Wednesday

Entering the era of LSST and SKA, astronomers are at the forefront of the "Big Data" revolution, and while many of our analysis techniques may seem astrophysically focussed, their applications can be far-reaching. Recently, the UK Government outlined a grand challenge as part of its Industrial Strategy to revolutionise the NHS. Their vision will use “AI and big data” to meet the needs of an “ageing society”. The role of AI and big data analysis in healthcare is clear: it has the potential to provide more accurate decisions faster, reducing cost and improving patient outcomes. Motivated by this challenge, and recognising the skill set of astronomers, we have developed a collaboration with the School of Cardiovascular Medicine & Science at King’s College Hospital (KCH). Out of Hospital Cardiac Arrest (OOHCA) occurs in over a quarter of a million patients each year and presents a major public health challenge. The rate of coronary artery disease in patients with OOHCA is high, but the current technique to guide the provision of an emergency coronary angiography remains a poor predictor of a culprit coronary artery lesion. Using Random Forest and Boosted Decision Trees, as currently deployed in many astronomical classification tasks, we demonstrate significantly improved predictions of a culprit coronary artery lesion in a large sample of OOHCA patients. In this talk, I will present both the results from our study and areas of cardiology in which astronomers may be able to contribute.