Author: Matthew Case, Richard Castillo, Sunil Dutta, Edgar Gelover, Katja M. Langen, Alexander Stanforth, Mingyao Zhu 👨🔬
Affiliation: Emory University 🌍
Purpose: To evaluate the application of LETd and track-end (TE) objective functions during the optimization of breast proton therapy plans to decrease the risk of lung toxicity and rib fractures.
Methods: Track-End and LETd-optimized plans were generated for 10 cases with measurable radiological lung changes and/or rib fractures and a schedule of 50 Gy in 25 fractions. The research version of a commercial TPS was employed. To evaluate the impact of TE/LETd optimization, clinical and TE/LETd-optimized plans were recalculated to generate biological dose (BD) maps using a linear LETd-based model (Beltran). Dose metrics scored were the ipsilateral lung mean dose, V20Gy[%], V5Gy[%] and ribs max dose D0.5cc. NTCP models for the lung (Tucker), and ribs (Wang) were incorporated into our TPS’s scripting environment. In addition, Dose-LETd volume histograms (DLVH) were used to extract VD,LET(d,l)% values, i.e. the volume receiving simultaneously a dose level, “d” and an LETd value, “l”.
Results: The risk probability of predicted grade 3+ radiation pneumonitis for TE/LETd-optimized plans was reduced by 7.5%. The predicted risk of rib fracture was reduced by 1.9% for the TE/LETd-optimized plans. All differences had p-values < 0.05. The improvements in BD metrics are 11.7%, 9.5%, 5.2 Gy, and 5.8 Gy for the lung V20, V5, mean dose, and rib max dose, respectively. All differences had p-values < 0.05. DLVH data show that on average the lung volume receiving V(20,7)% was reduced from 13.3% to 4.3%, and the ribs volume receiving V(47.5,4.5)% was reduced from 11.1% to 2.6% (p-values <0.001).
Conclusion: Our preliminary results show that using LETd and Track-End objectives has the potential to reduce the risk of pneumonitis by up to 7.5% and the risk of rib-fracture by ~2%. A shrinkage of ~9% in VD,LET(d,l)% values indicates that LETd reduction in relevant dose areas is also achievable.