Clinical evidence of variable proton relative biological effectiveness in locally advanced non-small-cell lung cancer patients

Background and PurposeTo derive and validate the proton variable relative biological effectiveness (RBE) from CT imaging changes in patients with locally advanced non-small-cell lung cancer. Materials and MethodsWe retrospectively analyzed data for patients previously treated on a prospective randomized trial with standard fractionated intensity-modulated photon therapy (IMPT) or passive scattering proton therapy (PSPT). CT image density change (IDC) was calculated in the normal ipsilateral lung between the planning and follow-up CT. Using IDC as the clinical evidence of response, we correlated dose-IDC relationships by fitting a modified Lyman-Kutcher-Burman NTCP model for each treatment modality. RBE was calculated with corresponding planned dose for the IMRT patients and planned physical dose for PSPT patients from the fitted NTCP model. To validate the measured clinical radiographic-based RBE values, we benchmarked the measured RBE values with those calculated from the McNamara, Wedenberg, Mairani, and Flint empirical models. In addition, we also fitted the McNamara and Wedenberg models using the measured RBE, LETd, and IDC values. ResultsThe IDC-based RBE values ranged from 1.3-2.3. Linear regression comparing the radiographic-based RBE value with the empirical models, returned R2 values of 0.025-0.72. The fitted McNamara and Wedenberg models returned pseudo-R2 values of 0.95 and 0.85, respectively. ConclusionRBE values can be computed from CT image density changes as the clinical endpoint and validated using published empirical models. This is the first study to demonstrate the potential of using a clinical radiographic-based endpoint to improve our understanding of the biological effects of proton therapy.