Recently, researchers have developed AI-powered tool that can predict long-term mortality from Chest X-ray analysis. Read on to know more about it…
Recently, researchers from Massachusetts General Hospital have developed new Artificial Intelligence (AI) powered tool that analyses chest X-rays for prognostic information on long-term health and mortality. Previously, AI was leveraged for automated diagnosis of pneumonia and tuberculosis using chest X-rays. The radiographs, used since the 19th century to detect specific abnormalities, could soon be repurposed to identify long-term mortality risk with a little help from Artificial Intelligence and Machine Learning (ML).
Research Findings
The research findings of this study, published in the journal JAMA Network Open, could help to identify patients most likely to benefit from screening and preventive medicine for heart disease, lung cancer and other conditions. In the latest research, scientists designed a convolutional neural network called CXR-risk to examine visual information. Using data from two large randomized trials, researchers have developed a convolutional neural network, called CXR-risk, that stratifies participants by all-cause mortality risk.
Michael Lu, from Massachusetts General Hospital (MGH) of Harvard Medical School and his colleagues developed a convolutional neural network – an AI tool for analysing visual information – called CXR-risk. The research team paired each image with data about the person’s survival over a 12-year period. This was intended to enable the AI technology to identify chest X-ray features that best predict health and mortality. The aim was for CXR-risk to learn the features or combinations of features on a chest X-ray image that best predict health and mortality. Michael Lu hopes that scores calculated using AI may incentivize high-risk individuals to lower their chance of dying with prevention, regular screening, and lifestyle modification.
According to the study findings, the neural network provided information predicting long-term mortality, independent of radiologists’ readings and other factors such as age and smoking status. A combination of the new technology with other risk factors, including genetics and smoking status, is expected to offer more accurate predictions, allowing earlier diagnosis, prevention and treatments.
The study found that CXR-risk provided information that predicts long-term mortality, independent of radiologists’ readings of the x-rays and other factors, such as age and smoking status.
Clinical Statistics
The convolutional neural network, CXR-risk, was trained using more than 85,000 chest X-rays from 42,000 clinical trial subjects. The research study involving chest X-rays of 16,000 participants from two prior trials showed that 53% (250 deaths among 472 individuals) of the people identified as ‘very high risk’ by CXR-risk died over 12.2 years follow-up in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening (PLCO) Trial; it was 33.9% (61/180) at 6.3 years in the National Lung Screening Trial (NLST). In cases of individuals identified as ‘very low risk’, less than 4% went on to die, said the team.
By score, PLCO mortality rates were 3.8% (97 deaths among 2543 participants) in the very-low-risk group, 7.8% (216 of 2769) in the low-risk group, 12.7% (339 of 2674) in the moderate-risk group, and 24.9% (500 of 2006) in the high-risk group.
