New PhD-led research University of Western Australia researchers find artificial intelligence using computers or machine learning can help reduce common cardiovascular complications after noncardiac surgery, such as heart attack and myocardial injury I have discovered that it is possible.
Dr. Janice Norde of the UWA School of Medicine and Royal Perth Hospital and an international research team evaluated data from more than 24,000 participants in the Vascular Events Cohort Evaluation (VISION) Study in Non-Cardiac Surgery Patients, the results of which were published in 2016. rice field. anesthesia.
Can researchers use machine learning and data to predict medical complications, particularly cardiovascular complications from surgery (other than cardiac surgery), before they occur, to better identify and treat vulnerable patients? I wanted to establish what
“Each year, more than 200 million patients worldwide undergo major noncardiac surgery, of which about 10 million experience a significant cardiovascular event within 30 days, leading to increased mortality, health It can lead to worsening of the condition and decreased long-term survival,” Dr. Norde said.
“The most common cardiovascular complications after surgery are heart attack and myocardial injury, but they are often difficult to detect because symptoms can be hidden or missed by routine testing.”
Using a sensitive laboratory test that measures a protein (troponin) that is released into the bloodstream when there is damage to the heart muscle, researchers found that one in six patients had levels in the first three days after surgery. was found to rise.
“This condition, known as post-noncardiac surgery myocardial injury, is associated with a significantly increased risk of death and other serious complications in the coming weeks, but the risk of death, age, physical fitness, underlying medical conditions, etc. It is difficult to predict because of the variables in , and we need to consider all the problems that arise during or early after surgery,” Dr. Norde said.
“Machine learning, and neural networks in particular, offer a promising approach because these techniques can analyze large amounts of data and identify complex patterns and relationships that are otherwise difficult to discover. It’s very adaptable, so it can be implemented and adjusted for different settings.”
Graham Hillis, Professor of Medicine at UWA and Head of Cardiology at Royal Perth Hospital, said the results of the study combine machine learning techniques with data collected regularly before, during and after surgery. suggested that this could be a promising method to better distinguish between the most at-risk and at-risk patients. Risk may increase over time.
“This may allow medical professionals to spot problems earlier and intervene earlier to reduce potential complications,” said Professor Hillis. “Further studies are planned to fine-tune these methods and incorporate them into routine care.”
For more information:
JM Nolde et al, Machine learning for predicting myocardial injury and death after non-cardiac surgery, anesthesia (2023). DOI: 10.1111/anae.16024
Magazine information:
anesthesia
