Early diagnosis of certain eye diseases and studying the early progression of the diseases has now become possible, thanks to the work carried out by a team of researchers from three institutes — IISER in Kolkata, L.V. Prasad Eye Institute in Hyderabad and BARC in Visakhapatnam.

The researchers used the retinal data captured by a well-established imaging method in ophthalmology (optical coherence tomography or OCT) and applied an algorithm based on a statistical biomarker tool for early detection of diabetic macular edema. The results were published in the Journal of Biomedical Optics.

Early diagnosis of eye diseases and quantification of disease progression has been a challenge. For instance, the human retina has 10 layers and subtle morphological changes in these layers do not lead to a change in thickness that can be detected by OCT imaging. But OCT images do contain data on subtle refractive index variations and the researchers have successfully teased out this information to help diagnose eye diseases early and study the early progression of disease.

Biological tissues have complex geometrical patterns, which are called multifractals. “The OCT images are light intensity-based and so the multifractal information gets hidden,” says Dr. Ashutosh Richhariya from Hyderabad’s L.V. Prasad Eye Institute and one of the authors.

There is a change in the density of the constituents in each of the 10 layers of the retina and this gets manifested as different refractive indices. Even submicron changes in refractive index of the retina can be detected by the algorithm. “As disease progresses there is a change in the refractive index of the medium which gets recorded in the OCT images. And this data, which is hidden in the OCT images, can be extracted using multifractal detrended fluctuation analysis (MFDFA),” says Sabyasachi Mukhopadhyay from the Indian Institute of Science Education and Research (IISER), Kolkata and one of the authors of the paper.

Using the software, the researchers are able to find a peak at the junction between two layers from the refractive index data extracted from the OCT images. The thickness of a layer can be calculated by measuring the distance between two successive peaks. When the thickness of a layer becomes more as the disease progresses the distance between successive two peaks increases. “Finding different layers in a diseased condition was a challenge. But now it is doable,” says Dr. Richhariya.

“The two-dimensional OCT images have information on depth and lateral direction depth. We are interested in depth-related information. So we first unfolded the two-dimensional image into one-dimensional images and analysed using multifractal detrended fluctuation analysis(MFDFA),” says Nandan Kumar Das from IISER Kolkata and the first author of the paper.

In the case of retina, different layers are affected by different diseases. For instance, the photoreceptor layer is affected in the case of diabetic macular edema, while the top layer (retinal nerve fibre layer) is affected in the case of glaucoma. “So that’s why it is important to know how the disease affects different layers,” says Dr. Richhariya.

“Diabetic macular edema is well characterised and we know what changes to expect. So we used this disease for establishing our technique. We can use our approach for other eye diseases such as age-related macular degeneration (AMD) too,” Dr. Richhariya says.

The researchers are planning to validate it on a larger number of people to be sure of its sensitivity and specificity. They are are also currently working on a smart phone-based diagnostic App so common people can utilise the benefits of our software for early diagnosis and early disease progression.

Published in The Hindu on January 8, 2017