S2946
RTT- RTT operational practice and workflow innovations
ESTRO 2026
WBRT. Automated and streamlined pathways are essential to reduce time burden and ensure consistency and accuracy. This study, led by a radiation therapist (RTT) aimed to robustly quality assure AI generated hippocampal contours using both structure comparison and dosimetry metrics, streamlining the pathway for the safe and effective introduction of hippocampus sparing radiotherapy. Material/Methods: Ten patient datasets were anonymised. Guideline based segmentation using Contour+ (MVISION AI Inc.) of left and right hippocampi on 1.25 mm post-contrast axial T1 MRI slices (Siemens Magnetom Sola 1.5T) was carried out. Four CNS radiation oncologists (ROs) reviewed and edited the AI generated hippocampal contours according to RTOG0933. The RO reviewed contours were compared against the MVISON raw AI contours using MVISION Guide & Verify (Surface Dice at 1mm, (sDCE1mm)), (Hausdorff distance, (HD)), and a python script (Average surface distance, (ASD)). The imported segmentations and MRI were then automatically registered to the patients CT dataset in Eclipse v18.1 (Varian Medical Systems, Palo Alto, CA, USA). Dose comparisons were made by overlaying the contours onto a standard HA-WBRT plan prescribed to 30Gy in 10 fractions using VMAT delivery. Results: RO agreement was strong for the AI hippocampal contours with contours measuring > 0.95 sDCE1mm (all but one > 0.99). ASD was < 1mm for all except one case, HD was 0 for all except one case (3 mm). Dosimetric results are summarised in Table 1. There were no or only minor adjustments made in the majority of cases (7/10). However, minor RO edits correlated with substantial dosimetric differences in three cases: Case 1, RO edits to superior left hippocampus added an extra 3.3 Gy to Dmax; Case 9, RO addition of one extra slice added an extra 3.0 Gy and 2.0 Gy to superior left and right hippocampi, respectively; Case 10, two ROs added one extra slice to superior left hippocampus adding 2.2 Gy and 2.4 Gy to Dmax.
Conclusion: Although minor differences in structure delineation can lead to notable dosimetric changes, automated hippocampal contours have been shown to be reliable and clinically acceptable with minimal/no manual intervention. This study highlights the sensitivity of dosimetric outcomes over geometric metrics. References: 1. Gondi, V., et al., Preservation of memory with conformal avoidance of the hippocampal neural stem- cell compartment during whole-brain radiotherapy for brain metastases (RTOG 0933): a phase II multi- institutional trial. J Clin Oncol, 2014. 32(34): p. 3810-6. Keywords: hippocampus, auto contouring Digital Poster 4283 Title: Measuring the impact of a Radiation Therapist-led multi-modal prehabilitation clinic for patients undergoing prostate radiotherapy Sarah Jane Flynn 1 , Denise Crone 1 , Ryan Newburn 1 , Katie Henry 1 , Thamra Ayton 2 1 Radiotherapy, BHSCT, Belfast, United Kingdom. 2 Physiotherapy, BHSCT, Belfast, United Kingdom Purpose/Objective: Reproducible bowel and bladder preparation for prostate cancer radiotherapy planning and treatment is essential, helping to optimise treatment accuracy, reduce side effects by safeguarding adjacent organs and enhance long-term outcomes. Poor bowel and/or bladder preparation may require repeat planning CT scans and additional cone beam CTs on-set thereby impacting scanner/linac capacity, delaying treatment planning and contributing addition radiation dose to the patient. Some patients find this physically and emotionally challenging. We wished to assess the impact of a patient-centred prehabilitation clinic on the rate of repeat planning scans and the patient satisfaction with a multi-modal prehabilitation clinic. Material/Methods: A retrospectively review of prostate rescan rates was
Made with FlippingBook - Share PDF online