S1828
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
For 4D vs 3D SCOPE2 comparison, significant differences were seen in the minimum dose to 1 cc and D99% (p < 0.01) and in D98% (p < 0.05), with lower doses observed in the SCOPE2 plan due to under- coverage of parts of the 4D PTV. Results:
automated left breast simultaneous integrated boost radiotherapy treatment planning’, Physics and Imaging in Radiation Oncology, vol. 28, Oct. 2023, doi: 10.1016/j.phro.2023.100492. Keywords: Autoplanning, breast treatment, ML dose prediction
Digital Poster Highlight 1309
Comparison of 3D versus 4D derived target volume margins in mid-oesophageal cancer radiotherapy planning: volumetric and dosimetric implications Ankita Menon 1 , Alison Hand 2 , Lubna Bhatt 1 , Tom Crosby 3 , Laura Forker 1 , Sarah Gwynne 4 , Hamid Sheikh 1 , Ganesh Radhakrishna 1 1 Clinical Oncology, The Christie NHS Foundation Trust, Manchester, United Kingdom. 2 Medical Physics and Engineering, The Christie NHS Foundation Trust, Manchester, United Kingdom. 3 Clinical Oncology, Velindre Cancer Centre, Cardiff, United Kingdom. 4 Oncology, wansea Bay UHB, Wales, United Kingdom Purpose/Objective: Respiratory motion may impact target coverage and normal tissue dose in oesophageal cancer radiotherapy. 4D-CT allows patient specific margins with motion to be accounted for in target volume delineation. We compared 3D versus 4D target delineation approaches for middle third oesophageal cancers to assess effects on target volume, dose coverage, and organ-at-risk (OAR) doses. Material/Methods: Ten patients with middle-third oesophageal cancer underwent both 4D-CT and 3D-CT simulation. 4D ITV was created as per the SCOPE 2 trial protocol accounting for target motion across respiratory phases with a 0.5 cm isotropic ITV–PTV margin. Retrospective 3D plans used (1) institution based margins (IBM) (1 cm isotropic) and (2) SCOPE2 3D margins (1 cm SI, 0.5 cm LR/AP). 4D ITV/PTV volumes were compared with 3D-derived volumes. Geographical miss analysis classified potential coverage failures using 3D margins: Type 1 – any 4D ITV outside 3D PTV; Type 2 – any 4D PTV outside 3D PTV; Type 3 – any 4D PTV not covered by the 95% isodose; Type 4 – any 4D PTV not covered by the 90% isodose. No Type 1 misses observed. However, SCOPE2 3D margins were associated with multiple Type 2–4 misses as seen in the table. Dose metrics (min dose to 1 cc, D50, D95, D98, D99) and OAR parameters (lung, heart, spinal cord (sc) PRV) were compared using t-tests and repeated-measures ANOVA. Paired t-tests demonstrated significant differences (p < 0.01) for all dose metrics between 4D and 3D IBM plans, with higher coverage in the 3D IBM plan reflecting the larger PTV used for optimisation.
Conclusion: 4D planning produced significantly smaller PTVs while maintaining target coverage and reducing OAR doses compared to conventional 3D margins, which resulted in significantly higher OAR doses. Although SCOPE2 margins generated PTVs comparable to 4D, Type 2–4 geographical misses occurred in several cases due to unaccounted motion. Nevertheless, all SCOPE2 plans met protocol-defined PTV coverage constraints, indicating that coverage remained clinically acceptable. 4D planning offers the optimal approach in mid-oesophageal tumours. Where 4D imaging is not feasible, 3D SCOPE2 provides clinically acceptable coverage and OAR doses. Further evaluation of motion mitigation with Abdominal Compression is underway. References: Bridges S, Thomas B, Radhakrishna G, Hawkins M, Holborow A, Hurt C, et al. SCOPE 2 – Still answering the unanswered questions in oesophageal radiotherapy? A randomised phase II/III trial of dose escalation in definitive chemoradiation with an embedded PET/CT sub-study. Clin Oncol 2022; 34(7): e269-e280 Keywords: 4D CT, 3D CT, Oesophagus Radiotherapy planning
Poster Discussion 1336
Probabilistic planning for non-small cell lung cancer, including clinical target map definition Matijs Geerts 1 , Ivar Bengtsson 2 , Johan Sundström 2 , Albin Fredriksson 2 , Edmond Sterpin 3,1 1 MIRO, UCLouvain, Brussels, Belgium. 2 Research, RaySearch Laboratories, Stockholm, Sweden. 3 ExpRT, KULeuven, Leuven, Belgium
Made with FlippingBook - Share PDF online