Abstract
Purpose: :
To enhance the diagnosis of glaucoma by mapping the inner retinal layer thickness using a new macular grid scan pattern for optical coherence tomography (OCT).
Methods: :
The cohort consisted of consecutive 99 patients (190 eyes) in the Advanced Imaging for Glaucoma Study (AIGS), a multi–center, prospective, longitudinal observational study. Eyes were assigned to one of three groups: 102 normal (N), 49 glaucoma suspect/pre–perimetric glaucoma (GSPPG), and 39 perimetric glaucoma (PG) eyes. Two scan patterns were performed using the Stratus OCT: 1) newly developed Macular Grid 7 (MG7), which consists of 16 concentric circular scans with diameters from 0.2mm to 3.5mm and center 0.75mm temple to fovea, and 2) standard peripapillary Fast Retinal Nerve Fiber Layer (RNFL) thickness scan with 3.4 mm diameter. The MG7 data was exported and processed by software we developed. The software maps both total retinal thickness (RT), and the thickness of inner retinal layer (IRL) which was the sum of nerve fiber, ganglion cell and inner plexiform layers. Peripapillary RNFL thickness were directly exported from Stratus OCT.
Results: :
Statistically significant thinning was found in the peripapillary RNFL, macular IRL, and macular RT in PG and GSPPG groups. With 5 percentile cutoff, peripapillary RNFL and macular IRL had better sensitivity in classifying PG and GSPPG subjects than macular RT (0.85 and 0.85 v 0.49 for PG group, 0.35 and 0.20 v 0.04 in GSPPG group). The combination of peripapillary RNFL and macular IRL increased the sensitivity with minimal loss of specificity for both PG and GSPPG groups (sensitivities of 0.92 and 0.45; specificities of 0.96 and 0.96, respectively for PG and GSPPG groups).
Conclusions: :
We have developed a new OCT grid scanning pattern for the macula (MG7) that covers a broader area likely to be damaged by glaucoma. MG7–derived IRL thickness has comparable to peripapillary RNFL thickness in diagnosing glaucoma. The combination of macular IRL thickness and peripapillary RNFL thickness maps improved sensitivity and specificity even further, as the two measurements appear to be complementary.
Keywords: imaging/image analysis: clinical • image processing • retina