Accuracy and Reliability of 3D Cephalometric Landmark Detection with Deep Learning
Research Square (Research Square),
Год журнала:
2025,
Номер
unknown
Опубликована: Янв. 13, 2025
Abstract
Objective:
Three-dimensional
(3D)
landmark
detection
is
essential
for
assessing
craniofacial
growth
and
planning
surgeries
such
as
orthodontic,
orthognathic,
traumatic,
plastic
procedures.
This
study
aimed
to
develop
an
automatic
3D
landmarking
model
oral
maxillofacial
regions
validate
its
accuracy,
robustness,
generalizability
in
both
spiral
computed
tomography
(SCT,
41
landmarks)
cone-beam
(CBCT,
14
scans.
Methods:
The
was
constructed
using
optimized
lightweight
U-Net
network
architecture.
Its
were
thoroughly
evaluated
validated
through
a
multicenter
retrospective
diagnostic
study.
internal
dataset
included
480
SCT
240
CBCT
cases.
For
external
validation,
320
150
cases
assessed
mean
radial
error
(MRE)
success
rate
within
2-,
3-,
4-mm
thresholds
the
primary
evaluation
metrics.
Error
analyses
along
each
coordinate
axis
performed.
Consistency
tests
among
index
observers
conducted.
Results:
average
MRE
consistently
below
1.3
mm
and,
notably,
1.4
complex
conditions
malocclusion,
missing
dental
landmarks,
presence
of
metal
artifacts.
No
significant
differences
SDR
at
2-4
observed
between
sets.
bone
landmarks
more
precise
than
ones,
with
no
difference
bone/soft
tissue
dental/soft
tissue.
exhibited
greater
precision
compared
landmarks.
A
detailed
analysis
across
axes
showed
that
coronal
had
highest
rates.
implementation
this
significantly
improved
proficiency
senior
junior
specialists
by
15.9%
28.9%,
respectively,
while
also
accelerating
process
factor
6
9.5
times.
Conclusions:
This
shows
AI-driven
delivers
high-precision
localization
structures,
even
scenarios.
can
aid
all
experience
levels
conducting
accurate
efficient
analyses,
owing
strong
clinical
utility,
generalizability.
Clinical
Relevance:
3D
cephalometric
crucial
diverse
surgical
procedures,
trauma,
aesthetic
interventions.
traditional
manual
identification
time-consuming
requires
expertise.
proposed
AI
method
provides
measurements
soft
hard
tissues,
streamlines
digital
planning,
decreases
reliance
on
expert
knowledge,
enhances
efficiency
treatments.
Язык: Английский
Relationship between Björk–Jarabak Cephalometrics Analysis Elements and Facial Profile for Yemeni Adult Samples
World Journal of Dentistry,
Год журнала:
2025,
Номер
15(10), С. 897 - 901
Опубликована: Янв. 27, 2025
Язык: Английский
Accuracy of cephalometric landmark identification by artificial intelligence platform versus expert orthodontist in unilateral cleft palate patients: A retrospective study
International Orthodontics,
Год журнала:
2025,
Номер
23(2), С. 100990 - 100990
Опубликована: Фев. 19, 2025
Язык: Английский
The Association Between Craniofacial Morphological Parameters and the Severity of Obstructive Sleep Apnea: A Multivariate Analysis Using the Apnea–Hypopnea Index and Nocturnal Oxygen Desaturation
Healthcare,
Год журнала:
2025,
Номер
13(8), С. 913 - 913
Опубликована: Апрель 16, 2025
Background:
Obstructive
sleep
apnea
(OSA)
is
characterized
by
repetitive
complete
or
partial
closure
of
the
upper
airway
during
sleep,
which
a
potentially
life-threatening
disorder.
A
cephalogram
simple
and
effective
examination
to
predict
risk
OSA
in
orthodontic
clinical
practice.
This
study
aims
analyze
relationship
between
craniofacial
characteristics
severity
using
polysomnography
data.
Gender
differences
these
parameters
are
also
investigated.
Methods:
included
112
patients
who
underwent
examination,
standard
study,
cephalometric
analysis
diagnose
obstructive
apnea.
divided
participants
into
male
female
groups
correlation
OSA.
The
involved
39
parameters.
was
evaluated
apnea–hypopnea
index
(AHI)
lowest
nocturnal
oxygen
saturation
(LSaO2).
Results:
final
assessment
adult
(male/female
=
67:45,
mean
age:
28.4
±
7.29
years,
28.8
7.62
27.8
6.79
years).
Multivariate
revealed
that
mandibular
position,
incisor
inclination,
facial
height,
maxillary
first
molar
position
were
strongly
associated
with
severity.
Gender-specific
predictors
identified,
distinct
correlating
AHI
LSaO2
males
females.
Notably,
demonstrated
stronger
associations
morphology
females
than
males.
Conclusions:
Cephalometric
can
be
assessing
based
on
AHI/LSaO2.
There
clear
difference
individuals.
gender-dependent
pattern
may
assist
personalized
diagnosis
management
Язык: Английский