Correlation between Normal Tissue Complication Probability for Radiation Pneumonitis with Dosimetric Parameters: In-Silico Evaluation Based on Published Radiobiological Models

Abdulhamid Chaikh, Jacques Balosso

Objectives: The first aim of this study is to evaluate the correlation between predicted normal tissue complication probability (NTCP) with dosimetric parameters. Then, the second goal is to evaluate the impact of published radiobiological models on NTCP prediction. Methods: Ten lung cancer patients have been included in this study. The dose was calculated using anisotropic analytical algorithm(AAA). Dose volume histograms (DVH) for healthy lung (lungs minus target volume) were calculated. Then, the dosimetric parameters as mean lung dose (MLD) and Vdose values (% lung volume receiving more than a threshold dose in Gy) as V20 and V30Gy were evaluated. The radiation pneumonitis (RP) was estimated using equivalent uniform dose (EUD) model, Lyman-Kutcher-Burman (LKB) model using radiobiological parameters assumed for AAA; and logistic fit using mean lung dose for RP (RPMLD). The statistical correlation was evaluated using Spearman’s correlation coefficient (ρ-value). Results: The average NTCP values were higher with RPMLD compared to EUD or LKB models, p < 0.05. The data showed a strong correlation between NTCP values with MLD. Conclusions: For a given set of patients, we assessed the NTCP using published radiobiological parameters as well as the correlation between predicted NTCP with dosimetric parameters. The considerable impact of radiobiological model on the NTCP values urges to tune NTCP parameters. Therefore, the good correlation could suggest the use of mean lung dose in NTCP model.

Full Text Attachment
Views : 360 Downloads : 16


Tell a Friend