Author: Alessandro Arrigo (Italy)
Co-authors: Cristian Perra, Emanuela Aragona, Maurizio Battaglia Parodi, Francesco Bandello
Purpose
Müller cells (MC) play a fundamental role in the homeostasis regulation of the vertebrate retina, are the most represented glial cells within retinal structure and are found involved in several retinal diseases. In the recent years, optical coherence tomography (OCT) techniques offered powerful ways to perform non-invasive, in vivo, detailed histology-like analyses of the human retina. Further information can be achieved by using post-processing algorithms. Through these methodologies, it is possible to use the light properties of each retinal component to perform dedicated analyses. The main purpose of this study was to propose a new post-processing technique to detect and analyze MC in the human retina and to compare our imaging findings with histologic analyses.
Setting/Venue
Clinical setting; Department of Ophthalmology, IRCCS San Raffaele Scientific Institute, Vita-Salute University, via Olgettina 60, Milan (Italy).
Methods
The study was designed as a cross-sectional, observational case series. Consecutive healthy human subjects and patients affected by ocular neoplasia, not involving the macular region and requiring surgical enucleation, to obtain histologic samples. The imaging protocol included high resolution 3x3mm macular OCT angiography (OCTA) acquisitions (DRI Triton Topcon; Topcon Corporation; Tokyo, Japan) (wavelength 1050nm; 100.000 A-Scans per second; in-depth resolution digital 2.6µm and optical function 8µm). The MC detection started with the assumption that extrafoveal MC are vertically oriented elements, not interfering with light passage and absorption. From this point of view, the analyses focused on the isolation of the non-reflective vertical signal included between the internal limiting membrane and the external limiting membrane. The method consists in a set of algorithms composed of the following blocks: data intermediate format, registration, layer extraction, MC counter. Blood flow signal, detected by OCTA, was excluded to improve data accuracy. With respect to the histologic analyses of the enucleated eyes, after the OCTA acquisition and the surgical enucleation, a 3x3mm area of the macular region was examined by means of Vimentin and glial fibrillary acidic protein. Imaging based reconstructions and histologic samples were qualitatively evaluated to assess morphologic agreement.
Results
Overall, 18 healthy subjects (10 men, mean age 35±10 years) and 2 patients (1 man, mean age 51±9 years) affected by peripheral intraocular melanoma, not involving the macular region, without any other ocular or systemic disorder, and underwent surgical enucleation were included. Best corrected visual acuity was 0.0±0.0 LogMAR (20/20 Snellen equivalent) for all the eyes. Our approach allowed to analyze, from structural OCT images, the trans-retinal, vertical, linear, hyporeflective signal, included between the internal limiting membrane and the external limiting membrane. This signal was characterized by a continuous “single-line” morphology. Remarkably, the signal isolated by structural OCT images showed high qualitative matching with histologic samples. MC signal resulted more detectable in the extrafoveal region than in foveal one. Our quantitative analysis showed an overall MC number of 42232±3478 cells localized in the macular volume of 3x3mm (~42.000/9 mm2 cells).
Conlusions
Most of MC studies are based on histologic analyses performed in animal models. Our methodology represents the first effort to perform dedicated MC analyses in-vivo, in the human retina. The qualitative comparison with histologic reconstructions showed a very good matching with respect to the vertical trans-retinal orientation of the MC in the extrafoveal region. Our analyses showed an underestimation of MC signal detection in the fovea than in the extrafoveal regions. We may advance two hypotheses to explain this finding. The first one is related with resolution limitations, making current structural OCT images not sufficient to detect foveal MC in their entirety. Indeed, foveal MC are more numerous within the fovea than in the rest of the retina, thus making assumable MC sizes below the 8µm threshold of structural OCT sensitivity. Moreover, MC show a very peculiar anatomy at the level of the fovea, following a z-shape morphology, making possible the change of their optical properties and interfering with the proper detection through our approach. The quantitative analysis showed good agreement with previous reports (~42.000/9 mm2 vs ~15.000 cells/mm2 found by previous histologic studies). In conclusion, our study proposed a new method to analyze MC in-vivo, in the human retina.
Financial Disclosure
N/A
Comments
-
