ESTRO 2020 Abstract book

S919 ESTRO 2020

magnetic field. We therefore conclude that this setup can also be used straightforward to validate adaptive workflows also for higher field strength or other motion management strategies, such as tracking. 1 C Baldock et al 2010 Phys. Med. Biol. 55 R1 PO-1597 Identifying inconsistent breathing patterns for increased monitoring during breath-hold treatment T. Nano 1 , M. Feng 1 , T.D. Solberg 1 , A. Witztum 1 1 University of California- San Francisco, Department of Radiation Oncology, San Francisco, USA Purpose or Objective To analyze spirometry-based breathing waveforms in patients undergoing voluntary breath-hold (BH) radiotherapy for detection of potential errors that may be missed clinically. Material and Methods 198 total breath-hold waveforms (average of 7.9 breath- holds per treatment fraction) from 16 patients with pancreas or liver disease undergoing stereotactic body radiation therapy in 3-5 fractions were included in the analysis. During each treatment, patients undergo multiple breath-holds using the SDX system (Dyn’r, Medical Systems, Aix-en-Provence, France). For each breath hold a breathing waveform is recorded (Figure 1A) from which the BH time is determined (Figure 1B). The average breath-hold time and range, number of missed holds (defined as BHs shorter than 2 seconds) and number of failed holds (defined as BHs longer than 2 seconds but shorter than the maximum BH minus 5 seconds) during each treatment fraction were extracted as quantitative metrics to assess treatment quality (Figure 1C). It is assumed the beam is not turned on during missed holds (since therapists wait >2 seconds to initialize beam), whereas failed holds require therapist intervention to abort the beam. A summary report is generated for each treatment (Figure 2). Results Out of 25 treatments, 14 were treated with inspiration breath-hold and 11 with expiration. For all treatments, breath-hold time was greater than beam-on time. Missed holds were observed in 2 BHs from 2 different patients, and failed holds were identified in 6 BHs from 4 different patients. For patients with multiple treatment information available (n=4), the average BH ranges between treatments were 0, 0.5, 0.4, and 0.6 seconds. Across all treatments the average (maximum) range of successful BHs was 1.4 (4.6) seconds. Conclusion Manual response to patients failing breath-hold is critical for safe treatment of patients undergoing SDX voluntary breath-hold and is necessary in the absence of full software integration with treatment delivery system. Identifying and documenting patients who are at increased risk of, or have previously fallen out of breath-hold, can better prepare therapists for future treatments.

sensitive PG volume with an additional 3mm margin. The PG was used as gating structure with an additional 3mm boundary margin and 10% of the target structure were allowed outside the boundary (as currently applied for patients treatments). PLP motion was based on a real patient breathing pattern resulting in a target motion amplitude of 2.5cm and a motion cycle frequency of 10 cycles/min. The motion was tracked during irradiation with a 2D trufi cine-MR (4 images/second, sagittal orientation). PG evaluation was carried out on a 3T MR system (PRISMA, Siemens) 48h after irradiation using a multi spin-echo sequence. Results were evaluated using the 3D -criterion of 3%/1mm and 3%/3mm (dose difference with respect to the local dose/distance-to-agreement, comparing gel dose with TPS dose) and taking only dose levels larger than 10 % of the maximum dose into account. Results The PLP showed an anthropomorphic image contrast both in MR and CT. Fine lung structures were visible and could be well separated from the heart and PG-based target structure (Figure 1). Although the PLP was moved between CT imaging and irradiation, the locations of inner structures were highly reproducible allowing for a good image registration between CT and MR. The gated treatment took 9:44 mins in total. The target structure could be tracked throughout the entire treatment session resulting in a successfully gated irradiation. Evaluation of the PG target showed no significant under- or overdosage throughout the entire evaluated volume resulting in high 3D γ passing rates of 95.9 % for 3%/3mm and 93.3 % for 3%/1mm (Figure 2).

Conclusion The presented study has shown to be an extremely valuable tool for the verification of intra-fractional motion management techniques in MRgRT. The PLP shows anthropomorphic image contrast, well comparable to that of real patients and the PG allows validating the applied dose in 3D, while the PG itself was not influenced by the