Objectives: To investigate the variability in fluorescein angiography interpretation for photodynamic therapy in age-related macular degeneration.
Methods: Eight graders, who included two TAP-certified ophthalmologists, three other retinal specialists, two fellows in vitreoretinal diseases, and a senior fundus photograph grader, evaluated fluorescein angiograms of six patients treated according to the Treatment for ARMD With Verteporfin (TAP) protocol at a single center. Each patient's baseline angiogram was evaluated to determine whether the CNV lesion was predominantly (> or =50%) classic. For each follow-up angiogram, at 3, 6, 12, and 24 months, the grader was required to determine whether fluorescein leakage was present. Six months after the initial gradings, each reader was again presented with the baseline angiogram for each patient and once again asked to determine whether the CNV lesion was predominantly classic without knowledge of the previous grading. All gradings were performed without knowledge of the clinical course.
Results: In grading initial visit and follow-up visit angiograms, the overall concordance rates were 81% and 82%, respectively. Concordance rates were not statistically different between the group as a whole when compared with the gradings of the two TAP-certified ophthalmologists. When initial visit angiograms were regraded, an intraobserver variability of 17% was noted. Overall, gradings were discordant with the majority opinion in approximately 19% of decisions.
Conclusions: Considerable variability can be expected in fluorescein angiography interpretation as the results of the TAP investigation are applied to clinical practice.