Artificial intelligence (AI) and machine learning are rapidly entering the healthcare space sparking great debate about their role in medicine. While AI has outperformed physicians at certain tasks in diagnostic fields like radiology and pathology, the greater power of AI is to bring novel tools into medicine that will augment, not replace, the physician. Likewise, systems biology, i.e. computational and mathematical modeling of complex biological systems, has recently been leveraged to analyze complex “omics” data and to simulate physiologic responses to injury and transfusion. With the availability of “big data” and advances in computational power, the field is now poised to transition away from the analysis of a handful of selected variables to modeling of high-dimensional data for the identification of predictive biomarkers and clinical prediction tools. This session will explain the basics of AI, present applications of machine learning in transfusion medicine, and demonstrate how systems biology can improve our understanding of red cell biology and physiologic responses during hemostatic resuscitation in trauma.
In the first session on AI and machine learning, Dr. Michelle Stram will provide an overview of machine learning techniques and will focus on the development and external validation of robust machine learning prediction tools for identifying which civilian trauma patients are most likely to require massive transfusion or significant hemostatic support in the early phases of resuscitation. She will also describe other areas of transfusion medicine that stand to benefit from the introduction of AI techniques.
There has been renewed interest in the use of whole blood during hemostatic resuscitation in trauma, since whole blood has a smaller volume of non-oxygen-carrying, non-hemostatic fluids compared with conventional component therapy with individual units of red cells, plasma and platelets. Dr. Jansen Seheult will present data from an in silico model of blood product resuscitation strategies used to predict their effects on body fluid compartments, important hemostatic factors and tissue oxygenation. He will also explore how in silico clinical trials can guide sample size calculations and selection of outcome measures to inform the design of randomized clinical trials of trauma resuscitation.
The red blood cell is the simplest human cell, making it an accessible starting point for the application of systems biology approaches. Dr. James Yurkovich will discuss how the use of systems biology has led to significant contributions in the identification of three distinct metabolic states that define the baseline decay process of red blood cells during storage and how the analysis of high-dimensional data led to the identification of predictive biomarkers of RBC metabolic health. Finally, he will explore broader applications of systems biology in transfusion medicine.
Explain the basics of artificial intelligence and machine learning.
Identify applications of machine learning in transfusion medicine.
Describe how systems biology can improve our understanding of red cell biology and physiologic responses during hemostatic resuscitation in trauma.