Case Study: Lund University achieves up to 10% higher five-year heart-transplant survival with MathWorks

Lund University Develops an Artificial Neural Network for Matching Heart Transplant Donors with Recipients

Lund University, together with Skåne University Hospital, faced a clinical challenge: improve long-term survival for heart-transplant recipients by better matching donors and recipients across dozens of nonlinear risk factors. The team turned to MathWorks tools—MATLAB, Deep Learning Toolbox, Parallel Computing Toolbox, and MATLAB Parallel Server—to replace slower, unstable open-source software that required weeks of computation and frequently crashed.

Using MathWorks software, researchers built and trained artificial neural networks, distributed experiments across a 56‑processor cluster, and ran Monte Carlo and simulated‑annealing analyses to rank 57 risk factors and automate donor–recipient matching. MathWorks-enabled workflows cut training time by more than two-thirds (from 30–60 minutes to 5–10 minutes), reduced simulation timelines from weeks to days, produced reliable results with no crashes, and in simulated trials would transplant ~20% more patients with a prospective five‑year survival advantage of 5–10%.

Lund University

Johan Nilsson,

Associate Professor in the Division of Cardiothoracic Surgery