A phone camera, twelve questions, and thirty seconds.

Oral cancer is survivable caught early and lethal caught late. Most of the people it kills live where the specialist equipment does not. OncoScan needs neither the equipment nor the specialist.

92%Detection accuracy

<30sPhoto to score

12Risk markers

0Extra hardware

With these twelve and one photo of your mouth, we can tell you whether to go see someone.

Tobacco use+

Smoking+

Alcohol use+

Betel quid+

HPV infection+

Chronic sun exposure+

Poor oral hygiene+

Low fruit and vegetable+

Family history+

Compromised immunity+

White or red patches+

Unexplained bleeding+

STANDBY

[ ]Intraoral capture0.0s

[ ]Field segmentation0.0s

[ ]CNN · 224×2240.0s

[ ]Clinical history model0.0s

[ ]Probability fusion0.0s

Awaiting scan

Screening a population, or building one of these?

Reach us →

Oral cancer diagnosis for everywhere a doctor isn't.

Vihaan Shringi

Founder

I built and patented OncoScan at seventeen, having decided the barrier to early oral cancer detection was the equipment budget rather than the science.

I am eighteen and studying computer science at Carnegie Mellon. Nine months before that I was an ML research intern at the Institute for Systems Biology, working on preeclampsia and HIV diagnostics.

Clinical Director

Oncologist

Reviews lesion criteria and the referral threshold above which a score becomes an appointment.

Clinical Director

Oncologist

Reviews the twelve-marker set and how each is weighted against the others.