Machine Learning Estimates Risk of Cardiovascular Death

first_img Machine Learning Translates Japanese Retro Games In Real TimeRobot Dog Astro Can Sit, Lie Down, and Save Lives Stay on target MIT researchers have developed a machine learning model that can estimate, based on the electrical activity of their heart, a patient’s risk of cardiovascular death.The system, dubbed “RiskCardio,” was created by the Computer Science and Artificial Intelligence Laboratory (CSAIL) with the intention of better predicting health outcomes.It focuses on folks who have survived acute coronary syndrome (ACS)—a range of conditions involving decreased flow to the heart. Just 15 minutes of a patient’s raw electrocardiogram (ECG) signal can produce a score that places people into different risk categories.“We’re looking at the data problem of how we can incorporate very long time series into risk scores, and the clinical problem of how we can help doctors identify patients at high risk after an acute coronary event,” lead study author Divya Shanmugam said in a statement.“The intersection of machine learning and healthcare is replete with combinations like this—a compelling computer science problem with potential real-world impact,” she added.This isn’t the first machine-learning attempt at risk metrics; previous experiments have used external patient information like age or weight, or knowledge and expertise specific to the system.RiskCardio, however, relies solely on patients’ raw ECG signal—with no added information.Say someone checks into the hospital after an ACS: A physician would normally estimate risk of cardiovascular death or heart attack using medical data and lengthy tests before choosing a course of treatment.MIT’s invention aims to improve that first step.The team tested their model using data from a study of past patients, separating each person’s signal into a collection of adjacent heartbeats. They then assigned a label—”risky” for those who died of cardiovascular complications and “normal” for those who survived—to each set of beats.When demonstrated on a new patient, CSAIL analysts were able to estimate whether someone would suffer from cardiovascular death within 30, 60, 90, or 365 days.Moving forward, the team hopes to make the dataset more inclusive to account for different ages, ethnicities, and genders, and evaluate how their system accounts for ambiguous cases.“Machine learning is particularly good at identifying patterns, which is deeply relevant to assessing patient risk,” Shanmugam said. “Risk scores are useful for communicating patient state, which is valuable for making efficient care decisions.”More on Scientists 3D Print Miniature Human HeartMIT’s Color-Changing Ink Lets You Customize Shoes, Phone CasesSlug-Inspired Slime Could Fix Your Broken Heartlast_img