ESTRO 2020 Abstract book

S204 ESTRO 2020

OC-0352 Increased accuracy in reduced time – surface guided RT for hypofractionated prostate cancer patients A. Mannerberg 1 , M. Kügele 1,2 , S. Hamid 2 , K. Petersson 2 , A. Gunnlaugsson 2 , S.Å. Bäck 2 , S. Engelholm 2 , S. Ceberg 1 1 Lund University, Medical Radiation Physics, Lund, Sweden ; 2 Skåne University Hospital, Hematology- Oncology and Radiation Physics, Lund, Sweden Purpose or Objective Ultra-hypofractionated radiotherapy of prostate cancer treats patients in fewer number of fractions and with higher fractionation dose compared to conventional fractionation schemes. The treatment efficiency is maintained and there is no increased risk for late toxicity (Widmark et al 2019). To further reduce patient treatment time, flattening filter free (FFF) beams are used. The aim of this study was to: • Evaluate if surface guided radiotherapy (SGRT) could further reduce the total treatment time by reducing the patient setup time. Material and Methods A total of 40 prostate cancer patients were enrolled in this study. All patients received a hypofractionated (42.7 Gy / 7 fr) 6MV-FFF VMAT treatment plan. Twenty patients were positioned with LBS and 20 patients were positioned using surface based setup (SBS). For LBS, the patients were positioned by aligning skin tattoos with in-room lasers. For the SBS, the optical surface scanning system Catalyst TM (C- Rad Positioning AB, Uppsala, Sweden) was used. A deformable algorithm was used to calculate the patient isocenter position. A colour map projected onto the patient skin with an 8 mm tolerance was used for posture correction. The daily patient setup was verified by two orthogonal kV images and internal gold fiducial markers were used for the matching. For comparison purposes, the deviation from the planned position and the online position was evaluated by the total vector offset, V , for both LBS and SBS. The displacement from the planned position in the lateral (lat), longitudinal (lng) and vertical (vrt) direction, respectively, was used to calculate V (Eq. 1). • Investigate if SGRT can improve prostate patient setup accuracy compared to conventional laser based setup (LBS). In total, 280 verification images were evaluated. The patient setup time was extracted from the linac log files for both LBS and SBS. Results Results showed that the patient setup time decreased with 24% using SBS compared to LBS. The patient setup time was on average reduced with 42 s for each treatment fraction (p < 0.01, Mann Whitney U test). A statistically significant improvement in the total vector offset was observed for SBS (p < 0.01, Mann Whitney U test). The median total vector offset for LBS was 5.2 mm (0.41 - 17.3 mm) and the corresponding number for SBS was 4.6 mm (0.0 - 10.4 mm). The results show that with SBS the largest setup deviations are detected and eliminated. Also, the SBS results follow the colour map tolerance, which indicate that with narrower tolerance even better setup results will be obtained. However, this needs to be further evaluated. (Eq. 1)

Fig 1. Dose prediction process for representative patient

Fig 2. Bland Altman plot of G2 rectal bleeding prediction errors Conclusion Research into deep learning for radiotherapy dose prediction is expanding, with neural networks proposed as solutions to expedite the RT treatment planning process. Within this work a DL dose prediction network was developed as a decision-support system for patient stratification. This preliminary work demonstrates that such a network could stratify high risk patients for rectal spacer insertion with > 80% accuracy and identify patients close to the stratification threshold. This approach could reduce the need for time and resource intensive treatment planning. Further model development, training and validation is warranted to maximise accuracy prior to clinical implementation.