Visual Computing for Industry Biomedicine and Art,
Journal Year:
2023,
Volume and Issue:
6(1)
Published: Dec. 1, 2023
Although
prognostic
prediction
of
nasopharyngeal
carcinoma
(NPC)
remains
a
pivotal
research
area,
the
role
dynamic
contrast-enhanced
magnetic
resonance
(DCE-MR)
has
been
less
explored.
This
study
aimed
to
investigate
DCR-MR
in
predicting
progression-free
survival
(PFS)
patients
with
NPC
using
(MR)-
and
DCE-MR-based
radiomic
models.
A
total
434
two
MR
scanning
sequences
were
included.
The
MR-
radiomics
models
developed
based
on
289
only
145
four
additional
pharmacokinetic
parameters
(volume
fraction
extravascular
extracellular
space
(ve),
volume
plasma
(vp),
transfer
constant
(Ktrans),
reverse
reflux
rate
(kep)
DCE-MR.
combined
model
integrating
DCE-MR
was
constructed.
Utilizing
methods
such
as
correlation
analysis,
least
absolute
shrinkage
selection
operator
regression,
multivariate
Cox
proportional
hazards
we
built
Finally,
calculated
net
reclassification
index
C-index
evaluate
compare
performance
Kaplan-Meier
curve
analysis
performed
model's
ability
stratify
risk
NPC.
integration
features
significantly
enhanced
compared
models,
evidenced
by
test
set
0.808
vs
0.729
0.731,
respectively.
improved
22.9%-52.6%
could
levels
(p
=
0.036).
Furthermore,
MR-based
feature
maps
achieved
similar
results
terms
reflecting
underlying
angiogenesis
information
Compared
conventional
showed
promising
delivering
more
accurate
predictions
provided
clinical
benefits
quantifying
monitoring
phenotypic
changes
associated
prognosis.
iScience,
Journal Year:
2023,
Volume and Issue:
26(6), P. 106932 - 106932
Published: May 19, 2023
Chemotherapy
remains
controversial
for
stage
II
nasopharyngeal
carcinoma
because
of
its
considerable
prognostic
heterogeneity.
We
aimed
to
develop
an
MRI-based
deep
learning
model
predicting
distant
metastasis
and
assessing
chemotherapy
efficacy
in
carcinoma.
This
multicenter
retrospective
study
enrolled
1072
patients
from
three
Chinese
centers
training
(Center
1,
n
=
575)
external
validation
(Centers
2
3,
497).
The
significantly
predicted
the
risk
metastases
was
validated
cohort.
In
addition,
outperformed
clinical
radiomics
models
terms
predictive
performance.
Furthermore,
facilitates
identification
high-risk
who
could
benefit
chemotherapy,
providing
useful
additional
information
individualized
treatment
decisions.
Clinical Hemorheology and Microcirculation,
Journal Year:
2024,
Volume and Issue:
87(4), P. 437 - 450
Published: March 12, 2024
At
present,
most
articles
mainly
focused
on
the
diagnosis
of
thyroid
nodules
by
using
artificial
intelligence
(AI),
and
there
was
little
research
detection
performance
AI
in
nodules.
Clinical Medicine Insights Oncology,
Journal Year:
2023,
Volume and Issue:
17
Published: Jan. 1, 2023
Previous
studies
have
shown
that
the
5-year
survival
rates
of
patients
with
nasopharyngeal
carcinoma
(NPC)
were
still
not
ideal
despite
great
improvement
in
NPC
treatments.
To
achieve
individualized
treatment
NPC,
we
been
looking
for
novel
models
to
predict
prognosis
NPC.
The
objective
this
study
was
use
a
deep
learning
network
structural
model
and
compare
it
traditional
PET-CT
combining
metabolic
parameters
clinical
factors.A
total
173
admitted
2
institutions
between
July
2014
April
2020
retrospective
study;
each
received
scan
before
treatment.
least
absolute
shrinkage
selection
operator
(LASSO)
employed
select
some
features,
including
SUVpeak-P,
T3,
age,
stage
II,
MTV-P,
N1,
III
pathological
type,
which
associated
overall
(OS)
patients.
We
constructed
prediction
models:
an
improved
optimized
adaptive
multimodal
task
(a
3D
Coordinate
Attention
Convolutional
Autoencoder
uncertainty-based
jointly
Optimizing
Cox
Model,
CACA-UOCM
short)
model.
predictive
power
these
assessed
using
Harrell
Consistency
Index
(C
index).
Overall
compared
by
Kaplan-Meier
Log-rank
tests.The
results
showed
could
estimate
OS
index,
0.779
training,
0.774
validation,
0.819
testing)
divide
into
low
high
mortality
risk
groups,
significantly
(P
<
.001).
However,
C-index
based
only
on
variables
0.42.The
18F-FDG
PET/CT
can
serve
as
reliable
powerful
tool
provide
therapeutic
strategies
individual
Nuklearmedizin - NuclearMedicine,
Journal Year:
2023,
Volume and Issue:
62(06), P. 361 - 369
Published: Nov. 23, 2023
Abstract
Aim
Despite
a
vast
number
of
articles
on
radiomics
and
machine
learning
in
positron
emission
tomography
(PET)
imaging,
clinical
applicability
remains
limited,
partly
owing
to
poor
methodological
quality.
We
therefore
systematically
investigated
the
methodology
described
publications
for
PET-based
outcome
prediction.
Methods
A
systematic
search
original
was
run
PubMed.
All
were
rated
according
17
criteria
proposed
by
authors.
Criteria
with
>2
rating
categories
binarized
into
“adequate”
or
“inadequate”.
The
association
between
per
article
date
publication
examined.
Results
One
hundred
identified
(published
07/2017
09/2023).
median
proportion
criterion
that
65%
(range:
23–98%).
Nineteen
(19%)
mentioned
neither
test
cohort
nor
cross-validation
separate
training
from
testing.
an
12.5
out
(range,
4–17),
this
did
not
increase
later
dates
(Spearman’s
rho,
0.094;
p
=
0.35).
In
22
(22%),
less
than
half
items
“adequate”.
Only
8%
published
source
code,
10%
made
dataset
openly
available.
Conclusion
Among
investigated,
weaknesses
have
been
identified,
degree
compliance
recommendations
quality
reporting
shows
potential
improvement.
Better
adherence
established
guidelines
could
significance
prediction
finally
lead
widespread
use
routine
practice.
