Skip to main content
  • University of Rochester Medical Center
    Comprehensive Ophthalmology, Retina/Vitreous

    Scientists from the University of Rochester and University of Pittsburgh have developed a new imaging technique that captures individual neurons within the retinal ganglion cell (RGC) layer.

    “In principle, this new approach might eventually allow us to detect the loss of single ganglion cells,” said David Williams, PhD, dean for Research in Arts, Sciences, and Engineering and the William G. Allyn Chair for Medical Optics at the University of Rochester. “The sooner we can catch the loss, the better our chances of halting disease and preventing vision loss.”

    Imaging individual human RGCs has not been done before, in part because they are nearly perfectly transparent. In a study published this week in the Proceedings of the National Academy of Sciences, the researchers showed how they used an existing technology – confocal adaptive optics scanning light ophthalmoscopy – to collect multiple images and then combined those images in a technique called multi-offset detection.

    Led by Ethan A. Rossi, PhD, assistant professor of Ophthalmology at the University of Pittsburgh, the researchers tested the technique in animals as well as volunteers with normal vision and patients with AMD.

    Rossi and his colleagues visualized not only individual RGCs, but structures within the animals’ cells, such as nuclei. Images from humans were of poorer quality than from the monkey for several reasons, mainly due to the lower light levels used to ensure patient safety.

    However, as they further develop the technique, the researchers believe it has the potential to diagnose glaucoma before the retinal nerve fiber thins, even before cell death, by detecting size and structure changes within the cell bodies.

    “This technique offers the opportunity to evaluate many retinal features that have previously remained inaccessible to imaging in the living eye,” said Rossi. “Not only RGCs, but potentially other nearly transparent cell classes as well.”