Room: AAPM ePoster Library
Purpose: This work investigates the incorporation of tumour visualization constraints into the optimization of volumetric modulated arc therapy (VMAT) plans using fiducial markers.
Methods: We investigated this approach on multiple patient disease sites (10 prostate, 5 liver, and 5 lung) using a radiotherapy optimization development software (MonArc – precursor to Eclipse RapidArc), where these new visualization constraints could be added to standard dosimetric constraints in the objective function. For all the investigated disease sites, three fiducial markers were implanted inside (prostate) or around (lung, liver) the planning target volume (PTV); and VMAT plans were created for each patient. We modified MonArc to analyze the multi-leaf collimator beam’s-eye-view (BEV) at all control points in the gantry arc, while including marker-based visualization constraints of type ‘hard’ (i.e. requiring 100% visualization of all markers, HC) and ‘soft’ (i.e. penalizes visualization for one [SC?] or two markers [SC??] only) in the optimization process. Dose distributions from the constrained plans (HC, SC?, and SC??) were compared to the non-constrained plan (NC) using metrics including conformity index, homogeneity index, mean PTV dose, and doses to organs-at-risk (OAR).
Results: Using the average index, there was a 7.1% difference for SC?, 10.6% for SC?? and 14.6% for HC respectively, compared to NC for the liver patients. However for lung patients, there was a 2.7% difference for SC?, 4.1% for SC?? and 6.7% for HC respectively. NC produced the best target conformity and the least OAR doses for the entire dataset, followed by the SC?, SC?? and HC, respectively. OAR doses depend on the constrained marker location.
Conclusion: We demonstrated that visualization constraints can be incorporated into the optimization together with dosimetric objectives to produce treatment plans satisfying both types of clinical objectives. This approach should ensure greater clinical success when applying real-time tracking algorithms while using VMAT delivery.
Funding Support, Disclosures, and Conflict of Interest: Funding support provided by Research Manitoba. No conflict of interest or disclosures.