Medical Image Analysis,
Journal Year:
2022,
Volume and Issue:
84, P. 102722 - 102722
Published: Dec. 15, 2022
Coronavirus
disease
(COVID-19)
has
caused
a
worldwide
pandemic,
putting
millions
of
people's
health
and
lives
in
jeopardy.
Detecting
infected
patients
early
on
chest
computed
tomography
(CT)
is
critical
combating
COVID-19.
Harnessing
uncertainty-aware
consensus-assisted
multiple
instance
learning
(UC-MIL),
we
propose
to
diagnose
COVID-19
using
new
bilateral
adaptive
graph-based
(BA-GCN)
model
that
can
use
both
2D
3D
discriminative
information
CT
volumes
with
arbitrary
number
slices.
Given
the
importance
lung
segmentation
for
this
task,
have
created
largest
manual
annotation
dataset
so
far
7,768
slices
from
patients,
used
it
train
segment
lungs
individual
mask
as
regions
interest
subsequent
analyses.
We
then
UC-MIL
estimate
uncertainty
each
prediction
consensus
between
predictions
slice
automatically
select
fixed
reliable
reasoning.
Finally,
adaptively
constructed
BA-GCN
vertices
different
granularity
levels
(2D
3D)
aggregate
multi-level
features
final
diagnosis
benefits
graph
convolution
network's
superiority
tackle
cross-granularity
relationships.
Experimental
results
three
datasets
demonstrated
our
produce
accurate
any
slices,
which
outperforms
existing
approaches
terms
generalisation
ability.
To
promote
reproducible
research,
made
datasets,
including
annotations
cleaned
dataset,
well
implementation
code,
available
at
https://doi.org/10.5281/zenodo.6361963.
Frontiers in Medicine,
Journal Year:
2021,
Volume and Issue:
8
Published: Sept. 30, 2021
Background:
Recently,
Coronavirus
Disease
2019
(COVID-19),
caused
by
severe
acute
respiratory
syndrome
virus
2
(SARS-CoV-2),
has
affected
more
than
200
countries
and
lead
to
enormous
losses.
This
study
systematically
reviews
the
application
of
Artificial
Intelligence
(AI)
techniques
in
COVID-19,
especially
for
diagnosis,
estimation
epidemic
trends,
prognosis,
exploration
effective
safe
drugs
vaccines;
discusses
potential
limitations.
Methods:
We
report
this
systematic
review
following
Preferred
Reporting
Items
Systematic
Reviews
Meta-Analyses
(PRISMA)
guidelines.
searched
PubMed,
Embase
Cochrane
Library
from
inception
19
September
2020
published
studies
AI
applications
COVID-19.
used
PROBAST
(prediction
model
risk
bias
assessment
tool)
assess
quality
literature
related
diagnosis
prognosis
registered
protocol
(PROSPERO
CRD42020211555).
Results:
included
78
studies:
46
articles
discussed
AI-assisted
COVID-19
with
total
accuracy
70.00
99.92%,
sensitivity
73.00
100.00%,
specificity
25
area
under
curve
0.732
1.000.
Fourteen
evaluated
based
on
clinical
characteristics
at
hospital
admission,
such
as
clinical,
laboratory
radiological
characteristics,
reaching
74.4
95.20%,
72.8
98.00%,
55
96.87%
AUC
0.66
0.997
predicting
critical
Nine
models
predict
peak,
infection
rate,
number
infected
cases,
transmission
laws,
development
trend.
Eight
explore
drugs,
primarily
through
drug
repurposing
development.
Finally,
1
article
predicted
vaccine
targets
that
have
develop
vaccines.
Conclusions:
In
review,
we
shown
achieved
high
performance
evaluation,
prediction
discovery
enhance
significantly
existing
medical
healthcare
system
efficiency
during
pandemic.
Scientific Reports,
Journal Year:
2022,
Volume and Issue:
12(1)
Published: Oct. 26, 2022
Healthcare
data
are
inherently
multimodal,
including
electronic
health
records
(EHR),
medical
images,
and
multi-omics
data.
Combining
these
multimodal
sources
contributes
to
a
better
understanding
of
human
provides
optimal
personalized
healthcare.
The
most
important
question
when
using
is
how
fuse
them-a
field
growing
interest
among
researchers.
Advances
in
artificial
intelligence
(AI)
technologies,
particularly
machine
learning
(ML),
enable
the
fusion
different
modalities
provide
insights.
To
this
end,
scoping
review,
we
focus
on
synthesizing
analyzing
literature
that
uses
AI
techniques
for
clinical
applications.
More
specifically,
studies
only
fused
EHR
with
imaging
develop
various
methods
We
present
comprehensive
analysis
strategies,
diseases
outcomes
which
was
used,
ML
algorithms
used
perform
each
application,
available
datasets.
followed
PRISMA-ScR
(Preferred
Reporting
Items
Systematic
Reviews
Meta-Analyses
Extension
Scoping
Reviews)
guidelines.
searched
Embase,
PubMed,
Scopus,
Google
Scholar
retrieve
relevant
studies.
After
pre-processing
screening,
extracted
from
34
fulfilled
inclusion
criteria.
found
fusing
increasing
doubling
2020
2021.
In
our
analysis,
typical
workflow
observed:
feeding
raw
data,
by
applying
conventional
(ML)
or
deep
(DL)
algorithms,
finally,
evaluating
through
outcome
predictions.
Specifically,
early
technique
applications
(22
out
studies).
multimodality
models
outperformed
traditional
single-modality
same
task.
Disease
diagnosis
prediction
were
common
(reported
20
10
studies,
respectively)
perspective.
Neurological
disorders
dominant
category
(16
From
an
perspective,
(19
studies),
DL
Multimodal
included
mostly
private
repositories
(21
Through
offer
new
insights
researchers
interested
knowing
current
state
knowledge
within
research
field.
Scientific Reports,
Journal Year:
2022,
Volume and Issue:
12(1)
Published: March 14, 2022
COVID-19
clinical
presentation
and
prognosis
are
highly
variable,
ranging
from
asymptomatic
paucisymptomatic
cases
to
acute
respiratory
distress
syndrome
multi-organ
involvement.
We
developed
a
hybrid
machine
learning/deep
learning
model
classify
patients
in
two
outcome
categories,
non-ICU
ICU
(intensive
care
admission
or
death),
using
558
admitted
northern
Italy
hospital
February/May
of
2020.
A
fully
3D
patient-level
CNN
classifier
on
baseline
CT
images
is
used
as
feature
extractor.
Features
extracted,
alongside
with
laboratory
data,
fed
for
selection
Boruta
algorithm
SHAP
game
theoretical
values.
built
the
reduced
space
CatBoost
gradient
boosting
reaching
probabilistic
AUC
0.949
holdout
test
set.
The
aims
provide
decision
support
medical
doctors,
probability
score
belonging
an
class
case-based
interpretation
features
importance.
Critical Care Medicine,
Journal Year:
2020,
Volume and Issue:
49(1), P. 102 - 111
Published: Oct. 28, 2020
OBJECTIVES:
To
identify
characteristics
that
predict
30-day
mortality
among
patients
critically
ill
with
coronavirus
disease
2019
in
England,
Wales,
and
Northern
Ireland.
DESIGN:
Observational
cohort
study.
SETTING:
A
total
of
258
adult
critical
care
units.
PATIENTS:
10,362
confirmed
a
start
between
March
1,
2020,
June
22,
whom
9,990
were
eligible
(excluding
duration
less
than
24
hr
or
missing
core
variables).
MEASUREMENTS
AND
MAIN
RESULTS:
The
main
outcome
measure
was
time
to
death
within
30
days
the
care.
Of
(median
age
60
yr,
70%
male),
3,933
died
As
July
189
still
receiving
further
446
acute
hospital.
Data
for
0.1%
7.2%
across
prognostic
factors.
We
imputed
data
ten-fold,
using
fully
conditional
specification
continuous
variables
modeled
restricted
cubic
splines.
Associations
candidate
factors
determined
after
adjustment
multiple
Cox
proportional
hazards
modeling.
Significant
associations
identified
age,
ethnicity,
deprivation,
body
mass
index,
prior
dependency,
immunocompromise,
lowest
systolic
blood
pressure,
highest
heart
rate,
respiratory
Pa
o
2
/F
io
ratio
(and
interaction
mechanical
ventilation),
lactate
concentration,
serum
urea,
platelet
count
over
first
hours
Nonsignificant
found
sex,
sedation,
temperature,
hemoglobin
concentration.
CONCLUSIONS:
patient
an
increased
likelihood
2019.
These
findings
may
support
development
prediction
model
benchmarking
providers.
Computational and Structural Biotechnology Journal,
Journal Year:
2021,
Volume and Issue:
19, P. 2833 - 2850
Published: Jan. 1, 2021
The
worldwide
health
crisis
caused
by
the
SARS-Cov-2
virus
has
resulted
in>3
million
deaths
so
far.
Improving
early
screening,
diagnosis
and
prognosis
of
disease
are
critical
steps
in
assisting
healthcare
professionals
to
save
lives
during
this
pandemic.
Since
WHO
declared
COVID-19
outbreak
as
a
pandemic,
several
studies
have
been
conducted
using
Artificial
Intelligence
techniques
optimize
these
on
clinical
settings
terms
quality,
accuracy
most
importantly
time.
objective
study
is
conduct
systematic
literature
review
published
preprint
reports
models
developed
validated
for
coronavirus
2019.
We
included
101
studies,
from
January
1st,
2020
December
30th,
2020,
that
AI
prediction
which
can
be
applied
setting.
identified
total
14
38
diagnostic
detecting
50
prognostic
predicting
ICU
need,
ventilator
mortality
risk,
severity
assessment
or
hospital
length
stay.
Moreover,
43
were
based
medical
imaging
58
use
parameters,
laboratory
results
demographic
features.
Several
heterogeneous
predictors
derived
multimodal
data
identified.
Analysis
data,
captured
various
sources,
prominence
each
category
was
performed.
Finally,
Risk
Bias
(RoB)
analysis
also
examine
applicability
setting
assist
providers,
guideline
developers,
policymakers.