requirements Have Had On Clinical Practice: Views From the Field

Moderator: Rebecca Marsh

University of Colorado School of Medicine, Aurora, CO

SA‐A‐Osceola BRC‐01

A. Hsu

Stanford Cancer Center

How the ACR Requirements have Changed Residency Programs and Future Physicists

SA‐A‐Osceola BRC‐02

B. Lofton

Colorado Assn in Medical Phys CAMP

Bridging the Gap: A Perspective from Full‐Service Consulting

SA‐A‐Osceola BRC‐03

M. Martin

Therapy Physics Inc.

Hiring Qualified Diagnostic Imaging Physicists: Before and After 2014

Over the past serval years, the American Board of Radiology (ABR) has added training requirements must be completed before one is eligible to take the board certification exams. At the same time, board certification is required by many employers and regulatory and accreditation agencies to provide clinical medical services. This session will discuss the impact these changes have had on our field, including effects on workforce supply, the reduction in alternative pathways into the profession, the role of M.S. physicists, and the quality of those entering clinical practice. Speakers will present views from academic and consulting and from therapy, imaging, and nuclear medicine environments.

Learning Objectives:

1. Understand how the certification and MOC processes have changed over recent years.

2. Understand the impact that current certification requirements have on hiring and staffing.

3. Learn how specialization in medical physics has been affected.

4. Consider if the additional training requirements have improved clinical care.

SA‐B‐Osceola BRC‐01

Toward a 3D Ultrasound Needle Guidance System for High‐Dose‐Rate Interstitial Gynecologic Brachytherapy Implant Placements

J Rodgers^1,*, J Bax¹, K Surry², W Hrinivich³, E Leung⁴, V Velker², D D'Souza², A Fenster¹

(1) Western University, London, ON, (2) London Regional Cancer Program, London, ON, (3) Johns Hopkins University School of Medicine, Baltimore, MD, (5) Odette Cancer Centre, Toronto, ON

Purpose: To design a 3D ultrasound US system for intraoperative implant assessment during high‐dose‐rate interstitial brachytherapy ISBT of gynecologic cancers, with the potential to improve implant quality and reduce risk to nearby organs‐at‐risk OAR.

Methods: The system incorporates two scanning geometries, 3D transrectal US (TRUS) and 360° 3D sidefire transvaginal US (TVUS), to account for the challenging presentations of gynecologic malignancies. The technique uses a 2D US probe, rotated to acquire a 3D image in <20 s. TRUS and TVUS modes were tested in five and six patients, respectively, where needle positions were compared against clinical post‐insertion CT images. An automatic segmentation algorithm using a modified randomized 3D Hough transform to simultaneously identify multiple needles was tested with one TVUS image containing eight needles as a proof‐of‐concept for needle localization in the TVUS images.

Results: Comparison of 58 needles between TRUS and registered CT images demonstrated a mean ± standard deviation tip position difference of 3.82 ± 1.62 mm and mean angular difference of 3.04° ± 1.63°. Similarly, the mean maximum positional difference was 2.36 ± 0.97 mm and mean angular difference was 1.95° ± 0.70° for 54 needles in the TVUS images. The algorithm identified all needles with a mean maximum position difference of 0.78 ± 0.17 mm and mean angular difference of 0.4°4 ± 0.19°. Additionally, nearby OAR were clearly visible in most images, including the rectum in the TVUS images and bladder in TRUS and TVUS images.

Conclusion: Despite the variable presentation of gynecologic cancers and necessity for precise placement of needles to deliver optimal dose during ISBT, there is currently no standard method to visualize the implant intraoperatively. The proposed system provides the potential for a versatile method to assess the implant during needle placement, allowing needles and nearby OAR to be visualized and localized.

This work was funded by the Ontario Institute of Cancer Research (OICR), the Canadian Institutes of Health Research (CIHR), and the Natural Sciences and Engineering Research Council of Canada (NSERC). We have no other conflicts of interest or disclosures.

SA‐B‐Osceola BRC‐02

Variability in Commercially Available DIR: A Multi‐Institutional Analysis Using Virtual Head and Neck Phantoms

A Kubli^1,*, J Pukala¹, P Kelly¹, R Manon¹, A Shah¹, K Langen², F Bova³, S Meeks¹

(1) UF Health Cancer Center at Orlando Health, Orlando, FL, (2) University of Maryland School of Medicine, Baltimore, MD, (3) University Florida, Gainesville, FL

Purpose: To evaluate the deformable image registration DIR performance of Velocity, MIM, and Eclipse, and to characterize inter‐algorithm and inter‐user variability for these algorithms using data submitted by multiple institutions.

Methods: Thirty‐five institutions performed DIR on the Deformable Image Registration and Evaluation Project (DIREP) virtual head and neck phantoms, using either Velocity, MIM, or Eclipse. The resulting deformation vector fields (DVFs) were compared to the ground‐truth DVFs, and the target registration error (TRE) was calculated for six ROIs: external, brainstem, cord, mandible, left parotid, and right parotid. Then, statistical analysis was performed in order to determine the algorithmic and user dependence on DIR performance.

Results: Overall registration accuracy was found to be 2.04 ± 0.35 mm for Velocity, 1.10 ± 0.29 mm for MIM, and 2.35 ± 0.15 mm for Eclipse. For this data, MIM produced results with an average TRE up to 1 mm smaller than either Velocity or Eclipse. Both Velocity and Eclipse produced similar results. User dependence within an algorithm was not found in most cases. When user dependence was identified, its impact on DIR performance was much smaller than the variability introduced by choice of algorithm.

Conclusion: Choice of DIR algorithm appears to have a larger impact than does the user dependence within a given algorithm. DIR commissioning should occur after software updates, as significant changes in DIR accuracy can be introduced. Because this can be time consuming, tools provided by vendors to expedite the commissioning process in accordance with TG‐132 recommendations would be beneficial. However, the tolerance for maximum registration error recommended by TG‐132 may not be achievable with current DIR software in some cases.

SA‐B‐Osceola BRC‐03

Using Artificial Intelligence to Auto‐Delineate Patient‐Specific Head‐And‐Neck Clinical Target Volumes Within a Fully Automated Treatment Planning System

C Cardenas^1,*, M Aristophanous², J Yang¹, L Zhang¹, R McCarroll³, K Kisling¹, D Rhee¹, S Ng¹, A Rao⁴, C Fuller¹, A Mohamed¹, A Garden¹, L Court¹

(1) MD Anderson Cancer Center, Houston, TX, (2) The University of Chicago, Chicago, IL, (3) University of Maryland Medical Center, Baltimore, MD, (4) The University of Michigan, Ann Arbor, MI

Purpose: To use artificial intelligence to automate clinical target volume CTV delineation for oropharyngeal cancer patients and, therefore, reduce inter‐physician variability which is the largest source of uncertainty in head‐and‐neck radiotherapy. This auto‐delineation would address a rate‐limiting step in fully automated treatment planning.

Methods: To establish a baseline, we assessed inter‐physician CTV delineation variability from a team of sub‐specialized head‐and‐neck radiation oncologists. Then, we developed models to auto‐delineate (1) high‐risk CTVs using a feedforward network that uses tumor and organs‐at‐risk information, and (2) low‐risk CTVs using a 3D convolutional neural network that uses the CT image and GTV. Each algorithm's auto‐delineations were compared to clinically used CTVs using overlap/distance metrics. These delineations were then used in our automated treatment planning system to evaluate the impact of CTV delineation on dosimetric distributions.

Results: The measured inter‐physician variability resulted in a median Dice Similarity Coefficient (DSC) of 0.75, translating to better agreement than commonly reported in literature. The auto‐delineations for high‐risk and low‐risk CTVs performed well in comparison to their respective clinically used delineations (median DSC values of 0.81 and 0.82, respectively). Auto‐delineated volumes provided improved delineation consistency compared to inter‐physician variability (P < 0.0001). Dosimetrically, coverage was acceptable (per RTOG 1016 guidelines) for 72% of auto‐delineated plans when evaluated on clinicallyused targets. The median per cent‐volume of clinically used CTVs receiving 98% of the prescribed dose was 98% for auto‐delineated CTV plans. We found no significant difference in normal tissue dose metrics between physician and auto‐delineated target plans using a paired‐Wilcoxon Rank‐Sum Test: Brainstem Dmax (P = 0.23), cord Dmax (P = 0.87), contralateral parotid Dmean (P = 0.41), and ipsilateral parotid Dmean (P = 0.59).

Conclusion: We demonstrated that artificial intelligence‐based algorithms can delineate high‐ and low‐risk CTV with less variability than radiation oncologists. This work addresses a major barrier to fully automated head‐and‐neck treatment planning by providing high‐quality patient‐specific CTVs.

SA‐B‐Osceola BRC‐04

Minimally Invasive Focal Liver Tumor Therapy Using a Mechanically Assisted 3D Ultrasound System for Geometrically Variable Imaging

D Gillies^1,*, J Bax¹, K Barker¹, L Gardi¹, D Tessier¹, N Kakani², A Fenster¹

(1) Robarts Research Institute, London, ON, (2) Manchester Royal Infirmary, Manchester, UK

Purpose: Minimally invasive focal ablations of liver tumors provide reduced patient complications and recovery times, but are currently associated with high local cancer recurrence. One source of error when performing therapy applicator guidance with 2D ultrasound US is the limited field‐of‐view, requiring mental reconstruction of the anatomy. Our solution to this limitation has been the development of a novel mechanically assisted 3D US imaging and guidance system capable of providing geometrically variable images.

Methods: A three‐motor mechanical mover was designed to provide adjustable linear, tilt, and hybrid geometries for variable 3D US fields‐of‐views. This mover can manipulate any clinically available 2D US transducer via transducer‐specific 3D‐printed holders to guide therapy applicator insertions intraoperatively. This mover is held by a portable counterbalanced mechanical system that features foot‐released electromagnetic brakes and encoders to track the position of the transducer. Optical tracking was performed with a mounted stylus to evaluate mechanical motions of the scanner. A string phantom with known dimensions was imaged and manually measured to validate image reconstruction geometry and preliminary images were acquired of a volunteer to assess clinical applicability.

Results: Optical tracking of the 3D scanner resulted in mean linear and angular motion errors of 0.21 mm (0.20%) and 0.23° (0.52%), respectively. Mean linear image measurement errors were <3% on the string phantom and human volunteer images were clinically applicable as the relevant anatomy could be visualized such as the portal vein bifurcation, gallbladder, and kidney.

Conclusion: The proposed mechanically assisted 3D US system for focal liver ablation performs image acquisition accurately and the three‐motor mover assembly allows for control over scan geometries to accommodate clinical anatomical variation. Current work is focused on mapping and correcting for the error in the tracking system encoders to calibrate the tracking system prior to a simulated image‐guided phantom procedure.

The authors would like to acknowledge funding support from the Canadian Institutes of Health Research (CIHR), Natural Sciences and Engineering Research Council (NSERC), and the Ontario Institute for Cancer Research (OICR).

SA‐B‐Osceola BRC‐05

Material Differentiation in Virtual Non‐Contrast Images and Iodine Maps of a Dual Energy CT (DECT) System

C Olguin*, C Schaeffer, N Correa, A Heshmat, N Quails, L Rill, M Arreola, I Barreto, S Leon

University of Florida, Gainesville, FL

Purpose: Hemorrhagic blood can appear hyperdense on CT images, complicating the differentiation of blood and contrast extravasation following a stroke. This study evaluates the limits of material differentiation in virtual non‐contrast VNC images and iodine maps of a dual energy CT DECT system, with focus on differing types of blood and mixtures of blood and iodine.

Methods: Limitations on detectability of iodine (2 mg/ml, 4 mg/ml), and blood (40 HU and 70 HU) will be assessed using a tissue equivalent multi‐energy CT (MECT) phantom (Gammex Model 1472). Reproducibility was tested by scanning at 80 and 135 kV three times daily for 6 non‐consecutive days thus far. Using the vendor generated head protocol, iodine maps, and VNCs were reconstructed in the raw data domain. Circular regions of interest at the central axial slice of the phantom were generated using the Gammex Rod Evaluate Software (v1.7). The mean pixel intensity [HU] and the standard deviation (SD) between the daily means of the ROI measurements were generated. These measurements quantified statistical limitations in iodine maps and VNCs while simulating clinical findings such as residual iodine contrast and hemorrhagic lesions.

Results: The coefficients of variation were calculated for the differing materials for the iodine map (0%–24%) and VNC image (2%–18%).In the VNC images, the mixture of iodine (4 mg/ml) and 40‐HU blood was not significantly different from a 70‐HU blood material (76 ± 4 HU and 68 ± 6 HU, respectively). The iodine map differentiated the blood material from the heterogeneous blood and iodine material (6 ± 4 HU and 109 ± 8 HU, respectively).

Conclusion: A limitation in the differentiation between blood and mixtures of iodine and blood was found in the VNC, but was resolved in the iodine map. The use of both image sets is recommended for proper material differentiation.

Has any company or organization whose products/services are the subject matter of your presentation provided you with any financial support for your research?

Yes, we have a Canon contract pending. We will receive funding to characterize Canon's DECT systems. Canon has also lent the Gammex MECT phantom for this purpose.

SA‐B‐Osceola BRC‐06

Interactive Analysis of a Web‐Based Quality Assurance System Using Python

D Dunkerley*, D Jacqmin

University of Wisconsin‐ Madison, Middleton, WI

Purpose: Incident learning systems in radiation oncology are an important quality and safety tool, and studies have shown that high participation in incident reporting is associated with good safety culture and quality care. However, evaluating, managing, and organizing a large incident reporting system is a time‐ and labor‐intensive process. The purpose of this work is to develop a graphical user interface GUI to facilitate the processing and review of in‐house web‐based quality assurance forms.

Methods: Python was selected as the programming language due to its flexibility and wide availability of libraries in addition to its ease of use for future adaptation. The GUI permits selecting and sorting the data according to different QA form data elements and custom date ranges. The GUI also includes a form browser and tools for cross‐correlational and longitudinal analysis.

Results: Reading data forms using the program is completely automated, takes 0.9 s per form, and removes the potential for transcription errors. The current database of 745 QA forms loads in approximately 10 min, and thereafter analysis is instantaneous. The previous data aggregation method (manually analysis) took hours and was limited in the types of analysis that could be done due to the time‐consuming nature of the work.

Conclusion: In contrast to reviewing QA forms individually, data aggregation and processing provide a broad overview to view trends. The use of a GUI allows those without programming experience to easily review data. The GUI developed in this work provides the ability for those interested in quality assurance to quickly and effectively analyze aggregated data for correlation and trends. The intuitive interface makes it easy to identify issues that happen at a higher frequency, which allows our clinic to direct quality improvement efforts to where they may have the largest impact.

SA‐B‐Osceola BRC‐07

Fully Automated Head‐And‐Neck Contouring and VMAT Planning with Integrated and Comprehensive QA

R McCarroll^1,*, B Beadle², P Balter³, H Burger⁴, C Cardenas³, S Dalvie⁵, D Followill³, K Kisling³, M Mejia⁶, K Naidoo⁷, C Nelson³, C Peterson³, K Vorster⁴, J Wetter⁴, L Zhang³, J Yang³, L Court³

(1) University of Maryland Medical Center, Baltimore, MD, (2) Stanford University, Stanford, CA, (3) MD Anderson Cancer Center, Houston, TX. (4) University of Cape Town, Groote Schuur Hospital, Cape Town, (5) University of the Free State, Bloemfontein, (6) University of Santo Tomas Hospital, Metro Manila, Manila, (7) Tygerberg Hospital, Bellville, Cape Town

Purpose: To develop and validate the automation of the entire head‐and‐neck treatment planning process.

Methods: In‐house and commercial software were integrated to automatically create head‐and‐neck VMAT plans in <30 min requiring only the treatment prescription, CT scan, and GTV contour. OARs and CTVs are automatically contoured using an in‐house independently verified multi‐atlas technique which has been implemented clinically. For QA, machine‐learning algorithms and deep‐learning are used to assess OARs and CTVs for accuracy and need for review. Plans are optimized using a modified commercially available RapidPlanᵀᴹ model. Eclipse API and dicom transfer techniques automate the process from CT import to dose calculation.

Results: On retrospective analysis, 98% (1903/1949) and 95%(381/400) of automatic OARs and CTVs were physician approved with no or minor edit. During clinical implementation, 50% of 1,415 OAR contours were not edited for clinical use and 86% had a maximum distance‐to‐agreement of <5 mm. In identifying contours requiring edit, random‐forest models detected >95% of autocontoured OARs with clinical edits exceeding a maximum distance‐to‐agreement of 10 mm. For CTV contour QA, the mean distance‐to‐agreement of autocontoured CTVs compared to those from an independent deep‐learning algorithm was significantly correlated to agreement with clinical CTVs (CCC = 0.83, P < 0.01). When autocontoured OARs were used for treatment planning only 5/898 clinical structures exceeded DVH criteria. Finally, comparing 30 autoplans to those from a clinical trial, the autoplans performed significantly better considering brainstem and spinal cord D(max), D(1 cc) of the high‐dose PTV, and D(mean) of contralateral submandibular and parotid glands (paired Wilcoxon rank‐sum test, P < 0.01). For no structure did autoplans perform significantly worse. Physicians from four institutions rated 90% of plans as clinically acceptable without edit, the remaining plans required only minor edit.

Conclusion: All planning processes have been successfully automated, with clinically acceptable plans created in less than 30 minutes. Independent validation of each step ensures safe plan delivery.

This work was funded by the NCI (UH2 CA202665). Additional support provided by Varian Medical Systems and Mobius Medical Systems.

SA‐B‐Osceola BRC‐08

Effective Automated Treatment Planning and Quality Assurance for Cervical and Breast Cancer for Limited‐Resource Clinics

K Kisling^1,*, L Zhang¹, B Anderson¹, D Anderson², P Balter¹, B Beadle³, H Burger², C Cardenas¹, M Du Toit⁴, N Fakie², R Howell¹, A Jhingran¹, J Johnson, R McCarroll⁵, K Schmeler¹, S Shaitelman¹, H Simonds⁴, T Thebe², C Trauernicht⁴, J Yang¹, L Court¹

(1) MD Anderson Cancer Center, Houston, TX, (2) University of Cape Town, Cape Town, South Africa, (3) Stanford University, Stanford, CA, (4) Stellenbosch University, Cape Town, South Africa, (5) University of Maryland Medical Center, Baltimore, MD

Purpose: To develop automation algorithms that increase the availability of high‐quality treatment plans in clinics with limited staff members and without access to advanced techniques, such as breath‐hold gating and VMAT.

Methods: We used atlas‐based segmentation, deep learning, optimization, and classification techniques to create automation algorithms that (1) create treatment plans from a patient CT and (2) perform plan QA. We automatically planned treatments using our algorithms (retrospective validation cohort). Plans were scored for clinical acceptability by radiation oncologists in the US and South Africa, and dose metrics were evaluated. To evaluate risk in our automated workflow, we performed Failure Modes and Effects Analysis (FMEA).

Results: For cervical cancer, 150 four‐field box treatments were automatically planned, and 89% were scored acceptable by physicians. The dose distributions were more homogeneous (P < 0.001) using automatically optimized beam weights compared with equal beam weights (commonly used in Africa). Our automatic QA technique flagged 90% of unacceptable plans (false‐positive rate: 16%). For breast cancer, we automatically planned three‐field treatments on free‐breathing CTs for 19 left‐sided, post‐mastectomy patients. Seventeen (89%) met constraints for lung and heart dose and target coverage. Physicians accepted the tangent beam angles for all plans. The inferior border required median changes of 4.0 mm (max: 34 mm), indicating that this should be identified manually. Otherwise, physicians accepted a majority of plans (12/19) with no or minor changes. Automatic QA of the dose flagged 6/7 unacceptable plans. Physicians indicated that 3/7 unacceptable plans needed a more complex treatment technique due to anatomy. FMEA showed that automated QA reduced risk (number of high‐risk failure modes was reduced by half) and that manual plan review is still vital for safety.

Conclusion: Automated treatment planning and QA was effective for the majority of patients tested. Our algorithms will be implemented clinically at our partner hospitals in South Africa next year.

This work was supported by a grant from the National Cancer Institute (UH2‐CA202665). We have received equipment and technical support from Varian Medical Systems and Mobius Medical Systems.

SA‐B‐Osceola BRC‐09

Radioluminescence Imaging of Dose Surrogates for Head and Neck Radiotherapy

D Alexander^1,*, I Tendler¹, P Bruza¹, D Gladstone^1,2,3, P Schaner^2,3, L Jarvis^2,3, B Pogue^1,2

(1) Thayer School of Engineering, Dartmouth College, Hanover, NH, (2) Geisel School of Medicine, Dartmouth College, Hanover, NH, (3) Norris Cotton Cancer Center, Dartmouth‐Hitchcock Medical Center, Lebanon, NH

Purpose: To characterize the optical emission from imaging both Cherenkov and scintillation signals underneath thermoplastic masks and bolus materials to assess feasibility of real‐time in vivo surface dosimetry during head and neck radiotherapy at the level of the mask, bolus, or tissue.

Methods: Radioluminescence emissions were imaged using a time‐gated intensified CMOS camera during head and neck radiotherapy treatment of a tissue equivalent body phantom. Emission spectra and relative emission intensity of different colored radiotherapy masks were measured to infer contribution to captured optical signals. Emission from small plastic scintillator targets was characterized underneath transparent bolus and thermoplastic mask material, and parameters such as air gap thickness were analyzed. Lastly, Cherenkov emission from a patient undergoing VMAT treatment for head and neck cancer was imaged through the mask.

Results: Cherenkov and scintillation intensity from underneath transparent bolus material can be quantified to estimate delivered surface dose. The transparent bolus material is found to transmit >80% of optical light at wavelengths of interest. Luminescence emission intensity and spectrum varies from different colored mask materials, and therefore the choice of mask color could impact measured signal. Imaging of plastic scintillators indicates that scintillation dose response linearity is maintained when imaged through transparent bolus material. Additionally, scintillator intensity is shown to drop up to a maximum of 6% in the presence of airgaps between the bolus and surface of up to 2 cm. Finally, Cherenkov images from patient treatment show potential for beam verification underneath mask material in head and neck cancer patients.

Conclusion: This work illustrates the potential for using time‐gated intensified imaging of optical emission during head and neck radiotherapy to measure local delivered surface doses. Tracking emission intensity across fractions could inform changes or abnormalities in dose delivery during treatment, and across treatments could help interpret or reduce skin reactions.

This work has been sponsored by NIH research grant R01EB023909. B Pogue is the president and the co‐founder of DoseOptics LLC, manufacturing the C‐Dose camera provided for this research. This work was not financially supported by DoseOptics.

SA‐B‐Osceola BRC‐10

Remote, Real‐Time and High‐Resolution Optical Imaging of Small Beamlets Relevant in Radiotherapy

M Ashraf^1,*, P Bruza¹, B Williams², D Gladstone², N Nelson², B Pogue¹

(1) Thayer School of Engineering, Dartmouth College, Hanover, NH (2) Dartmouth‐Hitchcock Medical Center, Lebanon, NH

Purpose: We propose a remote, real‐time optical imaging technique, for quality assurance, and commissioning of small beamlets relevant in radiotherapy. A pulsed intensified CMOS camera was used to capture optical photons generated due to radioluminescence in a cylindrical water tank doped with 1 g/L quinine sulfate.

Methods: A range of static beams and dynamic VMAT plans were simulated in a treatment planning system (TPS). For static beams, the percentage dose depth (PDD) and cross beam profiles (CBP) were obtained from the optical images and compared against TPS data. Gamma analysis was performed on all dynamic plans relative to the TPS image data. The technique was tested for sensitivity again common errors (MLC position, Gantry Angle) by inducing deliberate errors in VMAT plans per control point. The technique's detection limits for spatial resolution, the smallest beam size and the smallest number of control points that can be imaged reliably were also tested.

Results: The optical PDD's agree to within 2 % of the TPS data for small static square (5, 10 and 50 mm²) beams. For CBP's, we were able to achieve a gamma pass rate >98% for the 3%/1 mm criteria. All dynamic plans passed the 3%/3 mm criteria with a >90% passing rate. With the camera 2 m away from the isocenter, a theoretical spatial resolution 0.21 mm was achieved. Beams as small as 5 mm and a total of 20–10 Monitor Units can be reliably measured. The technique is sensitive to multileaf collimators (MLC) errors down to 2 mm. In future, the technique will be tested against other errors.

Conclusion: Optical imaging provides high resolution quality assurance compared to other QA devices used commercially. Ability to image down to 20‐10 MU potentially allows us to capture data per single control point for dynamic plans. The technique is sensitive to small offsets errors in MLC leaf positions.

SA‐B‐Osceola BRC‐11

Scripting Solutions for Lung Motion Management

A Shutler*, J McGlade, F Mourtada

Christiana Care Hospital, Newark, DE

Purpose: To determine the feasibility of using TPS built‐in registration assessment scripting tools, such as Dice and Image Similarity Coefficients (ISC), and distance to agreement (DTA). We assessed the quality of in‐house procedures used for motion assessment of both gantry‐mounted linac and Cyberknife units.

Methods: Python scripts generate ITVs for linac‐based treatment plans and provide tracking and margin recommendations for Cyberknife SBRT. Motion assessments are performed using Hybrid intensity and structure based deformable registrations to either generate an ITV or confirm that fiducial motion tracks with target motion. Target contours are first drawn on the planning CT and mapped to the individual phases of the 4DCT series. Accurate deformable registrations are critical for the success of this method, since target motion is evaluated based on the results of these mapped structures. Magnitude of target motion and, in the case of Cyberknife, the consistency of motion between fiducials and the target, are then examined via scripted reporting of the Dice and ISC.

Results: The ISC can range from 0 to 1, with 0 representing no agreement and 1 representing perfect pixel agreement for a given Region‐Of‐Interest (ROI). For the clinical cases examined, we report a range of 0.5 to 0.96. An ISC value of 0.8 or greater indicates a good fusion, resulting in negligible structure mapping error. Dice similarity coefficient and DTA show the volume of ROI overlap and magnitude of motion when comparing rigidly copied to deformed structures. Max DTA is used for tracking recommendation on Cyberknife plans (Synchrony versus non‐Synchrony).

Conclusion: ITV(s) and tracking recommendations for motion‐management can be generated automatically based on motion assessment of the 4DCT. A script to automate this process is feasible, and can include additional margin recommendations to account for deformable fusion error as well as contouring errors generated by gaps between 4DCT phase data.

SA‐C‐Osceola BRC‐00

Super Conductor or Attractive Nuisance? Real Talk About MR Safety…With No Spin

Moderators: Annie Hsu

Stanford Cancer Center, Stanford, CA and William Sensakovic, Florida Hospital, Orlando, FL

SA‐C‐Osceola BRC‐01

M. Cao

UCLA School of Medicine

Integration of MR in Radiation Therapy: Practical Safety Considerations

SA‐C‐Osceola BRC‐02

D. Jordan

University Hospitals Cleveland Medical Center:
Auditing and Evaluating MRI Facility Safety Programs

Expanding utilization and increased awareness of risks has recently made MR safety a topic of interest in the imaging community. Furthermore, MRI safety programs have come under increased attention from accrediting bodies. The ACR in particular now requires that the medical physicist review the safety program annually. In addition, there has been strong interest in integrating MRI into the radiation therapy workflow either as a stand‐alone unit or in new radiation delivery technologies. MR safety is one of the major challenges in incorporating MRI in therapy since most therapy physicists are not yet accustomed to a working environment with high magnetic fields. Many medical and immobilization devices used in radiotherapy also lack clear labels with regard to MR safety. This Joint session provides practical guidance for the physicist tasked with conducting annual MRI safety program evaluations as well as discusses strategies and provides guidance for establishing MR safety programs in the radiation oncology environment.

Learning Objectives:

1. Describe resources and references that provide benchmarks for the content of a facility’s MRI safety program.

2. Evaluate implementation of MRI safety programs through observation and review of documentation.

3. Identify the safety challenges of integration of MRI into radiation therapy and describe the strategies and references for establishing MR safety programs in radiotherapy

SA‐D‐Osceola BRC‐00

FUNdamentals ‐ The Black Box in Your TPS: Algorithms

Moderator: Annie Hsu

Stanford Cancer Center, Stanford, CA

SA‐D‐Osceola BRC‐01

K. Bush

Stanford University:
Algorithms‐ A Review of the Current State

SA‐D‐Osceola BRC‐02

X. Liang

University of Florida:
Algorithms‐ QA & Commissioning

Since the publication of the AAPM TG105 report, an increasing number of clinics have acquired advanced dose calculation capabilities through fast Monte Carlo or linear Boltzmann equation solver‐based algorithms provided by treatment planning system vendors.

Many clinical studies have demonstrated the dosimetric importance of applying such algorithms, in comparison with the convolution algorithms using point and pencil kernels. Clinical sites with significant tissue heterogeneity, such as lung, nasopharynx/oropharynx, and esophagus, especially stand to benefit from these advanced algorithms.

This session returns to the basics of the core engine of the external beam treatment planning system, namely dose calculation algorithms. Explanations of the commonly used current algorithms and their shortcomings as well as commissioning and QA will be visited.

Learning Objectives:

1. Understand the current state of commonly used external beam treatment planning algorithms.

2. Understand the clinical implementation process of treatment planning with the advanced algorithms.

SA‐D‐Orange Blossom BR‐00

The New IEC/FDA Regulatory Paradigm for X‐Ray Devices

Moderators: Matthew Hough

Florida Hospital, Altamonte Springs, FL and William Sensakovic, Florida Hospital, Orlando, FL

SA‐D‐Orange Blossom BR‐01

R. Sauer

Food Drug Administration:
FDA Prospective

SA‐D‐Orange Blossom BR‐02

L. Bush

Medical Imaging Technology Alliance MITA

MITA Prospective

SA‐D‐Orange Blossom BR‐03

W. Sensakovic

Florida Hospital:
Open Q&A

The FDA will no longer be updating the Code of Federal Regulations (CFR) for x‐ray devices. Instead, they will defer to IEC Standards, by reference. X‐ray device manufacturers will be allowed to decide whether each device adheres to a current IEC standard or the CFR, or parts of an IEC standard and parts of the CFR. There are merits to this initiative, but there are also serious downstream consequences to x‐ray facilities and their medical physicists.

Learning Objectives:

1. Understand the need and rationale for a new regulatory paradigm for x-ray devices.

2. Learn in detail the process that manufacturers will execute to make public all relevant IEC standard information.

3. Learn in detail the oversight and enforcement mechanisms in place to ensure that manufacturers adhere to required procedures.

SU‐A‐Osceola BRC‐00

Developments in Brachytherapy

Moderator: Brent Parker

University Texas Medical Branch of Galveston, Galveston, TX

SU‐A‐Osceola BRC‐01

B. Cai

Washington University School of Medicine:
Automation in Brachytherapy (1)

SU‐A‐Osceola BRC‐02

X. Jia

The University of Texas Southwestern Medical Center:
Automation in Brachytherapy(2)

SU‐A‐Osceola BRC‐03

Z. Grelewicz

Rush University Medical Center:
Intraoperative Radiation Therapy with Directional Brachytherapy (1)

SU‐A‐Osceola BRC‐04

I. Veltchev

Fox Chase Cancer Center:
Intraoperative Radiation Therapy with Directional Brachytherapy (2)

SU‐A‐Osceola BRC‐05

M. Studenski

University Miami:
Intraoperative Radiation Therapy with Directional Brachytherapy (3)

Brachytherapy has played an important role in cancer treatment for over 100 years. It is facing some great development opportunity as well as challenges associated with it. Our session aims to prepare the audience for the new stage of the brachytherapy by providing them tools to streamline the clinical processes as well as the knowledge of starting an IORT program using the newest developed unidirectional brachytherapy sources.

Automation in radiation therapy helps to avoid human error and to streamline the processes in patient care. This is especially critical in high dose rate (HDR) brachytherapy that requires quick turnaround and high precision. However, automation in brachytherapy is not as widely applied as in external beam radiation therapy (EBRT). Our speakers in the first half of the session will introduce their clinical use of the automatic systems in HDR brachytherapy cases, including image processing, automatic segmentation, treatment planning, optimization, and quality assurance.

Intraoperative radiation therapy (IORT) is another growing field with various systems and sources dedicated to providing radiation dose directly to the surgical margin. Implantable low dose rate (LDR) radiation sources embedded in a bioabsorbable substrate sutured directly to the surgical margin provide conformal radiation coverage directly to surgical margins suspicious for microscopic disease (i.e., positive surgical margin). IORT with LDR brachytherapy is especially beneficial to centers that do not have established IORT program. Directional brachytherapy has been demonstrated to be especially beneficial to locally recurrent patients that have already received EBRT. The second half of the session will explain how to initiate an IORT program with directional brachytherapy, perform treatment planning for directional brachytherapy, and control of IORT patient workflow. Initiating an IORT program requires training, meeting required regulatory and professional standards, preparing for receiving new sources, and commissioning of treatment planning system. Treatment planning for directional brachytherapy sources can be accomplished with commercially available 3D treatment planning software. There are some special considerations for directional sources to be adequately planned that will be reviewed. Control of IORT patient workflow can be cumbersome depending on the location of operating rooms relative to radiation oncology. Advice for working with surgeons to plan patients and order sources will be discussed. This session will discuss each step of implementing and maintaining a new IORT directional brachytherapy program.

Learning Objectives:

1. To learn the available applications of automation in clinical HDR brachytherapy.

2. To acknowledge the capability of automation in HDR brachytherapy.

3. To be able to implement these methods in clinics.

4. Introduce the participants to the required steps for clinical implementation of an IORT LDR Brachytherapy program.

5. The participants should receive adequate knowledge to be able to configure, commission, and use existing Treatment Planning Systems for Directional Brachytherapy dose calculations.

6. Understand the radiation safety considerations and the optimal patient workflow.

SU‐A‐Orange Blossom BR‐00

MR Safety and Quality Troubleshooting

Moderators: Matthew Hough

Florida Hospital, Altamonte Springs, FL and William Sensakovic, Florida Hospital, Orlando, FL

SU‐A‐Orange Blossom BR‐01

R. Stafford

UT MD Anderson Cancer Center:
Device Safety Testing in the MR Environment

SU‐A‐Orange Blossom BR‐02

A. Panda

Mayo Clinic, Arizona:
Troubleshooting MR Image Quality Issues

MR imaging is a rapidly evolving modality that is characterized by rapid advancement in hardware, software, pulse sequences, and post processing. The rapid evolution of the modality can make it difficult to understand safety implications of the technology and how to troubleshoot image quality issues. This session will provide information on basic principles of MR image characteristics, a structured path for troubleshooting common image quality issues, and a framework for addressing MR device safety issues that may arise in the clinic.

Learning Objectives:

1. Understand basic principles of MR image characteristics.

2. Learn practical methods for troubleshooting MR image quality.

3. Understand MR safety device profiles and how they are determined.

SU‐B‐Osceola BRC‐00

The Case Study: A Valuable Leadership Development Tool

Moderator: Matthew Meineke

Ohio State University, Columbus, OH

SU‐B‐Osceola BRC‐01

W. Ellet

University of Miami Business School: Part 1

SU‐B‐Osceola BRC‐02

J. Johnson

Houston, TX:
Part 2

Leadership is a relational process that involves noticing, deciding, acting, and reflecting on what has happened Impact International, 2016. Thus, leaders in any field need to learn to be attuned to facts and assumptions, uncover opportunities, appreciate problems, entertain options, and determine criteria to best resolve the predicament. An individual develops their leadership ability through practice of handling predicaments, either real or simulated, such as studying business cases.

The business case method is used throughout business schools to present real‐life business situations where a “protagonist” has a dilemma that must be resolved (i.e., “a real person with a real job is confronted with a real problem, (Stam, SPICE, 2018)“). Real problems are rarely simple and straightforward. However, with training leaders are more likely to choose the most optimal course of action.

This Medical Physics Leadership Academy (MPLA) session is to introduce the business case method to medical physicists. A medical physics‐related case study will be presented. If time allows, the case study will be analyzed and discussed in small groups.

Learning Objectives:

1. Explain what constitutes a case study.

2. Examine how business case method participants analyze, critique, make judgments, speculate, and express reasoned opinions.

3. Evaluate the effectiveness of the business case method in medical physics.

SU‐B‐Orange Blossom BR‐00

The Future of CT Testing

Moderators: Matthew Hough

Florida Hospital, Altamonte Springs, FL and William Sensakovic, Florida Hospital, Orlando, FL

SU‐B‐Orange Blossom BR‐01

J. Solomon, Duke University:

Update from Task Group 233

SU‐B‐Orange Blossom BR‐02

D. Bakalyar

Henry Ford Health System:
Update from Task Group 200

Modern CT scanners are rapidly advancing adding new technology to improve image quality and lower dose. Iterative reconstruction, tube current modulation, and other technology have been a benefit to the patient but also increase the complexity of the imaging chain. The implementation of this new technology necessitates advances in methods to accurately measure and track CT scanner function and output. This session will review proposed new methods to test modern CT scanners.

Learning Objectives:

1. Identify technology not accurately assessed by current QC methods.

2. Learn about new dosimetry phantoms and methods to implement them.

3. Learn about new metrics to assess CT scanners.

SU‐C‐Osceola BRC‐00

Talk to the Experts–Considerations in Establishing a Safety Program

Moderators: Annie Hsu

Stanford Cancer Center, Stanford, CA and Grace Gwe‐Ya Kim, University of California, San Diego, La Jolla, CA

SU‐C‐Osceola BRC‐01

G. Kim

University of California, San Diego:
A Brief Introduction

SU‐C‐Osceola BRC‐02

T. Pawlicki

UC San Diego:
Successfully Establishing a Safety Program in an Academic Environment

SU‐C‐Osceola BRC‐03

B. Schuller

SCL Health:
Unique Considerations in Establishing a Safety Program in a Community Setting

SU‐C‐Osceola BRC‐04

L. Schubert

University of Colorado Denver:
Methods for Overcoming Barriers and Techniques for Rolling Out Interdisciplinary Change

SU‐C‐Osceola BRC‐05

A. Hsu

Stanford Cancer Center:
Open Discussion with Audience

Establishing a safety culture and implementing safety initiatives, such as incident learning or prospective safety assessment techniques, is challenging to achieve in practice. It involves changes in staff member's practices, behaviors, and attitudes. What are the strategies to successfully implement safety initiatives and maintain staff members’ engagement?

In this interactive session, panel members will share their experiences implementing safety initiatives in their clinics. They will provide tips for success, and discuss challenges and lessons learned. Each panel member will represent various perspectives and experiences. The first speaker will describe how to successfully establish a safety program in an academic environment, and the second speaker will discuss the unique considerations in a community setting. The last speaker will discuss methods for overcoming barriers and techniques for rolling out interdisciplinary change.

Following the presentations, the moderators will facilitate an interactive dialogue among panelists, then open the discussion and incite questions and comments from the floor.

Learning Objectives:

1. Understand how to implement solutions to improve patient safety.

2. Discuss problems and issues in measuring and reporting safety.

3. Become aware of the range of tolls, solutions, and strategies for motivating changes to improve patient safety.

SU‐C‐Orange Blossom BR‐00

Medical Physics Practice Guidelines: Implementation Experience and New Reports

Moderator: Rebecca Marsh

University of Colorado School of Medicine, Aurora, CO

SU‐C‐Orange Blossom BR‐01

B. Parker

University Texas Medical Branch of Galveston:
Medical Physics Practice Guidelines: Past, Present, and Future

SU‐C‐Orange Blossom BR‐02

D. Gress

American College of Radiology:
Experiences and Challenges with MPPGs and Other Living Documents

The purpose of this session is to inform the audience of the AAPM Medical Physics Practice Guidelines (MPPG) initiative. The presentation will begin with the history of the Subcommittee on Practice Guidelines (SPG) and the purpose of starting the MPPG initiative. This will include the unique role of MPPGs, intended benefits to the AAPM membership, and location of SPG within the AAPM structure. The presentation will then review the publication history of existing practice guidelines and review those MPPGs currently in the creation and revision phases. The presentation will conclude by discussing future issues impacting the MPPG initiative. These include the new AAPM document review and publication process, integration of MPPGs with the AAPM TG‐100 initiative, and ongoing efforts to for regulatory and accreditation bodies to adopt MPPGs.

Learning Objectives:

1. Understand the purpose of the MPPG initiative.

2. Understand the role of SPG in related to MPPGs.

3. Understand the MPPG process, from concept to publication.

4. Review recent and current MPPGs.

5. Discuss future developments in the MPPG initiative.

SU‐D‐Osceola BRC‐00

Nomenclature and Big Data – TG 263 and the Future

Moderator: Brent Parker

University Texas Medical Branch of Galveston, Galveston, TX

SU‐D‐Osceola BRC‐01

C Hampton

Levine Cancer Institute/Atrium Health:

Practical Implementation of TG 263's Standardized Nomenclatures Across Multiple Radiation Oncology Practices

SU‐D‐Osceola BRC‐02

C. Mayo

University of Michigan:
Big Data ‐ How to Use the Past for the Future

TG 263's report describing standardized nomenclatures have been provided to physicist and clinics to, among other things, aid in enhancing conformity within and between clinics. Our institution consists of community‐hospital or free‐standing practices in nine different physical locations treating 350+ patients daily and utilizing three different commercially‐available electronic medical record databases and two different treatment planning systems. Additionally, the practices share second‐MU check software, dose volume histogram, and objective analysis software and a radiation oncology specific enterprise image management database. Even though the Task Group published its final report in 2017, many radiation oncology practices are just beginning to explore implementation of the group's recommended nomenclature with implications for clinics large and small. Our practices range from multiple single‐linac facilities to several multiple‐linac sites. Our experience implementing this change impacting so many users across multiple radiation oncology specific roles is instructional and metrics analyzing efficiencies realized during the treatment planning process will be presented. This lecture will present guidance to physicist for leading or contributing to an effective implementation strategy incorporating varied practice stakeholders.

Learning Objectives:

1. Describe Task Group 263's goals and recommendations.

2. Review clinical examples and receive guidance for gaining efficiencies across multiple software platforms used during the radiation therapy process across a multi-facility practice.

3. Explore the use of metrics designed to quantify efficiency gains realized by implementing the Task Group's recommendations.

SU‐D‐Orange Blossom BR‐00

Stereotactic Breast Imaging Biopsy Systems

Moderator: Nicole Ranger

Aspirus Wausau Hospital, Wausau, WI

SU‐D‐Orange Blossom BR‐01

T. Aufdemberge

Medical Physics Consultants:
Evaluating the Performance of Stereotactic Breast Imaging Biopsy Systems

Medical Physicists are faced with a wider variety of approaches to Stereotactic Breast Biopsy (SBB) in the clinic with different manufacturer user interfaces and features and corresponding Quality Control requirements. This presents a continuing challenge as this technology and the practice of medical physics evolves. The presenter is an experienced mentor to trainees learning SBB who will review testing requirements and efficient methods to achieve regulatory compliance that fulfills the ACR accreditation criteria while applying Medical Physics 2.0 criteria. Specific tips/strategies will be provided for the traditional Fisher/Siemens and Lorad prone SBB units and for the modern Hologic upright and prone SBB machines which can be adapted to other manufacturer's upright SBB add‐on attachments. This “No hands hands‐on” course will increase your confidence and efficiency in performing SBB physics evaluations.

Learning Objectives:

1. Acquire/Refresh understanding of the regulatory/accreditation requirements for testing of Stereotactic Breast Biopsy (SBB) systems.

2. Learn additional techniques to perform testing of SBB systems in an efficient manner.

3. Recognize opportunities to extend the scope of medical physics support to facilities utilizing SBB to achieve Medical Physics 2.0 objectives,

The author is an employee of Medical Physics Consulting, a consulting medical physics company.

MO‐A‐Osceola BRC‐00

Using Our Tools. The Future Is Now!

Moderator: Sotirios Stathakis

University of Texas Health, San Antonio, TX

MO‐A‐Osceola BRC‐01

C. Bojechko

University of California San Diego:
Use of EPIDs for Routine Linac and Patient QA

MO‐A‐Osceola BRC‐02

J. Moran

University Michigan Medical Center:
Automation of QA

During the last two decades, there have been numerous publications on the applications of EPID for both machine QA, pre‐treatment IMRT QA, and in vivo dosimetry. There is interest in how fully automated, EPID‐based QA methods may bring efficiencies to clinical programs and support quality and safety. While some institutions have developed in‐house solutions to integrate EPIDs into their machine QA program, the lack of commercially available solutions has hindered adoption by other clinics. Fewer clinics have integrated the use of EPIDs as an in‐vivo tool. Additional barriers include defining which tests from AAPM reports TG 142 and MPPG 8a are suitable for routine evaluation with an EPID, how to define action levels, and how to confirm problems identified from EPID QA data. Advantages of EPID‐based machine QA include improvements in the efficiency, standardization and documentation of input and output, and leveraging automation to further streamline QA tests. The course will present applications of EPIDs for automation of machine quality assurance along with the benefits and potential risks. The critical role of complementary non‐EPID technology to ensure adequate performance will also be discussed. The unique strengths of EPIDs for in vivo applications will also be presented.

Learning Objectives:

1. To recognize the current and future role of EPIDs, including automation, for machine and pre-treatment IMRT QA in radiation oncology.

2. To understand practical considerations in implementing automation EPID QA and how to best incorporate these techniques as part of a comprehensive QA program.

3. To understand the benefits and risk associated with patient specific quality assurance when performing EPID QA pre-treatment and in-vivo.

Jean Moran: Varian Medical Systems Grant; Casey Bojechko: UCSD grant.

MO‐A‐Orange Blossom BR‐00

Radiography Review

Moderators: Matthew Hough, William Sensakovic

Florida Hospital, Altamonte Springs, FL, Florida Hospital, Orlando, FL

MO‐A‐Orange Blossom BR‐01

R. Fisher

MetroHealth:
Acceptance Testing Converted CR to DR Systems

MO‐A‐Orange Blossom BR‐02

C. Willis

MD Anderson Cancer Center:
Protocol Conversion and Optimization When Converting From CR to DR

MO‐A‐Orange Blossom BR‐03

K. Hulme

The Cleveland Clinic:
EI Tracking

Digital Radiography (DR) has the potential to further optimize Radiographic imaging by improving image quality and/or lowering dose beyond that achievable with Computed Radiography (CR). The CMS's Payment Reduction for X‐rays Taken Using CR regulation that went into effect January 1, 2018 aims to push providers to realize this potential by implementing DR systems. However, this goal can only be achieved if conversion to Digital Radiography (DR) is done with a quality driven approach. This session will provide information on how to accomplish this goal; from acceptance testing of systems converted from CR to DR, to protocol optimization techniques for systems converted from CR to DR, to Exposure Index (EI) tracking to ensure proper operational implementation of imaging techniques and identification of problem areas that need to be addressed.

Learning Objectives:

1. Understand principles of acceptance testing of systems converted from CR to DR.

2. Learn protocol optimization techniques for systems converted from CR to DR.

3. Learn EI tracking methods to ensure proper protocol implementation and identify and address problem areas.

CW is a member of the Medical Advisory Panel for GE Medical Systems – X‐ray Imaging Systems.

MO‐B‐Osceola BRC‐00

Present and Future of IMRT QA

Moderator: Brent Parker

University Texas Medical Branch of Galveston, Galveston, TX

MO‐B‐Osceola BRC‐01

B. Hasson

Anne Arundel Medical Center:
Introduction and Background of Patient Specific QA

MO‐B‐Osceola BRC‐02

J. Carroll

Versant Medical Physics, Radiation Safety:
TG‐218 and Its Implementation Into the Clinic

MO‐B‐Osceola BRC‐03

J. Park

University of Florida:
Logfile Based Patient Specific QA

MO‐B‐Osceola BRC‐04

M. Miften

University of Colorado School of Medicine:
Future Directions in Patient Specific QA

To provide attendees with an understanding of present IMRT QA practices and tools, overview of patient specific QA considering Task Group 218 recommendations, logfile/electronic IMRT QA, and insights into research efforts of the future of patient‐specific IMRT QA.

A main cornerstone in Clinical Medical Physics is conducting quality assurance measures to insure accurate and reliable treatments. The field of Medical Physics is a dynamically changing field that requires continual evolution of the processes and methods to deliver the highest quality care. The introduction and clinical implementation of Intensity Modulated Radiation Therapy (IMRT) in the 1990s created a paradigm shift in radiation therapy 1. However, the QA processes for a dynamic treatment delivery had not been developed. In 1994 TG‐40 was introduced to update the existing QA standards that were in place since 1984 2,3. Though TG‐40 was comprehensive, it did not address the quality assurance for MLC or dynamic treatment deliveries. In 2001, TG‐50 introduced concepts for quality assurance measures with MLCs 4. The first AAPM report that specifically addressed IMRT quality assurance was 082 published in 2003, and the authors stated, “IMRT was one of the most significant technical advances in Radiation Therapy” 5. The first report to highlight specific QA measures for IMRT based delivery machines was TG‐1426. TG‐218, published in 2018, addresses the tolerance limits and measurement methods for IMRT/VMAT QA 6. TG‐218 provides guidance for the Medical Physicist, and a solid basis for standardization of IMRT QA across institutions. However, TG‐218 does not provide recommendations or tolerance limits for electronic (logfile) based QA. This session will provide an understanding of the evolution of IMRT and VMAT QA techniques, and highlight the importance of the Medical Physicist role in advancing technologies while providing the QA measures to insure accurate and reliable treatments in the future.

Learning Objectives:

1. Recognize characteristics of IMRT specific phantoms and detectors.

2. The difference in measurement techniques for IMRT QA and recommendations from TG-218.

3. Specific QA measures associated with IMRT QA.

4. IMRT QA processes and limitations of portal imagers.

5. Logfiles to compliment or possibly replace measurement based IMRT QA.

MO‐B‐Orange Blossom BR‐00

The New ACR Mammography QC Manual

Moderator: Nicole Ranger

Aspirus Wausau Hospital, Wausau, WI

MO‐B‐Orange Blossom BR‐01

D. Gress

American College of Radiology:
An Overview of the New ACR DBT QC Tests and Phantom

MO‐B‐Orange Blossom BR‐02

D. Pfeiffer

Boulder Community Health:
Physics Field Experience using the New ACR Manual

MO‐B‐Orange Blossom BR‐03

T. Moore

Alliance Medical Physics LLC:
The New ACR 2018 Digital Mammography QC Manual: An Overview and Early Adopter Perspectives

Full Field Digital Mammography (FFDM) systems with Digital Breast Tomosynthesis (DBT) now represent approximately half of the installed mammography systems in the United States. Due to the proliferation of different manufacturers and models of 2D Full Field Digital Mammography (FFDM) units, each with their own unique QC manual, the American College of Radiology (ACR) created a new unified QC manual in 2016. However, at the time of its publication the 2016 ACR Digital Quality Control Manual was not approved to include DBT and could not be implemented at facilities that had 2D FFDM units with DBT capability. In the intervening time many 2D systems have been converted to also perform DBT, and in the Fall of 2018, the ACR released an updated Digital Mammography Quality Control Manual which includes a DBT supplement. With the ability to implement this FDA approved alternative standard on both 2D FFDM and DBT systems, a significant number of sites may now be interested in transitioning to the new 2018 ACR Digital Quality Control Manual which presents a new set of challenges for the medical physicist, especially those operating in a consulting environment. These presentations will review the requirements of the new ACR Digital Mammography QC Manual and the practical implementation of this QC program in two different environments: a facility with a hospital‐based medical physicist and one using the services of a consultant medical physicist. Each scenario presents unique challenges; these are discussed and possible solutions are given.

Learning Objectives:

1. Gain an understanding of the key elements of the new 2018 ACR Mammography QC manual including radiologist, technologist, and medical physicist requirements.

2. Learn specific details of the medical physics testing protocols and requirements with a special focus on the new DBT supplement

3. Understand the challenges associated with implementation of the FFDM and DBT elements of the program for consulting medical physicists.

MO‐C‐Osceola BRC‐00

Medical Physicist Wellness: The Personal and Professional Toll of Burnout

Moderator: Matthew Meineke

Ohio State University, Columbus, OH

MO‐C‐Osceola BRC‐01

D. Jordan

University Hospitals Cleveland Medical Center:
Medical Physicist Wellness ‐ Part 1

MO‐C‐Osceola BRC‐02

T. Pawlicki

UC San Diego:
Medical Physicist Wellness ‐ Part 2

A “perfect storm” of factors has led to a decline in physician wellness and a crisis of stress and burnout. The same factors affect medical physicists and there is anecdotal evidence that levels of stress and burnout are increasing. Maintenance of medical physicist wellness through management of stress and burnout is crucial to safe and effective patient care, personal and professional satisfaction, and the prevention of errors. There are effective strategies to identify and combat burnout and to restore practitioners to a state of well‐being. Systematic problems must be recognized and addressed, beyond individual interventions designed to temporarily address stress (the symptom).

This symposium will present causes of burnout stemming from systematic problems that exist at departmental, institutional, and national levels in the healthcare system. Concepts and approaches for recognizing stress and burnout, acknowledging the root causes, and implementing short‐term survival tactics and longer term corrective strategies will be discussed. It is important for individual medical physics to realize that the problem is not caused by a lack of skill, ability, or effort on their part. At the same time, medical physicists must be able to recognize warning signs of burnout in themselves and others and to understand the adverse impact that a burned‐out professional can create in the quality of patient care and the effectiveness of their colleagues. Presenters will guide attendees in active exercises on personal wellness assessment and strategies for managing burnout.

Learning Objectives:

1. Define wellness, stress, and burnout in medical physics professional practice and briefly review relevant literature on physician burnout and managing worker wellness.

2. Identify systematic sources of stress and root causes of burnout arising from individual, departmental, organizational, and national factors, as well as characteristics of successful interventions that have been used at each level.

3. Describe strategies for recognizing signs of burnout in self and colleagues and “emergency” interventions to stabilize a professional who is potentially impaired due to stress and burnout.

MO‐D‐Osceola BRC‐00

Watching the Surface for What Lies Beneath ‐ SGRT

Moderator: Sotirios Stathakis

University of Texas Health, San Antonio, TX

MO‐D‐Osceola BRC‐01

D. Stanley

The University of Alabama at Birmingham:
Overview of Current Practice and Clinical Implementation

MO‐D‐Osceola BRC‐02

R. Foster

Atrium Health/LCI NorthEast:
Clinical Applications and Future

There is growing popularity in the use of Surface Guided Radiation Therapy (SGRT) for ensuring the accuracy and precision of patient positioning. Physicists implementing SGRT in their clinic need guidance for commissioning the system, for designing effective ongoing QA procedures and for appropriate clinical use.

This session will describe the commissioning, ongoing QA, current clinical use of SGRT and look toward the future of this rapidly expanding technology.

Speakers will describe their experience commissioning and using SGRT for a wide variety of disease sites and including an FMEA analysis to examine potential failure modes for SGRT for deep inspiration breath hold (DIBH) patients. They will give an overview and an update on Task Group 302.

Learning Objectives:

1. Provide physicists with commissioning and QA guidance for SGRT (based on TG302).

2. Review of the clinical implementation for various treatment sites.

3. Discuss potential future clinical applications of surface imaging.

MO‐D‐Orange Blossom BR‐00

Pediatric Nuclear Medicine

Moderators: Matthew Hough

Florida Hospital, Altamonte Springs, FL and William Sensakovic, Florida Hospital, Orlando, FL

MO‐D‐Orange Blossom BR‐01

N. Kasraie

Children's Mercy Hospitals clinics: 
Pediatric Nuclear Medicine

Nuclear medicine and PET imaging exams performed on adults can also be performed on children and infants; however, pediatric disease processes and the logistical and technical demands for imaging of children may markedly differ from those practiced in the adult population. On the other hand, images in pediatric nuclear medicine must be of exceptional diagnostic quality while maintaining very low doses adjusted to each child according to established standards. For example, radiopharmaceutical administered doses can be adapted to the specific diagnostic task under consideration of the appropriate choice of collimator, so that the desired diagnostic information can be obtained at reduced radiopharmaceutical administered activity. Thus the application of diagnostic nuclear medicine in children presents a particular set of challenges that call for special consideration in order to yield premium‐quality clinical information to the referring practitioner. Namely, the acquisition of image data, the choice of instrumentation, and the processing method. This session will provide an overview of the physical aspects of the acquisition, the equipment used, and the image processing that is applied to these studies in the context of pediatric nuclear medicine and PET, with emphasis on distinctions in the anatomy, physiology, or logistical approaches between children and adult protocols. Having or lacking a clear understanding of these choices can enhance or hinder the quality of the nuclear medicine imaging which is essential for this specialized class of patients.

Learning Objectives:

1. Become familiar with the pediatric specific challenges in nuclear medicine and PET imaging and their distinctions from the adult side of things.

2. Learn of novel technological approaches in protocol and instrumentation design that have arrived or are on the near horizon such as PET-MR studies or novel post processing methods in conventional modalities.

3. Learn how recent developments such as improved detectors, increased computing power, and novel reconstruction algorithms have allowed for significant reduction in injected doses in pediatric imaging, and understand how these advances have generally led to exams with improved sensitivity or image quality enhancements with lower counts, allowing for the extension of these technologies to smaller children or the reduction in radiation dose to patients while still providing high-quality clinical results.

TU‐A‐Osceola BRC‐00

SRS/SBRT

Moderator: Brent Parker

University Texas Medical Branch of Galveston, Galveston, TX

TU‐A‐Osceola BRC‐01

J. Roper

Sarah Cannon:
Special Considerations for SRS/SBRT Treatments Using the Single‐Isocenter for Multiple Targets Technique

TU‐A‐Osceola BRC‐03

J. Chang

Northwell Health:
New Features of Gamma Knife Icon Poses Challenges to the Workflow and Scheduling of a Traditional Gamma Knife Clinic

TU‐A‐Osceola BRC‐04

S. Lim

Memorial Sloan‐Kettering Cancer Center:
Commissioning of the Treatment Planning System for SRS/SBRT Using Small Fields

TU‐A‐Osceola BRC‐05

A. Goenka

Northwell Health:
New Opportunities for Fractionated Stereotactic Radiotherapy Using Gammaknife Icon

Ongoing clinical trials show promising outcomes for ablative stereotactic radiation therapy in patients with oligometastatic disease and there is a growing trend in clinical practice toward aggressive local control using stereotactic radiosurgery (SRS) or stereotactic body radiotherapy (SBRT) techniques. The increased use of SRS/SBRT technique poses new challenges to physicists and other clinical staff members due to the labor intensive and time consuming treatment.

Icon is the latest Gamma Knife (GK) model to address these challenges. It is equipped with an on‐board imager for cone beam computed tomography so that patient setup can be verified with three dimensional tomographic images in real time. With this capability, both framed and frameless setup/treatment can be performed on GK for patients requiring brain SRS/SBRT. However, the workflow and treatment planning system are modified considerably for GK Icon to accommodate the improved flexibility and efficiency. These changes might make it difficult for clinics familiar with the previous GK models to adopt this new technology.

The single isocenter for multiple targets (SIMT) technique has gained clinical acceptance for the linear accelerator based (linac‐based) SRS/SBRT treatment of multiple brain metastases as it allows multiple targets to be treated simultaneously, even at different dose levels, using radiation beams that are at most centered on only one of the targets. The potential benefits of a more efficient SIMT treatment, however, are not without potential pitfalls. There are important differences between SIMT and conventional SRS, which require different considerations during treatment planning and delivery. SIMT aperture shaping is more complex and has considerable influence on plan quality. Additionally, rotational setup errors that have only a minimal effect on target coverage with conventional SRS can have a substantial effect on SIMT coverage, in particular when targets are separated by large distances.

Another new challenge is that both modalities use small fields for SRS/SBRT treatment. Traditionally, linac‐based treatment planning systems are optimized for field size larger than 30x30 mm2. These models can potentially produce substantial dose calculation errors when the SIMT technique is used to treat small targets (< 10 mm in diameter). This effect can be amplified when the small targets are located further away from the isocenter. Similarly for GK, the field sizes are intrinsically small (maximal cone size is 16 mm in diameter). Special attentions are needed for commissioning the treatment planning system. With increased interests in implementing SRS/SBRT procedures, good understanding of small field dosimetry is becoming a crucial responsibility for clinical physicists.

The purpose of this symposium is to address these new challenges in SRS/SBRT. We will first present clinical advantages and potential pitfalls of GK‐based and linac‐based brain SRS/SBRT. This will be followed by a review of various treatment planning strategies and workflow managements developed for both modalities. Finally, commissioning of treatment planning system for SRS/SBRT using small fields will be discussed.

Learning Objectives:

1. Understand the clinical advantages and potential pitfalls of GK-based and linac-based brain SRS/SBRT.

2. Be acquainted with workflow managements developed for frameless treatment using GK.

3. Conceptualize the effects of rotational setup errors, quantify dosimetric consequences, and explore different strategies to compensate for these errors during treatment planning and delivery using SIMT technique.

4. Get familiar with the commissioning of treatment planning system for SRS/SBRT using small fields.

TU‐A‐Orange Blossom BR‐00

Physics and Clinical Perspectives On Tomosynthesis

Moderator: Nicole Ranger

Aspirus Wausau Hospital, Wausau, WI

TU‐A‐Orange Blossom BR‐01

A. Maidment

University Pennsylvania:
Physics and Technology Overview of Tomosynthesis Systems

TU‐A‐Orange Blossom BR‐02

E. Conant

The University of Pennsylvania Health System:
A Radiologist's Perspective On Tomosynthesis

Digital breast tomosynthesis (DBT) is the new, better mammogram based on observed increases in specificity and cancer detection when compared to digital mammography (DM) alone. However, when DBT is combined with 2D planar imaging, the dose of the DM/DBT study is more than double that of a DM‐only mammogram. Synthesized digital mammographic (SM) images, that are “2D‐like” yet reconstructed from the tomosynthesis acquisition, have been developed to maintain the benefits of planar 2D imaging while decreasing the dose received from dual acquisition. Estimates of dose reduction with SM imaging are from 39‐45% and both population based studies and reader studies have found comparable specificity and cancer detection when screening is performed with either DM/DBT or SM/DBT. The incorporation of DBT into clinical practice is also associated with improvements in workflow efficiency. DBT offers improved lesion conspicuity and the ability to localize lesions within the 3D image stack, allowing more direct patient flow to targeted ultrasound and improved reader confidence in lesion characterization. DBT‐guided biopsy devices also allow tissue sampling of lesions previously not accessible by stereotactic‐guided core biopsy, often at a reduced radiation dose. In this course, the fundamental physics of DBT imaging will be reviewed in light of commercial and future system designs, and the dosimetry and clinical physics of DBT systems will be discussed. Clinically, case‐based examples will be shown demonstrating improvements in specificity, cancer detection, and workflow for both DM/DBT and SM/DBT.

Learning Objectives:

1. Understand the physics and technology of tomosynthesis, available today and in the near future.

2. Appreciate the determinants of image quality, and gain insight into the source and presentation of image artifacts in DBT.

3. Compare the outcomes from screening with digital breast tomosynthesis to those of digital mammography alone.

4. Assess the benefits of digital breast tomosynthesis in the diagnostic evaluation of breast lesions.

5. Evaluate the impact of digital breast tomosynthesis on the breast imaging clinical workflow.

TU‐B‐Osceola BRC‐00

ABR, Regulatory, and Legislative Update

Moderator: Matthew Meineke

Ohio State University, Columbus, OH

TU‐B‐Osceola BRC‐01

M. Reiter

Capital Associates, Inc.:
A Post Midterm Look at the Legislative Environment

TU‐B‐Osceola BRC‐02

J. Elee

LA Dept of Environmental Qual:
Update on CRCPD's Medical Event Reporting Data and Review of other Current State Regulatory Issues

TU‐B‐Osceola BRC‐03

R. Martin

AAPM:
Trending Regulatory Issues Impacting the Medical Use of Radiation

TU‐B‐Osceola BRC‐04

K. Kanal

University Washington:
Getting Ready for the ABR Online Longitudinal Assessment (OLA)

Matt Reiter, AAPM Lobbyist Consultant—What Can We Expect from The New Congress? This session will provide an overview of the current federal legislative environment for healthcare issues. Mr. Reiter will talk about the post‐midterm Congress broadly and then look at policy changes that may impact the clinical work of medical physicists. The presentation will discuss AAPM's efforts to educate legislators on the benefits of the medical use of radioisotopes, and efforts to maintain patient access to radioisotopes. In addition, the session will address any draft bills introduced and/or anticipated that would be important to AAPM's advocacy priorities, including source security and funding for low‐dose radiation research. Mr. Reiter will offer some predictions for the year ahead and talk about how AAPM's advocacy strategies might change in the coming months.

Jennifer Elee—CRCPD/AAPM Collaborations, Medical Event Reporting Update, and Current State Regulatory Issues: This presentation will provide an update on some Conference of Radiation Control Program Directors (CRCPD) collaborations with AAPM. Ms. Elee will examine issues that state regulators are concerned about that may impact medical physicists. In addition, the presentation will look at the CRCPD's continuing efforts to create a national database of radiation medical events, providing a single point for all states to input events into a single database. The session will summarize recent data, offer analysis of radiation medical event data collected, and discuss the lessons learned so far. Ms. Elee has worked extensively on medical event reporting and incident learning. The presentation will address Ms. Elee's work with the International Atomic Energy Agency (IAEA) and other radiation safety organizations on this issue and talk about the path forward.

Richard Martin—Regulatory Activities Impacting Medical Physicists: This presentation will examine current and trending regulatory issues that are important to AAPM's advocacy at federal and state agencies, including radiation safety, source security, and patient access to radioactive sources. Mr. Martin will discuss the Government Accountability Office (GAO) report on source security and its anticipated regulatory impact. The presentation will examine U.S. Nuclear Regulatory Commission (NRC) activities, look at the NRC's reaction, if any, to the GAO report, and provide a status update of the NRC's re‐evaluation of Training and Experience requirements for Authorized Users (AUs). The session also will look at amendments to the Mammography Quality Standards Act (MQSA) regulations and discuss what the amendments will mean to the medical physics community. In addition, the presentation will address trending state regulatory issues and look at rulemaking interpreting “Qualified Medical Physicist” (QMP) and further defining scope of practice for medical physicists.

Kalpana Kanal, PhD, Trustee, American Board of Radiology ‐ Getting Ready for the ABR On‐line Longitudinal Assessment (OLA): Medical Physicists participate in Maintenance of Certification (MOC) which is an integral part of the quality movement in healthcare. In 2012, the American Board of Radiology (ABR) implemented a new MOC process, known as Continuous Certification, for all participating MOC diplomates. The Continuous Certification method uses an annual review in March to evaluate all four MOC parts and fees and render MOC participation status. The four parts of MOC are: Part 1: Professionalism and Professional Standing, Part 2: Lifelong Learning and Self‐Assessment, Part 3: Assessment of Knowledge, Judgment, and Skills and Part 4: Improvement in Medical Practice.

In this talk, I will focus on part 3. Part 3 requires passing the most recent summative decision for the Online Longitudinal Assessment (OLA). In May 2016, the ABR announced its plans to move away from the 10‐year exam to OLA. Within the OLA software, diplomates will be provided two question opportunities each week. Question opportunities will be available for four weeks to allow maximum diplomate flexibility. After opening a question, diplomates will be allowed a limited amount of time to answer the question and will learn immediately whether they answered correctly. In addition, they will receive a brief explanation of the correct answer, as well as a reference. I will demo the OLA software and how it works in this talk.

OLA is designed to have minimal impact on a diplomate's workday and requires no time away from work or travel expense. The potential for retesting areas of weakness provides a further opportunity for diplomates’ self‐assessment of their professional growth. OLA will be available to all diagnostic radiology diplomates in early 2019. OLA for radiation oncology, medical physics, and interventional radiology diplomates will follow in 2020.

Learning Objectives:

1. To understand the political environment impacting medical use of radiation, research funding, and health policy.

2. To learn about AAPM's federal legislative advocacy related to medical use of radiation and access to radioisotopes.

3. To understand the CRCPD/AAPM collaborative activities.

4. To learn about state regulatory issues that may impact the practice of medical physics.

5. To understand the CRCPD's event reporting activities and lessons learned from the CRCPD's event data.

6. To learn about current regulatory issues impacting the practice of medical physics.

7. To understand issues related to medical use of radiation that may impact patient access to radioactive sources.

8. To understand issues recently addressed by state regulatory agencies and learn about AAPM's state rulemaking advocacy efforts.

9. To learn about the important changes concerning the MOC Part 3: Assessment of Knowledge, and its transition to Online Learning Assessment (OLA).

10. To see a live demonstration of the OLA module in its current form and discuss its functionality and utility.