ESTRO 2020 Abstract Book

S177 ESTRO 2020

dosimetric differences between the planning CT scan and the total accumulated fraction CT dose to the bowel in the first 12 patients enrolled within the prospective STEAL trial, evaluating online adaptive stereotactic radiotherapy for abdominopelvic oligometastases in lymphnodes. Material and Methods Eligible oligometastatic patients with abdominal lymphadenopathies were enrolled in the STEAL trial, and treated with CyberKnife R with an in-room CT scan. Patients were all treated with a total dose of 45 Gy in 5 daily consecutive fractions of 9 Gy. For each patient, a library of 3 plans was created. Before each fraction, a CT-scan was made with the in-room CT scan (fraction CT scan). The fraction CT scan was forwarded to the planning system. Organs at Risk (OAR) from the planning CT scan were transferred to the fraction CT scan by deformable image registration (DIR). On basis of a protocol decision tree (PTV coverage and V35 of the bowel), the best of the 3 plans was used for treatment. Offline, the OARs were manually recontoured on all the pre-fraction CT scans (58 Ct scans) and a Gastrointestinal Organ (GIO) volume was obtained by summing stomach, duodenum and bowel. To verify the validity of using DIR in the clinical workflow, the dose in the OAR obtained by DIR and manual contouring was compared. The values for various dose-point volumes and D0.5 cc were extracted from the dose-volume histograms (DVHs) to see whether a significant difference existed between the two contouring modalities. The total accumulated fraction CT dose to the OAR was calculated by summing the OAR dose of each fraction. Results The plan choice based on DIR contours did correspond in 76% with the plan choice based on manual contours, although no statistically significant differences were found for dose-volumes from V25Gy to V40Gy. The total accumulated fraction CT dose to the GIO at higher doses levels (V25Gy, V30Gy, V35Gy, V40Gy) and the D0.5cc volume was significantly increased compared to higher doses levels and D0.5cc of the planning CT scan. Median overall PTV coverage was higher than 95% with the plan- of-the-day (POTD) choice on daily CT scans (96,53%). Conclusion The use of DIR contours of the bowel can be an effective tool when using online ART. With the use of DIR, no statistically significant differences were found for the higher dose-volumes parameters (V25Gy to V40Gy). The total accumulated fraction CT dose at the higher dose levels is higher to the GIO compared to the dose in the higher dose levels in the planning CT scan. OC-0337 First in Human: Patient Adaptive 4D Cone Beam CT (4DCBCT) R. Obrien 1 , O. Dillon 1 , A. George 2 , S. Smith 2 , A. Wallis 2 , S. Alnaghy 1 , J. Sonke 3 , P. Keall 1 , S. Vinod 2 1 University of Sydney, ACRF Image X Institute, Eveleigh, Australia ; 2 Liverpool Hospital, Radiation Oncology, Liverpool, Australia ; 3 Netherlands Cancer Institute, Radiation Oncology, Amsterdam, Netherlands Antilles Purpose or Objective Despite the widespread use of 4DCBCT for lung cancer, scan times are long (4 min), imaging doses are high and image quality is inconsistent from one patient to the next. To enable time and dose efficient 4DCBCT scans, we have developed a system that adapts the gantry rotation speed Proffered Papers: Proffered papers 18: Intra-fractional motion management

and projection acquisition in response to the patient’s real-time respiratory signal. In silico studies have suggested that scan times can be reduced from 4min to 60secs and imaging doses by 85% while maintaining consistent image quality. In this study, we report on the first patient scanned with adaptive 4DCBCT. Material and Methods In this ethics-approved study, called ADAPT, an Elekta Versa linear accelerator was modified to acquire patient adaptive 4DCBCT images by integrating our in-house circuitry to control the gantry rotation speed and suppress kV triggers to the kV generator during acquisition in response to changes in the patient’s real-time respiratory signal. The patient was treated as per clinical protocol and then a fast ADAPT scan that acquired 20 projections in each of 10 breathing phases across 20 breathing cycles was acquired (200 projections in total). The projections were reconstructed with the motion compensated FDK algorithm available in the reconstruction toolkit (RTK). Comparisons were made with a conventional 4 min scan that acquired 1320 projections. Results The first patient was acquired in October 2019. The ADAPT protocol reduced the imaging dose by 85% with scan times of 76 and 72 seconds for fraction 1 and 2 respectively (figure 1), corresponding to an average 69% reduction in acquisition time compared to the conventional scan. Image quality was considered acceptable for patient positioning in radiation oncology (figure 2).

Conclusion The first patient scanned with an adaptive 4DCBCT imaging protocol has been acquired showing large reductions in scan time (69%) and imaging dose (85%).