LeaRes turns a chest CT into a patient-specific functional twin of the lungs: anatomy reconstructed, ventilation and airflow simulated, in the browser, on your hospital’s GPU.
All MedLea solutions are browser-accessible and comply with GDPR · HIPAA · SOC2 · EU AI Act.
LeaRes is CE-marked medical device software (class I) for study, simulation and the training of healthcare professionals. It does not provide a diagnosis: the judgement stays with the clinician.
Lobes and airway tree reconstructed from the CT by AI segmentation, then set in motion: airflow simulated in every airway, regional ventilation in every lobe, on the patient’s own anatomy.
The breathing twin yields regional biomechanics: where the lung inflates, where it over-distends, how the two lungs share the breath. Simulated maps you can read on the familiar MPR views, for research and training.
In development at Medlea, not a function of the CE-marked device: pulmonary arteries and veins, heart and aorta reconstructed alongside the lungs, as the geometry a flow study needs for perfusion research.
Engineering · research & development
The engineering workspace takes the reconstructed anatomy, or an imported CAD geometry, sets materials and boundary conditions, and runs the flow. The full solver for the final answer; surrogates for the exploration.
A lattice-Boltzmann multiphysics solver on GPU computes airflow, pressure and particle transport in the patient’s own airways and vessels.
AI surrogates trained on full CFD runs return the same quantities in a fraction of the time, so settings and device variants can be explored interactively before the full solver confirms the final case.
Catheters, stents, flow diverters and ventilator circuits placed in the reconstructed anatomy or imported as CAD, with inlets, outlets and materials set in the browser.
Active Learning · pilot programme
Name the pattern, annotate inside LeaRes, train on your GPU, correct what the model proposes. Each round leaves less to correct. The data never leaves the hospital.
How the platform worksAn agentic workflow that takes a scan through segmentation, reconstruction and simulation.
LeaFlow is Medlea’s research and engineering workflow. AI agents carry a CT or MRI scan through segmentation, anatomical reconstruction and simulation set-up, step by step. At its core sits a multiphysics simulation engine: it models how blood, air, urine, drugs, contrast agents and particles move through an anatomy, and how a medical device — catheter, implant, inhaler, stent — changes that flow, so a device maker can test a prototype in a virtual anatomy before building it. LeaFlow is not intended for clinical decisions on individual patients.
Aerosol therapy, nebulisers and smart inhalers: LeaFlow injects the spray of an inhaler and follows the particles through the airways. A regional deposition model is in development.
Stents, grafts and valves: LeaFlow measures their effect on local flow, wall shear stress and downstream perfusion, inside the vessels of the individual patient.
Ureteral stents, catheters and artificial sphincters: LeaFlow shows how they change urinary flow, pressure gradients and the transport of sediments and particles.
CT and MRI scans become accurate 3D models: airways, vessels, renal tracts and other flow-relevant anatomy are identified and reconstructed automatically, case by case.
Patient models move in and out of standard CAD tools. Engineers place a stent, valve or catheter in a reconstructed anatomy and test the design.
LeaFlow runs in the cloud or on premises, on GPUs, for real-time or batch simulations, and every step is driven from an ordinary web browser.
