Optical Coherence Tomography Angiography Characteristics of Radiation Optic Neuropathy

Overview

Aim: To develop a suitable radiation optic neuropathy (RON) diagnostic model based on optical coherence tomography angiography (OCTA) and to reveal the correlation between endocrine and OCTA indexes in nasopharyngeal carcinoma (NPC) patients with radiation encephalopathy (RE). Methods: This retrospective cross-sectional study included 80 male NPC patients with RE following radiotherapy (40 non-RON and 40RON). Endocrine and OCTA indexes were collected. The macular and peripapillary vessel density (VD) were automatically analyzed using AngioVue 2.0 of the RTVue XR Avanti device. The OCTA indexes were included in the multivariable binary logistic regression model between non-RON and RON. For all RE patients, a multiple linear regression was performed between each of the OCTA indexes and the endocrine indexes.

Best corrected visual acuity (BCVA) and OCTA (Optovue, Inc., CA, USA) were performed in the RE patients. All examinations were performed by experienced ophthalmic examiners. AngioVue software 2.0 of the RTVue XR Avanti device (Optovue, Inc., CA, USA) was used to perform OCTA imaging. The split-spectrum amplitude-decorrelation angiography (SSADA) algorithm was applied to the scans. Images with a scan quality of not less than 6 were included, while images with obvious artifacts detected by a senior ophthalmologist were excluded. The vessel density (VD) and retinal thickness of the macular and optic nerve head (ONH) area were automatically analyzed in 6 × 6 mm angioretina mode and 4.5 × 4.5 mm angiodisc mode, respectively. The statistical analyzes were performed using SPSS 24.0 (SPSS Inc., IL, USA). Student's t-test was performed to compare normally distributed variables, while the Mann-Whitney test was performed to compare nonnormally distributed variables. Categorical variables were analyzed with the chi-squared test Statistical significance was set at p \< 0.05. The major statistically significant OCTA variables were included in the multivariable binary logistic regression model between NRON and RON. Then, thereceiver operating characteristic (ROC) curve was generated according to the model, and the area under the curve (AUC), sensitivity (SE) and specificity (SP) were reported. In addition, for all RE patients, backward multiple linear regression was performed between each of the above OCTA variables and the statistically significant endocrine indexes.