The age-related eye dieae tudy (ARED) ytem for claifying cataract from photograph: ARED report no. 4

PURPOE: To decribe the ytem for grading cataract from photograph in the Age-Related Eye Dieae tudy (ARED). METHOD: The ytem for grading cataract in ARED ue photograph taken in a tandardized fahion with pecially modified camera at 11 clinical center. The photograph are evaluated by grader for quality and cataract everity at a central reading center. The area of len involvement i ued to ae the everity of cortical and poterior ubcapular opacitie. Optical denity of nuclear opacity i graded againt a erie of even tandard photograph. Contemporaneou variability in grading i evaluated periodically by having a econd examiner regrade a ubet of the photograph. Temporal variability i aeed by annually regrading a ubet of photograph. REULT: Photograph of 925 eye, mot with no or early len opacitie, were regraded to ae intergrader reliability. For cortical opacitie, there wa an abolute difference of 10% or greater of area involved in 1.9% of the replicate grading. For poterior ubcapular opacitie an abolute difference of 5% of area involved wa noted in 2.8% of the regraded photograph. For nuclear opacitie, abolute difference of 1.5 or more tep were oberved in 0.6% of eye. There wa little evidence of temporal drift in grading any of the three type of opacity during four annual regrade. CONCLUION: We have demontrated a high degree of reliability in grading the everity of len opacitie in a large tudy cohort with motly early len change, the type of cohort mot likely to be entered in clinical trial involving cataract prevention. The Age-Related Eye Dieae tudy ytem for Claifying Cataract From Photograph could be ueful in tudie where there i a need to tandardize data collection over time and acro different data collection ite. Limitation of the ytem include the cot of implementation and, currently, the limited amount of data on grading reproducibility for more advanced len opacitie.