Comparing Different Normal Tissue Complication Probability (NTCP) Models of Radiation-Induced Temporal Lobe Injury after Intensity-Modulated Radiation Therapy for Nasopharyngeal Carcinoma.

L. Zeng,S. Liu,H. Deng,T. X. Lu,X. W. Deng,A. W. Liu,X. Wan
DOI: https://doi.org/10.1093/annonc/mdy438.019
2019-01-01
Abstract:To compare different normal tissue complication probability (NTCP) models to predict the incidence of radiation-induced temporal lobe injury (TLI) for nasopharyngeal carcinoma (NPC) patients who received intensity-modulated radiotherapy (IMRT). Data in 351 patients with nasopharyngeal carcinoma treated with IMRT were reviewed retrospectively. A large number of dose-volume effect models has been fitted to data on the occurrence of TLI after IMRT for NPC. The following models were considered: (1) Lyman model and (2) logit-formula with dose-volume histogram (DVH) reduced to generalized equivalent uniform dose (EUD), (3) serial reconstruction unit (RU) model, (4) Poisson-EUD model, and (5) mean dose model. The fitting was done with the maximum likelihood method. Twenty-nine of 351 patients (8.3%) developed TLI; 21 patients had unilateral TLI, and eight had bilateral TLI. As assessed qualitatively and quantitatively, the Lyman-EUD model fitted the data very well [a=∞, m=0.07 (0.011, 5.0), TD50=81.0 Gy (68.5,119.5Gy)], and was superior to the remaining models. Therefore, it demonstrates that temporal lobe is a serial organ. Temporal lobe mean dose did not correlate to TLI. Our study clearly confirms a dose effect for TLI after IMRT for NPC. The Lyman model was the most accurate predictor for the incidence of TLI. It suggests that delivery of a high dose of radiation to a small volume of the temporal lobe is unsafe.
What problem does this paper attempt to address?