OncoScan uses intraoral imaging and clinical risk data to deliver AI-powered oral cancer screenings in under 30 seconds — accessible to every provider, everywhere.
Three steps from patient to AI-generated risk score.
Capture and upload an intraoral photo. Any smartphone camera works — no specialist equipment needed.
Input patient demographics, tobacco/smoking history, and up to 12 clinical risk markers from a quick guided form.
AI fuses both models and returns a color-coded risk percentage in under 30 seconds — Low, Moderate, or High.
Two models. One score. Maximum accuracy.
History-Based Risk
Trained on structured clinical data — age, gender, tobacco use, smoking, alcohol, and 12 risk markers — to predict cancer likelihood from patient history alone.
Image-Based Risk
A convolutional neural network analyzes the intraoral image at 224×224 resolution to identify lesion patterns, color abnormalities, and high-risk visual markers.
Both model outputs are averaged into a single fused risk probability — maximizing precision by combining visual and clinical signals.
Models are deployed on a serverless SageMaker endpoint. Results are stored in DynamoDB. Images archived in S3. Zero infrastructure management.
OncoScan was founded with the belief that early cancer detection shouldn't be a privilege. Our tools are built for accessibility — not the hospital boardroom.
Vihaan founded OncoScan to bring AI-powered diagnostics to every corner of the world. His work sits at the intersection of machine learning, health equity, and compassionate care — with a singular focus on saving lives through early detection.
Advised by oncologists and AI researchers at the forefront of cancer detection and machine learning.
Connected with hospitals nationwide to bring AI-powered early detection to patients everywhere.
Supported by ACIC VGU, and AWS Activate — building at the cutting edge of health tech.
Access the screening portal
Combined Risk Score
Oncologist Feedback
Feedback submitted. Thank you.
Record Saved
Patient data submitted successfully.