Noncontact Specular Microscopy of Human Lens Epithelium

PURPOSE. To obtain in vivo specular images of human lens epithelial cells (LECs) from persons with or without age-related cataract (ARC); to identify features that describe individual aspects of these complex images; to develop feature scales to quantify the severity of each feature; and to study the association of these features with LEC count, age, Lens Opacity Classification System III (LOCS III) classifications and microscopic features of lens epithelium in ARC. METHODS. One hundred fifty-two individuals underwent ophthalmic examinations and LOCS III cataract classifications. Specular images of lenses were captured using a modified noncontact corneal specular microscope (SML-2; Konan, Hyogo, Japan). Enhanced images were graded in a masked fashion, and the presence or absence and severity of each of four features in the specular image (“columnar organization,” “linear furrows,” “puffy clouds,” and “black holes”) was graded on a four-step scale. The generalized linear model with intraclass correlation was used to ascertain the statistical significance of associations between age, sex, LOCS III grade, cell count, and feature grade. Capsulorrhexis specimens from 29 patients were studied with correlative light and electron microscopy. RESULTS. LEC density declined with age and was inversely correlated with the scalar grade for puffy clouds and for the size and number of black holes. The scalar grade for columnar organization was inversely associated with the severity of posterior subcapsular and nuclear cataracts, which was the only feature associated with the LOCS III grade of ARC. No statistically significant associations were found between average cell count and LOCS III grade. CONCLUSIONS. With the use of the corneal specular microscope excellent in vivo specular images of the LECs were obtained, the features in these images that correlated well with microscopic findings were classified, and cell density in vivo was estimated. (Invest Ophthalmol Vis Sci. 2000;41: 474 ‐ 481)