09-29-2021 Primetime Living - Flipbook - Page 5
A Special Advertising Section of Baltimore Sun Media Group | Wednesday, September 29, 2021 5
AI health care, continued from previous page
M.S., professor of information sciences
and technology, and his team have developed an application that uses abnormalities in a patient’s speech ability and facial
muscular movements within minutes with
the same accuracy – 79% – of a trained
and experienced emergency department
physician.
“Computer learning leads to better accuracy,” says Wang. “The more
data we have, which in our case are
comprised of actual stroke patients’
facial motions, muscle movements, and
speech patterns, the more accurate we
can be. When someone suffers a stroke,
time is critical because parts of the brain
are dying. Once at an ED, tests have to
be performed for an accurate diagnosis.
That takes time a victim can ill-afford.
Eventually, our app can be employed by
first responders, identifying a stroke in
a few minutes. It saves time by taking
the patient to a hospital with a comprehensive stroke center or one that is
thrombectomy-capable (able to remove
the clot causing the stroke) and the
stroke team has been alerted, ready to
start treatment.”
In collecting data, one requirement
Wang and his team needs is a variety of
faces of all ethnic groups – Hispanics,
African Americans and other non-whites
– for across-the-board precision. So far,
they have 150 scans and are enlisting
additional hospitals to expand their dataset. “Since we’re in EDs, we’re dependent on actual patients willing to be
scanned. Our immediate goal is to have a
couple thousand faces before we deploy
the app.
As it stands today, clinicians can’t
always figure out the bottom line. “If we
can identify ways for earlier interventions,
figure out what therapies will work best,”
Moukheiber adds, “it means we can treat
someone more quickly. And this is just
from a diagnostic standpoint. That’s what
machine learning does.”
Identifying Stroke Part 2
As mentioned above, time is critical for treating stroke victims. RapidAI
(www.rapidai.com), advanced imaging
software first launched 10 years ago,
allowed doctors to see a stroke happening in real time. Since each person’s
stroke progresses differently, it changed
the information physicians had to make
patient treatment decisions resulting in
a huge advancement for both doctors
and patients. For those having an ischemic stroke, the most common type,
RapidAI identified regions of the brain
with reduced blood flow and volume,
indicated parts of the brain with irreversible injury, and parts that could be saved.
It led to a major reduction in stroke disability and mortality.
Rachel Witalec, vice president of product and strategy at RapidAI, talks about
the company’s next steps. “The time
issue we addressed with stroke, along
with all we’ve learned from emergency
and specialist teams of doctors, nurses
and clinicians, can now be applied to
other similar life-threatening emergencies, such as a pulmonary embolism
(a blood clot in the lung that can lead
to heart failure), aneurysm, STEMI and
trauma. We also want to address optimizing the care team workflow so that they
know what needs to be done even before
a patient’s arrival.”
Witalec sees future problems being
solved by multidisciplinary teams,
incorporating technology and Artificial
Intelligence (AI). “These are tools that
can be used to better share information
in real-time while improving efficiency
of care teams. We’re always looking
forward, asking ‘How many more people
can we save?’ And for how many health
systems can we help reduce costs? Our
products have approval in markets globally and are now in use at over 2,000
hospitals worldwide.”
The Future
“Soon,” Moukheiber prognosticates,
“doctors and technology will be working more closely together. Human ability is finite, while machines can enhance
our accuracy, reducing medical errors, a
huge advance in patient care. Then we’ll
have the best of both worlds.”
AI health care,
continued on page 20
Have confidence
IN EVERY CONVERSATION.
Do phone calls ever sound muffled or hard to follow? Captioned Telephone can
help! Similar to a standard telephone—but with one big difference—a Captioned
Telephone displays captions of your conversations so that you can listen to what
is being spoken, while reading captions on the device’s screen. As a service of
Maryland Relay, Captioned Telephones can be purchased directly from a thirdparty provider. You can find more information at mdrelay.org.
The Maryland Accessible Telecommunications (MAT) program also offers
Captioned Telephones, provided by the State, to those who qualify. The
MAT program includes a variety of equipment options to support clearer
communication, including amplified phones. Visit our website to see if you or a
loved one qualify for no cost assistive equipment through the MAT program.
CapTel ® is a registered trademark of Ultratec, Inc.
For more information:
800-552-7724 | 410-767-6960 (Voice/TTY)
443-453-5970 (Video Phone) | mdrelay.org
