Journal of Personalized Medicine,
Journal Year:
2024,
Volume and Issue:
14(10), P. 1047 - 1047
Published: Oct. 9, 2024
Apparent
Diffusion
Coefficient
(ADC)
maps
in
prostate
MRI
can
reveal
tumor
characteristics,
but
their
accuracy
be
compromised
by
artifacts
related
with
patient
motion
or
rectal
gas
associated
distortions.
To
address
these
challenges,
we
propose
a
novel
approach
that
utilizes
Generative
Adversarial
Network
to
synthesize
ADC
from
T2-weighted
magnetic
resonance
images
(T2W
MRI).
Medical Image Analysis,
Journal Year:
2024,
Volume and Issue:
97, P. 103276 - 103276
Published: July 17, 2024
Radiation
therapy
plays
a
crucial
role
in
cancer
treatment,
necessitating
precise
delivery
of
radiation
to
tumors
while
sparing
healthy
tissues
over
multiple
days.
Computed
tomography
(CT)
is
integral
for
treatment
planning,
offering
electron
density
data
accurate
dose
calculations.
However,
accurately
representing
patient
anatomy
challenging,
especially
adaptive
radiotherapy,
where
CT
not
acquired
daily.
Magnetic
resonance
imaging
(MRI)
provides
superior
soft-tissue
contrast.
Still,
it
lacks
information,
cone
beam
(CBCT)
direct
calibration
and
mainly
used
positioning.
Adopting
MRI-only
or
CBCT-based
radiotherapy
eliminates
the
need
planning
but
presents
challenges.
Synthetic
(sCT)
generation
techniques
aim
address
these
challenges
by
using
image
synthesis
bridge
gap
between
MRI,
CBCT,
CT.
The
SynthRAD2023
challenge
was
organized
compare
synthetic
methods
multi-center
ground
truth
from
1080
patients,
divided
into
two
tasks:
(1)
MRI-to-CT
(2)
CBCT-to-CT.
evaluation
included
similarity
dose-based
metrics
proton
photon
plans.
attracted
significant
participation,
with
617
registrations
22/17
valid
submissions
tasks
1/2.
Top-performing
teams
achieved
high
structural
indices
(≥0.87/0.90)
gamma
pass
rates
(≥98.1%/99.0%)
(≥97.3%/97.0%)
no
correlation
found
accuracy,
emphasizing
when
assessing
clinical
applicability
sCT.
facilitated
investigation
benchmarking
sCT
techniques,
providing
insights
developing
radiotherapy.
It
showcased
growing
capacity
deep
learning
produce
high-quality
sCT,
reducing
reliance
on
conventional
planning.
Physics and Imaging in Radiation Oncology,
Journal Year:
2025,
Volume and Issue:
33, P. 100719 - 100719
Published: Jan. 1, 2025
Synthetic
Computed
Tomography
(sCT)
is
required
to
provide
electron
density
information
for
MR-only
radiotherapy.
Deep-learning
(DL)
methods
sCT
generation
show
improved
dose
congruence
over
other
(e.g.
bulk
density).
Using
30
female
pelvis
datasets
train
a
cycleGAN-inspired
DL
model,
this
study
found
mean
differences
between
deformed
planning
CT
(dCT)
and
were
0.2
%
(D98
%).
Three
Dimensional
Gamma
analysis
showed
of
90.4
at
1
%/1mm.
This
accurate
sCTs
(dose)
can
be
generated
from
routinely
available
T2
spin
echo
sequences
without
the
need
additional
specialist
sequences.
Computers in Biology and Medicine,
Journal Year:
2025,
Volume and Issue:
192, P. 110160 - 110160
Published: April 22, 2025
The
generation
of
Synthetic
Computed
Tomography
(sCT)
images
has
become
a
pivotal
methodology
in
modern
clinical
practice,
particularly
the
context
Radiotherapy
(RT)
treatment
planning.
use
sCT
enables
calculation
doses,
pushing
towards
Magnetic
Resonance
Imaging
(MRI)
guided
radiotherapy
treatments.
Moreover,
with
introduction
MRI-Positron
Emission
(PET)
hybrid
scanners,
derivation
from
MRI
can
improve
attenuation
correction
PET
images.
Deep
learning
methods
for
MRI-to-sCT
have
shown
promising
results,
but
their
reliance
on
single-centre
training
dataset
limits
generalisation
capabilities
to
diverse
settings.
creating
centralised
multi-centre
datasets
may
pose
privacy
concerns.
To
address
aforementioned
issues,
we
introduced
FedSynthCT-Brain,
an
approach
based
Federated
Learning
(FL)
paradigm
brain
imaging.
This
is
among
first
applications
FL
MRI-to-sCT,
employing
cross-silo
horizontal
that
allows
multiple
centres
collaboratively
train
U-Net-based
deep
model.
We
validated
our
method
using
real
multicentre
data
four
European
and
American
centres,
simulating
heterogeneous
scanner
types
acquisition
modalities,
tested
its
performance
independent
centre
outside
federation.
In
case
unseen
centre,
federated
model
achieved
median
Mean
Absolute
Error
(MAE)
102.0
HU
across
23
patients,
interquartile
range
96.7-110.5
HU.
(interquartile
range)
Structural
Similarity
Index
(SSIM)
Peak
Signal
Noise
Ratio
(PNSR)
were
0.89
(0.86-0.89)
26.58
(25.52-27.42),
respectively.
analysis
results
showed
acceptable
performances
approach,
thus
highlighting
potential
enhance
generalisability
advancing
safe
equitable
while
fostering
collaboration
preserving
privacy.
Deleted Journal,
Journal Year:
2024,
Volume and Issue:
1(1)
Published: Jan. 1, 2024
Abstract
Generative
artificial
intelligence
(AI)
has
enabled
tasks
in
radiology,
including
tools
for
improving
image
quality.
Recently,
new
hotspots
have
emerged,
such
as
intra-
or
inter-modal
translation,
task-specific
synthesis,
and
text
generation.
Advances
generative
AI
facilitated
the
move
towards
low-dose,
cost-effective,
high-quality
radiological
acquisition.
Large
language
models
can
aid
radiologists
by
generating
professional
answers
facilitating
patient-physician
communications.
However,
must
be
aware
of
potential
inaccuracies
generated
content
should
only
use
after
rigorous
validation
their
performance.
