Related Experiment Video
Updated: May 8, 2026
Using Retinal Imaging to Study Dementia
Published on: November 6, 2017
Evaluation of retinal nerve fiber layer thickness measurements for glaucoma detection: GDx ECC versus spectral-domain
Francesca Bertuzzi1, Eleonora Benatti, Giulia Esempio
1*Department of Ophthalmology, Policlinico di Monza, University of Milano-Bicocca, Monza †Mario Negri Institute, Milan, Italy.
Background:
To assess the ability of retinal nerve fiber layer (RNFL) thickness measurements obtained using GDx-enhanced corneal compensation (ECC) or spectral-domain optical coherence tomography (RTVue), and that of ganglion cell complex (GCC) scan available on RTVue, to detect glaucoma.
Methods:
One randomly selected eye of 205 subjects (70 normal, 65 ocular hypertension, and 70 glaucoma) underwent a complete clinical and instrumental examination. RTVue spectral-domain optical coherence tomography was used to assess RNFL thickness and GCC parameters, GDx ECC to assess RNFL thickness. Areas under the receiver operating characteristic curves (AUCs) and sensitivity of the RNFL and GCC parameters were calculated at a fixed specificity of 95%, and the diagnostic abilities of the RNFL values obtained using the 2 instruments were compared. We also compared the results obtained in the normal, ocular hypertensive, and glaucomatous subjects.
Results:
Best GDx RNFL parameter was nerve fiber indicator (NFI) (AUC 0.99, sensitivity 96%); the best RTVue parameters were average (AUC 0.98, sensitivity 90%), inferior-temporal (AUC 0.97, sensitivity 89%), and superior-temporal RNFL thickness (AUC 0.96, sensitivity 87%). There were no significant differences between the 2 devices (P>0.05). Best GCC parameters were focal loss volume (AUC 0.98, sensitivity 91%) and global loss volume (AUC 0.96, sensitivity 87%).
Conclusions:
GDx ECC and RTVue show a very good diagnostic ability to detect glaucoma. Most of the RNFL parameters had high AUCs and sensitivities. The diagnostic validity of GCC was comparable with that of the RNFL parameters, and they may be very useful in detecting RNFL damage.
