Page 18 - BB Neuromedicine Highlights 2019

A single radiology image requires
dozens of critical assessments
and considerations before making
a final interpretation. These
processes are even more complex
for some imaging types like brain
MRIs. The capacity of neurora-
diology AI platforms to accurately
manage these tasks is well beyond
the scope of current systems.
training neurology Ai systems
and other radiology machine
learning platforms are therefore
a major challenge for business-
es today. Likewise, the use of
artificial intelligence in radiology
requires system validation, data
preparation, and incorporation
into normal workflows.
WhaT arTIFIcIal INTellI- In this way, radiologists are at AI platform might screen head CT For these reasons, bold busi-
geNce IN raDIOlOgy lOOks the center of advancing deep scans for emergent stroke evalu- nesses are developing artificial
lIke learning algorithms for artificial ations before formal radiologist intelligence in radiology to
As with any technological intelligence in radiology software evaluations. This artificial intelli- augment imaging interpretation
disruption, the use of artificial systems. gence in radiology systems are not quality and efficiency. In doing so,
intelligence in radiology might be From a practical perspective, this advanced yet, but the future neuroradiology AI platforms and
a bit unnerving if you’re a radiolo- what does this mean? The obvi- looks very bright in this regard. others can progressively improve
gist. It, however, is not the case in ous application of this artificial in- while boosting existing healthcare
neuroradiology AI, chest radiol- telligence in radiology systems will OvercOmINg challeNges services.
ogy AI, or other similar systems. be as decision support structures IN NeurOraDIOlOgy aI aND
Using a people-centric approach, for radiologists. As these systems OTher areas emergINg arTIFIcIal INTellI-
businesses are developing arti- improve in their deep learning A number of challenges exist geNce IN raDIOlOgy leaDers
ficial intelligence in radiology by capacities, they can provide when it comes to using artificial As you might imagine, business-
recruiting radiologists to facilitate screening detection and oversight intelligence in radiology. For one, es interested in providing artificial
development. In essence, these for radiologist interpretations. radiologists do much more than intelligence in radiology services
systems enable radiologists to For example, a neuroradiology to simply interpret a single image. are growing. Incredible advances
collect image data, validate find-
ings, and actually train AI systems.