Diagnostics,
Journal Year:
2025,
Volume and Issue:
15(6), P. 786 - 786
Published: March 20, 2025
Background:
Generalizability
and
domain
dependency
are
critical
challenges
in
developing
predictive
models
for
healthcare,
particularly
medical
diagnostics
radiation
oncology.
Predictive
designed
to
assess
tumor
recurrence
rely
on
comprehensive
high-quality
datasets,
encompassing
treatment
planning
parameters,
imaging
protocols,
patient-specific
data.
However,
dependency,
arising
from
variations
dose
calculation
algorithms,
computed
tomography
(CT)
density
conversion
curves,
modalities,
institutional
can
significantly
undermine
model
reliability
clinical
utility.
Methods:
This
study
evaluated
differences
the
head
neck
cancer
plans
of
19
patients
using
two
systems,
Pinnacle
9.10
RayStation
11,
with
similar
algorithms.
Variations
grid
size
CT
curves
were
assessed
their
impact
dependency.
Results:
Results
showed
that
had
a
more
significant
influence
within
than
Pinnacle,
while
curve
introduced
potential
discrepancies.
The
findings
underscore
role
precise
standardized
enhancing
modeling
assessment.
Conclusions:
Incorporating
such
as
distribution
target
volumes,
explicit
features
training
mitigate
enhance
prediction
accuracy.
Solutions
multi-institutional
data
harmonization
adaptation
techniques
essential
improve
generalizability
robustness.
These
strategies
support
better
integration
into
workflows,
ultimately
optimizing
patient
outcomes
personalized
strategies.
Expert Review of Anticancer Therapy,
Journal Year:
2025,
Volume and Issue:
unknown, P. 1 - 22
Published: April 5, 2025
The
World
Health
Organization's
2021
classification
of
central
nervous
system
neoplasms
incorporated
molecular
and
genetic
features
for
classifying
gliomas.
Classification
gliomas
located
in
deep-seated
structures
became
a
clinical
conundrum
given
the
absence
crucial
pathological
data.
Advances
noninvasive
imaging
modalities
offered
virtual
biopsy
as
novel
solution
to
this
problem
by
identifying
surrogate
radiomic
signatures.
Liquid
biopsies
blood
or
cerebrospinal
fluid
provided
another
enormous
opportunity
genomic,
metabolomic
proteomic
We
summarize
appraise
current
state
evidence
with
regards
liquid
care
patients
PubMed,
Embase
Google
Scholar
were
searched
on
7/30/2024
relevant
articles
published
after
year
2013
English
language.
A
large
body
preclinical
preliminary
suggests
that
is
possible
combined
use
multiple
conjunction
machine
learning
radiomics.
Likewise,
focused
ultrasound
may
be
valuable
tool
obtain
genomic
data
regarding
glioma
minimally
invasive
manner.
These
will
likely
become
an
integral
part
future.
Life,
Journal Year:
2025,
Volume and Issue:
15(4), P. 606 - 606
Published: April 5, 2025
Differentiating
tumor
progression
from
radionecrosis
in
patients
with
treated
brain
glioma
represents
a
significant
clinical
challenge
due
to
overlapping
imaging
features.
This
study
aimed
develop
and
evaluate
machine
learning
model
that
integrates
radiomics
features
T2*-weighted
Dynamic
Susceptibility
Contrast
MRI
perfusion
(DSC
MRI)
parameters
improve
diagnostic
accuracy
distinguishing
these
entities.
A
retrospective
cohort
of
46
(25
confirmed
radionecrosis,
21
progression)
was
analyzed.
From
lesion
segmentation
on
DSC
MRI,
851
were
extracted
using
PyRadiomics,
alongside
seven
(e.g.,
relative
cerebral
blood
volume,
time
peak)
obtained
time–intensity
curves
(TICs).
These
combined
into
single
dataset
14
classification
algorithms
evaluated
GroupKFold
cross-validation
(k
=
4).
The
top-performing
selected
based
predictive
area
under
the
curve
(AUC)
yield.
Logistic
Regression
classifier
achieved
highest
performance,
an
AUC
0.88,
followed
by
multilayer
perceptron
AdaBoost
values
0.85
0.79,
respectively.
precision
72%,
74%,
78%
for
three
models,
respectively,
while
63%,
70%,
71%.
Key
variables
included
like
wavelet-HHH_firstorder_Mean
mean
normalized
TIC
values.
Our
approach
integrating
shows
strong
potential
progression.
However,
further
validation
larger
cohorts
is
essential
confirm
generalizability
this
approach.