Diagnostic Performance for Detection of Glaucomatous Structural Damage Using Pixelwise Analysis of Retinal Thickness...

Area of Science:

• Ophthalmology
• Medical Imaging
• Glaucoma Diagnostics

Background:

• Glaucoma diagnosis relies on detecting structural changes in the retina.
• Optical coherence tomography (OCT) provides high-resolution cross-sectional images of the retina.
• Accurate measurement of macular retinal layers is crucial for early glaucoma detection.

Purpose of the Study:

• To compare the diagnostic accuracy of individual and combined macular retinal layer thickness measurements.
• To discriminate between glaucomatous, glaucoma suspect, and healthy eyes using pixel-based analysis.
• To evaluate the effectiveness of different retinal layers and their combinations in glaucoma diagnosis.

Main Methods:

• Retrospective analysis of OCT scans from glaucomatous, suspect, and healthy eyes.
• Manual correction of segmentations to measure thickness of nerve fiber layer (NFL), ganglion cell layer (GCL), inner plexiform layer (IPL), ganglion cell complex (GCC), and total neural retina (TR).
• Creation of receiver operating characteristic (ROC) curves using abnormal pixel counts to assess diagnostic accuracy.

Main Results:

• The nerve fiber layer (NFL) and ganglion cell layer (GCL) showed the highest diagnostic accuracy among individual layers.
• The ganglion cell complex (GCC) demonstrated high area under the ROC curve (AUC) (94.3%), comparable to GCL (93.4%) and NFL (92.3%).
• A Boolean combination of NFL or GCL achieved the highest AUC (95.4%) with 85.1% sensitivity at 95% specificity, outperforming GCC.

Conclusions:

• Pixel-based thickness measurements of NFL and GCL are highly accurate for glaucoma diagnosis.
• GCC performance is equivalent to NFL and GCL, but combined NFL or GCL offers superior sensitivity.
• Assessing multiple retinal layers independently provides complementary information for comprehensive glaucoma diagnosis.