S. Tung

SU‐FF‐T‐94: Clinical Evaluation of Direct Machine Parameter Optimization Algorithm for Head and Neck IMRT Treatment

Purpose: Because many critical structures are in close proximity to target volumes, cancers of the head and neck (H&N) are often suited for treatment with IMRT. However, the time required to generate and deliver a clinically acceptable IMRT plan can be significantly longer than a conventional plan. This study evaluated a new inverse planning algorithm, DMPO (direct machine parameter optimization), with attention to parameter settings, plan quality and treatment efficiency for H&N cancers. Method and Materials: The Pinnacle treatment planning system version 7.4 was used. The DMPO allows users to limit the number of total MLC segments (N) for treatment. After a user-defined number of iterations (n) for pencil beam optimization, the DMPO generates MLC segments for each field for dose calculations using a convolution algorithm. Both the MLC leaf positions and the weight of each segment are then optimized until cost tolerance or iteration number is reached. Treatment plans generated using DMPO were compared with H&N cases that were previously treated using an older version (6.2). The plan quality was compared using cost functions and DVHs of target volumes and critical structures. The total monitor units and MLC segments for treatment were compared for different combinations of n and N. Results: The DMPO provided plans of DVHs similar to clinical cases with significantly less planning time. More importantly, the total MU and MLC segments for treatment delivery were reduced by 40% to 50%. Cost functions changed only slightly on n and N and total MU increased as n increased, but was independent of N. Our preliminary data indicated a combination of n=10–15 with 10 segments per field appeared to be optimal for most H&N cases. Conclusion: The DMPO algorithm generated more efficient plans while providing equal or better quality than the previous plans for IMRT treatment.

SU‐FF‐T‐179: Accuracy of CT Numbers and Its Effect On Dose Calculations

Purpose: The CT‐to‐density conversion table can vary for various CTscanners or when using different scanning parameters (such as KVp, field‐of‐view etc.). These CT number variations could be up to −/+100 for the same material. However, the dosimetric impact of these CT number variations was not well studied in the past. This study is to report the variations of CT numbers and their clinical impact on dose calculations. Method and Materials: The CT‐to‐density conversion table was measured on six different scanners. Two special tables were created with +/− 100 CT‐number shifts from the default conversion table in our clinic. Data showed that these two tables represented two bounding situations to accommodate the variations in different CTscanners or scanning parameters. Treatment plans using lung, head and neck, prostate, and breast cases were recalculated using these new CT conversion tables for both photon and electron beams.Results: There were noticeable differences across different density ranges for different CTscanners. Using the two extreme CT‐to‐density conversion tables, it was shown that the differences in dose calculations were −1.6% to 0.3% for head & neck (6MV), −2.5% to 0.9% for lung (6MV), −4.0% to 3.0% for prostate (6MV), and −2.3% to 1.7% for prostate (18MV), respectively. It showed that the impact was more notable for treatment targets at deeper depths or at lower beam energy. For electron beams, the distal 90% fall‐off edge was changed approximately 2 mm for commonly used electron energies. Conclusion: Dose calculations seem to be not sensitive to different CT‐density conversion tables from different CTscanners or using different CT scanning parameters. Considering most scanners were calibrated to water, our results could be the worst case scenario. A single CT‐to‐density conversion table could be used for all CTscanners.

SU-F-T-453: Improved Head and Neck SBRT Treatment Planning Using PlanIQ

PURPOSE Treatment planning for Head and Neck(HN) re-irradiation is a challenge because of ablative doses to target volume and strict critical structure constraints. PlanIQ(Sun Nuclear Corporation) can assess the feasibility of clinical goals and quantitatively measure plan quality. Here, we assess whether incorporation of PlanIQ in our SBRT treatment planning process can improve plan quality and planning efficiency. METHODS From 2013-2015, 35 patients (29 retrospective, 6 prospective) with recurrent HN tumors were treated with SBRT using VMAT treatment plans. The median prescription dose was 45 Gy in 5 fractions. We retrospectively reviewed the treatment plans and physician directives of our first 29 patients and generated score functions of the dosimetric goals used in our practice and obtained a baseline histogram. We then re-optimized 12 plans that had potential to further reduce organs-at-risk (OAR) doses according to PlanIQ feasibility DVH and plan quality analysis and compared them to the original plans. We applied our new PlanIQ-assisted planning process for our 6 most recently treated patients and evaluated the plan quality and planning efficiency. RESULTS The mean plan quality metric(PQM) and feasibility adjusted PQM(APQM) scores of our initial 29 treatment plans were 77.1±13.1 and 88.7±11.9, respectively (0-100 scale). The PQM and APQM scores for the 12 optimized plans improved from 75.9±11.0 and 85.1±10.2 to 80.7±9.3 and 90.2±8.0, respectively (p<0.005). Using our newly developed PlanIQ-assisted planning process, the PQM and APQM scores for the 6 most recently treated patients were 93.6±6.5 and 99.1±0.6, respectively. The planning goals were more straightforward to minimize OAR doses during optimization, thus less planning and revision time were used than before. CONCLUSION PlanIQ has the potential to provide achievable planning goals and also improve plan quality and planning efficiency.

SU-F-T-406: Verification of Total Body Irradiation Commissioned MU Lookup Table Accuracy Using Treatment Planning System for Wide Range of Patient Sizes

PURPOSE To verify the accuracy of total body irradiation (TBI) measurement commissioning data using the treatment planning system (TPS) for a wide range of patient separations. METHODS Our institution conducts TBI treatments with an 18MV photon beam at 380cm extended SSD using an AP/PA technique. Currently, the monitor units (MU) per field for patient treatments are determined using a lookup table generated from TMR measurements in a water phantom (75 × 41 × 30.5 cm3). The dose prescribed to an umbilicus midline point at spine level is determined based on patient separation, dose/ field and dose rate/MU. One-dimensional heterogeneous dose calculations from Pinnacle TPS were validated with thermoluminescent dosimeters (TLD) placed in an average adult anthropomorphic phantom and also in-vivo on four patients with large separations. Subsequently, twelve patients with various separations (17-47cm) were retrospectively analyzed. Computed tomography (CT) scans were acquired in the left and right decubitus positions from vertex to knee. A treatment plan for each patient was generated. The ratio of the lookup table MU to the heterogeneous TPS MU was compared. RESULTS TLD Measurements in the anthropomorphic phantom and large TBI patients agreed with Pinnacle calculated dose within 2.8% and 2%, respectively. The heterogeneous calculation compared to the lookup table agreed within 8.1% (ratio range: 1.014-1.081). A trend of reduced accuracy was observed when patient separation increases. CONCLUSION The TPS dose calculation accuracy was confirmed by TLD measurements, showing that Pinnacle can model the extended SSD dose without commissioning a special beam model for the extended SSD geometry. The difference between the lookup table and TPS calculation potentially comes from lack of scatter during commissioning when compared to extreme patient sizes. The observed trend suggests the need for development of a correction factor between the lookup table and TPS dose calculations.

SU‐E‐T‐389: Effect of Interfractional Shoulder Motion On Low Neck Nodal Targets for Patients Treated Using Volume Modulated Arc Therapy (VMAT)

PURPOSE To quantify the dosimetric impact of interfractional shoulder motion on targets in the low neck for head and neck patients treated with volume modulated arc therapy (VMAT). METHODS Three patients with head and neck cancer were selected. All three required treatment to nodal regions in the low neck in addition to the primary tumor. The patients were immobilized during simulation and treatment with a custom thermoplastic mask covering the head and shoulders. One VMAT plan was created for each patient utilizing two full 360° arcs. A second plan was created consisting of two superior VMAT arcs matched to an inferior static AP supraclavicular field. A CT-on-rails alignment verification was performed weekly during each patients treatment course. The weekly CT images were registered to the simulation CT and the target contours were deformed and applied to the weekly CT. The two VMAT plans were copied to the weekly CT datasets and recalculated to obtain the dose to the low neck contours. RESULTS The average observed shoulder position shift in any single dimension relative to simulation was 2.5 mm. The maximum shoulder shift observed in a single dimension was 25.7 mm. Low neck target mean doses, normalized to simulation and averaged across all weekly recalculations were 0.996, 0.991, and 1.033 (Full VMAT plan) and 0.986, 0.995, and 0.990 (Half-Beam VMAT plan) for the three patients, respectively. The maximum observed deviation in target mean dose for any individual weekly recalculation was 6.5%, occurring with the Full VMAT plan for Patient 3. CONCLUSION Interfractional variation in dose to low neck nodal regions was quantified for three head and neck patients treated with VMAT. Mean dose was 3.3% higher than planned for one patient using a Full VMAT plan. A Half-Beam technique is likely a safer choice when treating the supraclavicular region with VMAT.

SU-E-T-522: A Multi-Isocenter VMAT Technique for Cranio-Spinal Irradiation.

PURPOSE Develop a matching VMAT field technique and investigate planning feasibility for treating the entire central nervous system (CNS) using Cranio-Spinal Irradiation (CSI) . METHODS Two patients diagnosed with acute myeloid leukemia (AML) presented with CNS involvement, received CSI, and were included in this study. The patients were treated with the traditional CSI technique: prone position, opposing lateral brain fields, two posterior fields (upper and lower spine), and 5mm junction shifts to improve dose uniformity. The patients were retrospectively re-planned using volumetric arc therapy (VMAT). The spine and brain were contoured to create the clinical target volume (CTV) as well as normal tissues including kidneys, lung and heart for optimization. Three isocenters were used for planning: brain, upper and lower spine. The beams were allowed to overlap by approximately 10cm. Entire 360 degree rotations were used for the brain fields and posterior 120 degree arcs were used for the spine fields. The dosimetric coverage of the target between the VMAT and traditional plans was compared, as well as the dose to normal tissues. RESULTS Both VMAT plans achieved improved dose uniformity in the CTV (standard deviation < 2%), and reduced hot spots (<110%). Dose to the heart was reduced, with the V10 being 12.7% and 28.2%, compared to 44.6% and 50.2%, respectively, for the traditional plan. Dose to the total lung V5 increased for the VMAT plans for both patients (21.6% and 27.8% compared to 12% and 13% respectively). The results for the kidneys were mixed with the mean dose increasing for one patient and decreasing for the other . CONCLUSION The efficacy of planning CSI treatments using a matching VMAT technique was demonstrated. The developed technique has the potential to improve dose uniformity to the target while at the same time reduce the risk of under or over dosing the spine.

SU-E-T-404: Simple Field-In-Field Technique for Total Body Irradiation in Large Patients

PURPOSE A simple Field-in-Field technique for Total Body Irradiation (TBI) was developed for traditional AP/PA TBI treatments to improve dosimetric uniformity in patients with large separation. METHODS TBI at our institution currently utilizes an AP/PA technique at an extended source-to-surface distance (SSD) of 380cm with patients in left decubitus position during the AP beam and in right decubitus during the PA beam. Patients who have differences in thickness (separation) between the abdomen and head greater than 10cm undergo CT simulation in both left and right decubitus treatment positions. One plan for each CT is generated to evaluate dose to patient midline with both AP and PA fields, but only corresponding AP fields will be exported for treatment for patient left decubitus position and PA fields for patient right decubitus position. Subfields are added by collimating with the x-ray jaws according to separation changes at 5-7% steps to minimize hot regions to less than 10%. Finally, the monitor units (MUs) for the plans are verified with hand calculation and water phantom measurements. RESULTS Dose uniformity (+/-10%) is achieved with field-in-field using only asymmetric jaws. It is dosimetrically robust with respect to minor setup/patient variations inevitable due to patient conditions. MUs calculated with Pinnacle were verified in 3 clinical cases and only a 2% difference was found compared to homogeneous calculation. In-vivo dosimeters were also used to verify doses received by each patient with and confirmed dose variations less than 10%. CONCLUSION We encountered several cases with separation differences that raised uniformity concerns - based on a 1% dose difference per cm separation difference assumption. This could Resultin an unintended hot spot, often in the head/neck, up to 25%. This method allows dose modulation without adding treatment complexity nor introducing radiobiological variations, providing a reasonable solution for this unique TBI situation.

SU‐E‐T‐592: Investigation of Volumetric Modulated Arc Therapy for Total Body Irradiation with Lung Blocks

PURPOSE Uncertainties from the use of lung blocks during total body irradiation (TBI) are present due to the process of block cutting and block mounting required for treatment. We investigated the use of volumetric modulated arc therapy (VMAT) to improve clinical TBI practice. METHODS Two pediatric patients (Pat-1, Pat-2) and one adult patient (Pat-3) had CT scans from head to upper thigh. The treatment targets (12Gy) included bony structures (all patients) and lymph nodes+spleen (LNS) (Pat-1, Pat-3). Three plans were generated in Pinnacle for each patient: TBI plans with AP/PA fields at 380cm SAD with and without lung blocks and a VMAT plan at 100 SAD with 3-4 couch positions. Target coverage V12Gy and dose to lung V6Gy were compared among three plans for each patient. Treatment times as well as doses to other critical structures were also compared for different techniques. RESULTS for AP/PA plans, lung blocks reduced lung V6Gy to 28.7%, 6.3% and 24.9% for these three patients from 100% (without block), however, target V12Gy decreased to 72.5%, 68.6%, and 88.7% from 84.9%, 86.3%, 95.5% (without block), respectively. VMAT plan achieved 44.3%, 53.7% and 71.5% for lung V6Gy with excellent target coverage of 91.7%, 96.2% and 95.3%. While lung blocks only be used 3/8 treatments for AP/PA TBI, VMAT plan lung mean doses were 21-35% lower than AP/PA+block technique. Meanwhile, average mean dose to the brain, heart, kidneys and liver was also reduced 40%, 43.7%, 41.1% and 40.6% using VMAT plans. VMAT planning was more challenging when targets include LNS and/or for large patients (shoulder>40cm, Pat-3). Setup and beam-on time for VMAT were comparable to or less than AP/PA technique for Pat-1, Pat-2, but were much longer for Pat-3. CONCLUSION VMAT plans have potential to reduce uncertainties for treating small-medium sized patients who need lung blocks in TBI treatments.

SU‐E‐T‐804: Total Scalp Irradiation Using Volumetric Modulated Arc Therapy and Comparison with IMRT and Conventional Photon/Electron Field Combination

Purpose: Irradiation of total scalp with conventional photon/electron fields has been challenging. The unique target shape adds complexity to the beam setup and poses difficulties in achieving a homogeneous plan. Patient setup can be cumbersome and time consuming. IMRT provides excellent dose distributions and simplifies the process of field boundary matching; however beam delivery time is still long due to large number of fields. VMAT has the capability to deliver a homogeneous dose with reduced set‐up and treatment time. VMAT will be compared to other techniques in this work.Methods: Seven patients have been planned for total scalp irradiation using conventional photon/electron planning, IMRT, and SmartArc VMAT. After optimized plans have been achieved using Pinnacle3 planning system, dosimetric comparisons were made between different methods on each patient. Results: Compared with conventional therapy, both IMRT and VMAT yield more homogeneous plan on all the patients. V100% target coverage can easily increase to 98% or higher from approximately 75%. Maximum dose reduces to less than 108% of prescribed dose from 150%. The conventional method performs better in protecting critical organs. However, brain volume at D50%, maximum and mean doses to the cord and brainstem can all be reduced by more than 50% compared to IMRT and VMAT. As a comparison between VMAT and IMRT, in all the cases, IMRT plans are slightly more uniform, while VMAT generally demonstrates better critical organ protection. Different dose levels can be easily achieved without boost using IMRT or VMAT. An added bonus of VMAT is shortened beam delivery time, which is 2.5 minutes at maximum with double arc. Conclusions: VMAT offers a viable alternative to conventional techniques in treatment of the total scalp. Advantages are improved dose distributions to the target volumes compared to conventional, and shorter treatment times compared to IMRT.

SU‐E‐T‐456: Comparison of Dosimetric Effect of Translational Setup Errors for Total Scalp Irradiation Using Conventional Photon‐Electron Combination, IMRT and VMAT Techniques

Purpose: IMRT and VMAT showed superior total scalp coverage and uniformity over conventional photon/electron combination technique. This study is to investigate and compare the dosimetric effects of translational setup errors for these techniques. Methods: Three total scalp irradiation patients were studied. For each patient, three treatment plans were generated using conventional, IMRT and SmartArc VMAT technique. In order to evaluate the dosimetric effect due to translational setup errors, 36 additional plans were calculated for each patient on the planning‐CT image, which included isocenter shift of 2mm and 5mm in lateral (LR), anterior‐posterior (AP) and superior‐inferior (SI) directions for each of the three plans. The dose distributions for target and organs at risk (OAR) (brain, brainstem and spinal cord) from the 36 plans were compared to the original plans and among the three techniques using dose‐volume histograms, tumor‐volume coverage of prescription dose and the maximum OAR dose. Results: Conventional plans were robust to lateral shift (less than 0.8% tumor coverage drop and less than 0.8cGy brain mean dose increases for up to 5 mm isocenter shift) because only lateral beams were used. Tumor coverage drops were apparent when isocenter moved towards posterior and superior scalp. At 2mm shift, tumor coverage drop for VMAT plans were comparable to conventional plans (LR 0.4%, AP 1.4%, SI 2.3%), while for IMRT plans, the coverage dropped 3.3%, 3.0% and 2.6% in LR, AP and SI direction, respectively. At 5mm shift, both IMRT and VMAT plans had significant higher coverage drop than conventional plans with IMRT plans the worst in all directions. OAR dose increases were less than 50cGy for conventional plans, and were up to 150cGy for IMRT and VMAT plans at 2mm isocenter shift. Conclusions: The dosimetric effect were comparable between VMAT and conventional technique when translational setup errors were within 2mm. IMRT plans showed worse effect than VMAT.