Doctors Turn To AI For Faster, More Accurate Stroke Diagnosis Amid Delay In Patient Arrival

When blood supply to the brain is disrupted during a stroke—a life-threatening medical emergency—every passing minute becomes crucial. In a significant step towards strengthening emergency care, hospitals in India are increasingly adopting advanced artificial intelligence (AI)-based imaging software such as RapidAI to enable faster, more precise diagnosis and treatment decisions in stroke management.

The initiative reflects a growing global shift—seen in many hospitals in developed countries—where AI tools are being deployed to reduce delays and improve outcomes in stroke care.

“One in four people is likely to suffer a stroke during their lifetime, underscoring the scale of the crisis. However, stroke is not a condition that gives you time. Every minute, millions of brain cells die,” said Dr. Sumit Goyal, Director and Head of Neurosurgery and Group Director of Neurointervention.

“Unfortunately, most patients reach us after 10 to 12 hours, when a significant portion of the brain is already damaged.”

He said that the latest integration of RapidAI at their hospital marks a shift in how stroke is assessed and treated. Instead of merely confirming the diagnosis, doctors are now able to determine how much of the brain has already suffered irreversible damage and how much can still be salvaged—a factor that directly guides treatment decisions.

“When a patient reaches us, we assess not just whether a stroke has occurred, but the extent of damage and the amount of salvageable brain tissue. That is the most critical factor in deciding treatment,” Dr. Goyal said.

Widely used in developed countries for over a decade, RapidAI automates the analysis of CT perfusion and MRI scans, providing quantitative insights into blood flow in the brain. Within minutes of completing the scan, doctors receive detailed results on their mobile devices.

“Within about three minutes of completing the scan, we receive a comprehensive analysis,” Dr. Goyal explained. “It tells us exactly how many millilitres of brain tissue are already lost and how much can still be saved. Earlier, we relied largely on estimation. Now, we have objective data.”

The software works by tracking the flow of blood mixed with contrast through different regions of the brain. It identifies areas with reduced or delayed circulation, allowing clinicians to distinguish between permanently damaged tissue and the surrounding “at-risk” zone that may still recover with timely intervention.

This level of precision has been central to several international clinical trials that have reshaped stroke treatment protocols. These studies have shown that, in carefully selected patients, treatment can be effective even between 6 and 24 hours after symptom onset—extending beyond the earlier standard window of 4.5 to 6 hours.

“Many of the global guidelines we follow today are based on studies that used this software. That is why its credibility is very high,” Dr. Goyal noted.

Importantly, the technology also helps avoid unnecessary procedures. “In some cases, the scan shows that no brain tissue can be salvaged. In such situations, invasive interventions will not benefit the patient and may only increase complications and cost,” he said.

Another advantage is the reduction of variability in interpretation. Early stroke changes on CT scans are often subtle and may be read differently by different radiologists. AI-based analysis standardises this process, ensuring faster and more consistent results.

However, experts emphasise that AI remains a support tool rather than a replacement for clinical judgement. “All reports are cross-checked by radiologists, and final decisions are always clinical,” said Dr. Rahul Sharma, senior consultant in neurosurgery at Yatharth Hospital.

Dr. Vikas Gupta, Senior Consultant Neurosurgeon at Kailash Deepak Multispecialty Hospital, Delhi NCR, agreed, saying that artificial intelligence is a support system, not a substitute for clinical judgement. “AI doesn’t replace doctors; it supports them. AI-assisted tools enhance precision during both diagnosis and intervention,” he said.

AI tools also assist in planning procedures such as mechanical thrombectomy by helping doctors determine the most effective approach for clot removal. During the intervention, advanced imaging systems provide real-time visualisation, guiding catheters and enabling clinicians to monitor the restoration of blood flow.

Despite these advances, the biggest challenge remains delayed recognition of symptoms. Stroke typically presents suddenly—with slurred speech, weakness on one side of the body, facial drooping, or loss of coordination. Yet, these warning signs are often ignored or misattributed, particularly among elderly patients.

Estimates suggest that nearly 90% of stroke patients in India do not reach hospitals within the optimal treatment window. “This is evident in the fact that, in the entire month of March, we did not receive a single patient within the ideal window,” he added, matter-of-factly.

With stroke emerging as a major public health concern—driven by rising risk factors such as diabetes, alcohol consumption, and sedentary lifestyles—experts stress that awareness, speed, and timely intervention remain critical.

“Time is brain,” emphasised Dr. Akshya Shirodhar, consultant in neurosurgery at the hospital. “Recognising symptoms early and reaching the hospital immediately can make the difference between recovery and lifelong disability.”

With approximately 1.5 to 1.8 million new cases annually, the stroke burden in India is increasing, often affecting younger individuals, with 15–20% of cases occurring under 45 years.