ORIGINAL ARTICLE

Two-dimensional vs 3-dimensional
comparison of alveolar bone over
maxillary incisors with A-point as a
reference
Theodore J. Kula, III,a Ahmed Ghoneima,a George Eckert,b Edwin T Parks,c Achint Utreja,a and Katherine Kulaa
Indianapolis, Ind

Introduction: Our objectives were to compare, relative to A-point, (1) bone thickness over the most forward
maxillary incisor (MFMI) in 2 dimensions vs 3 dimensions, and (2) bone thickness and inclination of each maxil-
lary incisor in 3 dimensions. Methods: Thirty-four cone-beam computed tomography (CBCT) images were
coded, and 2-dimensional (2D) cephalograms were derived from each image using Dolphin software (Dolphin
Imaging and Management Solutions, Chatsworth, Calif). A-point and the MFMI crown were located. After
reliability tests, alveolar bone buccal to 3 points on the MFMI root, bone to reference line Frankfort horizontal
(FH)–A-point, and incisor inclination were measured. This procedure was repeated on the 3-dimensional (3D)
CBCT images comparing MFMI with all maxillary incisors. The 2D and 3D measurements were compared
using paired t tests, and 3D measurements were compared with analysis of variance. A 5% significance level
was used for all tests. Results: The MFMI's buccal bone thickness at the root apices and the distance between
buccal bone and FH–A-point line at 2 root points were significantly greater in 2 dimensions than in 3 dimensions.
In 3 dimensions, bone thickness at MFMI's root apex and the distance from FH–A-point line at all root points were
significantly greater than those of the lateral incisors. Bone buccal to MFMI was significantly smaller than at the
lateral incisors 3 mm from the cementoenamel junction. Conclusions: Evaluation of 2D CBCT derivations can
result in overestimation of alveolar bone buccal to the maxillary incisor root apices compared with 3D evalua-
tions. The anterior nasal spine obscures bone measurements over the maxillary incisors in 2 dimensions.
(Am J Orthod Dentofacial Orthop 2017;152:836-47)

A
-point was originally called Point A by Downs.1
On a lateral cephalogram, it is the deepest midline
point on the premaxilla between the anterior
nasal spine (ANS) and prosthion,1 and it is used to deter-
mine the most forward position of the maxilla. However,
a
Department of Orthodontics and Oral Facial Genetics, School of Dentistry,
Indiana University, Indianapolis, Ind.
b
Department of Biostatistics, School of Medicine, Indiana University, Indianap-
olis, Ind.
c
Department of Oral Pathology, Medicine and Radiology, School of Dentistry,
Indiana University, Indianapolis, Ind.
All authors have completed and submitted the ICMJE Form for Disclosure of
Potential Conflicts of Interest, and none were reported.
Funding from the Indiana University School of Dentistry Graduate Research
Committee and the Jarabak Endowed Professorship.
Address correspondence to: Katherine Kula, Department of Orthodontics and
Oral Facial Genetics, Indiana University School of Dentistry, 1121 W Michigan
St, Indianapolis, IN 46202; e-mail, kkula@iu.edu.
Submitted, June 2016; revised and accepted, May 2017.
0889-5406/$36.00
Ó 2017 by the American Association of Orthodontists. All rights reserved.
http://dx.doi.org/10.1016/j.ajodo.2017.05.030
836
there are a number of problems with using A-point as a
landmark. For example, A-point can be difficult to iden-
tify reliably in 2 dimensions because of overlying struc-
tures and positioning of the head.2,3 A-point can be
affected by the movement of the anterior incisors.4 It is
influenced by the position of the ANS.5 Van der Linden5
believed that the position of ANS caused the deepest point
on the concavity to be more superior than the alveolar
bone of the maxilla. Since ANS is a midline structure, it
could mask the amount of bone over the apical portion
of the incisor root when viewed in 2 dimensions, thus
obscuring the amount of labial bone present for labial
root movement.
Another problem in measuring the amount of bone
over the incisors is the identification of the actual root
of the most forward maxillary incisor (MFMI) using a 2-
dimensional (2D) radiograph. Overlying structures such
as the roots of the other incisors, the canine, and even
other alveolar bone—ie, the canine eminence—can obscure
the root outline of the MFMI. Also, the amount of bone
Kula et al
over the other incisors cannot be determined using 2D
radiographs. Several authors have reported the amount
of bone over the maxillary incisors using cone-beam
computed tomography (CBCT) images,6-9 but there is
little information concerning the location of A-point
and the amount of bone over the maxillary incisors.
Little information comparing 2D with 3-dimensional
(3D) alveolar measures exists. Lack of recognizing the
amount of bone over the maxillary incisors can
compromise treatment when the roots are orthodonti-
cally moved through the bone.
The purposes of the study were to compare the
amount of bone located over the MFMI in 2 dimensions
vs 3 dimensions and to determine the influence of A-point
on the measurements by relating the distance of the alve-
olar bone to A-point. Another purpose was to compare in
3 dimensions the amount of bone and the inclination of
each maxillary incisor and the relationship with A-point.
MATERIAL AND METHODS
This retrospective study was approved by the Indiana
University Institutional Review Board. After a sample
size calculation, 35 CBCT images (27 male, 7 female;
white ethnicity; age range, 18-37 years) were randomly
selected from the pool of initial research records avail-
able in the university's graduate orthodontic clinic. All
3D CBCT images were taken with the same machine
(iCAT; Imaging Sciences International, Hatfield, Pa),
set for a full 13-cm field of view, 8.9-second scanning
time, 120 kV(p), 18 mA, and a resolution of 0.3 mm voxel
size. Inclusion criteria included eruption of all perma-
nent teeth anterior to the first molar. All malocclusions
(20 Class I, 6 Class II, 8 Class III, based on Angle classifi-
cations) were accepted. Exclusion criteria included
impacted maxillary anterior teeth, craniofacial abnor-
malities, severely resorbed roots, previous orthodontic
treatment, and noticeable periodontal diseases. Notice-
able periodontal disease was based on vertical bone
defects or alveolar bone greater than 3 mm from the
cementoenamel junction (CEJ), as diagnosed from the
constructed panoramic radiograph. During the study, 1
CBCT image was discarded because of poor resolution
in the area of measurement, resulting in 34 images.
Radiographs were coded and randomized so that
the investigator (T.J.K.) was blinded to their identity.
The CBCT files (DICOM) were exported into Dolphin Im-
aging (version 11.8; Dolphin Imaging and Management
Solutions, Chatsworth, Calif) for measurements.
A 2D cephalogram was created from the CBCT image
in Dolphin. The 2D derivations of all CBCT images were
standardized using the same Dolphin Mask Filter 1. A
standardized horizontal position (Frankfort Horizontal
American Journal of Orthodontics and Dentofacial Orthopedics
837
Fig 1. Two-dimensional image of A-point and the FH–A-
point line with the MFMI showing measurements of overlying
bone over the 3 points on the root and the measurement of
the distance from FH–A-point to the edge of the overlying
bone on a line parallel to the FH–A-point line.
[FH]) was constructed using the patient's right porion
and right orbitale on the cephalogram.1 A perpendicular
line was drawn from FH through A-point (FH–A-point
line) as a vertical reference for tooth and bone measures.
The MFMI was located and the root measured from the
labial CEJ to the apices. Three points along the root (root
apex, half the length of the root, and 3 mm gingival to
the CEJ) were determined. The thickness of the alveolar
bone buccal to the 3 points on the roots was measured
on a line horizontal to the FH. Then the distance along
the same lines from the edge of the bone to the FH–A-
point line was measured (Fig 1). The inclination of the
incisor was measured as an angle between a line from
the incisor tip through its apex to the FH.
On the 3D CBCT image (Fig 2), the FH line was
marked, using the right porion and orbitale. Perpendic-
ular lines were made (as follows) to standardize
measures and relate A-point to all the incisors. The
radiograph was oriented frontally with the FH parallel
to the floor, and a vertical orientation (midsagittal
plane) was made with the skeletal midline through
nasion perpendicular (Fig 3). A-point was identified on
the right lateral view of the 3D volumetric image and
transferred to the frontal view. From the frontal view,
a perpendicular line through A-point was then made
to the FH (FH–A-point line). A line horizontal to the
FH and perpendicular to the FH–A-point line at A-point
December 2017  Vol 152  Issue 6
838
Fig 2. Three-dimensional image of the axial orientation of a skull along the FH.
was drawn across the frontal view. A-point was marked
at each maxillary incisor. The MFMI was identified and the
midsagittal line transferred to the center of its crown. Root
length was measured from the labial CEJ to the root apex.
On a sagittal 0.3-mm section through the midsagittal line
of the incisor crown, 3 points were identified (root apex,
half of the root length, and 3 mm apical to the CEJ),
similar to the 2D cephalogram. A perpendicular line was
drawn from the FH through A-point (Fig 4) via the
Dolphin software. The FH–A-point line was transferred
via software to the midsagittal section of each incisor.
The thickness of the alveolar bone buccal to the 3 points
on the roots was measured on a line horizontal to the FH
(Fig 5, A). Then the distance from the edge of the bone to
the FH–A-point line was measured along the same lines
(Fig 5, B). The inclination of the incisor was measured
as the angle between a line from the incisal tip, extending
through the root apex, to the FH (Fig 6). The other maxil-
lary incisors were located using the previously mentioned
points, and the same measurements as described for the
MFMI were performed. A total of 34 central incisors
that were not the most forward incisor and 68 lateral
incisors (34 right, 34 left) were measured. The uncertainty
of the delineation of the bone edge on 1 MFMI affecting 2
parameters led to a decision to exclude those values,
resulting in only 33 measurements for the MFMI at (1)
December 2017  Vol 152  Issue 6 American Journal of Orthodontics and Dentofacial Orthopedics
Kula et al
root apex to bone and (2) A-point to bone at root apex.
Only the Dolphin magnification function was used to
enhance the image of the incisor during measurement.
Before data collection, a reliability test was
performed using both 2D and 3D radiographs. One
investigator (T.J.K.) measured 10 randomly selected ra-
diographs twice, 2 weeks apart.
Statistical analysis
Intraclass correlation coefficients (ICCs), Bland-
Altman plots, and measurement errors were calculated
to evaluate the reliability of the 2D and 3D radiograph
measurements and to check for patterns of disagree-
ment in the reliability measures. Normality of the data
was examined and determined to be acceptable. Com-
parisons between the 2D and 3D measurements were
made using paired t tests. To determine whether sex
and occlusion affected the comparisons between 2
and 3 dimensions, analysis of variance (ANOVA) was
used, with fixed effects for 2 dimensions/3 dimensions,
sex, occlusion, and their interactions. A random subject
effect was used to create the 2D-3D pairing in ANOVA.
Comparisons between the MFMI and the other maxillary
incisor roots were made using ANOVA tests with a fixed
effect for maxillary incisor root; a random subject effect
was included to allow correlations among the roots
Kula et al 839

Fig 3. Three-dimensional image of the midsagittal orientation of a skull along the skeletal midline and
nasion.

Fig 4. Three-dimensional image of a sagittal slice showing the FH–A-point line extended from the
facial midline to the middle of a central incisor.

within a subject. To control the overall significance
level, pair-wise tests between the roots were only con-
ducted when the overall effect was significant in AN-
OVA. Pearson correlation coefficients were calculated
to evaluate the associations of the amount of bone
with the incisor inclination. A 5% significance level
was used for all tests.
Before the study, sample size calculations showed
that with a sample of 35 subjects, with 2D and 3D
images evaluated for each subject, the study would

American Journal of Orthodontics and Dentofacial Orthopedics December 2017  Vol 152  Issue 6
840 Kula et al

Fig 5. Three-dimensional image of sagittal slices with the FH–A-point line extended from the facial
midline to the middle of a central incisor: A, measurement of overlying bone over the 3 points on the
root, and B, measurement of the distance from FH–A-point to the edge of the overlying bone on lines
parallel to the FH–A-point line.

Fig 6. Three-dimensional image of the inclination of the incisors to the FH.

have 80% power to detect an effect size of 0.5 for the
difference between the 2D and 3D measurements,
assuming 2-sided tests, each conducted at a 5% signif-
icance level. However, near the end of the study, 1 sub-
ject was dropped because of resolution of the 3D image,
resulting in 34 subjects.
RESULTS
In 2 dimensions (Table I), 5 measurements had excel-
lent ICC values, 0.90 or greater, and 1 was 0.79. Two had
moderate-strength ICC values of 0.59 to 0.70. In 3
dimensions, 6 measurements had strong ICC values,

December 2017  Vol 152  Issue 6 American Journal of Orthodontics and Dentofacial Orthopedics
Kula et al 841

Table I. Reliability measures for 2 dimensions vs 3 dimensions

2D 3D
1
Measurement ICC (95% CI) ME ICC (95% CI) ME
Root length 0.99 (0.97-1.00) 0.23 0.71 (0.40-1.00) 0.53
Root apex to bone 0.96 (0.92-1.00) 0.42 0.86 (0.69-1.00) 0.41
One-half root to bone 0.79 (0.54-1.00) 0.27 0.43 ( 0.09 to 0.95) 0.28
Root 3 mm from CEJ to bone 0.66 (0.30-1.00) 0.28 0.66 (0.30-1.00) 0.30
A-point to bone at root apex 0.59 (0.17-1.00) 0.20 0.85 (0.67-1.00) 0.28
A-point to bone at 1/2 root 0.90 (0.77-1.00) 0.35 0.88 (0.74-1.00) 0.28
A-point to bone 3 mm from 0.99 (0.97-1.00) 0.16 0.88 (0.74-1.00) 0.43
CEJ
Incisor angle to FH 0.98 (0.95-1.00) 0.92 0.99 (0.99-1.00) 0.47
ME, Measurement error (standard deviation of the within-image repeats representing the estimate of the measurement error; the confidence in-
tervals are approximately 6 2 measurement errors from the estimated difference between the repeated measurements).

greater than 0.70, and 2 had moderate strength.
Measurement error was also determined (Table I).
Measurement error was the standard deviation of the
within-image repeats, and 95% confidence intervals
(CI) were considered approximately 6 2 measurement
errors from the estimated difference between the
repeated measurements. Bland-Altman plots showing
the 95% CI had no underlying patterns of disagreement
in the reliability data (Supplemental material, Figs 7 and
8). We concluded that there were no significant
differences between the 2D and 3D ICC values.
Summary statistics were developed (Table II). The
root length of the MFMI was significantly greater in
the 2D cephalograms than in the 3D ones (Table III).
The thickness of bone at the root apices was significantly
greater in 2 dimensions than in 3 dimensions. The dis-
tance from A-point to the labial aspect of the bone
was significantly greater in 2 dimensions vs 3
dimensions at the root apex and at 3 mm from the
CEJ. No other 2D vs 3D measures were significantly
different.
Occlusion affected the 2D vs 3D comparison (Table
IV). The inclination of the MFMI was significantly
greater in 2 dimensions than in 3 dimensions for Class
III subjects (P 5 0.0029), but not for Class I
(P 5 0.15) or Class II (P 5 0.10) subjects. The 2D vs
3D comparison results were not affected by occlusion
for any other measurement or by sex for any measure-
ment.
The distance for the MFMI was significantly greater
from the root apex to the outer labial bone and from
the A-point line to bone at all 3 root points. The
distance of the root 3 mm from the CEJ to bone was
significantly smaller for the MFMI than for either lateral
incisor (Tables V and VI). No other differences were
found between the MFMI and the other maxillary
incisor roots.
Inclination (Table VII) was positively correlated with
the bone thickness labial to the root apex in 2 and 3
dimensions and the distance of A-point to bone 3 mm
from the CEJ in 2 and 3 dimensions, but the
correlations were weak to moderate. No other correla-
tions with inclination were statistically significant
(P .0.16).
DISCUSSION
The thickness of bone buccal to the root apices of
the MFMI is greater when measured on a 2D cephalo-
gram derived from the CBCT image compared with
the 3D CBCT image. This difference is probably related
to the ANS projecting from the midline of the maxilla
and obscuring in 2 dimensions the bone labial to the
root apices. The significant differences between 2D
and 3D measures of the distance from the transposed
FH–A-point line to the labial bone plate reinforces
this concept. In 2 dimensions, the bone is essentially
contiguous with A-point at the root apex of the
MFMI, whereas in 3 dimensions, the labial bone is
significantly posterior to A-point at the root apex. At
the root apices, only 1 subject had alveolar bone ante-
rior to A-point in 3 dimensions whereas in 2 dimensions
only 1 subject had bone posterior to A-point. This
suggests that clinicians who are concerned with the
amount of bone at incisor root apices if performing
labial root torque should be cautious when analyzing
the 2D rendition of a CBCT image.
Fuentes et al7 measured the bone buccal to the
maxillary incisors of Chileans at 5 points from the root
apex to the bone crest near the CEJ using CBCT but
did not select the MFMI. Similar to our study, the
average bone thickness buccal to all maxillary incisors
in their study was greatest at the root apex and then
decreased toward the crown with the thickness of bone
varying at midroot. The small differences in bone

American Journal of Orthodontics and Dentofacial Orthopedics December 2017  Vol 152  Issue 6
842 Kula et al

Table II. Summary statistics of 2D and 3D measures of root length, thickness of alveolar bone labial to root, distance
from bone to A-point along maxillary incisors, and incisor angulation
Measurement Location n Mean SD Minimum Maximum
Root length (mm) 2D most forward crown 34 14.9 1.9 11.8 20
3D most forward crown 34 13.7 1.6 10.2 16.2
3D left lateral incisor 34 13.3 1.7 9.8 17.4
3D left central incisor 17 13.9 1.8 9.8 16.8
3D right central incisor 17 13.8 1.8 10.6 17
3D right lateral incisor 34 13.5 1.6 10 17.8
Root apex to bone (mm) 2D most forward crown 34 5.5 1.9 1.3 9.3
3D most forward crown 33 3.2 1.1 1.5 6.9
3D left lateral incisor 34 2.3 0.9 1.1 4.7
3D left central incisor 17 3.3 1.3 1.9 7.1
3D right central incisor 17 3.2 1.1 1.8 5
3D right lateral incisor 34 2.2 0.7 1 4.1
1/2 root to bone (mm) 2D most forward crown 34 2.0 0.8 0.5 4.5
3D most forward crown 34 1.8 0.6 0.8 3.3
3D left lateral incisor 34 1.9 0.7 0.8 4.9
3D left central incisor 17 1.7 0.6 0.9 3.3
3D right central incisor 17 1.7 0.3 1.2 2.3
3D right lateral incisor 34 2.0 0.7 0.8 4.5
Root 3 mm from CEJ to bone (mm) 2D most forward crown 34 1.3 0.4 0.4 2.2
3D most forward crown 34 1.4 0.4 0.6 2.2
3D left lateral incisor 34 2.0 0.7 0.5 4
3D left central incisor 17 1.4 0.4 0.7 1.9
3D right central incisor 17 1.4 0.4 0.8 2
3D right lateral incisor 34 1.9 0.6 1 3
A-point to bone at root apex (mm) 2D most forward crown 34 0.3 0.3 0.4 1.2
3D most forward crown 33 1.9 1.1 4.1 0.2
3D left lateral incisor 34 5.1 1.4 7.9 0.6
3D left central incisor 17 1.5 1.4 3.8 0.6
3D right central incisor 17 1.8 1.4 4.4 0.6
3D right lateral incisor 34 5.0 1.3 7.5 1.9
A-point to bone at 1/2 root (mm) 2D most forward crown 34 1.7 1.0 0.2 4.1
3D most forward crown 34 1.3 1.0 1.5 3.7
3D left lateral incisor 34 1.1 1.1 4.2 1.3
3D left central incisor 17 1.7 1.1 0.3 3.8
3D right central incisor 17 1.0 0.9 0.4 2.5
3D right lateral incisor 34 0.9 1.4 4.3 2.1
A-point to bone at 3 mm from CEJ (mm) 2D most forward crown 34 3.6 1.4 1.1 6.7
3D most forward crown 34 3.1 1.2 1.1 5.9
3D left lateral incisor 34 0.8 1.1 1.2 3.5
3D left central incisor 17 3.5 1.3 1.1 5.9
3D right central incisor 17 2.7 0.9 1.2 3.9
3D right lateral incisor 34 0.9 1.4 1.2 4.2
Angle to FH ( ) 2D most forward crown 34 116.4 7.8 98.2 133.8
3D most forward crown 34 116.1 7.5 97.5 131.5
3D left lateral incisor 34 116.0 7.1 94.1 127
3D left central incisor 17 115.2 9.2 91.6 127.1
3D right central incisor 17 113.1 6.1 102.8 122.3
3D right lateral incisor 34 115.7 7.4 96.8 127.7

thickness between our studies might be explained by the
angle at which the measurement was taken. Fuentes et al
measured bone thickness perpendicular from the root
surface to the bone surface, whereas we measured
bone thickness from the root on a perpendicular to a
standardized line. Also, Fuentes et al excluded CBCT
images with buccal bone over 3 mm as potentially path-
ological, whereas we did not exclude subjects based on
bone thickness. Lee et al10 also reported similar changes
in the buccal bone over the roots of the maxillary incisors
of Koreans. Nowzari et al6 reported similar findings for
central incisor roots. Their results showed considerable
variability that could be explained by the inclusion of
multiple ethnic groups in their sample.
Similar to the study of Sherrard et al,11 who measured
the tooth from a frontal perspective using periapical

December 2017  Vol 152  Issue 6 American Journal of Orthodontics and Dentofacial Orthopedics
Kula et al 843

Table III. Comparisons between 2D and 3D most forward crown with paired t tests
Measurement Mean difference SD 95% CI P value
Root length 1.16 1.01 0.81 1.51 \0.0001*
Root apex to bone 2.30 1.74 1.70 2.91 \0.0001*
1/2 root to bone 0.23 0.77 0.04 0.50 0.0901
Root 3 mm from CEJ to bone 0.09 0.39 0.23 0.04 0.1704
A-point to bone at root apex 2.18 1.11 1.79 2.57 0.0001*
A-point to bone at 1/2 root 0.36 1.17 0.05 0.77 0.0837
A-point to bone 3 mm from CEJ 0.51 1.00 0.17 0.86 0.0051*
Angle to FH 0.31 3.41 0.88 1.50 0.6008*
The table shows the means, standard deviations, and 95% confidence intervals for the differences between 2 and 3 dimensions, calculated as 2
dimensions minus 3 dimensions, along with the P value from the test, n 5 34.
*Significant.

radiographs, the MFMI root length was longer in 2
dimensions than in 3, possibly because the 2D derivation
of the CBCT image was a composite of multiple sections
whereas the 3D image evaluated only one 0.3-mm
section of the root. Typically, incisor roots tip slightly
to the distal aspect, allowing them to be identified as
right or left, and a midline section of the tooth might
have not have included a small amount of root apex.
Extrapolation from the data was not considered accept-
able because of the limited number of data points and
unknown changes in unmeasured bone coverage. This
standardization of method could be considered a limita-
tion of our study. The 3D data show that the maxillary
incisors had similar length roots, which could easily
overlay the root of the MFMI and cause problems with
identification in 2 dimensions.
Incisor inclination was not significantly different
between 2D and 3D images overall. However, the incli-
nation of the incisor was increased for 2D vs 3D images
for subjects with Class III molars. The positive correla-
tion of inclination with the amount of bone both at
the root apices and 3 mm from the CEJ was significant,
but of moderate strength. Our findings corroborated
the relationships between incisor inclination and
bone thickness along the incisor roots in Chinese sub-
jects.9 The findings suggest that in both 2 and 3 dimen-
sions, proclined incisors have more buccal bone at the
root apices of the MFMI than do upright or reclined in-
cisors, thus allowing more labial root inclination during
treatment.
In 3 dimensions, the distance from the A-point line to
bone increased apically at 3 points along the MFMI root,
meaning that ANS affected the measurement of primar-
ily the bone over the root apices. However, 2 subjects had
measurable amounts of bone posterior to A-point at the
midroot point, suggesting that ANS can affect 2D radio-
graphic interpretation in some patients more than
others. The labial surface of bone at the root apex was
posterior to A-point for all maxillary incisor and lateral
incisor roots more so than for central incisor roots.
Similar to the results of Lee et al,10 the MFMI had less
bone over the root 3 mm above the CEJ when compared
with the lateral incisors. Fuentes et al7 found more bone
over the central incisors compared with the lateral inci-
sors, but the MFMI was not measured. These findings
suggest that ANS hides buccal bone thickness when 2D
cephalograms are evaluated and could potentially affect
treatment decisions.
The lateral incisors frequently need additional labial
root torque compared with the central incisors because
of their lingual position in the arch. This can present a
problem, since the lateral incisor roots have less bone
buccal to the root apices than do the central incisors.
The lateral incisors are more posterior from A-point
compared with the central incisors, and 2D radiographs
might allow the interpretation that these roots could
undergo more torque. In actuality, a dehiscence or
fenestration of the buccal bony plate could occur.
Based on tracing of 2D lateral cephalograms, Jacob-
son and Jacobson2 indicated that A-point should be
situated 3 mm anterior to the junction of the upper third
and lower two thirds of the central incisor. Similarly, our
study shows that in young adult white subjects, A-point
is not at the root apices of the MFMI. Rather, A-point is
coronal to the maxillary incisor root tip and generally
apical to the midpoint of the MFMI root. This research
suggests that, at least in white patients, the anterior
portion of the maxilla over the incisors recedes from
A-point.
Our decision to measure alveolar bone at a distance
of 3 mm from the CEJ was based on reports that buccal
bone generally is 3 mm from the CEJ.6,10 Similarly, all
our subjects had alveolar bone within 3 mm of the
CEJ. A wider cuff of bone exists near the CEJ of the
lateral incisors than at the central incisors, although
this was not tested statistically.10
Reliability in 2 dimensions was excellent for some
measures but was not as good when the measures

American Journal of Orthodontics and Dentofacial Orthopedics December 2017  Vol 152  Issue 6
844 Kula et al

Table IV. ANOVA table for 2D vs 3D comparisons with sex and occlusion factors (n 5 34)
Measurement Effect Num df Den df F value P value
Root length 2Dv3D 1 28 31.84 \0.0001
Occlusion 2 28 1.96 0.1602
2Dv3D*Occlusion 2 28 3.59 0.0409
Sex 1 28 7.77 0.0094
2Dv3D*Sex 1 28 2.05 0.1632
Occlusion*Sex 2 28 2.01 0.1534
2Dv3D*Occlusion*Sex 2 28 0.13 0.8829
Root apex to bone 2Dv3D 1 27 36.34 \0.0001
Occlusion 2 27 0.58 0.5673
2Dv3D*Occlusion 2 27 2.28 0.1215
Sex 1 27 4.33 0.0471
2Dv3D*Sex 1 27 1.06 0.3122
Occlusion*Sex 2 27 0.65 0.5278
2Dv3D*Occlusion*Sex 2 27 0.71 0.4990
1/2 root to bone 2Dv3D 1 28 5.68 0.0242
Occlusion 2 28 0.04 0.9653
2Dv3D*Occlusion 2 28 0.30 0.7465
Sex 1 28 0.73 0.3999
2Dv3D*Sex 1 28 2.13 0.1555
Occlusion*Sex 2 28 1.64 0.2127
2Dv3D*Occlusion*Sex 2 28 1.32 0.2826
Root 3 mm from CEJ to bone 2Dv3D 1 28 0.03 0.8662
Occlusion 2 28 0.16 0.8550
2Dv3D*Occlusion 2 28 0.07 0.9348
Sex 1 28 0.06 0.8053
2Dv3D*Sex 1 28 1.94 0.1746
Occlusion*Sex 2 28 1.66 0.2085
2Dv3D*Occlusion*Sex 2 28 0.06 0.9463
A-point to bone at root apex 2Dv3D 1 27 53.49 \0.0001
Occlusion 2 27 1.34 0.2790
2Dv3D*Occlusion 2 27 2.82 0.0775
Sex 1 27 2.12 0.1572
2Dv3D*Sex 1 27 5.46 0.0271
Occlusion*Sex 2 27 2.45 0.1057
2Dv3D*Occlusion*Sex 2 27 2.75 0.0820
A-point to bone at 1/2 Root 2Dv3D 1 28 0.08 0.7758
Occlusion 2 28 0.65 0.5276
2Dv3D*Occlusion 2 28 0.45 0.6428
Sex 1 28 3.63 0.0672
2Dv3D*Sex 1 28 1.16 0.2907
Occlusion*Sex 2 28 1.85 0.1753
2Dv3D*Occlusion*Sex 2 28 1.33 0.2795
A-point to bone from 3 mm from CEJ 2Dv3D 1 28 2.95 0.0971
Occlusion 2 28 4.84 0.0156
2Dv3D*Occlusion 2 28 0.63 0.5426
Sex 1 28 1.01 0.3241
2Dv3D*Sex 1 28 0.01 0.9251
Occlusion*Sex 2 28 3.93 0.0314
2Dv3D*Occlusion*Sex 2 28 1.04 0.3684
Angle to FH 2Dv3D 1 28 7.12 0.0125
Occlusion 2 28 0.61 0.5478
2Dv3D*Occlusion 2 28 7.11 0.0032
Sex 1 28 0.06 0.8121
2Dv3D*Sex 1 28 3.97 0.0560
Occlusion*Sex 2 28 3.08 0.0617
2Dv3D*Occlusion*Sex 2 28 2.14 0.1365
Den, denominator; Num, numerator.

December 2017  Vol 152  Issue 6 American Journal of Orthodontics and Dentofacial Orthopedics
Kula et al 845

Table V. ANOVA table for tests among 3D locations

Measurement Effect Num df Den df F value P value
Root length Location 4 98 1.67 0.1621
Root apex to bone Location 4 97 13.60 \0.0001
1/2 root to bone Location 4 98 1.95 0.1074
Root 3 mm from CEJ to bone Location 4 98 12.78 \0.0001
A-point to bone at root apex Location 4 97 110.95 \0.0001
A-point to bone at 1/2 root Location 4 98 91.62 \0.0001
A-point to bone from 3 mm from CEJ Location 4 98 142.48 \0.0001
Angle to FH Location 4 98 2.41 0.0544
Den, denominator; Num, numerator.

Table VI. Pairwise tests for 3D images for other incisor locations against most forward crown
Measurement Incisor Difference SD 95% CI P value
Root length Left central 0.11 1.27 0.44 0.67
Left lateral 0.41 1.27 0.85 0.02
Right central 0.10 1.27 0.46 0.66
Right lateral 0.22 1.27 0.65 0.22
Root apex to bone Left central 0.11 1.12 0.37 0.60 0.6474
Left lateral 0.98 1.12 1.36 0.59 \0.0001*
Right central 0.09 1.12 0.58 0.39 0.7063
Right lateral 1.03 1.12 1.41 0.64 \0.0001*
1/2 root to bone Left central 0.13 0.60 0.39 0.13
Left lateral 0.11 0.60 0.10 0.31
Right central 0.05 0.60 0.21 0.31
Right lateral 0.20 0.60 0.01 0.40
Root 3 mm from CEJ to bone Left central 0.07 0.63 0.34 0.20 0.6053
Left lateral 0.59 0.63 0.37 0.80 \00001*
Right central 0.06 0.63 0.21 0.33 0.6357
Right lateral 0.49 0.63 0.27 0.70 \0.0001*
A-point to bone at root apex Left central 0.17 1.28 0.39 0.73 0.5434
Left lateral 3.24 1.28 3.68 2.80 \0.0001*
Right central 0.24 1.28 0.32 0.80 0.4038
Right lateral 3.12 1.28 3.56 2.68 \0.0001*
A-point to bone at 1/2 root Left central 0.04 1.00 0.48 0.40 0.8505
Left lateral 2.42 1.00 2.76 2.08 \0.0001*
Right central 0.09 1.00 0.35 0.53 0.6881
Right lateral 2.21 1.00 2.55 1.86 \00001*
A-point to bone from 3 mm from CEJ Left central 0.12 0.77 0.46 0.22 0.4748
Left lateral 2.30 0.77 2.56 2.04 \0.0001*
Right central 0.08 0.77 0.26 0.41 0.6598
Right lateral 2.19 0.77 2.46 1.93 \0.0001*
Angle to FH Left central 1.42 4.25 3.28 0.44
Left lateral 0.04 4.25 1.49 1.40
Right central 2.37 4.25 4.23 0.50
Right lateral 0.38 4.25 1.83 1.07
The table shows the model-based means, standard deviations, and 95% confidence intervals for the differences between the other incisors and the
most forward crown, calculated as other incisor minus most forward crown. P values for comparisons of root length, angle to FH, and 1/2 root to
bone are not shown because the overall F test showed no significant differences among locations.
*Significant (P 5 0.05).

were to the edge of the alveolar bone. The reliability of
3D measures also varied for similar measures. One
contributing factor to lower reliability might be that
the labial cortical plate at the midroot of maxillary
incisor roots might not be as thick as the alveolar bone
over the root apices, making it difficult to identify the
edge effect of the cortical plate. Since the root apices
were farther from the labial bony edge, the cortical plate
could be thicker and easier to identify. A voxel size of
0.3 mm also adds to the potential for lower reliability.
However, most parameters showed greater differences
between 2D and 3D measures than the reliability

American Journal of Orthodontics and Dentofacial Orthopedics December 2017  Vol 152  Issue 6
846 Kula et al

mask is used to increase the resolution of an image or

Table VII. Correlations (Pearson) with incisor angle to
sharpen an image that tends to blur because of pixila-
FH
tion.14 The gray of the pixels at the edge of 2 structures
2D 3D tends to be an average of the 2 structures and causes
P P
blur. Unsharp mask tends to change the opaqueness
Measurement Correlation value Correlation value of those pixels pending the percentage of each of the
Root length 0.19 0.2958 0.19 0.2756 2 structures. This can sharpen the edge but, potentially,
Root apex to bone 0.42 0.0137* 0.39 0.0258* can cause edge problems that relate to measurements.
1/2 root to bone 0.24 0.1665 0.12 0.4940 Although it is possible that this outline affected the
Root 3 mm from 0.16 0.3716 0.18 0.2978
CEJ to bone
measures at the bone, we did not think it caused major
A-point to bone 0.01 0.9599 0.15 0.3957 problems because our measures were to the bony edge
at root apex and, if anything, might make the 2D measures at the
A-point to bone 0.22 0.2071 0.17 0.3502 bone smaller and more similar to the 3D measures.
at 1/2 root The bone near the CEJ often bulged compared with
A-point to bone 0.37 0.0290* 0.37 0.0282*
the bone over the middle of the root. A slight change,
3 mm from CEJ
perhaps even a voxel-size change, in the placement of
*Significant correlation, P \0.05. the midsagittal slice on each tooth could cause a differ-
ence in measurement due to identifying the edge of a
measures. None of the authors reporting the thickness of
different part of the bulging bony curve. Similarly,
bone labial to the maxillary incisors included intrainves-
Baumrind and Frantz3 found that landmarks on a curve
tigator reliability studies in their published articles so
were not reliable in 2D conventional cephalograms.
that it is difficult to compare reliabilities.6,7,9,10
Although magnification is a factor in conventional 2D
Moreover, Timock et al12 showed that reliability of alve-
radiographs such as periapicals leading to increased
olar bone thickness measures using CBCT images,
root length measures compared with CBCT images, the
although excellent, was not as good as measures of alve-
2D radiographs in our study were extracted from CBCT
olar bone height. Sun et al13 suggested that the similar-
data and were not conventional or digital 2D radio-
ity between bone and cementum radiodensity can
graphs.14 Although the differences were small, Sherrard
negatively influence landmark identification more for
et al11 reported a greater difference when repeated mea-
bone thickness than for bone height when the bone is
sures were taken at 2 times vs repeated measures the
compared with soft tissue.
same day.
In addition, a dark line was noted, particularly on
The lack of a conventional 2D cephalogram of the same
the printed images, at the periphery of the crowns of
person is a limitation of this study. A prospective study us-
some teeth and for a short distance on the bone of
ing conventional films and an additional CBCT image of
some subjects. This shade also appeared to be sharper
the same patient would increase radiation exposure and
and more delineated on the printed image, particularly
could result in positioning problems. Some clinicians stan-
when magnified considerably, than on the monitor.
dardly take only 3D CBCT images and derive the 2D infor-
This might be explained by the fact that the resolution
mation usually from panoramic, cephalometric, and
of the printed and saved images is different from the
periapical radiographs from the CBCT image. This study
resolution of the computer monitor where actual
shows the advantage of using strictly 3D images. Although
measurements were performed. The outline was more
conventional 2D cephalograms should have better resolu-
obvious on some 2D constructed cephalograms more
tion than 3D CBCT images, they do not allow isolation of
so than others and appeared to be heaviest around
various areas such as a maxillary incisor, a nonmedian
the crown of a tooth compared with the bony edge. It
structure. Magnification would also be an issue with
is possible also that the lips were not touching the
a conventional 2D radiograph compared with a 3D image.
crowns of the teeth in all patients, and this lack of
contact is shown as a radiolucency (black), but the These data strongly support the use of 3D CBCT
black lines were not noted in the 3D CBCT images. images to evaluate the thickness of bone over maxillary
These shades exist only in the CBCT constructed ceph- incisor roots rather than 2D cephalograms derived from
alograms, and they are related to the algorithm that is 3D CBCT images.
used to convert 3D into 2D images. By default, the
CONCLUSIONS
Dolphin x-rays module includes 6 mask filters. In our
study, we standardized measurements by using the When 2D cephalograms extracted from 3D CBCT
same mask filter (Dolphin 1) across the sample. Unsharp images are compared with 3D images, (1) the

December 2017  Vol 152  Issue 6 American Journal of Orthodontics and Dentofacial Orthopedics
Kula et al
thickness of alveolar bone overlying the root apices of
the maxillary central incisors appears to be overesti-
mated in 2D cephalograms compared with 3D CBCT
images; (2) ANS interferes with bone measurements
over maxillary incisors in 2 dimensions; and (3) as
the inclination of the incisors increases, the thickness
of alveolar bone overlying the maxillary incisor root
apices increases.
REFERENCES
1. Downs WB. Variations in facial relationships; their significance in
treatment and prognosis. Am J Orthod 1948;34:812-40.
2. Jacobson RL, Jacobson A. Point A revisited. Am J Orthod 1980;77:
92-6.
3. Baumrind S, Frantz RC. The reliability of head film measurements.
1. Landmark identification. Am J Orthod 1971;60:111-27.
4. Bicakci AA, Cankaya OS, Mertoglu S, Yilmaz N, Altan BK. Does
proclination of maxillary incisors really affect the sagittal position
of point A? Angle Orthod 2013;83:943-7.
5. van der Linden FP. A study of roentgenocephalometric bony land-
marks. Am J Orthod 1971;59:111-25.
6. Nowzari H, Molayem S, Chiu CH, Rich SK. Cone beam computed
tomographic measurement of maxillary central incisors to deter-
mine prevalence of facial alveolar bone width $2 mm. Clin
Implant Dent Relat Res 2012;14:595-602.
7. Fuentes R, Flores T, Navarro P, Salamanca C, Beltran V, Borie E.
Assessment of buccal bone thickness of aesthetic maxillary region:
American Journal of Orthodontics and Dentofacial Orthopedics
847
a cone-beam computed tomography study. J Periodontal Implant
Sci 2015;45:162-8.
8. Behnia H, Motamedian SR, Kiani MT, Morad G, Khojasteh A.
Accuracy and reliability of cone beam computed tomographic
measurements of the bone labial and palatal to the maxillary
anterior teeth. Int J Oral Maxillofac Implants 2015;30:
1249-55.
9. Tian YL, Liu F, Sun HJ, Lv P, Cao YM, Yu M, et al. Alveolar bone
thickness around maxillary central incisors of different inclination
assessed with cone-beam computed tomography. Korean J Orthod
2015;45:245-52.
10. Lee SL, Kim HJ, Son MK, Chung CH. Anthropometric analysis of
maxillary anterior buccal bone of Korean adults using cone-
beam CT. J Adv Prosthodont 2010;2:92-6.
11. Sherrard JF, Rossouw PE, Benson BW, Carrillo R, Buschang PH.
Accuracy and reliability of tooth and root lengths measured on
cone-beam computed tomographs. Am J Orthod Dentofacial
Orthop 2010;137(Suppl):S100-8.
12. Timock AM, Cook V, McDonald T, Leo MC, Crowe J,
Benninger BL, et al. Accuracy and reliability of buccal bone height
and thickness measurements from cone-beam computed tomog-
raphy imaging. Am J Orthod Dentofacial Orthop 2011;140:
734-44.
13. Sun Z, Smith T, Kortam S, Kim D, Tee BC, Fields H. Effect of bone
thickness on alveolar bone-height measurements from cone-beam
computed tomography images. Am J Orthod Dentofacial Orthop
2011;139:e117-27.
14. Molen AD. Considerations in the use of cone-beam computed
tomography for buccal bone measurements. Am J Orthod Dento-
fac Orthop 2010;137(Suppl):S130-5.
December 2017  Vol 152  Issue 6
847.e1 Kula et al

SUPPLEMENTAL MATERIAL

Fig 7. Bland-Altman plots for reliability of 2D measures of MFMI: a, length of root; b, root apex to bone;
c, half-way point on root to bone; d, 3 mm from CEJ on root of bone; e, A-point to bone at root apex; f, A-
point to bone at half-way point on root; g, A-point to bone at 3 mm from CEJ; h, incisal angle to the FH.

December 2017  Vol 152  Issue 6 American Journal of Orthodontics and Dentofacial Orthopedics
Kula et al 847.e2

Fig 8. Bland-Altman plots for reliability of 3D measures of MFMI: a, length of root; b, root apex to bone;
c, half-way point on root to bone; d, 3 mm from CEJ on root to bone; e, A-point to bone at root apex; f, A-
point to bone at half-way point on root; g, A-point to bone at 3 mm from CEJ; h, incisal angle to the FH.

American Journal of Orthodontics and Dentofacial Orthopedics December 2017  Vol 152  Issue 6