K Deschesne

SU‐GG‐T‐88: A Multi‐Institutional Retrospective Study On Clinical IMRT Treatment Delivery Efficiency

Purpose: To better understand IMRT delivery efficiency in daily clinical operation we conducted a five‐institution retrospective study on clinical IMRTtreatment delivery time and IMRT MUs as functions of a number of variables — the number of fields per IMRTtreatment,treatment site, total MUs, and total number of segment fields for MLC‐IMRT treatment. The goal is to identify major contributing factor(s) for IMRT delivery time for each IMRT approach for improvement. Method and Materials: Data from more than 350 patients were extracted from IMPAC/LANTIS R&V system. A 5‐day average value was taken for any time data. The IMRTtreatment delivery time represents the portion of total patient delivery time that is specific to IMRT delivery technique and is defined as the time elapsed between the beam‐ON of the first field/segment and the beam‐OFF of the last field/segment in an IMRTtreatment. The accelerators studied are Siemens, Varian, and Elekta. Treatment planning systems used are PLanUNC, ADAC Pinnacle, and CMS XiO. The four IMRT delivery approaches studied are: segmental MLC‐IMRT on Siemens accelerator, segmental MLC‐IMRT on Elekta accelerator, dynamic MLC‐IMRT on Varian accelerator, and compensator‐IMRT on Siemens accelerators. Results: Our initial results show that for MLC‐IMRT treatments the IMRTtreatment delivery time is closely correlated with the number of segment fields and less correlated with the total number of MUs and treatment site. There is a large variation in IMRTtreatment delivery time for IMRTtreatment of a given number of fields, depending on the accelerator and IMRT approach used. Conclusion: Our five‐institution retrospective study on clinical treatment delivery data shows that manual compensator‐IMRT treatment is among the fastest of the five IMRT delivery approaches studied. In average all IMRT delivery approaches spent only approximately 20% of the IMRT delivery time (as defined) on actual radiation delivery.

SU-FF-T-119: Comparison of Compensator-IMRT and Segmental MLC-IMRT Techniques: A Retrospective Study On Treatment Time, Monitor Units, and Dosimetry in Clinical Application

Purpose: There are more than 150 radiotherapy centers in the US using compensators to deliver IMRT treatments. However, compensator‐IMRT is still not a well‐understood technique to many; better knowledge of the technology can encourage acceptance and proper use of this valuable IMRT delivery technology. We will present our compensator‐IMRT experience in clinical application using retrospective patient treatment data from eleven years (1100 patients) of compensator‐IMRT experience and five years (500 patients) of segmental MLC‐IMRT in parallel. Method and Materials:IMRT plans are designed by the in‐house IMRT TPS PLanUNC. The resulting continuous intensity maps are used for compensator design. The intensity maps are converted to discrete maps for MLC segment generation if segmental MLC‐IMRT technique is used. Compensators are fabricated using a Par Scientific milling machine and granular compensator material. This study uses data from the treatment RV 2) both IMRT delivery techniques use similar monitor units; and 3) the high spatial resolution compensator‐IMRT generally has a similar or better dosimetric quality compared to the segmental MLC‐IMRT technique. Conclusion: We evaluated both the compensator‐IMRT and segmental MLC‐IMRT delivery techniques in terms of treatment delivery time, treatment monitor units, and dosimetric quality (DVH and EUD). Our experience demonstrated that the compensator‐IMRT technique delivered high quality IMRTdosimetry, fast IMRT treatments, and similar monitor units compared to the segmental MLC‐IMRT technique.

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

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

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

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

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

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

SU‐GG‐J‐109: A Study on Image Guided Radiotherapy Using Cone‐Beam Computed Tomography (CBCT) for Head and Neck Cancer IMRT Patient Setup

Purpose: To use our CBCT shift results to determine how much of the CTV to PTV margin is necessary to compensate for the daily setup uncertainty inherent in the treatment of head and neck cancer patients. Method and Materials: To accommodate the daily setup uncertainly, a 3–6 mm uniform margin is used for head and neck IMRT patients to expand CTV to PTV. In order to verify patient setup accuracy, CBCT is taken prior to the treatment. Our protocol calls for CBCT for the first five consecutive fractions, followed by two CBCT every week. Each CBCT is registered to the planning CT to determine the corresponding translational shift. The registration is based on the region‐of‐interest—the rigid bones near the tumor. On the days no CBCT is taken, the average of the previous three calculated translational shifts is used instead. No rotational correction is considered. Results: A total of 22 head and neck IMRT patients were studied with a Siemens Artiste linear accelerator using an MVision CBCTscanner. Each patient had approximately 14 sets of CBCTimages taken over the entire course of treatment. The average interfraction shift was 0.6±1.6, 1.4±1.5, 0.1±1.6 mm in the anterior‐posterior, left‐right, and superior‐inferior dimensions, respectively. These data indicate that in the absence of CBCTimage guided radiotherapy, a CTV to PTV margin of 2.5, 4.7 and 1.3 mm is needed in the anterior‐posterior, left‐right, and superior‐inferior dimensions, respectively. Conclusion: Our results have shown that asymmetric margins of 2.5, 4.7 and 1.3 mm can be used for CTV to PTV expansion for head and neck cancerradiotherapy if no IGRT presented. Our CBCT protocol will allow a reduction in the CTV to PTV margin. Further investigation needs to be done to find out the specific CTV to PTV margin with different CBCT frequency.

SU-FF-I-90: A Clinical Evaluation of the M-Rep-Based Automatic Prostate Segmentation

Purpose: To evaluate the clinical application readiness of a statistically trainable deformable shape model, called an m‐rep, for automatic segmentation of the prostate from CTimages by comparison with manual contouring. Method and Materials: ConStruct, an in‐house automatic prostate segmentation tool featuring m‐rep (medial representations) models, was used in this study. This study is to test the robustness of the m‐rep models for prostate treatment planning segmentation in routine use. For each patient case, an m‐rep was initialized and automatically deformed to segment the prostate. Ground truth was taken to be the contour set used for treatment planning. Two radiation oncologists each manually created an additional contour set. The average distance between m‐rep segmentation (or manual segmentation) and the ground truth was calculated. Retrospective clinical validation was also carried out in all cases. Two radiation oncologists blindly scored all the contour sets including the ground truth. Score 1: perfect contour—no corrections needed; 2: minor correction—less than one third of the slices require correction; 3: major correction—more efficient than redrawing all the contours from scratch; 4: not acceptable. Results: Averaged over five patients, the average distances between m‐rep, set 1 and 2 manual contour sets and the ground truth are 0.51, 0.45, and 0.42 cm, respectively. The average scores for m‐rep, set 1, set 2 contour sets, and the ground truth are 2.4, 2.4, 2.0, and 2.1, respectively. Conclusion: The qualitative and clinical validations were performed on m‐rep‐based automatic segmentation by comparing it to manual contouring. M‐reps appeared to produce reasonable and acceptable prostate contour sets when compared to those generated by clinicians. We propose that m‐reps be used to improve clinical efficiency by automatically contouring the prostate as a starting shape and then manually editing when needed. Further validation on a larger patient cohort is indicated.

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

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

SU‐FF‐T‐137: IMRT Treatment Delivery Efficiency — A Multi‐Institutional Retrospective Study

Purpose: To compare daily clinical IMRT delivery efficiency in 6 institutions using different accelerators, delivery techniques, treatment planning systems, and clinical practice environments; to deduce key contributing factors for IMRT delivery efficiency improvement. Method and Materials:IMRTtreatment MUs, daily treatment delivery time (time elapsed between beam‐ON of the 1st field and beam‐OFF of the last field), and other parameters from Record & Verify systems for 421 patients using accelerators from 4 different vendors and 4 different accelerators treatment planning systems are retrospectively analyzed.IMRTtreatments are delivered using compensator‐IMRT on Siemens, segmental MLC‐IMRT on Siemens, Elekta and Varian accelerators, and via TomoTherapy. RESULTS: In average the shortest average IMRT delivery times are associated with TomoTherapy (7.3 min.), plans using the least MLC segments (10.1 min.) and compensator‐IMRT (11.3 min.) (times quoted for 9‐fld IMRT). Longer delivery time is not mainly due to more MUs as beam‐ON time for all LINAC‐based IMRT were all < 2 min., which is only < 20% of the IMRT delivery time. The majority of the IMRT delivery time is spent on preparation of delivery such as segment field formation and verification. LINAC‐based IMRT MUs are < 7 times TomoTherapy‐IMRT MUs. The average MU ratio of LINAC‐based IMRT to non‐IMRT treatments is less than 2:1. CONCLUSION: Major improvement in the MLC‐IMRT delivery time requires significant reduction in MLC leaf motion and verification time. In meantime the most efficient approach to delivery time reduction is to reduce the total number of MLC segment. The retrospective study shows that compensator‐IMRT delivery time is comparable to the most efficient MLC‐IMRT, and TomoTherapy uses the shortest IMRT delivery time. Conflict of Interest: Thomas Mackie has financial interest in TomoTherapy Inc. Sha Chang has a research grant (unrelated to this work) from Siemens.

SU‐GG‐T‐09: Dose Accumulation From Film‐Based Brachytherapy Planning and CT‐Based External Beam Radiotherapy Planning

Purpose: Radical radiation therapy that combines external beam therapy (EBRT) and brachytherapy (BRT) is effective in managing local‐regional confined cervical cancer. Although CT is widely used for EBRTtreatment planning (TP), traditional 2D film is still commonly used today in many institutions for BRT TP. The incompatible image information between the BRT and EBRT leads to great difficulty in computation of the cumulative radiationdose from both treatments. To date, doses to target and critical structures are approximated by adding the point doses of BRT to the EBRT plan. In this project, we propose to register the orthogonal films of BRT and CT of EBRT so as to accumulate the doses in a more accurate way. Method and Materials: Five patients treated with both EBRT and low dose rate BRT were retrospectively used in the study. 3D dose grids from the BRT and EBRTtreatment planning were merged by a film‐to‐CT registration, which was accomplished by creating DRRs from EBRT‐CT in the same projections of the BRT films. A landmark based image registration tool was developed to register films and DRRs of CT. The calculated shift, rotation and scaling were applied on the dose grid of BRT. Once the registration was completed, dose distributions from EBRT and BRT were merged. Results: A composite plan with accumulated doses of BRT and EBRT was created for each patient. The dose accuracy was verified at relevant points such as “A” points. The isodose curve and DVHs were analyzed, and appear reasonable. Conclusion: We developed a method to accumulate the dose distributions from film‐based brachytherapy and CT‐based external beam radiotherapy. This yielded a more realistic estimate of the cumulative dose received by the patient from both treatments.