Rian Sanghi, a Grade 12 student at Scottish High International School, Gurgaon, has developed OroScan, an AI-powered prototype designed to help screen images of the mouth for signs that may require further medical examination. As someone who thrives on engineering and loves experimenting with AI and technology, he has always been drawn to ideating concepts where practical problems meet engineering solutions. The idea for OroScan came to him when he noticed a gap between oral cancer testing and accessibility. Oral cancer is treatable when caught early, but in a lot of communities, routine oral examinations simply don’t take place. By the time something is detected, it’s often escalated well past the point where early intervention and management could have made the biggest difference. The outcome ends up far from what it could have been, had it been caught sooner. That gap became the focal point for this device. From the get-go, Rian was careful about how he framed OroScan. It was not a replacement for a diagnosis by a medical professional, but rather a preliminary screening and referral tool. The model in no way supersedes a medical professional’s opinion, but is simply trying to be the reason someone goes to see one sooner than they otherwise would have. It provides something very useful: A first opinion. What OroScan Actually Does At its core, OroScan pairs an endoscopy camera with a custom-built PCB and a CNN-based AI model trained to analyse images from inside the mouth. The model classifies each image as suspicious or clear. Rather than giving a simple verdict, it generates a heatmap showing exactly which region of the image influenced its assessment; this ensures the result isn’t a black box. It also has an onboard display, so you can view the image and result directly on the device. This simplifies the process and accessibility without needing to plug it into a laptop or send data elsewhere. Getting the Model Right Getting a model that was both sensitive and precise took a lot of iteration. OroScan recorded 96% accuracy, 98% recall, and 94% precision on a held-out test dataset. Within these testables, recall mattered most to the student. He explained further that in a screening tool, missing a genuine case is a far more serious failure than flagging something that turns out to be clear. He continued, saying he’d rather send someone for an additional follow-up than let something slip through. But a good model on a laptop isn’t a usable device. So the project moved beyond a software demonstration to a working prototype. A 3D-printed enclosure, the custom PCB, the endoscopy camera, and the onboard screen, all built to function as one self-contained unit that could realistically be carried into a screening camp. Taking It Beyond the Lab OroScan was shortlisted for Samsung Solve for Tomorrow, received a Gold award at the International Greenwich Olympiad, was presented at IIT Delhi, and has been acknowledged by the Ministry of Health and Family Welfare, Government of India. The recognition was overwhelming for this young student, but his real test came when the device reached people outside a lab. That’s why running a screening camp with TCTF was so integral. It moved OroScan out of a controlled testing environment and into a community setting, where it had to hold up under the realities it aimed to work with. Through this initiative and outreach, OroScan reached more than 1,500 people, with this number still growing. What Comes Next As of now, Rian is working to expand the programme through partnerships with additional NGOs and community organisations. He is also running more camps, gathering feedback directly from healthcare professionals, and refining the device to make it as user-friendly and accurate as possible.
Ultimately, he doesn’t envision OroScan as a replacement for diagnosis in the existing system. But Rian hopes it can help bridge the gap that exists by providing an early first opinion.