Comparison of fluorescein angiography with microvascular anatomy of macaque retinas.

Recent anatomic work has shown that the capillary network of the fovea is multilaminar. We have identified the elements of this network that are visualized by fluorescein angiography and those that are missed. Fluorescein angiograms of monkey retinas (Macaca fascicularis) with good visualization of individual capillaries were obtained by standard clinical techniques. Retinal whole mounts were prepared from the same animals. Anatomic drawings made from the whole mounts were used to identify which parts of the capillary network were visualized angiographically. Angiographic estimates of dimensions of the foveal avascular zone corresponded closely to the anatomy. Capillary visibility declined rapidly from near perfect visualization at the edge of the foveal avascular zone to less than 40% by 900 microns eccentricity. While all the widest capillary segments (diameter 6.1-7.0 microns) were visualized, only 43% of the modal group of capillary segments (diameter 4.1-4.5 microns) were detected. When a relatively homogeneous population of capillaries was analyzed (diameters limited to the narrow range of 4.0-5.0 microns), visualization declined monotonically with depth in the retina. Capillary segments in the nerve fiber plane were visualized more than four times as effectively as segments of comparable diameter in the deepest vascular plane. High quality angiograms accurately delineate the foveal avascular zone, but they visualize only a fraction of the adjacent multilaminar network. Therefore, current techniques may not detect the earliest nonperfusion of capillaries in vaso-occlusive diseases. Capillary visibility is a joint function of diameter and of retinal depth. The decline in visualization with retinal depth implies that light scattering in the retina degrades the angiographic image.