ESTRO 38 Abstract book

S534 ESTRO 38

dose calculation. The application of 4D dosimetry in combination with the log file analysis underlined the dependence of the dosimetric accuracy of an individual plan delivery on beam delivery variations and patient- specific parameters. PO-0978 Accurate positioning with decreased treatment time using surface guided tomotherapy A. Haraldsson 1,2 , S. Ceberg 2 , C. Crister 2 , S. Engelholm 1 , S.Å.J. Bäck 1,2 , P.E. Engström 1 1 Skåne University Hospital, Department of Hematology- Oncology and Radiation Physics, Lund, Sweden ; 2 Lund university, Medical Radiation Physics, Lund, Sweden Purpose or Objective Accurate, reproducible, and fast setup of the patient is important for a successful radiotherapy treatment, especially for complex treatment techniques such as helical tomotherapy. Daily imaging with mega voltage computer tomography (MVCT) is accurate but increases the imaging dose and the total patient time on the treatment couch. Thereby additional intra-fraction position uncertainty caused by patient movements are introduced. In this study, we quantified the setup deviation and time gain when using surface scanning for daily setup with weekly MVCT imaging for reference as compared to daily MVCT. We analysed 16 835 treatment fractions from our TomoTherapy HD (Accuray Inc., Madison, USA) installed with a Sentinel optical surface scanning system (C-RAD Positioning AB, Uppsala, Sweden). Material and Methods The surface of the patient was used to calculate the position using rigid registration against a reference surface. The patient setup was performed using in-room lasers, surface scanning and MVCT for the first three fractions. A new reference surface was acquired on the third fraction after a MVCT-based couch correction. On the following fractions, we used in-room lasers for setup, followed by daily surface scanning and weekly MVCT. The MVCT image setup correction vector, which was defined as the 3D translational correction as assessed by image registration, was used to evaluate the setup based on surface scanning. For each plan, the correction vector for one randomly selected fraction was used for the analysis. The imaging time, which was the time from imaging start to beam-on including any following procedures. The imaging time for one fraction was randomly selected per plan and imaging modality and multiplied with the number of fractions. Typical PTV margins at our clinic are 5 mm for head and neck patients and 10 mm for targets in the thorax and abdomen. Results A total of 894 plans were analysed with a treatment date from 2012 to 2018. Of the setup correction vectors, 90 % were within 2.3 mm for CNS (N=284), 2.9 mm for H&N (N=254), 8.7 mm for thorax (N=144) and 10.9 for abdomen (N=134) patients. The setup deviations were divided in a random and a systematic component according to Yan et al. 1997 (Table 1). The difference in imaging time was assessed as total imaging time per treatment plan, modality, and treatment site (Figure 1). The difference in total imaging time per fraction was significant for all sites (p<0.005) as assessed by linear regression and the Mann- Whitney U test. Conclusion The setup deviation was small compared to the standard PTV margins for all sites but the abdomen. In addition, daily surface scanning with weekly MVCT was significant faster than using daily MVCT. We therefore conclude that daily surface scanning with weekly MVCT is an accurate and fast alternative to daily MVCT for positioning of CNS, H&N and thorax patients receiving treatments on a TomoTherapy unit. For patients treated in the abdomen, daily surface scanning should proceed with caution.

defined beam extraction event enabling reproducible measurements. BD parameters, recorded with a time- resolution of 5.0µs, were retrospectively correlated time- wise with the dosimetric data (figure 1 (top)). Results Comparing corresponding spills of the treatment deliveries, the intensity variation was up to 8% for TV1 and up to 24% for TV2 (see table 1). This resulted in accumulated treatment time differences of up to 11 and 39 sec, respectively. The total dose differences (DDs) between deliveries of the same plan were within 0.3% for individual PPs for static measurements for both TV sizes. 0.6cm target motion resulted in DDs of up to 1% for TV1 and up to 2% for TV2. Motion of 0.6cm did not introduce much larger contributions to the dose perturbations within tumor region than those coming from accelerator dependent intensity fluctuations. DDs were more pronounced for a target motion of 2 cm. The dose deviated more for the smaller volume, i.e. by 16, 7, 5 and 16 % for PP1-4 (figure 1 (bottom)), while for TV2 it remained below 9% for PPs within the tumor volume (PP1-4). Measurements in the penumbral region (PP5) for TV2 showed a high sensitivity to motion for both amplitudes (15% relative SD for 0.6cm motion and 42% for 2.0cm motion).

Conclusion Time-resolved dose measurements and their correlation with the BD log files showed the relation of the BD time structure and the spill intensity, which is the basis for 4D