We build a digital twin of a patient’s liver and run it forward in time.

1 · The product

One patient, one liver, followed through time.

The Px Twin workspace follows a patient’s liver through time, and shows where it is heading.

Cholangiography, measured: per-stricture calibre, stenosis grading, tracking over serial studies.

Ten years of one liver: lesions tracked segment by segment as disease progresses.

120 months of trajectories: stricture calibre, hepatic volume, and the prognostic score panel.

Demonstration data shown, not a real patient.

2 · What a forecast unlocks

A forecast changes what a clinician and a drug developer can do.

Drug developers come first: liver-disease trials are long and expensive precisely because nobody can measure progression well. A clinician-facing service follows regulatory approval.

3 · Where we start

We start where a forecast is worth the most.

Our first disease is autoimmune sclerosing cholangitis (ASC), the form of primary sclerosing cholangitis (PSC) that begins in childhood. The bile ducts scar and narrow. No therapy is approved to slow it, and no trial in the disease has an approved surrogate endpoint. Prognosis today rests on scores computed at a single moment, and none of them says where the patient is going.

The work is clinically grounded from the first day: it is advised by a consultant transplant hepatologist at a national liver transplant centre.

4 · Team

The team.

Jay LiuFounder

Twelve years in AI research, with award-winning work in the pharma space. NIHR artificial intelligence grant assessor. Overall project lead: clinical strategy, the regulatory route and commercial relationships.

Adrien VernyFounder & technical lead

Fifteen years of award-winning work across AI research and production. Designed the foundational hepatology AI models commissioned by EASL, the European Association for the Study of the Liver. Manages the engineering and research team.