Primetime Living 11.26.25 - Flipbook - Page 8
8
A Special Advertising Section of Baltimore Sun Media Group | Wednesday, November 26, 2025
HEALTH
Better Tools
to Find Cancer
Does AI hold the key for
better cancer detection?
By Margit B. Weisgal, Contributing Writer
A
recent press release crossed my desk touting a new tool with significant
improvements in accuracy and reliability using Artificial Intelligence – AI – for
early cancer detection.
“AI is going to
be the key to
understanding
and solving many
of the world’s most
complex problems.”
— Satya Nadella,
CEO of Microsoft
Joshua T. Vogelstein, Ph.D., associate professor of biomedical engineering at Johns Hopkins and a lead
investigator, explained what this new
method is and how it will help medical
practitioners in the future. He specializes in working with AI to address diagnostic tools for detecting and treating
cancers.
“A major difficulty doctors face is
accuracy and reliability with the tests
they use to determine whether or not
a patient has cancer,” says Vogelstein.
“No one ever wants a false positive,
since those cause great anxiety for
patients and physicians. At the same
time, more diagnostic tests and tools
are using AI as a means to determine
results.”
Vogelstein is correct. One area
already using AI to improve results is
mammography, screening for breast
cancer. In June of 2024, the Susan G.
Komen organization highlighted the voluntary option of adding an AI screening
to one’s mammogram. Currently, mammograms are 87% accurate.
He is part of a team that published
an article in Cancer Discovery in its
October 1, 2025, issue: “Translating
Artificial Intelligence Breakthroughs
into Cancer Diagnostic Breakthroughs,”
by Juan M. Lavista Ferres; Elliot K.
Fishman; Ed Catmull; Bert Vogelstein
Corresponding Author; and Joshua T.
Vogelstein.
The article summary points out that
“The revolution of artificial intelligence
has yet to find its way into clinical
practice in oncology. We highlight eight
specific challenges, focused on diagnostics, that will enable this translation
once they are adequately addressed.
“In simple terms, these eight challenges include the false expectation
that AI tools need to be flawless before
they’re considered useful; the need to
present results as probabilities rather
than simple yes-or-no answers; making sure AI predictions match realworld probabilities; ensuring results
are reproducible; training models on
diverse populations; explaining how AI
makes decisions; recognizing how test
accuracy can change when diseases
are rare; and avoiding over-reliance
on computer-generated recommendations.”
Vogelstein explained where they are
in the process. “Our first step was
to create MIGHT, which stands for
Multidimensional Informed Generalized
Hypothesis Testing.
“MIGHT is an AI method that we
created to meet the high level of confidence needed for AI tools used in
clinical decision making. Anything used
in clinical practices had to simplify the
testing and be very accurate.
“We also wanted MIGHT to have a
high sensitivity for cancer, while ensuring low false-positive results, so we
incorporated data from actual patients
– usually patients at Johns Hopkins and
other institutions.
“In tests using patient data, MIGHT
consistently outperformed other AI
AI Cancer Detection
Continued on page 22
