Diagnostic performance of artificial intelligence in detection of renal cell carcinoma: a systematic review and meta-analysis
BMC Cancer,
Год журнала:
2025,
Номер
25(1)
Опубликована: Янв. 27, 2025
The
detection
of
renal
cell
carcinoma
(RCC)
tumors
in
the
earlier
stages
is
great
importance
for
more
effective
treatment.
Encouraged
by
key
role
imaging
management
RCC,
we
conducted
a
systematic
review
and
meta-analysis
studies
that
made
use
artificial
intelligence
(AI)
RCC
to
quantitatively
determine
performance
AI
distinguishing
related
lesions.
PubMed,
Scopus,
CENTRAL,
Embase
electronic
databases
were
systematically
searched
November
2024
identify
applied
or
classification
RCC.
We
evaluate
diagnostic
utilized
algorithms.
Moreover,
meta-regression
was
over
suspected
covariates
potential
sources
inter-study
heterogeneity.
Publication
bias
quality
assessment
also
done
included
studies.
Sixty-four
this
review,
which
31
selected
meta-analysis.
assessing
algorithms'
on
internal
validation
showed
pooled
sensitivity
specificity
85%
(95%
confidence
interval
[CI],
82
87)
76%
CI,
70
80),
respectively.
externally
validated
Al
algorithms
had
80%
73
84)
90%
84
93),
Studies
performed
clinician
79%
72
85)
60%
49
70).
findings
present
study
validate
acceptable
when
contrasted
with
medical
professionals
identification
categorization
Nevertheless,
presence
heterogeneity
between
absence
coherence
results
underscore
necessity
cautious
interpretation
these
additional
prospective
Язык: Английский
Differentiation of solid and cystic small renal masses: the role of multiphase CT markers in predicting malignant histology, subtype, and grade
Polish Journal of Radiology,
Год журнала:
2025,
Номер
90, С. 239 - 252
Опубликована: Май 21, 2025
Purpose
This
study
aimed
to
assess
the
diagnostic
performance
of
multiphase
contrast-enhanced
computed
tomography
(MCECT)
in
differentiating
benign
and
malignant
solid
cystic
small
renal
masses
(SRMs),
predicting
histologic
subtypes,
grading,
using
signal
intensity
(SI)
tumour-to-cortex
(TCSI)
ratio.
Material
methods
A
retrospective
analysis
was
conducted
on
181
patients
with
SRMs
(≤
4
cm).
MCECT
imaging
across
phases
(non-contrast,
corticomedullary,
nephrographic,
excretory)
performed.
SI
TCSI
values
were
measured,
their
evaluated
receiver
operating
characteristic
(ROC)
analysis.
Solid,
Bosniak
IIF,
III,
IV
underwent
histopathological
confirmation.
Results
Among
SRMs,
excretory
phase
achieved
an
area
under
curve
(AUC)
0.848
for
RCC
from
other
100%
sensitivity
61.3%
specificity.
For
distinguishing
cell
carcinoma
(RCC)
most
effective
parameter
ratio
obtained
(88.6%
sensitivity,
52.4%
specificity,
0.763
AUC).
IIF
cysts,
corticomedullary
provided
AUC
0.902,
93%
87.5%
subtyping
showed
distinct
characteristics
phases,
particularly
clear
RCC.
Nephrographic
differentiated
low-
versus
high-grade
RCC,
0.901,
90.2%
86.4%
Conclusions
MCECT-derived
biomarkers,
TCSI,
are
non-invasive
tools
characterising
aiding
differentiation
lesions,
histological
tumour
grades.
Their
integration
advanced
radiomics
could
further
enhance
accuracy.
Язык: Английский