ESTRO 2025 - Abstract Book

S4093

RTT - Patient care, preparation, immobilisation and IGRT verification protocols

ESTRO 2025

Purpose/Objective: To evaluate the accuracy of different automatic image registration methods versus manual registration in Cone Beam Computed Tomography (CBCT) for patients undergoing Stereotactic Body Radiation Therapy (SBRT) for prostate cancer. Material/Methods: 103 prostate cancer patients treated with SBRT on a VersaHD Linac were retrospectively analyzed. Two dose regimens were represented: 36.25 Gy in 5 fractions (52 patients) and 42.7 Gy in 7 fractions (51 patients). Each patient underwent a standardized preparation for rectal emptying and bladder filling before each session. The CBCTs acquired during treatment to ensure patients’ set up and anatomical reproducibility were retrospectively re-matched by an expert operator to define the manual registration values by recording translational displacements along the X, Y, and Z axes. The Elekta XVI system provided a Dual Registration Tool with clipbox and mask algorithms, focusing on bone structures and soft tissues. Automatic registrations were obtained using three different masks, created from the Clinical Target Volume contour by adding 0 mm, 5 mm, and 10 mm. Statistical values, including mean and standard deviation, of the variations between automatic and manual registrations were calculated for each axis and dose regimen. Bland-Altman plots were employed to determine the agreement between the registrations. Results: A total of 620 CBCT scans were acquired, and each scan was re-evaluated using the different registration methods, resulting in a comprehensive analysis of 2480 registrations. Bland-Altman plots revealed high concordance between manual and automatic registration, with most differences falling within clinically acceptable limits. The automatic registration using a 5 mm mask margin provided the highest agreement with manual registration, with mean translational discrepancies of less than 1 mm across all axes. The lowest variability was along the X-axis (left-right) due to reduced anatomical motion, while wider deviations were observed along the Z-axis (anterior-posterior), reflecting the prostate sensitivity to variations in bladder and rectal filling. Significant discrepancies were particularly observed in scans with non-standard patient preparation or the presence of double metallic prostheses, which introduced artifacts that challenged automatic alignment. Conclusion: The Elekta XVI can achieve high alignment accuracy in SBRT for prostate cancer when using a 5 mm mask margin, closely approximating the results of an expert manual registration. However, manual intervention remains crucial in cases of suboptimal patient preparation or metal artifacts, which limit the effectiveness of automatic alignment. Adopting a hybrid approach—initial automatic alignment followed by manual adjustment—could improve treatment precision and reliability.

Keywords: Prostate cancer, CBCT, IGRT

2691

Digital Poster Designing, printing and assessing 3D printed wedges for reflector block placement Fiona Mary Milliken, Christopher Hamill-Taylor Radiotherapy, NHS Tayside, Dundee, United Kingdom

Purpose/Objective: Reproducibility and accuracy of the DIBH reflector block (RB) inter-fractionally from the CT planning scan to each treatment appointment is crucial. Many centres have implemented DIBH for breast cancer patients using Varian’s Real-time Position Management (RPM) system[1]. This system uses an infrared camera to track an externally placed RB with four reflective dots, placed on the patient’s upper abdomen. The RB is required to be in a level position during free breathing, however, when gauze is used to level RB there is the potential for it to be different at