Purchase this article with an account.
Robert James Casson, Chelsea Guymer, Glyn Chidlow, John P M Wood, Lloyd Damp; Highly Accurate, Fully-Automated Batch Processing of Retinal Ganglion Cell Counts from Retinal Flatmounts. Invest. Ophthalmol. Vis. Sci. 2019;60(9):640. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The ability to accurately count retinal ganglion cells (RGCs) in animal models of glaucoma is a highly desirable quantitative technique in neuroprotection-based research. Traditionally this has been performed by manual or semi-automated counting of RGCs. This is a notoriously labor intensive, time consuming and tedious process with considerable inter and intra-observer variability. Here we describe and validate fully automated, novel software that accurately and efficiently counts immunostained RGCs. It has the unique capacity to differentiate individual cells in a cluster, batch process images with the same immunostain and export the results in tabular format to a spreadsheet to expediate data analysis.
The software was validated by comparison with manual RGC flatmount counts from both normal Sprague-Dawley rat eyes and eyes that had received a sustained laser-induced intraocular pressure elevation (n = 8 in each group). Retinal flatmounts were immunostained with RGC markers, Brn3a and NeuN; 4 peripheral and 1 central photograph (150 μm x 150 μm) were obtained from each retina. The RGCs in each photograph were manually counted by two independent observers. The photographs were also processed using the novel automated software. The RGC counts using the manual and automated techniques were compared and the level of agreemnet assesed by Bland-Altman plots.
The automated batch processing was at least 50 times faster than the manual counting. There was good interobserver agreement on the manual counts and a very strong level of agreement between the manual and automated methods. On Bland-Altman analysis, the bias (mean difference) between methods was only 0.74% (95% CI, -4.1 - 5.5%). The software performed equally well for control and injured retinas and produced consistent results.
This innovative fully-automated cell counting software can help accelerate data acquisition and improve research productivity. The automated software demonstrated excellent accuracy comapred to manual counting in both normal and inured retinas tested with two different immunostains. It has the potential for wider application with other cell labelling and cell counting protocols.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.
Highly accurate automated imaging with novel software
This PDF is available to Subscribers Only