Nick MaxfieldThe van der Schaar Lab’s eleventh Revolutionizing Healthcare engagement session for clinicians took place virtually on October 26, 2021.
In this session, Mihaela van der Schaar and a panel of 3 international clinicians expanded on the agenda outlined in the previous session: the importance of bridging the gap between ideas and implementation in machine learning for healthcare.
Our panel for this session consisted of:
- George Mells, MD PhD (MRC Clinical Academic Research Partner, Academic Department of Medical Genetics, University of Cambridge)
- Pierre Marquet, MD PhD (Head of Pharmacology-toxicology Department at Limoges University Hospital Team lead, Pharmacology & Transplantation, University of Limoges and Inserm)
- Vincent Gnanapragasam, MD PhD (Professor of urology, University of Cambridge; Honorary consultant, Cambridge University Hospitals NHS Foundation Trust)
In the first part of the session, Mihaela and the clinicians showed how machine learning can assist dynamic time-to-event analysis and temporal phenotyping, using a working demonstrator to show real-world case studies from prostate cancer. The latter part of the session featured a clinician roundtable, during which the panelists discussed different approaches to ensuring that clinicians can make full use of machine learning models.
Introduction – 0:00
Declaration of interests – 0:47
Session overview – 1:58
Meet the roundtable panelists – 5:05
Presentation by Mihaela: “Getting ML-powered tools into the hands of clinicians (part 2)” – 5:55
Vincent Gnanapragasam: time-to-event analysis and temporal phenotyping – 21:04
Vincent Gnanapragasam: prostate cancer studies shown in an ML-powered demonstrator – 38:25
Evgeny Saveliev: building robust ML tools for healthcare – 47:23
Roundtable topic 1: the current state of affairs in phenotyping – 51:04
Roundtable topic 2: how ML for phenotyping can help serve clinicians – 53:49
Audience question – 1:06:10
Intro to next sessions and note on CPD credits – 1:09:33
NOTE: This information was up-to-date at the time of the presentation but does not take into account material published since then.
Sign up for our upcoming sessions here.
Andreas Bedorf
Tim Schubert
Tim Oosterlinck
