Intra- and inter-grader agreement in grading of coverage of panretinal photocoagulation by ultra-wide field color fundus images

Abstract

Purpose
Panretinal photocoagulation (PRP) is the gold‐standard treatment for proliferative diabetic retinopathy (PDR), but evaluation of the coverage of PRP after treatment can be difficult and is not standardized. Ultra‐wide field (UWF) color fundus images may offer an objective alternative to subjective evaluation of PRP‐coverage. We propose a new model to quantify the coverage of PRP treatment in each retinal quadrant by UWF images. The purpose of the study was to evaluate the intra‐ and inter‐grader reliability of the model.

Methods
The retinal coverage of PRP in each quadrant of the retina on 20 UWF color fundus images in patients treated for PDR was quantified by two independent graders (grader A and B), and intra‐ and inter‐grader association (Kappa) were calculated. In each retinal quadrant, OPTOS UWF color fundus images (Optomap, Optos PLC., Dunfermline, Scotland, UK) were used, and each grader evaluated the total retinal area covered by laser. The graders categorized the amount of laser visible on the retina into four categories (no visible laser spots, <50%, 51‐75%, 76‐100%).

Results
The total retinal area covered by laser was 76‐100% (for both graders). Intra‐grader association was 0.715 and 0.464 for Grader A and Grader B, respectively, and inter‐grader association was substantial at 0.627.

Conclusions
Using Optos UWF images, we were able to achieve high intra‐ and inter‐grader reliability for the quantification of the retinal area covered with PRP in patients with PDR. This might aid clinicians when comparing PRP coverage between patients.