Mechanistic core
Models built from known biology
Simulations of physiology and the immune response, written from the underlying mechanisms rather than learned only from data.
Medical digital twins for critical care
A digital twin for every patient.
Metis Biotech builds digital twins of critically ill patients. We combine mechanistic models of human physiology and immunology with AI, so each twin is grounded in biology and calibrated to the individual patient.
The problem
Two ICU patients with the same diagnosis can need very different care.
Critical illnesses such as sepsis involve immune responses that are complex, shift over hours, and vary widely from person to person. A treatment that helps one patient can do nothing for the next, or cause harm.
Clinicians make these decisions under time pressure, with partial information, guided by results averaged across populations. The patient in the bed is not an average.
Our approach
Biology first, then AI.
Our twins start from mechanistic models of how the body and immune system work, then use AI to fit those models to each patient.
AI augmentation
Calibrated to the individual
Machine learning tunes each model to a patient's own measurements and explores the range of ways that patient might respond.
Living twin
Updated as care unfolds
As new vitals and labs arrive, the twin updates, so it keeps reflecting the patient as their condition changes.
What twins make possible
Test the plan on the twin before it reaches the patient.
- Compare treatment strategies in simulation and see how a specific patient is likely to respond.
- Anticipate how a patient's course may evolve, so the care team can act earlier.
- Design clinical trials around the patients most likely to benefit, using simulated cohorts.