ESTRO 2020 Abstract Book

S879 ESTRO 2020

Conclusion The TFT strongly influences the ITV reliability in the case of abdominal lesions, while it shows a limited impact on thoracic lesions. The ITV strategy developed using MIP image can be considered strongly reliable for thoracic lesions if TFT≤9 min, moderately reliable for abdominal lesions when TFT ≤ 3 min. A larger cohort of patients is required to validate these preliminary results. PO-1614 Breath hold characteristics and intra-fraction motion during nasal high flow therapy assisted RT G. Vilches Freixas 1 , A. Vaniqui 1 , F. Vaassen 1 , C. Hazelaar 1 , M. Öllers 1 , S. Canisius 1 , D. Tissen 1 , E. Van Enckevort 1 , E. Rousch 1 , K. Verhoeven 1 , J. Van Loon 1 , D. De Ruysscher 1 , G. Bosmans 1 , W. Van Elmpt 1 , I. Rinaldi 1 , S. Peeters 1 1 MAASTRO Clinic, Department of Radiotherapy-GROW- School for Oncology and Developmental Biology- Maastricht University Medical Centre, Maastricht, The Netherlands Purpose or Objective In order to reduce breathing motion, a clinical study (NCT03729661) investigated the feasibility of using a nasal high flow therapy (NHFT) device (AIRVO®) to achieve moderate voluntary deep inspiration breath hold (mDIBH) during radiotherapy. This NHFT device may have the potential to increase the breath-hold (BH) duration and, thus, treatment time efficiency; and the BH stability compared to a standard voluntary BH. This technique could also be beneficial for proton therapy by reducing uncertainties for moving targets. Material and Methods Ten left-sided breast cancer patients and ten stage III NSCLC patients were prospectively included and treated with photon radiotherapy during mDIBH using the NHFT device. Heated and humidified airflow of 40 l/min was administered through a nasal interface with 80% oxygen. An optical surface scanning system (CatalystHD, C-RAD®) was used to monitor the voluntary breath holds (BH) with a 3 mm gating window and visual feedback to the patient using augmented reality goggles. Preceded by a training session, a planning CT scan and a treatment plan were performed in mDIBH. Treatment was on a Truebeam linac (Varian) with daily kV-kV imaging for breast patients and daily CBCT imaging in mDIBH for lung patients. Treatment plans consisted of 2 to 3 partial arcs or hybrid technique for lung patients, and 4 to 5 beams using a hybrid technique for breasts. CatalystHD is able to monitor and calculate isocenter shifts instantly during treatment using a non-rigid registration and volumetric deformation model, which we used to assess the intra-fraction motion during mDIBH. Results The first, middle and last NHFT treatment fractions of 3 lung and 3 breast cancer patients (76 BH in total) that completed the treatment with the NHFT device were analyzed. The mean BH length was 52s ± 32s, with a maximum of 135s. Regarding the BH stability, the fluctuation of the respiratory signal within one BH (i.e. standard deviation over one BH) was on average 0.4 mm (range 0.1-2.3 mm). In 92% of the BHs of the breast patients, the median isocenter shift (composite vector) during BH was within 5 mm (setup tolerance); and for the lung patients in 86% was within 8 mm (setup uncertainty). We observed a learning curve because patients get used to the device. In 69% of the fractions, patients were able to maintain the BH during kV or CBCT imaging. In 50% of the fractions, the whole treatment delivery could be given in ≤ two long BHs (Figure 1). The CBCT image quality for matching was judged to be qualitatively higher during BH compared to free breathing CBCT imaging (Figure 2). The patient reported excellent subjective tolerance of the NHFT device.

whole RT treatment. The validity of this hypothesis could be strongly dependent by the duration of treatment fraction time (TFT) and the tumor’s location. With the introduction of MR-guided RT, it has become possible to continuously monitor the tumor motion during the treatment. Aim of this study was to investigate the reliability of the ITV approach related to variations of TFT, in case of abdominal and thoracic lesions. Material and Methods A total of 10 thoracic and 12 abdominal lesions were analysed. Before treatment, a 10-phase 4DCT was acquired and the maximum intensity projection (MIP) image was reconstructed. The lesion was contoured on the single breathing phase image (SBPI) where the lesion was most visible, and ITV was delineated on the MIP image. The ITV margins were estimated in craniocaudal (CC) and anteroposterior (AP) direction as the maximum difference in extension between ITV and the tumor delineated on SBPI. All the patients received an MR-guided RT treatment in free-breathing, with the tumor position directly monitored on a sagittal plane at 4 frames/sec. The ITV margins were projected on the tumor trajectory and the percentage of treatment time in which the tumor was inside the ITV (%TT) was measured. The ITV approach was considered moderately reliable when %TT ≥ 90% and strongly reliable when %TT ≥ 95%. Five fractions were analyed per patient and the mean values in terms of %TT were reported, supposing that each treatment fraction would take 1,2,3,4,5,6,7,8,9 or 10 min. The additional ITV margin required to obtain %TT ≥ 95% was also estimated to vary of TFT. Results Figure 1 shows the mean value of %TT in function of TFT. For thoracic lesions the ITV strategy can be considered strongly reliable for TFT ≤ 9 min. For abdominal lesions, the ITV strategy can be considered only moderately reliable if TFT ≤3 min. Figure 2 shows the additional margin required to make the ITV strategy based on 4DCT prior to treatment strongly reliable (%TT ≥ 95%). For thoracic lesions, the ITV margins are conservative, and might be reduced with 1 mm in AP and CC if TFT is maintained ≤1 min. For abdominal lesions, a 2 mm additional margin in AP is recommended to make the ITV approach strongly reliable, maintaining the TFT ≤ 4 min.