Purpose: The aim of this study is to evaluate the impact of normal tissue complication probability (NTCP)-based radiobiological models on the estimated risk for late radiation lung damages. The second goal is to propose a medical decision-making approach to select the eligible patient for particle therapy.
Materials and methods: Fourteen pediatric patients undergoing cranio-spinal irradiation were evaluated. For each patient, two treatment plans were generated using photon and proton therapy with the same dose prescriptions. Late radiation damage to lung was estimated using three NTCP concepts: the Lyman-Kutcher-Burman, the equivalent uniform dose (EUD) and the mean lung dose according to the quantitative analysis of normal tissue effects in the clinic QUANTEC review. Wilcoxon paired test was used to calculate p-value.
Results: Proton therapy achieved lower lung EUD (Gy). The average NTCP values were significantly lower with proton plans, p < .05, using the three NTCP concepts. However, applying the same TD50/5 using radiobiological models to compare NTCP from proton and photon therapy, the ΔNTCP was not a convincing method to measure the potential benefit of proton therapy. Late radiation pneumonitis estimated from the mean lung dose model correlated with QUANTEC data better.
Conclusions: Treatment effectiveness assessed on NTCP reduction depends on radiobiological predictions and parameters used as inputs for in silico evaluation. Because estimates of absolute NTCP values from LKB and GN models are imprecise due to EUD ≪ TD50/5, a reduction of the EUD value with proton plans would better predict a reduction of dose/toxicity. The EUD concept appears as a robust radiobiological surrogate of the dose distribution to select the optimal patient's plan.
Keywords: Equivalent uniform dose; mean lung dose; normal tissue complication probability.