ESTRO 2020 Abstract book

S914 ESTRO 2020

PO-1590 Spirometer guided breathhold breast VMAT on Halcyon verified with portal images and surface tracking L. Delombaerde 1,2 , S. Petillion 1 , C. Weltens 1,2 , T. Depuydt 1,2 1 University Hospital Gasthuisberg, Department of Radiation Oncology, Leuven, Belgium ; 2 KU Leuven, Department of Oncology, Leuven, Belgium Purpose or Objective The fast imaging and radiation delivery capabilities of the Halcyon linac (Varian Medical Systems) are perfectly suited for breath-hold gated treatments. In this study we treated breast cancer patients, ages 70 and above, in deep inspiration breath-hold (DIBH) with the SDX spirometer system (Dyn’R). The stability and reproducibility of the spirometer-assisted breath-holds were assessed via portal images acquired at ‘tangential’ angles and by intra-bore surface monitoring using an in-house developed structured-light-based prototype. Material and Methods Four (preliminary data) left-sided breast cancer patients were treated using a triple-arc VMAT-SIB protocol in breath-hold with the SDX system (Dyn’R) for 21 fractions. Portal image frames were acquired at 15 Hz during DIBH VMAT. Three portal images are available per fraction (one per arc at 301° gantry angle) for verification of the anatomy during breath-hold. Concurrently the in-house developed surface scanning system monitored the movement of the patients’ chest surface in 6 degrees-of- freedom (DoF) during the entire treatment fraction, from completion of the setup until the end of radiation delivery. Offline, the portal image taken during the first arc was manually registered to the ribs and breast contour, delineated on the DRR generated from the planning CT. The superior-inferior (SI) and pseudo anterior-posterior (“AP”) systematic (Σ) and random (σ) errors were determined. The vertical chest displacement during breath-holds was extracted from the surface data. The intra fraction variability was calculated by subtracting the minimal from the maximal breath-hold per fraction. The inter fraction variability was calculated by subtracting the mean breath-hold per fraction from the mean breath-hold of the first fraction. The total number of breath-holds per fraction and the total time spent on couch was recorded. Results Portal image analysis (81 images) showed systematic and random errors for the SI direction of Σ = 0.9 mm and σ = 1.0 mm and for the “AP” direction of Σ = 0.7 mm and σ = 1.2 mm. The mean (± standard deviation) intra fraction vertical breath-hold variation was 1.4 (± 1.0) mm and the inter fraction variation was 1.0 (± 0.9) mm (56 complete fractions monitored). Median number of breath-holds per fraction was 4 (range 4 – 17) viz. one breath-hold for the CBCT and one breath-hold per arc. Median total time patients spent on the couch was 8 min (range 6 min – 36 min).

5% outmost navigator positions). This resulted in 10 respiratory correlated 3D images. From the 4DMRIs, the breathing amplitude was obtained by manually registering (in cranio-caudal direction only) each of the 10 image to a reference (i.e. end-exhale image), based on the right hemidiaphragm (Figure 1). Breathing amplitudes during FB and RB obtained in the two MR sessions were compared to investigate the reproducibility of the distribution of RB. Breathing amplitudes of FB and RB from both MR session combined were compared to investigate the effect of breathing regularization by ventilation. Results From the 12 MRIs (six volunteers, two sessions), three breathing motion amplitudes could not be reconstructed due to poor image quality and a missing acquisition. Overall, RB was well tolerated by all volunteers. The breathing amplitude was smaller for RB than for FB for each of the 9 measurement pairs (Figure 2). We pooled the data of both MRI sessions to analyze differences between FB and RB, demonstrating a significant smaller median amplitude of 11.8 (range 8.8–19.0) mm for RB as compared to 20.3 (range 10.6–24.3) mm for FB, tested with a Wilcoxon signed-rank test, p<0.01. The median relative reduction of amplitude was 42% (range 15–54%). Between the two MRI sessions, the variation of amplitude differed from 0.4-13.2 mm for FB and 1.3-5.3 mm for RB (n=4). Conclusion Regularized breathing by non-invasive ventilation significantly reduced the motion of the right hemidiaphragm compared with free breathing. Reproducibility of regularized breathing needs further investigation.