| Hello, readers! This is Sy. Google parent Alphabet’s life sciences arm Verily—the firm running an ambitious study to map a “baseline” of human health—may have an intriguing new way to assess heart disease risk: Through your eyes. As with pretty much all things Verily, the method is centered on machine learning, as detailed in a report published in the journal Nature Biomedical Engineering. The Verily algorithm was trained using a dataset of about 300,000 patients, including eye scans and other medical information. Via neural networks, it then made connections between disparate data points—including between eye scans and the biggest risk factors for cardiovascular disease. (If you’re wondering, why eye scans?, the answer lies in the interior wall of your eye, or the “fundus,” which contains enough blood vessels to draw educated inferences about things like blood pressure.) So how did the program fare? 70% of the time, it was able to correctly identify which of two patients had a “cardiovascular event” like a heart attack within five years of the analyzed eye scan. That’s just a couple of percentage points lower than the accuracy of common current methods that involve blood tests. “Our results indicate that the application of deep learning to retinal fundus images alone can be used to pr edict multiple cardiovascular risk factors, including age, gender, and [blood pressure],” wrote the study authors. “That these risk factors ar e core components used in multiple cardiovascular risk calculators indicates that our model can poten tially predict cardiovascular risk directly.” Still, plenty of more testing will need to be done before Verily can move the algorithm into a clinical setting. Read on for the day’s news. | 
No comments:
Post a Comment