AI for screening of multiple retinal and optic nerve diseases, Health News, ET HealthWorld

AI for screening of multiple retinal and optic nerve diseasesBy Dr S Natarajan

Artificial intelligence refers to software that appears intelligent. An AI algorithm is used to detect multiple optic diseases using deep learning. This AI system is intended to help general physicians and non-ophthalmic healthcare providers who need an accurate and immediate assessment of their diseases. The article introduces existing AI methods, the ophthalmic imaging modalities, and a summary of major eye diseases and the application of AI to screen them and facilitate promising AI projects in ophthalmology. The most intensively studied are Diabetic Retinopathy, Glaucoma, and AMD, which are explained further.

Diabetes affects around 415 million people globally, or one out of every eleven adults. It’s a long-term diabetes consequence and vasculopathy that affects one-third of diabetic people and can result in permanent blindness. Further, to address this global disease, AI can be used to predict DR risk and progression in diabetic individuals. Utilising 52 optical coherence tomography (OCT) pictures, a DL-based computer-aided method detects DR with an AUC of 0.98. Even with the good outcomes in the cross-validation process, the system needs further validation in larger patient cohorts. A CAD system based on CML algorithms that used optical coherence tomography angiography (OCTA) pictures to autonomously identify non-proliferative DR (NPDR) also had good accuracy and an AUC.

In the industrialised world, AMD is the major cause of irreversible blindness among the elderly. The aim of using ML algorithms is to automatically identify AMD-related lesions to improve AMD diagnosis and treatment. ML has been used to detect drusen, fluid, reticular pseudodrusen, and geographic atrophy in SD-OCT and fundus images. The accuracy is typically greater than 80per cent, and agreement between models and retina specialists can exceed 90per cent. Multiple CML approaches have been used for the automated diagnosis and grading of AMD. However, the most significant work has been done in the last two years using DL approaches. The DCNN appears to play a screening role in these studies, and its performance is comparable to that of physicians. Exudates, macular edema, drusen, and choroidal neovascularisation have all been detected automatically using DL algorithms.

Glaucoma is the third-largest sight-threatening eye disease in the world and has a critical impact on global blindness. Glaucoma patients have high intraocular pressure, optic nerve head (ONH) damage, retina nerve fibre layer (RNFL) defect, and progressive vision loss. Automatically detecting features related to glaucoma has great significance in its timely diagnosis. So far, DL-based glaucomatous diagnosis models have been built using fundus pictures, VFs, and wide-field OCT scans, and DL is better than CML at detecting preperimetric open-angle glaucoma (OAG) eyes.

The AI system is intended to help general physicians and non-ophthalmic healthcare providers who need an accurate and immediate assessment of diseases and contribute significantly to the field of ophthalmology.

By Dr S Natarajan, Chief, Clinical Services, Aditya Jyot Eye Hospital, A unit of Dr. Agarwal’s Eye Hospital.

(DISCLAIMER: The views expressed are solely of the author and ETHealthworld does not necessarily subscribe to it. shall not be responsible for any damage caused to any person / organisation directly or indirectly)

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: