APOC - Automated machine learning based perioperative monitoring of cardiac function

Aims of this project are first, to develop and validate a robust and feasible, non-harming, ultrasound-based application for automated monitoring of the LV utilizing artificial intelligence (AI). Secondly, we will test this application for added clinical value as compared to standard care during major surgery, and in critically ill patients.

Bildet viser stegene fra hvordan en datamaskin tar ut informasjon under en operasjon til hvordan legene kan lese av resultatet.
​Monitoring of vital signs is mandatory during surgery, in the post-operative state, and in unstable hemodynamic situations.  Conventionally vital signs monitoring comprises measurement of blood pressures, heart rhythm and arterial O2 saturation. These parameters do only indirectly represent the most important factor for hemodynamic stability, i.e. left ventricular (LV) function. Direct visualization of the LV by trans-oesophageal echocardiography (TOE) has therefore emerged as valuable method, but depends on subjective image interpretation of ultrasound images deviating the physician’s attention from other important tasks in the acute or operative setting. Recently, miniature TOE probes designed for long-term (days) use, have been developed. 

Aims of this project are first, to develop and validate a robust and feasible, non-harming, ultrasound-based application for automated monitoring of the LV utilizing artificial intelligence (AI). Secondly, we will test this application for added clinical value as compared to standard care during major surgery, and in critically ill patients. We have demonstrated the feasibility of AI combined with TOE for automated measurement of mitral plane systolic excursion (MAPSE) and regional LV systolic function by strain-measurements. 

Masters- eller PhD-kandidater kan ta direkte kontakt med prosjektlederen om forskningsmuligheter.

Gabriel Hanssen Kiss

E-post: Gabriel.Kiss@stolav.no

Telefon: 918 97 945

Sist oppdatert 01.08.2022