ESTRO 2020 Abstract book

S1040 ESTRO 2020

prostate cancers including elective lymph node (LN) regions. Material and Methods Our standard practice for pelvic groups is CT-based, MR- guided delineation using MR images to aid target definition. Patients are treated with VMAT and daily imaging with cone beam CT (CBCT). MRCAT images were acquired along with standard MR and CT images for validation purposes taking care to avoid coils touching the abdomen or pelvis. Before moving to clinical treatment, we tested the MRCAT suite in validation studies focusing on differences in delineations based on MRI and sCT, in treatment plan dose distributions, and on the feasibility of daily CBCT based patient positioning. After these studies we performed a pilot study where 4 patients (one prostate without LN and one prostate, one rectal and one anal cancer each with LN) were treated based on MRCAT scans. Results In the validation studies 2 clinicians delineated a set of patients on both CT and sCT. Generally, delineation variations between sCT and CT were no larger than between clinicians. Dose distributions were tested both on plans optimized on standard CT and recalculated on sCT and the opposite. Dosimetric comparison showed that differences between CT and sCT were small (mean CTV difference < 0.5 Gy). CBCT registration was tested offline for automated bone match (Elekta XVI 5.04). The automated bone match algorithm yields similar registrations with the sCT, and evaluation of soft tissue was no different from ordinary CT. In the pilot study the anal cancer patient failed sCT reconstruction and had to be excluded. The 3 remaining patients were planned and treated without incident, and daily image registration could quickly be transferred to the treating RTT’s with only reviews by physics staff. Conclusion Based on initial evaluation of the fidelity of dose calculations and delineation on MRCAT sCT compared to standard CT, we felt confident in running a clinical pilot study using MRCAT sCT for treating general pelvis cancers. The pilot study was completed with satisfactory results, increasing our confidence in the workflow, and we are now moving forward with MR-only treatment in a larger group of patients to improve workflow and further quantify differences in delineated volumes. PO-1775 Texture analysis of Tomotherapy planned sinograms to assess plans modulation T. Santos 1,2 , T. Ventura 2 , J. Mateus 2 , M. Capela 2 , M.D.C. Lopes 2 1 University of Coimbra, Physics Department, Coimbra, Portugal ; 2 Instituto Portugues de Oncologia de Coimbra, Medical Physics Department, Coimbra, Portugal Purpose or Objective The purpose of this work was to explore the use of texture analysis to evaluate the degree of modulation of helical Tomotherapy (HT) planned sinograms. Material and Methods Four texture features, namely contrast, correlation, energy and homogeneity were calculated in MATLAB (Mathworks, Natick, MA, USA) for sinograms extracted from the DICOM RT plans. To investigate the variation of each textural feature with the degree of modulation, four test cases were first considered (Fig.1). The adopted displacement distances (d) between the pixel of interest and its neighbors were 1, 3 and 5 and the corresponding offset angles were 0°, 45°, 90° and 135°. After this previous study, the sinograms of 20 head and neck (H&N) HT treatment plans created in VoLO (Accuray Inc., Sunnyvale, CA, USA) were analysed. The deliverability of the selected H&N plans was also evaluated using the local pre-treatment verification protocol. Plans were recalculated in the TomoTherapy Phantom (Cheese phantom) and the Dosimetry Check

resources are allocated according to a Failure mode and effects analysis (FMEA) [Huq, 2008; AAPM TG 100]. Our aim is to define a decision method regarding QA needs, tolerances and periodicities for Surface Guided Radiotherapy (SGRT), dependent on our clinical workflows, treatment techniques, and equipments. The method will be initially established for a single SGRT application. Material and Methods For the clinical process of monitoring and treating breast patients under Deep-Inspiration Breath-Hold (DIBH), through SGRT, a process tree and corresponding FMEA were defined, and to each potential risk event a risk priority number (RPN) was calculated. The RPN was the product of three parameters, likelihood of occurrence, severity, and lack of detectability, defined from 1 to 5. In parallel, a set of QA tests was specified according to current guidelines [AAPM TG 147, 2012]. To each eventual risk event QA-tests that might impact it were assigned. The periodicity and exigency of each test was based on the scored RPN of the corresponding risks. Results In our SGRT workflow, applied to DIBH breast patients, 38 sources of failure events were found, with 34% of them corresponding to possible sources of equipment miscalibration/failures, while the remaining correspond to human errors, only possible to reduce through future software or workflow changes. System events were ranked according to the RPN, from the highest to the lowest, and defined as requiring daily, monthly or yearly/after-Service QA actions, respectively (figure 1). The attributed QA checks and corresponding parameters, showed to be comparable with the existing guidelines. Conclusion This method showed to be feasible and representative of our clinical workflow. The obtained periodicity was similar to the existing guidelines, hence it may be extended to other modalities (e.g. SRS, SBRT), where different RPN- weights will result in a different set of tests/tolerances. However, the subjectivity of this method should not be neglected. PO-1774 First clinical experiences with MR-only treatment planning and delivery for pelvic cancers H.D. Nissen 1 , S.P. McIlroy 1 , B.M. Havelund 2 , C. Krog 1 , J. Pløen 2 , M. Berg 1 1 Vejle Hospital, Medical Physics, Vejle, Denmark ; 2 Vejle Hospital, Oncology, Vejle, Denmark Purpose or Objective The CT scan is the backbone of treatment planning and daily positioning in radiotherapy providing densities for dose calculation and imaging for target delineation and daily patient positioning. However, MRI provides superior soft tissue contrast and has become a standard for aiding target delineation in pelvic cancers especially where precise definition of tumor volumes is necessary to minimize radiation dose to organs-at-risk (OAR). To minimize resources and patient time using only MRI would be optimal, therefore our department works on testing and validating the Philips MRCAT pelvis suite (Philips Healthcare) for generating synthetic CT (sCT) scans from MRI. Here we present experiences from validation and a pilot study of clinical treatment of rectal, anal and