BACKGROUND
The
integration
of
multimodal
capabilities
in
GPT-4V
represents
advancement
AI's
application
to
clinical
fields,
particularly
neuroradiology.
Despite
preliminary
evidence
capability
medical
imaging
interpretation,
questions
remain
about
its
performance
complex
scenarios
requiring
integrated
analysis
history
and
findings.
OBJECTIVE
To
evaluate
GPT-4V's
diagnostic
on
neuroradiology
board-style
multiple-choice
questions,
integrating
both
data
imaging.
METHODS
Twenty-nine
cases
from
the
RSNA
Case
Collection,
each
including
vignette
CT/MRI
images,
were
presented
GPT-4V.
model
evaluated
studies
data,
selecting
options
while
quantifying
relative
influence
image
versus
text
decision-making.
RESULTS
achieved
75.86%
accuracy,
with
contributing
an
average
66.9%
final
answers.
relied
more
heavily
incorrectly
answered
(75%
image-based)
compared
correct
ones
(61.74%).
CONCLUSIONS
findings
suggest
potential
over-reliance
where
context
is
crucial.
Our
results
highlight
need
for
improved
AI
models,
future
development
focusing
refining
decision-making
processes
enhance
accuracy.
Abstract
With
the
increasing
demand
for
precision
medicine
in
cancer
patients,
radiogenomics
emerges
as
a
promising
frontier.
Radiogenomics
is
originally
defined
methodology
associating
gene
expression
information
from
high‐throughput
technologies
with
imaging
phenotypes.
However,
advancements
medical
imaging,
omics
technologies,
and
artificial
intelligence,
both
concept
application
of
have
significantly
broadened.
In
this
review,
history
enumerated,
related
five
basic
workflows
their
applications
across
tumors,
role
AI
radiogenomics,
opportunities
challenges
tumor
heterogeneity,
immune
microenvironment.
The
positron
emission
tomography
multi‐omics
studies
also
discussed.
Finally,
faced
by
clinical
transformation,
along
future
trends
field
npj Digital Medicine,
Год журнала:
2025,
Номер
8(1)
Опубликована: Фев. 12, 2025
Abstract
Recent
advancements
in
large
language
models
(LLMs)
have
created
new
ways
to
support
radiological
diagnostics.
While
both
open-source
and
proprietary
LLMs
can
address
privacy
concerns
through
local
or
cloud
deployment,
provide
advantages
continuity
of
access,
potentially
lower
costs.
This
study
evaluated
the
diagnostic
performance
fifteen
one
closed-source
LLM
(GPT-4o)
1,933
cases
from
Eurorad
library.
provided
differential
diagnoses
based
on
clinical
history
imaging
findings.
Responses
were
considered
correct
if
true
diagnosis
appeared
top
three
suggestions.
Models
further
tested
60
non-public
brain
MRI
a
tertiary
hospital
assess
generalizability.
In
datasets,
GPT-4o
demonstrated
superior
performance,
closely
followed
by
Llama-3-70B,
revealing
how
are
rapidly
closing
gap
models.
Our
findings
highlight
potential
as
decision
tools
for
challenging,
real-world
cases.
International Endodontic Journal,
Год журнала:
2025,
Номер
unknown
Опубликована: Март 31, 2025
Abstract
The
integration
of
artificial
intelligence
(AI)
into
education
is
transforming
learning
across
various
domains,
including
dentistry.
Endodontic
can
significantly
benefit
from
AI
chatbots;
however,
knowledge
gaps
regarding
their
potential
and
limitations
hinder
effective
utilization.
This
narrative
review
aims
to:
(A)
explain
the
core
functionalities
chatbots,
reliance
on
natural
language
processing
(NLP),
machine
(ML),
deep
(DL);
(B)
explore
applications
in
endodontic
for
personalized
learning,
interactive
training,
clinical
decision
support;
(C)
discuss
challenges
posed
by
technical
limitations,
ethical
considerations,
misinformation.
highlights
that
chatbots
provide
learners
with
immediate
access
to
knowledge,
educational
experiences,
tools
developing
reasoning
through
case‐based
learning.
Educators
streamlined
curriculum
development,
automated
assessment
creation,
evidence‐based
resource
integration.
Despite
these
advantages,
concerns
such
as
chatbot
hallucinations,
algorithmic
biases,
plagiarism,
spread
misinformation
require
careful
consideration.
Analysis
current
research
reveals
limited
endodontic‐specific
studies,
emphasizing
need
tailored
solutions
validated
accuracy
relevance.
Successful
will
collaborative
efforts
among
educators,
developers,
professional
organizations
address
challenges,
ensure
use,
establish
evaluation
frameworks.
JMIR Medical Informatics,
Год журнала:
2025,
Номер
13, С. e64963 - e64963
Опубликована: Апрель 25, 2025
Abstract
Background
With
the
rapid
development
of
artificial
intelligence
(AI)
technology,
especially
generative
AI,
large
language
models
(LLMs)
have
shown
great
potential
in
medical
field.
Through
massive
data
training,
it
can
understand
complex
texts
and
quickly
analyze
records
provide
health
counseling
diagnostic
advice
directly,
rare
diseases.
However,
no
study
has
yet
compared
extensively
discussed
performance
LLMs
with
that
physicians.
Objective
This
systematically
reviewed
accuracy
clinical
diagnosis
provided
reference
for
further
application.
Methods
We
conducted
searches
CNKI
(China
National
Knowledge
Infrastructure),
VIP
Database,
SinoMed,
PubMed,
Web
Science,
Embase,
CINAHL
(Cumulative
Index
to
Nursing
Allied
Health
Literature)
from
January
1,
2017,
present.
A
total
2
reviewers
independently
screened
literature
extracted
relevant
information.
The
risk
bias
was
assessed
using
Prediction
Model
Risk
Bias
Assessment
Tool
(PROBAST),
which
evaluates
both
applicability
included
studies.
Results
30
studies
involving
19
a
4762
cases
were
included.
quality
assessment
indicated
high
majority
studies,
primary
cause
is
known
case
diagnosis.
For
optimal
model,
ranged
25%
97.8%,
while
triage
66.5%
98%.
Conclusions
demonstrated
considerable
capabilities
significant
application
across
various
cases.
Although
their
still
falls
short
professionals,
if
used
cautiously,
they
become
one
best
intelligent
assistants
field
human
care.
medRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Июль 16, 2024
Previous
studies
evaluated
the
ability
of
large
language
models
(LLMs)
in
medical
disciplines;
however,
few
have
focused
on
image
analysis,
and
none
specifically
cardiovascular
imaging
or
nuclear
cardiology.
GPT-4
Vision
(GPT-4V)
represents
a
significant
advancement
in
multimodal
artificial
intelligence,
enabling
text
generation
from
images
without
specialized
training.
This
marks
the
transformation
of
ChatGPT
as
large
language
model
(LLM)
into
GPT-4's
promised
(LMM).
As
these
AI
models
continue
to
advance,
they
may
enhance
radiology
workflow
and
aid
with
decision
support.
technical
note
explores
potential
GPT-4V
applications
evaluates
performance
for
sample
tasks.
capabilities
were
tested
using
web,
personal
institutional
teaching
files,
hand-drawn
sketches.
Prompts
evaluated
scientific
figure
analysis,
radiologic
image
reporting,
comparison,
handwriting
interpretation,
sketch-to-code,
artistic
expression.
In
this
limited
demonstration
GPT-4V's
capabilities,
it
showed
promise
classifying
images,
counting
entities,
comparing
deciphering
However,
exhibited
limitations
detecting
some
fractures,
discerning
change
size
lesions,
accurately
interpreting
complex
diagrams,
consistently
characterizing
findings.
Artistic
expression
responses
coherent.
While
eventually
assist
tasks
related
radiology,
current
reliability
gaps
highlight
need
continued
training
improvement
before
consideration
any
medical
use
by
general
public
ultimately
clinical
integration.
Future
iterations
could
enable
virtual
assistant
discuss
findings,
improve
reports,
extract
data
provide
support
based
on
guidelines,
white
papers,
appropriateness
criteria.
Human
expertise
remain
essential
safe
practice
partnerships
between
physicians,
researchers,
technology
leaders
are
necessary
safeguard
against
risks
like
bias
privacy
concerns.
