Clinical Validation of a Novel Imaging Method for Conjunctival Hyperemia: Red Intensity Coefficient Analysis Using the MR-6000 System
Objectives: Grading of conjunctival hyperemia is subjective and prone to interobserver variability. This study evaluated the Red Intensity Coefficient (RIC), an automated image-analysis algorithm implemented in the Tomey MR-6000 multifunction unit, as an objective method for assessing conjunctival hyperemia. Methods: RIC grading thresholds were established using a development dataset of 313 standardized anterior segment images and subsequently applied without modification to an independent validation cohort of 62 eyes. Three expert ophthalmologists graded all images according to the Efron grading scale. Agreement between automated RIC-based grading and the expert consensus reference standard was assessed. Results: Agreement between the expert consensus and the MR-6000 was moderate (linearly weighted Cohen κ=0.592; 95% CI, 0.432–0.721). Exact agreement was observed in 54.8%, and 95.2% differed by no more than one Efron grade. For binary classification of conjunctival hyperemia, sensitivity was 91.2%, specificity 60.0%, and overall accuracy 88.7%. Robustness analyses accounting for inclusion of both eyes yielded comparable agreement estimates. Conclusions: The predefined RIC algorithm demonstrated moderate agreement with expert consensus grading in an independent validation cohort. Automated RIC-based grading may support standardized assessment and documentation of conjunctival hyperemia. Larger multicenter studies are warranted to evaluate repeatability, external validity, and longitudinal clinical utility.
Authors
- Martin Emesz
- Michael Amon
- Hideo Takata (ORCID: https://orcid.org/0009-0000-2165-2137)
- Julian Bernhart (ORCID: https://orcid.org/0009-0009-2423-1843)
- Michal Ptak
Publication Details
- Journal
- Eye & Contact Lens Science & Clinical Practice
- Published
- 2026-11-19
- DOI
- https://doi.org/10.1097/icl.0000000000001325
- Primary Topic
- Ocular Surface and Contact Lens
- Type
- article
- Field-Weighted Citation Impact
- 0.00