Transcript Rio de Janeiro Corneal Tomography and Biomechanics Study Group

Agreement of Artificial Intelligence
Topographic Diagnosis From
Placido and Scheimpflug Systems
Leonardo N. Pimentel, MD1,2;
Isaac Ramos, MD1,2 ; Bruno Valbon, MD1,2
Ana Laura C. Canedo, MD1,2
Renato Ambrósio Jr., MD, PhD1,2
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group
1 – Instituto de Olhos Renato Ambrósio (Brazil)
2 – Rio de Janeiro Corneal Tomography and
Biomechanics Study Group (Brazil)
Dr. Ambrósio is consultant for Oculus Optikgeräte GmbH (Wetzlar, Germany)
Purpose
To compare artificial intelligence in the axial
curvature maps (sagittal), obtained by the
topographic survey of Placido reflection system and
those obtained using the rotating Scheimpflug system.
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group
Methods
Retrospective study.
449 patients (898 eyes) with axial topography by the
Nidek Corneal Navigator System using the Magellan
Placido’s Corneal Topographer and axial topography
by the Pentacam using the Scheimpflug system.
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group
Placido’s Corneal Topographer
The corneal topography based on the Placido disk
technology is a method that analyzes the anterior
curvature assisted by a computer through the recording,
capture and processing of the reflected rings that are
illuminated and projected onto the cornea.
Analysis of geometric shape of the reflected rings
provides twenty two thousands measuring points that
allow the construction of a topographic map.
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group
Scheimpflug Corneal
Topographer
The Scheimpflug scanning provides considerably larger area
measurement of the cornea compared with Placido’s
topography systems.
It is also not sensitive from irregularities on the corneal
reflex and interference from the tear film. The exam generates
curvature maps from the front and back corneal.
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group
Methods
To compare the artificial diagnoses, the
corresponding numbers were used, then the
percentage of agreement was observed.
Was used Artificial intelligence by Nidek Corneal
Navigator .
CASO #
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group
N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
A
KCS
KC
PMD
PKP
MRS
HRS
OTH
99
98,4
0,6
99
94,2
1,2
98,1
97,4
96,7
M
4,8
97,8
99
0,9
1,6
2,3
99
99
99
77,8
21,2
99
99
99
0
0
3
3
2
2
0
0
0
0
0
0
0
2
0
0
Artific
ial
Intelli
gence
N ou
AST
KCS
KC ou
PMD
PKP,
MRS
ou
HRS
OTH
Numb
er
corres
pondi
ng
0
0,5
1
3
2
Methods
CASE
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group
TKC
1 #2 #3 #4 # poss,
5 # poss,
6 # poss,
7 # 3-4
8 # 3-4
9 # 3,PMD?
10 # 3,PMD?
11 # Abnorm,
12 # Abnorm,
13 # Abnorm,
14 # Abnorm,
15 # C,Surg.?
16 # C,Surg.?
17 # C,Surg.?
Was used TKC system by Pentacam (Oculus).
The pacients that made hypermetropic Lasik
were excluded from the Study, because TKC
system didn`t have program for this cases.
Results
Observed agreement between diagnoses Artificial
by Maguellan (Nidek Cornea Navigator) and TKC
from Pentacam (Oculus) in 525 eyes (58%) and
disagreement in 373 eyes (48%). The agreement
between diagnoses Artificial by maguellan and TKC
system by Pentacam was better in normal and
keratoconus cases, while was worse in suspect cases.
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group
42% 58%
373 Eyes
525 Eyes
Conclusions
The axial topography by the Pentacam using
Scheimpflug system is not sensitive from irregularities
on the corneal reflex and interference from the tear
film.
Especially in suspicious cases, is necessary
additional data from tomography or biomechanics
study to enhance the sensitivity and specificity to
detect ectasia.
Rio de Janeiro
Corneal
Tomography
and
Biomechanics
Study Group