Purchase this article with an account.
Boris Busov, Cagri G Besirli; Optimization of ImageJ for automated image analysis to assess for photoreceptor cell death in retinal tissue sections.. Invest. Ophthalmol. Vis. Sci. 2014;55(13):4365.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To develop an automated, accurate way of analyzing different parameters of retinal sections to determine the effects of retinal detachment.
Experimental retinal detachments were created in adult C57/B16 mice. Eyes were obtained 1 and 2 months post-retinal detachment and processed for histologic analysis. Retinal sections were stained and photographed. Sections were examined using ImageJ, a Java-based image-processing program available by the National Institutes of Health (NIH). Automated cell counting was performed in the outer nuclear layer (ONL) of retinal sections where photoreceptor nuclei are located. Mean size of photoreceptor nuclei in attached and detached samples were compared. Parameters in ImageJ were adjusted to improve the accuracy of automated counting. Photoreceptor nuclei area was found. The number of nuclei in the ONL was determined with ImageJ and compared to manual counting. Two other measurements were also assessed: the ratio of ONL area to total retinal section area and the ratio of ONL thickness to total retinal thickness.
Detached retinal sections had significantly lower photoreceptor nuclei counts than attached retinal sections and automated cell counting produced accurate results. Average cell area was calculated and plotted. The cell area was used to obtain a more accurate value for photoreceptor nuclei count—this was accomplished by first determining the average cell area and a standard deviation which were used to construct a confidence interval to eliminate any outliers. Furthermore, the ratio of ONL area to total retinal area showed a reduction in retinal area for the detached retinal sections and the ratio of ONL to total retinal thickness showed that detached retina had significantly reduced ONL thickness.
Automated analysis using an image analysis program available publicly from the NIH is accurate and may be used in place of manual evaluation to reduce the time needed for data collection considerably.
This PDF is available to Subscribers Only