{"id":21493,"date":"2018-12-04T06:30:18","date_gmt":"2018-12-04T14:30:18","guid":{"rendered":"https:\/\/insidebigdata.com\/?p=21493"},"modified":"2018-12-05T08:51:33","modified_gmt":"2018-12-05T16:51:33","slug":"ai-in-medicine-radiology-genomics","status":"publish","type":"post","link":"https:\/\/insidebigdata.com\/2018\/12\/04\/ai-in-medicine-radiology-genomics\/","title":{"rendered":"AI in Medicine: Developments in Radiology, Genomics & Diagnostics"},"content":{"rendered":"

This\u00a0insideBIGDATA special report series<\/a>\u00a0explores the intersection of AI and healthcare. This entry highlights some of the latest developments in AI in medicine, including progress in radiology and genomics.\u00a0<\/em><\/p>\n

\"AI<\/a>

Download the full report.<\/a><\/em><\/p><\/div>\n

Radiology study<\/strong><\/p>\n

Next-generation radiology tools may remove the need for tissue sampling and greatly expand the area of delivery. Lucas A. \u201cLuke\u201d Wilson is a data scientist and artificial intelligence researcher in the Dell EMC HPC and AI engineering group, focusing on developing hardware configurations and software solutions for deep learning problems. He has been involved in the development of models for using AI to diagnose pneumonia, emphysema and other thoracic pathologies from chest X-rays, using the Stanford University CheXNet model.<\/p>\n

Commenting on a recent project at the Dell EMC HPC and AI Innovation Lab to train an AI radiologist, Wilson highlighted that, \u201cIn the last five to six years, we have reached the point where we have enough computational power to really think about processing images in this way.<\/p>\n

“We have shifted to where we are allowing a computer to learn a set of features it thinks is important, instead of us telling it what we think is important.<\/p>\n

\u201cThis is new ground for everyone. We can train an AI radiologist model with specialized skills, replicate that model 100 times over, and even shrink it down to where it can work on a cell phone or low-powered PC. This enables patients to get specialized predictions and opinions in areas where a trained radiologist might not easily get to.\u201d Wilson added that in locations where physicians have access to an X-ray but not necessarily a radiologist, the algorithm can help provide triage and guidance on what course of action to take. \u201cIn the end,\u201d he continued, \u201canything that involves making a subjective decision will always involve a human doctor, but AI can augment that with specialist capabilities. Those capabilities can be deployed where access to specialized care is difficult or impossible.\u201d<\/p>\n

Center for Rare Childhood Disorders TGen discovery<\/strong><\/p>\n

Key results: <\/strong><\/p>\n