Repeatability and inter-observer variability of in vivo retinal cone imaging using a modified Heidelberg Retinal Angiography (HRA2) in normal subjects

Abstract 

 

Purpose

 

To determine whether the modified narrow-angle Heidelberg Retina Angiograph (HRA2) is repeatable and reliable for acquiring good quality in vivo cone images, and the future potential to develop such an instrument for routine high resolution imaging in primary and secondary care.

 

Methods

 

3 x 3 degrees retinal images were acquired at an eccentricity of six degrees from the fovea, in a group of thirty healthy subjects (21 - 65 years of age), using the 680nm laser arm of the modified Heidelberg Retinal Angiograph (HRA2). Two different operators performed the cone imaging and one operator repeated the measurements over two separate sessions. The images were analysed and cone counts generated by using customized software.

 

Results

 

The mean cone count was 4797 cones per scan area (7054 cones/mm²) ranging from 4258 to 5125 cones per scanned area. The cone counts acquired by the two operators were on average different by 12 cones (<1%) within the scan window with a coefficient of repeatability of 212 cones (1.96 x SD). Cone counts from images acquired by the same observer on two different occasions ranged from 4258 to 5120 with mean bias of 46 cones and coefficient of repeatability of 284 cones per scan area.

 

Conclusions

 

The narrow-scan HRA2 showed very good inter- and intra-examiner repeatability in comparison with alternative adaptive optics cone imaging devices.¹ However, our data were acquired in a younger population (<66 years) and therefore more data are needed to determine the level of repeatability in an older population.<br /> The modified HRA2 could be used in patients with diseases affecting retinal cone density to determine how density relates to visual function at the early stages of the disease. The method has a potential for diagnosis and monitoring of retinal disease in a clinical setting.