advertisement
A landmark paper reporting the first successful imaging of retinal ganglion cells in the living human eye
Ethan Rossi, with a team of investigators led by David Williams at the University of Rochester Medical Center, published earlier this year a landmark paper reporting the first successful imaging of retinal ganglion cells in the living human eye.1 This is a major milestone with significant relevance to glaucoma research. Because neuronal cell bodies within the retina produce very low backscatter, they are nearly transparent and extremely difficult to detect - even by advanced ocular imaging modalities such as adaptive optics scanning light ophthalmoscopy (AOSLO). Rossi et al. overcame this challenge by modifying the AOSLO configuration to have an off-axis detection scheme, in which light is collected from portions of the image plane adjacent to the confocal position. The theoretical explanation for the consequential benefit, as first proposed by Elsner, Burns and colleagues,2,3 is that light passing through the confocal aperture along the optical axis represents predominantly backscatter, which masks the fainter contrast produced by multiply scattered light. However, areas of the image plane away from the optical axis contain relatively more light that has undergone multiple scatter within the retina - and thus, additional information. Previous studies cited by Rossi et al. had shown that off-axis detection methods could be used to enhance contrast of otherwise faint retinal structures (e.g., cells of the outer retina and blood vessel walls), but none had successfully detected neurons of the inner retina. Rossi et al. in their study used an offset aperture approach and systematically evaluated the effects of offset distance, offset direction and aperture size as well as various combinations of images from different aperture positions to enhance contrast of retinal cells, including those within the ganglion cell layer.The structures they observed within the retinal ganglion cell layer using the 'multi-offset' approach matched closely the size of retinal ganglion cells known from post mortem histopathology studies. The investigators provided further evidence by imaging anesthetized macaque monkeys simultaneously with the multi-offset AOSLO and a two-photon technique designed to detect intrinsic fluorescence from ganglion cells. Also, the higher light levels used for multi-offset AOSLO imaging in the monkey eye enabled the investigators to detect even subcellular structure in some of the putative retinal ganglion cells.
While these exciting results offer a view with potentially transformative impact for glaucoma research, numerous hurdles remain to be addressed before this remarkable achievement enjoys wide spread implementation. For example, the authors pointed out that the specific combination of offset aperture positions producing the best contrast enhancement in the ganglion cell layer varied in unpredictable ways between imaging locations, as well as within the field of view (1.2° to ~1.5° square) of each imaging area. This study was also limited to retinal locations where the overlying nerve fiber bundles were thin, sparse and ganglion cell bodies were spread in a single layer such as along the temporal raphe, in a small number of eyes. More work is needed to optimize the array of offset positions and image combinations, as well as to image ganglion cells stacked several layers thick nearer to the fovea.
Numerous hurdles remain to be addressed before this remarkable achievement enjoys wide spread implementation
Remarkably, there are already indications that these challenges are being met by other means. Indeed, it was exhilarating to see the results presented during ARVO 2017 by Zhoulin Liu and a group of investigators led by Don Miller at Indiana University,4 who were also able to image retinal ganglion cells in the living human and quantify their diameter and density in three dimensions using AO-OCT running at an A-line acquisition rate of 500 KHz. Using this approach, Liu and colleagues were able to image ganglion cells even at locations where they were stacked four to five rows deep or beneath thicker nerve fiber bundles.
Collectively, the recent report by Rossi et al. and the emerging results of Liu et al. point to a bright future when direct imaging of retinal ganglion cells will be possible in a clinical setting.