Abstract
Purpose :
To assess a new modification of a previously published photographic four-step severity grading scale (S1-S4) for trachomatous scarring (TS).
Methods :
A cross-sectional study was conducted of adult women in Kongwa, Tanzania. Images of the everted eyelids were taken and graded for the presence and severity of TS. The previously published severity scale for TS defines S3 as ‘conjunctival scarring occupying at least one third but less than 90% of the upper eyelid’. In the new modification, S3 was subdivided into two categories: 1) S3A - scarring covering at least one third but less than 50% of the upper eyelid conjunctiva, and 2) S3B - scarring covering at least 50% but less than 90% of the upper eyelid conjunctiva. The feasibility, reliability, and ease of use of this new grading modification were evaluated. This new categorization was then applied to images taken of the same women 3 years prior as part of a longitudinal study, to evaluate whether it could help detect prior undetected progression.
Results :
A total of 5397 eyes of 2712 subjects were included in this study. We report on a subset of these eyes, 4412 eyes of 2217 subjects. Of the eyes graded, 761 (17.5%) were graded as S3. 87 of the S3 eyes were re-graded using the modified grading scheme; 29 (33.3%) were S3A, and 58 (66.7%) were S3B. 64 eyes were noted to also have been graded as S3 3 years prior. To evaluate whether splitting the S3 category might help detect more progression, these eyes were re-graded using the modified grading scheme; 22% of the eyes were noted to have progressed from S3A to S3B. Interobserver agreement using the modified grading system on a set of 36 images was an unweighted kappa of 0.60.
Conclusions :
The modified photographic TS grading system is both easy to implement and reliable; it is also able to help detect more progression of scarring. This more nuanced modified TS grading system will be beneficial for future studies analyzing trachomatous scarring in photographs.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.