ESTRO 2020 Abstract book

S1106 ESTRO 2020

Results Figure 1 shows a typical output from Philips' geometric QA analysis algorithm, with colours inverted for clarity. Iso- contours of systemic distortion are drawn on top of MR- images of the phantom at the isocentre, ± 6 cm, ± 13 cm, and ± 20 cm.

Conclusion Total delivered dose was acceptable for all LNs with 5mm and 10mm PTV plans even for the patient with considerable swelling (up to 13mm). For swelling <10mm, the fractional dose difference was <5% for LNs which were not located in the dose build up region (d≤2mm). In fractions with considerable LN displacement, the dose could drop by up to 12%. Therefore, early and persistent swelling could potentially have relevant dosimetric impact. Monitoring by RTTs of swelling on daily CBCT could be relevant. Furthermore, our results motivates for testing if margin reduction from 10mm to 5mm PTV margin of inguinal LN boosts could be possible in vulvar RT. PO‐1887 Reproducibility of Geometric QA in MR‐RT C. Krog 1 , G. Grimnisdottir 1 , H.D. Nissen 1 1 Vejle Hospital, Department of Medical Physics, Vejle, Denmark Purpose or Objective We are planning to introduce MR-only workflow for pelvic cancers based on Philips’ MRCAT. Geometric fidelity is of utmost importance when moving from CT-based delineation to MR-only delineation. Here, we investigate the geometric performance of our scanner and variations related to daily set-up using a vendor-supplied phantom. Material and Methods Using a Geometric QA Phantom on an Ingenia 1.5 T MR-RT- scanner (both: Philips Healthcare, Best, The Netherlands), a series of QA measurements were made over six weeks. Three times a week, three sets of measurements were taken. The measurements were performed on Tuesday mornings and Thursday mornings and afternoons. First, the RTT positioned the phantom on the MR-RT couch top over marker H1, further guided by two strips of tape on the couch top. The Ingenia system’s lasers were used to place the phantom at the scanner’s isocentre. Philips’ geometric fidelity QA scan sequence was run, yielding data from the isocentre, ± 6 cm, ± 13 cm, and ± 20 cm distance from the isocentre. Second, the RTT repositioned the phantom using the Ingenia system’s lasers and repeated the geometric fidelity scan. Third, the RTT removed the phantom rack from the couch top, removed the geometric phantom from the phantom rack, and then set it up again as before to produce the final set of measurements. Data were analysed with Philips' automatic geometric QA algorithm and in-house python scripts.

Figure 2 shows distance from the centre point to the 2mm iso-contour in all 54 QA scans, at the isocentre. The same kind of plot was made for slices at other positions. Each week is plotted in one shade of colour, and each set of measurements within a week has a different line style. The graphs were used to estimate whether a slice passed the QA, as an iso-contour in each slice is compared with an ellipse of vendor-prescribed dimensions. The 2mm contour was compared to the three central slices, the 3mm contour to the ± 13 cm slices, and the 5mm contour to the ± 20 cm slices. Only two slices, a month apart, did not pass the QA. They were deemed intermittent fluctuations, as subsequent scans passed the QA.

Conclusion The geometric QA was found to be highly reproducible over a six-week period, with only two scans, over a month apart, showing non-passable anomalies in only one slice each. Graphs as shown in Figure 2 were found to be useful for estimating whether there were significant deviations from typical scanner performance over time, and which scans required more detailed evaluation. Further verification, using the External Laser Positioning System (ELPS), which is used in positioning patients for MRCAT, is required. Geometric QA using a 3D phantom, the abdominal coil, and the sequences used for MRCAT is on the agenda.