[...] New cognitive technologies have the potential to substantially improve CAD for radiology images and also those from pathology labs, and to combine them with other diagnostic data. These technologies are advancing quickly in research labs, but have yet to make their way into medical practice. A relatively new center at Partners Healthcare - the Center for Clinical Data Science (CCDS) - is focused on bringing these technologies to the clinical world. Based at the highly-ranked institutions Massachusetts General Hospital (MGH) and Brigham & Women's Hospital (BWH) in Boston, the CCDS is a joint effort of MGH and BWH. Its goal is to employ machine learning and other artificial intelligence technologies to improve the healthcare delivery system; in particular, a key CCDS objective is to improve the effectiveness of imaging-based diagnosis.
The CCDS is pursuing a variety of machine learning approaches, but the primary technology that it is employing is deep neural networks (also known as deep learning). These technologies have already led to breakthroughs in other areas of image recognition, and many researchers expect that they eventually will do so with medical images. A recent article in the New England Journal of Medicine, "Translating Artificial intelligence into Clinical Care," expressed hope that this type of machine learning will lead to a breakthrough in care. As Dr. Keith Dreyer, Partners' Chief Data Science Officer, puts it:
We've had CAD for a couple of decades, but deep learning is a much better technology. It will provide much higher sensitivity and specificity than we have today, and radiologists will trust it. Integrating it with clinical practice offers many potential benefits.
The diagnosis of a lumbar spine injury, for example, might involve up to 300 MRI images and various other test results in an electronic medical record system. A deep learning application could quickly identify the most important images for a radiologist to review and recommend treatment alternatives. The technology could save substantial time for critically injured trauma patients and could leverage the radiologist's time for all patients.
Source: Forbes (View full article)
Posted by Dan Corcoran on November 6, 2017 07:23 AM
Post a comment