Validation of the Precision of a Novel Representation of Images Obtained by the Optos P200Tx Ultra Wide-Field Imaging System

Abstract

 

Purpose

 

To validate the ability of a novel representation of retinal images that correct for geometrical distortion of images obtained by the Optos P200Tx ultra-wide field system.

 

Methods

 

In this cross sectional study, thirteen paired (early and late) fluorescein angiography images of normal fellow eyes of subjects who participated in a clinical trial were analyzed. The images underwent transformation using Optos image projection software that removes warping of geometrical dimensions that occurs on ultra wide-field images. Each image was divided into 8 zones, zones 1-4 in the macula and zones 5-8 in the periphery, within each quadrant of the eye. Image J Version 1.45s (National Institutes of Health, Bethesda, Maryland, USA) was used to determine the same area in each zone for each pair in projected and non-projected (original) images. And the difference in area in each pair was calculated. For projected optomaps, the area can be calculated in mm2, whereas for optomaps this cannot be done and therefore pixels are used.

 

Results

 

Six image pairs showed good registration (degree of agreement when superposition of one image over the other was performed) and seven image pairs showed poor registration beyond seven disc diameters. The table shows the difference in percentage of area measured between pairs of optomaps and pairs of projected optomaps. The results are reported by peripheral vs central and by with and without distortion as measured by registration. Less difference was found in non-distorted to distorted due to less warping under variation of eye alignment.

 

Conclusions

 

No significant differences exist in the change in area measured in pairs of FA frames from the same sequence for optomap vs projected optomap. Hence, the new representation adds no averse distortions. Further research is needed if the difference in area measured on the projected optomap is to be decreased, for instance by compensating for patient alignment variance. The Standard Deviations of the current results are too large to make statistical significant conclusions.

 

 

Keywords: 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • 549 image processing • 688 retina