This new technique can detect human eye diseases early

With the coming of new technology, a software analysis can now highlight the subtle differences between a diseased and healthy retina.

Kolkata: With the coming of new technologies, a software analysis can now highlight the subtle differences between a diseased and healthy retina.

Researchers have developed a new technique that shows promise of early detection of diseases of the human eye and could also help design a smart phone-based retinal diagnostic app.

This new technique may also prove useful for identifying other types of tissue abnormalities from OCT images.

Optical coherence tomography (OCT) is a non-invasive imaging test that enables physicians to monitor alterations in the thickness of the retinal layer as disease progresses in the human retina.

Researchers at IISER-Kolkata, L.V. Prasad Eye Institute in Hyderabad and Bhabha Atomic Research Centre in Mumbai have developed a software analysis that was able to parse through the differences between a healthy retina and retinal layers of diabetic macular edema (DME) subjects from OCT images.

With OCT, each of the retina's distinctive layers can be seen, allowing ophthalmologists to map and measure their thickness.

These measurements help with diagnosis and provide treatment guidance for glaucoma and retinal diseases, such as age-related macular degeneration and diabetic eye disease.

"However, subtle changes in the retinal layers due to early disease progression are often not discernible. Our analysis can overcome that drawback and pinpoint early disease progression. The software can directly tell us," N.K. Das, IISER-Kolkata, told IANS.

"In the future, we can develop affordable and portable devices for early detection, including a smart-phone app where images can be fed through an optic fibre," Das added.

Published in the Journal of Biomedical Optics, the study is co-authored by Nirmalya Ghosh, Sabyasachi Mukhopadhyay, Jay Chhablani, Ashutosh Richhariya, Kompalli Divakar Rao and Naba Kishore Sahoo.

(With IANS inputs)