Aravind Eye Hospital’s Pondicherry unit has made significant strides in diagnosing glaucoma using artificial intelligence (AI). The hospital found that AI technology delivered more accurate results than teleophthalmologists, prompting the expansion of the system across all its vision centers in the region.
The hospital conducted a study at six of its centers, screening 303 individuals for glaucoma. The results were then compared with those provided by teleophthalmologists using the Fundus camera, which captures retinal images for analysis. The findings revealed that AI technology surpassed teleophthalmologists in diagnostic accuracy.
R. Venkatesh, chief medical officer at Aravind Eye Hospital, explained that the AI technology was developed by a Bengaluru-based company and allows the retina’s image to be captured on an iPhone. “This AI can function offline as well. We check for eye pressure and optic nerve changes, which are key indicators of glaucoma. The system is trained to detect these changes and flag glaucoma cases,” he said. The AI system can refer the detected changes to a doctor for further examination.
While AI has already been widely used to detect diabetic retinopathy, this is the first time it has been employed to diagnose glaucoma. The AI achieved 91% sensitivity in identifying glaucoma cases and 93% specificity in confirming normal cases, which are critical measures in medical testing. Sensitivity refers to the system’s ability to correctly identify patients with glaucoma, while specificity measures its accuracy in declaring a patient disease-free.
In comparison, teleophthalmologists demonstrated a higher rate of false positives when referring patients for further diagnosis. This means that AI not only provides a more accurate initial diagnosis but also reduces the chances of unnecessary referrals, which can be costly and stressful for patients.
To validate the AI system’s accuracy, the hospital reinvestigated patients by performing further diagnostic tests at the hospital. Patients diagnosed with glaucoma were invited to the hospital for follow-up investigations, while those identified as having normal vision were visited at their respective centers. The results of these follow-up investigations confirmed the AI’s diagnostic precision.
Dr. Venkatesh emphasized the importance of the AI’s offline functionality, stating, “Operating offline and on-the-edge, the AI system offers a cost-effective solution for glaucoma screening, particularly in areas with limited internet access.” He added that this technology has the potential to significantly improve early detection of glaucoma, reduce healthcare costs, and enhance overall eye care in underserved communities.
Swathi Upadhyaya, the chief author of the study titled Diagnostic Performance of the Offline Medios Artificial Intelligence (AI) for Glaucoma Detection in a Rural Tele-ophthalmology Setting, published in Ophthalmology Glaucoma, also praised the breakthrough. Dr. Venkatesh served as the mentoring author of the study.
Encouraged by the success of this study, Aravind Eye Hospital plans to deploy the AI system across all 13 of its vision centers in the region, aiming to revolutionize glaucoma screening and improve patient outcomes.
Reference:
1. Activity report. Accessed September 23, 2024. https://aravind.org/wp-content/uploads/2024/08/Activity-Report-2023_2024.pdf.
(Input from various sources)
(Rehash/Ankur Deka/MSM)