Biomedicines,
Journal Year:
2023,
Volume and Issue:
11(11), P. 2880 - 2880
Published: Oct. 24, 2023
This
study
harnessed
machine
learning
to
forecast
postoperative
mortality
(POM)
and
pneumonia
(PPN)
among
surgical
traumatic
brain
injury
(TBI)
patients.
Our
analysis
centered
on
the
following
key
variables:
Glasgow
Coma
Scale
(GCS),
midline
shift
(MSB),
time
from
emergency
room
arrival
(TIE).
Additionally,
we
introduced
innovative
clustered
variables
enhance
predictive
accuracy
risk
assessment.
Exploring
data
617
patients
spanning
2012
2022,
observed
that
22.9%
encountered
mortality,
while
30.0%
faced
(PPN).
Sensitivity
for
POM
PPN
prediction,
before
incorporating
clustering,
was
in
ranges
of
0.43-0.82
0.54-0.76
Following
sensitivity
values
were
0.47-0.76
0.61-0.77
Accuracy
0.67-0.76
0.70-0.81
prior
clustering
0.42-0.73
0.55-0.73
after
clustering.
Clusters
characterized
by
low
GCS,
small
MSB,
short
TIE
exhibited
a
3.2-fold
higher
compared
clusters
with
high
TIE.
In
summary,
leveraging
offers
novel
avenue
predicting
TBI
Assessing
amalgamated
impact
characteristics
provides
valuable
insights
clinical
decision
making.
Research Square (Research Square),
Journal Year:
2024,
Volume and Issue:
unknown
Published: March 21, 2024
Abstract
Background
Postoperative
pulmonary
complications
(PPCs)
remain
a
prevalent
concern
among
elderly
surgical
patients,
with
notably
higher
incidence
observed
in
the
undergoing
thoracic
surgery.
This
study
aimed
to
construct
nomogram
predict
risk
of
PPCs
this
population.
Methods
A
total
2963
patients
who
underwent
surgery
were
randomly
enrolled
and
divided
into
training
cohort
(80%,
n
=
2369)
validation
(20%,
593).
Univariate
multivariate
logistic
regression
analyses
conducted
identify
factors
for
PPCs,
was
developed
based
on
findings
from
cohort.
The
used
validate
model.
predictive
accuracy
model
evaluated
by
receiver
operating
characteristic
curve
(ROC),
area
under
ROC
(AUC),
calibration
decision
analysis
(DCA).
Results
918
(31.0%)
reported
PPCs.
Nine
independent
identified:
preoperative
presence
chronic
obstructive
disease
(COPD),
elevated
leukocyte
count,
partial
pressure
arterial
carbon
dioxide
(PaCO2)
levels,
location
surgery,
thoracotomy,
intraoperative
hypotension,
blood
loss
>
100
mL,
duration
180
min
malignant
tumor.
AUC
value
0.739
(95%
CI:
0.719–0.762),
that
0.703
0.657–0.749).
P
values
Hosmer-Lemeshow
test
0.633
0.144
cohorts,
respectively,
indicating
good
fit.
DCA
showed
could
be
applied
clinically
if
threshold
between
12%
84%,
which
found
8%
82%
Conclusions
underscores
pressing
need
early
detection
exhibited
promising
efficacy
individuals
enabling
identification
high-risk
consequently
aiding
implementation
preventive
interventions.
Research Square (Research Square),
Journal Year:
2024,
Volume and Issue:
unknown
Published: June 4, 2024
Abstract
Background
Respiratory
complications
after
thoracic
surgery
are
common
and
can
lead
to
increased
perioperative
morbidity
mortality.
Although
several
clinical
risk
scores
for
the
prediction
of
respiratory
have
been
proposed,
these
not
specific
surgery.
In
addition,
few
adopted
in
practice
due
lack
external
validation.
Our
thoracic-specific
score,
CARDOT
showed
good
predictive
performance
postoperative
during
score
development.
This
study
aimed
validate
an
dataset
determine
including
neutrophil-lymphocyte
ratio
(NLR)
as
additive
predictor.
Methods
A
retrospective
cohort
consecutive
surgical
patients
at
a
single
tertiary
hospital
northern
Thailand
was
conducted.
The
development
validation
datasets
were
collected
between
2006
2012
from
2015
2021,
respectively.
Six
prespecified
factors
identified,
formed
(chronic
obstructive
pulmonary
disease,
American
Society
Anesthesiologists
physical
status,
right-sided
operation,
duration
surgery,
oxygen
saturation,
thoracotomy),
calculated.
evaluated
terms
discrimination
by
using
area
under
receiver
operating
characteristic
(AuROC)
curve
calibration.
Results
incidence
15.7%
(171
1088)
24.6%
(370
1642),
dataset.
had
discriminative
ability
both
(AuROC
0.789
(95%
CI
0.753–0.827)
0.758
0.730–0.787),
respectively).
calibration
datasets.
high
NLR
(≥
4.5)
significantly
(P
<
0.001).
AuROC
with
greater
power
than
that
alone
=
0.008).
Conclusions
consistent
tool
may
be
beneficial
settings
where
preoperative
function
tests
routinely
performed.
Translational Lung Cancer Research,
Journal Year:
2024,
Volume and Issue:
13(6), P. 1318 - 1330
Published: June 1, 2024
Background:
Sleeve
lobectomy
is
a
challenging
procedure
with
high
risk
of
postoperative
complications.
To
facilitate
surgical
decision-making
and
optimize
perioperative
treatment,
we
developed
stratification
models
to
quantify
the
probability
complications
after
sleeve
lobectomy.
Methods:
We
retrospectively
analyzed
clinical
features
691
non-small
cell
lung
cancer
(NSCLC)
patients
who
underwent
between
July
2016
December
2019.
Logistic
regression
were
trained
validated
in
cohort
predict
overall
complications,
major
specific
minor
The
impact
prognostic
was
explored
via
Kaplan-Meier
method.
Results:
Of
included
patients,
232
(33.5%)
including
35
(5.1%)
197
(28.5%)
respectively.
showed
robust
discrimination,
yielding
an
area
under
receiver
operating
characteristic
(ROC)
curve
(AUC)
0.853
[95%
confidence
interval
(CI):
0.705–0.885]
for
predicting
complication
0.751
(95%
CI:
0.727–0.762)
specifically
risks.
Models
also
achieved
good
performance,
AUCs
ranging
from
0.78
0.89.
Survival
analyses
revealed
significant
association
poor
prognosis.
Conclusions:
Risk
could
accurately
severity
NSCLC
following
lobectomy,
which
may
inform
future
patients.
Canadian Respiratory Journal,
Journal Year:
2022,
Volume and Issue:
2022, P. 1 - 8
Published: Sept. 4, 2022
This
study
is
conducted
to
investigate
the
correlation
between
perioperative
fractional
exhaled
nitric
oxide
and
postoperative
pneumonia
(POP)
feasibility
of
FeNO
for
predicting
POP
in
surgical
lung
cancer
patients.Patients
who
were
diagnosed
with
non-small-cell
(NSCLC)
prospectively
analyzed,
relationship
was
evaluated
based
on
patients'
basic
characteristics
clinical
data
hospital.There
218
patients
enrolled
this
study.
Finally,
183
involved
study,
19
them
group
164
non-POP
group.
The
had
significantly
higher
(median:
30.0
vs.
19.0
ppb,
P
<
0.001)
as
well
change
10.0
0.0
before
after
surgery.
For
receiver
operating
characteristic
(ROC)
curve,
a
cutoff
value
25
ppb
(Youden's
index:
0.515,
sensitivity:
78.9%,
specificity:
72.6%)
4
0.610,
84.2%,
76.8%)
selected.
Furthermore,
according
bivariate
regression
analysis,
FEV1/FVC
(OR
=
0.948,
95%
CI:
0.899-0.999,
P=0.048),
POD1
1.048,
1.019-1.077,
P=0.001),
1.087,
1.044-1.132,
associated
occurrence
POP.This
prospective
revealed
that
high
(>25
ppb),
an
increased
(>4
may
have
potential
detecting
patients.
Journal of Pain Research,
Journal Year:
2023,
Volume and Issue:
Volume 16, P. 2251 - 2256
Published: July 1, 2023
Chronic
postsurgical
pain
(CPSP)
is
a
common
complication
after
thoracic
surgery
and
associated
with
long-term
adverse
outcomes.
This
study
aims
to
develop
two
prediction
models
for
CPSP
video-assisted
(VATS).This
single-center
prospective
cohort
will
include
total
of
500
adult
patients
undergoing
VATS
lung
resection
(n
=
350
development
n
150
external
validation).
Patients
be
enrolled
continuously
at
The
First
Affiliated
Hospital
Soochow
University
in
Suzhou,
China.
validation
recruited
another
time
period.
outcome
CPSP,
which
defined
as
the
numerical
rating
scale
score
1
or
higher
3
months
VATS.
Univariate
multivariable
logistic
regression
analyses
performed
based
on
patients'
data
postoperative
day
14,
respectively.
For
internal
validation,
we
use
bootstrapping
technique.
discrimination
capability
assessed
using
area
under
receiver
operating
characteristic
curve,
calibration
evaluated
curve
Hosmer-Lemeshow
goodness-of-fit
statistic.
results
presented
model
formulas
nomograms.Based
models,
our
contribute
early
treatment
VATS.Chinese
Clinical
Trial
Register
(ChiCTR2200066122).
Biomedicines,
Journal Year:
2023,
Volume and Issue:
11(11), P. 2880 - 2880
Published: Oct. 24, 2023
This
study
harnessed
machine
learning
to
forecast
postoperative
mortality
(POM)
and
pneumonia
(PPN)
among
surgical
traumatic
brain
injury
(TBI)
patients.
Our
analysis
centered
on
the
following
key
variables:
Glasgow
Coma
Scale
(GCS),
midline
shift
(MSB),
time
from
emergency
room
arrival
(TIE).
Additionally,
we
introduced
innovative
clustered
variables
enhance
predictive
accuracy
risk
assessment.
Exploring
data
617
patients
spanning
2012
2022,
observed
that
22.9%
encountered
mortality,
while
30.0%
faced
(PPN).
Sensitivity
for
POM
PPN
prediction,
before
incorporating
clustering,
was
in
ranges
of
0.43-0.82
0.54-0.76
Following
sensitivity
values
were
0.47-0.76
0.61-0.77
Accuracy
0.67-0.76
0.70-0.81
prior
clustering
0.42-0.73
0.55-0.73
after
clustering.
Clusters
characterized
by
low
GCS,
small
MSB,
short
TIE
exhibited
a
3.2-fold
higher
compared
clusters
with
high
TIE.
In
summary,
leveraging
offers
novel
avenue
predicting
TBI
Assessing
amalgamated
impact
characteristics
provides
valuable
insights
clinical
decision
making.
