Abstract
Purpose :
To evaluate the diagnostic accuracy of peripapillary retinal nerve fiber layer (pRNFL) thickness, peripapillary radial capillary plexus perfusion density (pRCP-PD), and their combination to discriminate among healthy controls, glaucoma suspects (GS), and glaucoma patients.
Methods :
Patients were prospectively enrolled from 09/2016 to 08/2019 underwent standard automated perimetry, structural OCT, and 6x6 mm OCT angiography (OCTA). Average and quadrant pRNFL thickness and pRCP-PD, and corresponding OCT and OCTA regions were combined with a multivariable logistic model. Diagnostic accuracy of structural parameters, vascular parameters, and their combination was evaluated with age-adjusted clustered ROC curves.
Results :
104 eyes of 59 glaucoma patients, 60 eyes of 33 GS patients, and 78 eyes of 41 healthy subjects were included. Glaucoma patients (mean±SD age: 63±12 years) were older than GS (54±17 years, p=0.007) and healthy subjects (53±14, p<0.001), and had worse MD (median[IQR]: -2.3[-5.7 to -0.9] dB) than GS (-0.9[-2.3 to -0.2] dB, p=0.009) and healthy subjects (-0.1[-1.8 to 0.7], p=0.01). Gender, ethnicity, and structural and vascular signal strength did not differ among the groups. Average and all quadrants pRNFL values (Fig 1) were significantly lower in glaucoma patients compared to GS and healthy subjects (p=0.017 for temporal quadrant, p<0.001 for all the others), with average, superior, and inferior quadrants showing the best separation among the groups. Differences in pRCP-PD values (Fig 1) were less pronounced among the groups, and only average, superior, and inferior pRCP-PD were significantly reduced in glaucoma patients compared to healthy subjects and GS (p=0.006, p=0.014, and p<0.001, respectively). Average, superior, and inferior pRNFL had significantly higher AUROC values compared to the corresponding pRCP-PD parameters (p<0.01, p<0.05, and p<0.001, respectively), showing a higher ability to discriminate between glaucoma patients and either healthy subjects or GS (Fig 2). The ROC curves for the index integrating structural and vascular data (Fig 2) did not perform better than structural OCT alone at any quadrant (p>0.2).
Conclusions :
Peripapillary OCT-A does not add any additional benefit to OCT in the diagnosis of glaucoma. Further studies are warranted to clarify if different OCT-A angiocube sizes and algorithms may improve glaucoma diagnosis.
This is a 2020 ARVO Annual Meeting abstract.