Abstract
Purpose :
Several risk calculators based solely on OCT data have been proposed to screen for, or diagnose, glaucoma. The purpose of this study was to evaluate the benefits of adding clinical data such as age, intra-ocular pressure and central corneal thickness to such a calculator.
Methods :
Data from a convenience sample of patients being seen at the Illinois College of Optometry (ICO) was reviewed to identify subjects with OCT data from Maestro2 (Topcon, Tokyo) as well as other clinical information, including intra-ocular pressure, central corneal thickness and visual fields. We limited analysis to subjects with definite glaucoma or suspect glaucoma, as assessed clinically.
The OCT-based risk score (ORS) used was Model 3 from Fukai et al, TVST 11(8), 2022.
A logistic regression model was developed that included the ORS for both eyes and the IOP and CCT for both eyes as well as the age of the subject.
Receiver operating characteristic analysis was performed for both the ORS and the clinically-based risk score (CRS) to try to distinguish subjects with definite glaucoma from suspects.
Results :
A total of 106 (55 glaucoma, 51 suspect) patients were included in the dataset, 75% African-American, 64% female. Figure 1 shows the receiver operating curves for the standalone ORS model (AUC = 0.85) as well as a generalized linear model based on the ORS and the additional clinical factors from both eyes and the age of the patient (AUC = 0.88).
Conclusions :
The addition of age, IOP, PSD, and CCT improve the performance of a calculator at distinguishing subjects with glaucoma from those with only suspicion of glaucoma or no glaucoma.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.