A second pair of eyes at the front door of oral cancer care.
Clinician-in-the-loop AI triage for oral mucosal lesions — runs in any browser, with NICE NG12 hard rules built in.
Clinician-in-the-loop AI triage for oral mucosal lesions — runs in any browser, with NICE NG12 hard rules built in.
Over 7,500 new oral cancers are diagnosed in the UK each year, and stage at presentation has barely moved in two decades. Most cases pass through a GDP or GP first. The diagnosis is rarely the hard part — the triage is.
OralPath sits at the front door. It does not diagnose. It surfaces risk, applies NICE NG12 hard rules, and forces an escalation pathway when red flags fire.
A clinician walks through a 7-step capture: demographics, history, graded risk factors (pack-years, alcohol, betel, HPV), trauma context, anatomical site picker, photograph.
A purpose-built vision system analyses the lesion against a curated, clinician-labelled reference library, combined with a graded clinical risk layer. Inference runs on-device — no images leave the room.
Any 2WW-eligible scenario forces a Refer 2WW output regardless of the model score. Belt and braces — a neuro-symbolic safety layer above the model.
Risk score 0–100, a pathway recommendation (Refer 2WW / Refer routine / Watchful waiting / Reassure), and pre-filled 2WW form fields.
OralPath does not replace the clinician. It does not see the histology. It does not touch the EHR. It is a structured second opinion that surfaces risk earlier than a 10-minute appointment can.
Trust R&D leads, primary care networks and digital-first dental groups: book a 30-minute demo and see the live tool.