September 15, 2017
I’ve talked about the triple aim, developing standards and innovation happening in Healthcare with imaging in my previous posts. Let’s dive deeper now into the data from these imaging systems. I believe that data is the next big resource and key to unlocking another quantum jump in capabilities.
The data in imaging starts with generation of a medical image usually depicting the inside of a body. The image is derived from either a shadow in X-rays or measuring time gap between two photons in Nuclear Medicine or echoes from sound waves in Ultrasound or a radio signal from Hydrogen molecules excited by a magnetic field in an MRI system. These signals, when converted into pixels, combine to give us an image of the body or organ being studied. Highly trained Radiologists then interpret the image to locate or confirm a condition like a broken bone or tumor. This leads to a diagnosis to guide appropriate treatment or further monitoring as the case may be.
As imaging capability increases to measure smaller signals or compute more data to derive usable information, the clarity and depth of the scan improves. The radiologists look for comparisons between a known healthy or diseased anatomy and what is presented to them.
The diagnosis is derived from current data sets being compared to historical records. What makes it complex is the uniqueness of human anatomy. A human heart or brain is so similar yet so unique for each person that you cannot easily program a computer to look for condition X in location Y. That was the case until now, enter machine learning and artificial intelligence which may make automated imaging diagnosis possible.
There is plenty of interest to enhance medical imaging by artificial intelligence algorithms, see this list of 12 Startups Diagnosing Medical Images With AI. Some predictions may seem dire but we have a pretty long ways to go and plenty of room to improve what lies ahead, Scanning The Future, Radiologists See Their Jobs At Risk.
Rad@Sami 