MathWorks today announced that Lund University, one of the world’s leading research universities, used MATLAB, Neural Network Toolbox, Parallel Computing Toolbox, and MATLAB Distributed Computing Server to improve long-term survival rates for heart-transplant recipients by identifying optimal recipient and donor matches.
Researchers at Lund University and Skåne University Hospital explored the complex relationships among multiple transplant variables, including the weight, gender, age, and blood type of both donor and recipient, and the time during a transplant when there is no blood flow to the heart. Analyzing the six variables requires the simulation of 30,000 different combinations, and simulating all these combinations for 50,000 patients took weeks using an open-source software package that proved to be unstable and inaccurate.
To address the speed and reliability challenges, the researchers employed MATLAB and Neural Network Toolbox to develop predictive artificial neural network (ANN) models. The ANN models were built with donor and recipient data from two global databases: the International Society for Heart and Lung Transplantation (ISHLT) registry and the Nordic Thoracic Transplantation Database (NTTD).
Lund researchers used Parallel Computing Toolbox to program parallel applications and MATLAB Distributed Computing Server to scale those applications to a cluster to accelerate the simulation of more than 200,000 ANN configurations. They then evaluated the results to find the best-performing configuration. The models showed that the prospective five-year survival rate for the ANN-selected patients was 5–10% higher than those matched with the criteria physicians use today.
“Many of the techniques we use are computer-intensive and time-consuming,” said Dr. Johan Nilsson, Associate Professor in the Department of Cardiothoracic Surgery at Lund University. “Working with MathWorks tools, we completed experiments that regularly took three to four weeks in about five days. Being able to access and analyze tremendous amounts of data at a fast pace helped us build and use our research models quickly.”
“The gains realized by Lund University are a prime example of how high-performance computing enables teams to develop more reliable, complex models in less time,” said Silvina Grad-Freilich, senior manager, parallel computing marketing at MathWorks. “Engineers and scientists want to solve their problems faster, and over the past decade, the ability to available hardware effectively has been a barrier to their efforts. With tools such as Parallel Computing Toolbox and MATLAB Distributed Computing Server, MathWorks is addressing this obstacle.”