Visual Field Estimation in X-Linked Retinitis Pigmentosa Associated with Retinitis Pigmentosa GTPase Regulator (RPGR)...

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• X-linked retinitis pigmentosa (RP) is a genetic eye disease causing progressive vision loss.
• Accurate estimation of visual field (VF) defects is crucial for monitoring RP progression.
• Current methods for VF assessment can be time-consuming and may not capture subtle changes.

Purpose of the Study:

• To develop an efficient artificial intelligence (AI)-driven approach for estimating visual field (VF) parameters in patients with X-linked retinitis pigmentosa (RP).
• To correlate structural Optical Coherence Tomography (OCT) parameters with functional VF measurements.
• To validate the use of AI in analyzing macular OCT scans for predicting VF outcomes.

Main Methods:

• Retrospective analysis of 332 OCT-VF pairs from male patients with RPGR-associated RP.
• AI-powered segmentation and quantification of macular ellipsoid zone width (EZW) and area (EZA).
• Joint analysis of OCT imaging and VF data (mean sensitivity, Hill of Vision parameters) acquired within a 1-month range.

Main Results:

• Ellipsoid zone area (EZA) showed the strongest association with mean sensitivity (MS) and central 20° volume (V20).
• Significant correlations were found between MS and EZW (P=0.00176), and MS and EZA (P=0.0009).
• AI efficiently extracted structural OCT parameters, enabling robust structure-function predictions across a range of disease severities.

Conclusions:

• AI-based analysis of OCT scans provides an efficient method for estimating VF parameters in X-linked RP.
• This approach facilitates research and enables accurate structure-function predictions.
• The method is applicable to patients with varying disease severity, including milder forms of RP.