Professor Rayaz Malik, a British-Pakistani scientist leading research in diabetic neuropathy and neurodegenerative diseases at Weill Cornell Medicine-Qatar, has developed an artificial intelligence-powered eye scan capable of detecting dementia and diabetic nerve damage years before any clinical symptoms appear. The scan takes just two to three minutes to complete and uses a technology called corneal confocal microscopy, a tool traditionally used by ophthalmologists to diagnose surface-level eye infections, but which Professor Malik and his colleagues have repurposed to detect microscopic nerve fibre damage linked to conditions far beyond the eye itself. The technology is being described as a potential turning point in how neurodegenerative diseases are identified and managed, shifting the diagnostic window from late-stage confirmation to early intervention at a point when treatment can still make a meaningful difference.
The scientific basis of the technology lies in the cornea’s unique biological properties. The cornea contains the densest concentration of sensory nerves anywhere in the human body, a feature that evolved to protect vision, and which also makes it a remarkably sensitive indicator of what is happening in the nervous system more broadly. When a patient presents with memory loss and receives a dementia diagnosis, the underlying nerve damage has typically been developing for ten to fifteen years, by which point standard treatments have limited effectiveness and MRI scans may only then be showing abnormalities. Corneal confocal microscopy, however, is able to identify abnormal corneal nerves during the mild cognitive impairment stage, which precedes a full dementia diagnosis by as many as five years in many cases. For diabetic neuropathy, the detection advantage is equally significant, with corneal confocal microscopy able to identify nerve damage up to five years before conventional diagnostic methods would flag a problem, giving patients and clinicians a window to initiate nerve repair through weight loss, blood glucose control, lipid management, and blood pressure reduction.
Artificial intelligence has transformed what was already a promising diagnostic tool into one of considerably greater clinical power. Where a human clinician reviewing a corneal nerve image might identify three or four distinguishing features, artificial intelligence systems can analyse more than 2,500 features within the same image and identify patterns associated with specific diseases in seconds rather than the extended periods previously required for manual analysis. The result is a diagnostic accuracy of 90 to 95 percent for neurodegenerative diseases, with studies involving diabetic neuropathy and Parkinson’s disease achieving sensitivity and specificity approaching 100 percent. The technology faced initial resistance from neurologists sceptical that an endocrinologist could use an eye-based scan to diagnose neurological conditions, and for years was also constrained by the fact that only a single German manufacturer produced the required corneal confocal microscopy machines. The entry of a Chinese manufacturer into the market has significantly improved equipment availability and affordability, opening the door to wider adoption across the developing world, including in Pakistan, where rising diabetes rates make early neuropathy detection an urgent public health priority.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
