Purchase this article with an account.
Yesha Raval, John Rodriguez, Keith Lane, Paulo Gomes; Confounding Factors on Computerized Assessment of Conjunctival Redness Induced by Conjunctival Allergen Challenge (CAC). Invest. Ophthalmol. Vis. Sci. 2013;54(15):2553.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The Conjunctival Allergen Challenge model (CAC) is the standard regulatory model for evaluation of ocular therapeutic agents and is responsible for 19 new drug approvals. Itching is subjectively assessed while redness of three vessel beds (ciliary, conjunctival, episcleral) is evaluated by a trained clinician using a slit lamp in accordance with standardized and FDA accepted scales. In an effort to add precision to redness quantification, a computer program originally design to capture injection patterns due to dry eye, was adapted to capture redness levels in allergic conjunctivitis. Allergy subjects frequently present with chemosis. This condition causes optical distortion in a digital image due to swelling of the conjunctiva, producing a blanching and opaque effect over the visualization.
Conjunctival photography was taken as part of a 60 subject phase II clinical trial studying the effects of an alpha-adrenergic agonist. Inclusion criteria incorporated a CAC response ≥ 2 for both ocular itching and clinician graded ocular redness within 10 minutes of instillation. Photography was taken of the conjunctiva immediately following inclusion into the study. The Photography was analyzed for redness with a computer program developed with the Opencv computer vision library for calculating redness intensity and the output was compared to the clinician grades in an effort to verify the ability of the software to capture injection in an allergic eye.
The correlation between the clinician grade and the software output was 0.27. This level of accuracy was insufficient to predict clinical graded scores to less than one unit and thus unable to replicate scores of an expert grader.
The unexpected low correlation between these two methods of assessing redness of the conjunctiva is likely due to the chemosis that results from an allergic response. A clinician grading the amount of redness with a standardized scale can overcome the resulting optical distortion with the aid of a slit lamp and the ability to scan all different angles with a slit beam. While this computerized technique has proven useful for dry eye redness pattern dissertation, it failed to produce precision with allergic conjunctival injection.
This PDF is available to Subscribers Only