Computer Analysis May Predict Mortality
According to a study conducted at the School of Public Health at the University of Adelaide in Australia, advances in precision medicine—defined by the National Institutes of Health as "treatment that takes into account individual variability in genes, environment, and lifestyle"—could predict patient 5-year mortality with up to 70 percent accuracy.
For the study, researchers gathered more than 15,000 CT scan images of internal organs and tissues—including the heart and lungs—from patients over the age of 60 and identified imaging features linked to 5-year mortality. Then they programmed this information into a computer, “teaching” the machine to analyze and understand the CT scan images and predict outcomes.
When researchers used the computer to analyze chest CT scans of 48 additional patients, it predicted 5-year mortality at an accuracy rate (69 percent) comparable to that of health care professionals. According to Dr. Luke Oakden-Rayner, lead study author, more research—involving tens of thousands of patients—is the next step.