Improving diagnostic and therapeutic outcomes with AI
As the volume of clinical data increases, so does the amount and complexity of the work needed for a doctor to interpret them. Magnus Boman researches on how AI can be employed for healthcare-improving analyses. Meet one of Karolinska Institutet's new professors who will participate in this year's installation ceremony at Aula Medica on 3 October.
Text: Karin Tideström, for KI’s installation ceremony booklet 2024
What are you researching?
“I want to contribute to patient benefit and better healthcare by looking at how learning algorithms, AI and machine learning can support decision-making and data-driven analyses. In purely practical terms, I use AI methods to translate data from different sensors or texts into clinically useful information. The methods sift out medically interesting data from the noise. AI can also be used for tasks normally done by people, such as segmenting organs in medical images.”
Why is this important?
“AI can help to improve diagnostics and treatment. The workload of a radiologist is getting heavier and heavier both in terms of the number of images they study every day and how complex and difficult to interpret they are. Given the incredible amount of data that modern imaging techniques collect, the complexity of epidemiological studies with different types of data, collected with or without a particular purpose, is growing. Learning algorithms can also plough through 21 million PubMed articles and quickly prepare an oncologist for a tumour board by producing relevant background material.”
How do you hope that your results will be used?
“This question has a brief answer. I want my research to be used clinically so that it can give quantifiable patient benefit. Clinicians are always coming to me with a problem, wondering if machine learning can be used. I don’t ever formulate any clinical problems myself. When a problem comes in, I can usually see if it will be easy or hard to solve and if AI can be of use. My colleagues and I then model a solution.”
About Magnus Boman
Professor of Artificial Intelligence (AI) in Health at the Department of Medicine, Solna
Magnus Boman was born in 1963 in Stockholm. He took his PhD in 1993 in computer and systems sciences at Stockholm University, shortly after which he entered a collaboration with Karolinska Institutet in which he helped to establish the multidisciplinary Stockholm Group for Epidemic Modeling (S-GEM). He is also an honorary professor of psychiatry at University College London. In 2020 has was made AI consultant to the Karolinska Institutet president. Magnus Boman was appointed Professor of Artificial Intelligence (AI) in Health at Karolinska Institutet on 1 January 2024.