ESTRO 2021 Abstract Book

S688

ESTRO 2021

Conclusion RT in deep inspiration breath hold reduced the estimated radiation doses to the heart without compromising the dose to the target in a cohort of patients with gastric lymphoma. The procedure was feasible in all patients. PD-0855 First patient experience with real-time EPID-based breath hold monitoring during breast radiotherapy E. Vasina 1 , P. Greer 2 , J. Baeza Ortega 1 , T. Kron 3 , J. Ludbrook 2 , D. Thwaites 4 , J. Lehmann 5 1 University of Newcastle, School of Mathematical and Physical Sciences, Newcastle, Australia; 2 Calvary Mater Newcastle, Department of Radiation Oncology, Newcastle, Australia; 3 Peter MacCallum Cancer Centre, Peter MacCallum Cancer Centre, Melbourne, Australia; 4 Institute of Medical Physics, University of Sydney, School of Physics, Sydney, Australia; 5 University of Newcastle, School of Mathematical and Physical Sciences, Newcastle, Australia Purpose or Objective Real time analysis of continuously-acquired treatment field images, using an electronic portal imaging device (EPID), has the potential to aid the assessment of breath hold quality during deep inspiration breath hold (DIBH) breast cancer treatments. A system to monitor breath hold quality in this way, via measurement of the lung depth (LD) and distance from the skin to the posterior field edge (skin distance, SD) in the images, has been developed and now is being tested on patients on the study. This is the first patient experience of the system. Materials and Methods The system consists of: (i) an image grabber application which acquires and saves MV frames in real time and (ii) image-processing software which analyses the row profiles of the frames and measures LDs and SDs at user specified locations. The system was developed using >1000 images of breast tangents from a number of clinics and has shown sub-millimeter accuracy in phantom experiments [1]. The latency of the system is <350 ms. During the first clinical trial, 3 LD and 3 SD values (superior, midline, and inferior, as shown schematically in Figure 1A) were measured in real time in MV EPID images of a left breast tangent treatment. Expected LD and SD values were obtained via MV images predicted for 6 MV using the planning CT images and in-house software ArtPlanGenerator [2]. The patient’s breath hold was monitored with Varian’s RPM system with a 5 mm breathing window. The treatment included 4 beams (6 and 18 MV). The gantry angle for beams 1 and 2 was 316 degrees, and 136 degrees for beams 3 and 4. Dynamic MLC was used near the end of beams 1 and 3. Results Figures 1B and 1C show LDs and SDs measured during one treatment fraction (beams are separated by black vertical lines). The mean differences of the LD and SD measurements from their expected values were calculated and are plotted along with their standard deviations for each fraction of the patient’s treatment in Figure 2. The overall maximum variation found was about 20 mm (beam 3). Figure 1. (A) LD and SD are shown by red and blue lines. (B) Measured. Expected values in mm (superior, midline, inferior). Beams 1 and 2: LD (13, 13, 5); SD (107, 105, 83). Beams 3 and 4: LD (13, 13, 2), SD (112, 113, 89). Figure 2. The mean of differences between measured LDs and SDs and their predicted values for every treatment fraction.