Identifying an optimal proton planning strategy for varying bowel density in complex neuroblastomas
Allison Toltz,
United Kingdom
PO-1497
Abstract
Identifying an optimal proton planning strategy for varying bowel density in complex neuroblastomas
Authors: Allison Toltz1, Steven Court1, Vasilis Rompokos1, J.E. Gains2, Mark Gaze2, Andrew Poynter1, Y.-C. Chang2, P.S. Lim2
1University College London Hospitals NHS Foundation Trust, Department of Radiotherapy Physics, London, United Kingdom; 2University College London Hospitals NHS Foundation Trust, Department of Oncology, London, United Kingdom
Show Affiliations
Hide Affiliations
Purpose or Objective
Proton beam therapy (PBT) is particularly sensitive to anatomical changes resulting in dosimetric degradations to the target volume coverage and normal tissue sparing. Young children with neuroblastoma treated under general anaesthesia have been shown to have significant bowel gas variations throughout their course of radiotherapy, posing a major challenge for the safe delivery of PBT.
This work aims to identify the optimal planning approach to ensure pencil beam scanning (PBS) PBT plans generated for paediatric abdominal neuroblastomas maintain their dosimetric robustness against uncertainties from inter-fractional variations in bowel filling.
Material and Methods
A case with a complex target volume was selected for this study: a large midline target overlapping the adjacent bowel and ipsilateral kidney, posing a planning challenge due to kidney tolerances. Plans were created in Eclipse (PCS v15.6.05).
Planning aims were to deliver a prescribed dose of 21 Gy to the CTV in 14 fractions, limiting the contralateral kidney V14Gy < 10% and the adjacent vertebrae V20Gy > 95%. The beam arrangement comprised an anterior (gantry angle 0°) and two posterior oblique fields (gantry angles 215° and 145°) and was kept consistent throughout the study. The anterior field was intentionally included as it had the benefit of reducing the kidney dose and allowed evaluation of the plan robustness to varying bowel filling scenarios.
Plan optimisation using NUPO was performed on four different bowel density scenarios: 1) the original planning CT, 2) air-filled (bowel override to –1000 HU), 3) semi-solid-filled (bowel override to 120 HU) and 4) average (mean HU of structure) (bowel override to –293 HU). Each of these optimised plans was then recalculated on the respective remaining bowel density scenarios (Fig 1), and the dosimetric consequences to clinical objectives were evaluated.
Results
Single field optimization (SFO) achieved better dose objectives compared with multifield optimisation (MFO) when plan optimisation on various bowel density scenarios were performed. When dose recalculations were performed, the SFO plan (20% anterior weighting) optimised on the original planning CT was found to best maintain the intended dosimetry to the target coverage and normal tissues despite different bowel density scenarios.
Figure 1: Diagram of bowel filling scenario used for planning optimisation and recalculation assessment.
Conclusion
SFO plans generated on the nominal planning CT without any bowel density overrides provide the best robustness against intra-fractional bowel density changes.
Future work will evaluate this strategy on an expanded patient cohort to determine if this is a reliable approach for implementation within the clinical PBT planning pathway. Plans will be recalculated using Monte Carlo techniques to benchmark the dose distribution accuracy against Eclipse proton convolution superposition dose calculation.