Purchase this article with an account.
F. Knezevich, D. M. Moscaritolo, H. Jampel, R. Zeimer; Pilot Evaluation of a New Non-Mydriatic Semi-Automated Camera Designed for Glaucoma Screening in Primary Care Offices. Invest. Ophthalmol. Vis. Sci. 2008;49(13):4222. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To assess the adequacy of optic disc images acquired with a new non-mydriatic semi-automated camera designed for glaucoma,diabetic retinopathy and macular degeneration screening in primary care offices (DigiScope GenIII).
Glaucoma patients and normal volunteers were recruited from a glaucoma clinic and their fundi were imaged. A series of ten consecutive fundus images were acquired at different focal settings of the optics. An automated software algorithm developed on a Matlab © platform identified the image with the best focus and the operator was asked to accept the algorithm’s determination or to select another image. The optics were adjusted automatically for best focus and four images of the disc were acquired with a CCD camera (12 bit,696x520 px, 80ms per frame)under illumination by a green laser with a power density of 2mW/cm². An image of the pupil was acquired in infrared during the imaging session and was used to measure the pupil size immediately before and after the fundus image acquisition. The images were stored digitally for later processing. A masked operator chose the best of the four images and performed image enhancement with routine tools provided by using Adobe Photoshop ©. Masked readers were then asked to evaluate the quality of the images and categorize each as good, adequate, fair or poor. The same images were presented to a masked glaucoma specialist who determined whether they could be used to detect glaucoma damage and classified them as yes, no , maybe, or unreadable.
We studied 38 eyes of 31 glaucoma patients aged 40 to 92 years and 22 eyes of 14 normal volunteers aged 25 to 71. The quality of the images were determined to be of the following quality grades: 37% good, 31% fair, 15% adequate and 17% poor. Poor contrast which made it difficult to distinguish between the cup and the disc, and blurring of the image were the apparent cause of a poor quality image. 98% of the images with better than poor quality were judged adequate for identification of glaucoma damage. The minimum pupil size of interpretable images was 2.1mm. The mean pupil size pre imaging was 3.7mm, and post imaging was 3.0mm.
This pilot study indicates that, without drug induced dilation, this semi-automated camera can provide useful optic disc images in 84% of a cohort of normal individuals and patients with glaucoma. This ratio may differ with technical improvements and the nature of the populations in primary care environments.
This PDF is available to Subscribers Only