Purpose: To explore the interobserver agreement in the diagnosis of retinal angiomatous proliferation (RAP) using fluorescein (FA) and indocyanine green angiographies (ICGA) and to detect which morphologic features of the neovascular lesion are associated with RAP diagnosis.
Methods: In this cross-sectional study, consecutive patients with newly diagnosed neovascular age-related macular degeneration (AMD) evaluated in 8 retina centers were considered. The FA and ICGA were obtained in all centers according to a standard protocol, both performed either as a static or as a dynamic examination. All images were graded by 2 observers from different institutions.
Results: A total of 201 eyes with neovascular AMD of 155 consecutive patients (mean age 76±8 years) were considered. Overall RAP prevalence was 30% using FA and 26% using ICGA. Patients studied with dynamic angiography were twice as likely to be diagnosed with RAP as those using static angiography. Interobserver agreement for the overall detection of RAP was high using FA (kappa: 0.868; 95% confidence interval [CI]: 0.793-0.944) and very high using ICGA (kappa: 0.905; 95% CI 0.836-0.974). The agreement between the 2 observers tended to be higher for the truncated vessel than for the anastomosis in FA as well as in ICGA, but no comparison yielded statistical significance (p=0.258 and p=0.584, respectively).
Conclusions: The interobserver agreement for RAP detection was very good both using FA and ICGA, but the overall detection of RAP was higher for dynamic strategy compared with static one.