This new technology can help predict if you're going to die

wo of the images of the proximal left anterior descending coronary  artery used in the study. The mortality cases (left side) display  prominent visual changes of emphysema, cardiomegaly, vascular disease  and osteopaenia, while the survival cases on the right appear visually  less diseased and frail. (Credit: University of Adelaide)       

Want to know if you’ll be dead in five years? Just let a computer look at your organs.  New research has indicated that “future” predicting  computers could be coming to hospitals in the near future. Researchers  are hoping that the technology could be used to predict serious  illnesses and medical conditions such as heart attacks. For the study, five year–old medical images of 48  patient’s chests were analyzed by artificial intelligence. From these  images alone, the system was able to predict (with 69 percent accuracy)  whether or not a patient would die within five years. It was also able  to predict medical outcomes by analyzing large volumes of data and  discovering subtle patterns. This new exam–by–AI has proven to be more  effective than a physical exam from a doctor, though apparently it’s  really not a fair fight. MICROSOFT, OTHERS SPEARHEADING A 'SMART CITIES' INITIATIVE FOR EVERYONE “Human doctors are not trained to predict mortality,  so the comparison is a bit unfair,” study leader Dr. Luke Oakden–Rayner  of the University of Adelaide told Fox News. Oakden–Rayner added that  previous research using clincial data such as age, sex or physical  fitness had between 65 percent and 75 percent accuracy, so the new study  "compare[s] favorably, especially considering we excluded factors like  age and sex from our analysis." Currently the system is only trained to predict death  within five years, although, according to Oakden–Rayner, with the right  dataset it should be trivial to extend the technique to other time  scales.  To predict mortality at ten years, for example, the  system would need to analyze CT scans performed over ten years ago so  that Oakden–Rayner’s team could have the follow up results. The  researchers also couldn’t tell what exactly it was the computers were  seeing in the images to make their assessment, though they did find a  strong relationship between the prediction of mortality and the presence  of visible illnesses such as emphysema and congestive heart failure.  Oakden–Rayner explained that certain techniques could  be applied to visualize how the computers were seeing the scans, but  the study was too small to use them this go–round. “You can identify the regions of the images that  contributed to the prediction, and you can ‘hallucinate’ images that  exaggerate the features that the system uses (generating exemplars of  "survival" and "mortality" scans),” he said. “We couldn't apply these  techniques effectively due to our small dataset, but are currently  applying them to a much larger group consisting of tens of thousands of  patients.” ZAPPING YOUR BRAIN WITH ELECTRICITY CAN AID CREATIVITY -- BUT THERE'S A CATCH For this next round of testing, Oakden–Rayner and his  team will be incorporating highly predictive clinical information like  age and sex into their models, which they expect will improve prediction  accuracy. Similar medical AI news has been cropping up lately: a  startup in China revealed an AI system that can help doctors identify  lung cancer by examining CT scans, and IBM now has AI in hospitals  (called Watson) that can answers patient questions. However, if  perfected, this new medical AI could be the most exciting development  yet. “We will start looking at predicting other medical  events before they happen, like strokes, cancer and heart attacks,”  Oakden–Rayner said. The study can be found in "Scientific Reports."