Send to

Choose Destination
Am J Ophthalmol. 1994 Jul 15;118(1):16-23.

Reproducibility of retinal nerve fiber layer evaluation by dynamics scanning laser ophthalmoscopy.

Author information

Department of Ophthalmology, University of Milan, S. Paolo Hospital, Italy.


Scanning laser ophthalmoscopy is a laser-based image acquisition technique, which greatly improves the quality of the examination of the fundus and the retinal nerve fiber layer. To assess retinal nerve fiber layer imaging by scanning laser ophthalmoscopy and evaluate intra- and interobserver reproducibility in the classification of retinal nerve fiber layer defects, three independent observers evaluated on two separate occasions the videotaped images of 150 eyes of 80 consecutive patients with ocular hypertension or glaucoma. Ophthalmoscopy was performed using argon blue light (488 nm), confocal apertures of 3 to 1 mm, and 40-degree and 20-degree field angles. Of 150 eyes, 20 (13.3%) were excluded from the study because of the poor quality of the images (clinically significant cataract or myopic peripapillary atrophy). The retinal nerve fiber layer was evaluated qualitatively according to a standard classification: normal pattern, slit, wedge, and diffuse defects. Intraobserver reproducibility, evaluated by kappa statistic, was excellent (> or = 0.75): observer A = 0.78 (95% confidence limits, 0.67-0.88); observer B = 0.84 (95% confidence limits, 0.72-0.96); and observer C = 0.79 (95% confidence limits, 0.67-0.91). Interobserver reproducibility was also excellent in all cases: observers A-B = 0.84 (95% confidence limits, 0.71-0.98); observers A-C = 0.76 (95% confidence limits, 0.65-0.87); and observers B-C = 0.80 (95% confidence limits, 0.69-0.92). Kappa values ranged between 0.59 and 0.69 for intraobserver reproducibility and between 0.55 and 0.69 for interobserver reproducibility when using only those eyes in which abnormalities were noted by at least one observer.(ABSTRACT TRUNCATED AT 250 WORDS).

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center