Proceedings of the Southwest State University Series IT Management Computer Science Computer Engineering Medical Equipment Engineering,
Journal Year:
2024,
Volume and Issue:
14(3), P. 104 - 120
Published: Nov. 15, 2024
The
purpose
of
the
research
is
to
analyze
efficiency
U-net
neural
network
architecture
in
decision
support
systems
for
glioma
diagnostics
and
segmentation
brain
areas
affected
by
it
on
MRI
images.
Methods.
To
conduct
experimental
studies,
a
training
dataset
was
generated
data
normalized.
A
software
implementation
U-Net
performed
using
Keras
framework
Python
programming
language.
model
trained.
Results.
series
experiments
were
conducted,
during
which
error
classification
matrices
obtained,
trained
"Tumor"
"No
tumor"
classes
assessed
metrics
such
as
Recall,
Precision
F1-measure,
quality
glioma-affected
test
set
assessed.
IoU
metric,
reflects
ratio
bounding
boxes
used
assess
accuracy
spatial
correspondence
predicted
segmented
highlighted
masks.
Based
results
testing
solving
problem
segmenting
glioma,
average
value
metric
0.812,
an
acceptable
result.
Conclusion.
showed
that
based
able
effectively
diagnose
presence
with
values
metrics,
indicates
possibility
this
medical
diagnostics,
well
its
However,
advisable
refine
reduce
number
false
negative
results,
critically
important
diagnostics.
Reviews in the Neurosciences,
Journal Year:
2024,
Volume and Issue:
unknown
Published: Nov. 12, 2024
Glioblastoma
is
a
brain
cancer
with
poor
prognosis.
Failure
of
classical
chemotherapy
and
surgical
treatments
indicates
that
new
therapeutic
approaches
are
needed.
Among
cell-free
options,
exosomes
versatile
extracellular
vesicles
(EVs)
carry
important
cargo
across
barriers
such
as
the
blood-brain
barrier
(BBB)
to
their
target
cells.
This
makes
an
interesting
option
for
treatment
glioblastoma.
Moreover,
can
comprise
many
cargos,
including
lipids,
proteins,
nucleic
acids,
sampled
from
special
intercellular
compartments
origin
cell.
Cells
exposed
various
immunomodulatory
stimuli
generate
enriched
in
specific
molecules.
Notably,
secretion
could
modify
immune
response
innate
adaptive
systems.
For
instance,
glioblastoma-associated
(GBex)
uptake
by
macrophages
influence
macrophage
dynamics
(e.g.,
shifting
CD
markers
expression).
Expression
critical
immunoregulatory
proteins
cytotoxic
T-lymphocyte
antigen-1
(CTLA1)
programmed
death-1
(PD-1)
on
GBex
direct
crosstalk
these
nano-size
system.
The
present
study
reviews
role
system
cells,
B
T
natural
killer
(NK)
dendritic
cells
(DCs),
well
novel
technologies
field.
Proceedings of the Southwest State University Series IT Management Computer Science Computer Engineering Medical Equipment Engineering,
Journal Year:
2024,
Volume and Issue:
14(3), P. 104 - 120
Published: Nov. 15, 2024
The
purpose
of
the
research
is
to
analyze
efficiency
U-net
neural
network
architecture
in
decision
support
systems
for
glioma
diagnostics
and
segmentation
brain
areas
affected
by
it
on
MRI
images.
Methods.
To
conduct
experimental
studies,
a
training
dataset
was
generated
data
normalized.
A
software
implementation
U-Net
performed
using
Keras
framework
Python
programming
language.
model
trained.
Results.
series
experiments
were
conducted,
during
which
error
classification
matrices
obtained,
trained
"Tumor"
"No
tumor"
classes
assessed
metrics
such
as
Recall,
Precision
F1-measure,
quality
glioma-affected
test
set
assessed.
IoU
metric,
reflects
ratio
bounding
boxes
used
assess
accuracy
spatial
correspondence
predicted
segmented
highlighted
masks.
Based
results
testing
solving
problem
segmenting
glioma,
average
value
metric
0.812,
an
acceptable
result.
Conclusion.
showed
that
based
able
effectively
diagnose
presence
with
values
metrics,
indicates
possibility
this
medical
diagnostics,
well
its
However,
advisable
refine
reduce
number
false
negative
results,
critically
important
diagnostics.
