ORIGINAL ARTICLE

Apical root resorption: The dark side of the root

Marcio Jose
da Silva Campos,a Karine Simo
~ es Silva,b Marco Abdo Gravina,c Marcelo Reis Fraga,a
d
and Robert Willer Farinazzo Vitral
Juiz de Fora, Minas Gerais, Brazil

Introduction: The diagnosis of apical root resorption is usually based on routine radiographs. However, these
methods are limited because the images reflect the superimposition of the whole root structure and can lead to
underestimation of the extent of apical root resorption. In this study, we aimed to determine the lengths of the
labial and lingual surfaces of incisors with apical root resorption and compare them with the longest radicular
length obtained on sagittal images of cone-beam computed tomography, and to create a qualitative visual
scale of the different patterns of apical root resorption. Methods: Eighty-two incisors with apical root resorption
from 25 patients had their labial and lingual root surfaces and the longest radicular lengths determined in the
sagittal plane and compared. Five orthodontists, at 2 times, classified the images of each incisor according to
a visual scale developed by the authors. Results: There was no significant difference between the labial and
lingual surfaces; however, the longest radicular length was significantly greater than the shortest surface length.
The visual scale showed intraobserver agreement of 0.615 and interobserver agreements of 0.74 and 0.52 at
both times, respectively. Conclusions: The difference between the longest and shortest root lengths suggests
that radiographic superimposition underestimates the extent of the resorption lesion. The proposed visual scale
showed a frequency of agreement above 65% and a coefficient of reproducibility varying from moderate to sub-
stantial. (Am J Orthod Dentofacial Orthop 2013;143:492-8)

A
pical root resorption is an undesirable effect, com-
mon in orthodontic treatment, that leads to per-
manent loss of dental structure.1,2 The clinical
diagnosis of this condition is predominantly based on
routine radiographic procedures such as periapical or
panoramic radiographs.2,3 However, these methods are
limited because they produce bidimensional images of
a tridimensional structure, leading to labiolingual
superimposition of the whole root structure.2,4 This
limitation can jeopardize the diagnosis of root resorption
because radiographic images might underestimate the
extent of the apical root resorption lesion.2,5
The advent of tridimensional images brought about
the perspective of a precise quantitative evaluation
From the Department of Orthodontics, Juiz de Fora Federal University, Juiz de
Fora, Minas Gerais, Brazil.
a
Professor.
b
Postgraduate student.
c
Associate professor.
d
Associate professor and chair.
The authors report no commercial, proprietary, or financial interest in the prod-
ucts or companies described in this article.
Supported by Fundaç~ao de Amparo a Pesquisa do Estado de Minas Gerais
(FAPEMIG).
Reprint requests to: Marcio Jos
e da Silva Campos, R Guaçui 530/204, Juiz de
Fora MG 36025 190, Brasil; e-mail, Drmarciocampos@hotmail.com.
Submitted, August 2012; revised and accepted, October 2012.
0889-5406/$36.00
Copyright Ó 2013 by the American Association of Orthodontists.
http://dx.doi.org/10.1016/j.ajodo.2012.10.026
492
that permits differentiating the levels of root resorption
in various root surfaces.6,7
Cone-beam computed tomography (CBCT) is an effi-
cient technique to obtain data of dental structures.6 This
method provides images in slices of the dental roots that
eliminate superimposition of structures and show differ-
ent levels of resorption on the labial and lingual surfaces,
demonstrating it to be a precise tool in the diagnosis of
root resorption lesions.7 In addition, it has the advantage
of a much lower radiation burden to the patient when
compared with conventional computed tomography.2,6-8
In 1982, Malmgren et al9 created a quantitative evalu-
ation index of apical root resorption made up of the fol-
lowing scores: irregular apical contour, apical root
resorption less than 2 mm (minor resorption), apical root
resorption from 2 mm to a third of the original root length
(severe resorption), and apical root resorption exceeding
a third of the original root length (extreme resorption).
However, qualitative or quantitative classifications appli-
cable to sagittal radiographic images of apical root resorp-
tion lesions could not be found in the literature.
Since the visualized images on periapical radiographs
reflect the superimposition of the whole root structure,
underestimating the extent of the resorption lesion, in
this study, we aimed to determine, in sagittal images
of CBCT, the lengths of the labial and lingual surfaces
of the incisors with a diagnosis of apical root resorption
by comparing them with the longest root length, and to
da Silva Campos et al 493

Fig 1. Multiplanar reconstruction visualization mode screen on sagittal cut.

create a qualitative visual scale for the classification of

the images of apical root resorption.

MATERIAL AND METHODS

The sample comprised 82 incisors with apical root re-
sorption from 25 patients (14 female, 11 male; age range,
13-41 years; mean age, 20.4 years) having orthodontic
treatment with fixed edgewise appliances and selected
based on the diagnosis of apical root resorption in at least
1 maxillary incisor by means of follow-up periapical ra-
diographs. In the pretreatment records, these patients
did not have maxillary incisors with apical root resorption,
root morphologic alterations, or history of trauma. This
study was approved by the Human Research Ethics Com-
mittee of Juiz de Fora Federal University in Brazil.
The images were obtained with CBCT (i-CAT; Imag-
ing Sciences International, Hatfield, Pa), operated at
120 kVp and 3-8 mA, voxel size of 0.4 mm, and a field Fig 2. Sagittal cut of the maxillary incisor showing line
of view of 160 3 100 mm, and analyzed with the soft- CE, points L, Li, and A, and their respective orthogonal
ware i-CAT Vision (Imaging Sciences International) with projections on line CE.
0.5-mm-thick slices on the multiplanar reconstruction
visualization mode. and Li and their orthogonal projections on line CE. In
The sagittal image selected for the analysis of each the same way, the longest root length was determined
incisor was the one obtained from the sagittal cut of by the distance between point A and its orthogonal pro-
the mesiodistal center of the root and perpendicular to jection on line CE (Fig 2). The shortest root length was
the incisal border (Fig 1). Three points of reference determined by the dimension of the shortest of either
were determined: L, the most apical point of the labial the labial or the lingual root surface.
surface of the root from which apical root resorption be- All root measurements of 15 randomly selected inci-
gins; Li, the most apical point of the lingual surface of sors were evaluated at 2 times in a 3-week interval and
the root from which apical root resorption begins; and compared for determination of the average errors.
A, the most apical point of the tooth root. A horizontal We created a qualitative visual scale for classifying the
line (line CE) was drawn passing through the median sagittal appearance of the apical root resorption lesions
point of the labial and lingual cementoenamel junctions (Fig 3), and its reproducibility was then determined. Sag-
and perpendicular to the long axis of the tooth root. The ittal cut images of the 82 incisor were selected, and the
lengths of the labial and lingual surfaces were deter- apical root resorption lesions were classified by 5 ortho-
mined, respectively, by the distance between points L dontists according to a visual scale at 2 times, with an

American Journal of Orthodontics and Dentofacial Orthopedics April 2013
Vol 143
Issue 4
494 da Silva Campos et al

Fig 3. Visual scale representing the classifications of apical root resorption of the incisors in sagittal
cuts: type 1, flat apical root resorption; type 2, flat apical root resorption associated with superficial re-
sorption on the lingual surface; type 3, flat apical root resorption associated with superficial resorption
on the labial surface; type 4, apical root resorption associated with superficial resorption on the labial
and lingual surfaces; type 5, diagonal apical root resorption in the labiolingual direction; type 6, diagonal
apical root resorption in the lingual-labial direction.

Table I. Normality tests of the evaluated variables

Labial Lingual LRL SRL

Incisors n Test P Test P Test P Test P

Maxillary
Central incisorsy 37 0.122 0.177 0.141 0.177 0.087 0.200 0.124 0.162
Lateral incisorsy 26 0.176 0.038* 0.112 0.200 0.140 0.200 0.160 0.087
Mandibular
Central incisorsz 9 0.930 0.482 0.908 0.302 0.843 0.062 0.922 0.411
Lateral incisorsz 10 0.981 0.968 0.877 0.120 0.924 0.390 0.961 0.793
Totaly 82 0.055 0.200 0.075 0.200 0.067 0.200 0.066 0.200

LRL, Longest root length; SRL, shortest root length.

*Nonnormal distribution pattern; yKolmogorov-Smirnov test; zShapiro-Wilks test.

average interval of 30.6 days between them. The ob- RESULTS

servers received no additional information or previous
training and did not participate in the construction of
the visual scale and the measurement of the incisors.
During the second evaluation, the order of the radio-
graphs was randomly changed in relation to the first one.
Statistical analysis
The average error of measurement of the root length
was calculated by the intraclass correlation coefficient.
The distribution pattern (normal or nonnormal) of the
variables was determined by the Kolmogorov-Smirnov
and Shapiro-Wilks tests. Root lengths were compared
by using paired Student t and Wilcoxon tests. The repro-
ducibility of the visual scale for the apical root resorption
lesions was evaluated by the kappa coefficient for the in-
terobserver and intraobserver analyses. The level of sig-
nificance of 0.05 was used for all statistical analyses.
Tests were made with SPSS software (version 17.0.0;
SPSS, Chicago, Ill).
The intraclass correlation coefficient for the root
lengths was 0.986, indicating excellent agreement.10
Among the 82 incisors evaluated, 9 (11%) were man-
dibular central incisors, 10 (12.2%) mandibular lateral
incisors, 36 (32.9%) maxillary central incisors, and 27
(43.9%) maxillary lateral incisors. Table I shows the
values of the normality tests of the lengths of the labial
and lingual surfaces and the longest and shortest root
lengths for each group of incisors. All variables had nor-
mal distributions, except the length of the labial surface
of the maxillary lateral incisors.
Table II shows the average lengths of the labial and lin-
gual surfaces and the longest and shortest root lengths in
each group of incisors. There was no statistically signifi-
cant difference between the labial and lingual surfaces.
However, the longest root length was significantly longer
than the shortest root length in all groups of incisors.
Table III demonstrates the absolute and relative dis-
tributions of the incisors evaluated according to the

April 2013
Vol 143
Issue 4 American Journal of Orthodontics and Dentofacial Orthopedics
da Silva Campos et al 495

Table II. Average lengths of labial and lingual surfaces, longest and shortest root lengths, and their comparisons
Labial Lingual LRL SRL

Incisors Mean SD Mean SD P Mean SD Mean SD P

Maxillary
Central incisors 10.89 2.25 10.93 1.73 0.876 11.78 1.85 10.37 1.93 0.000
Lateral incisors 11.06 2.08 10.64 2.12 0.059 11.68 1.90 10.32 2.10 0.000
Mandibular
Central incisors 9.69 1.75 10.22 1.82 0.372 10.58 1.67 9.33 1.69 0.007
Lateral incisors 10.50 1.43 10.70 1.65 0.528 11.27 1.15 10.00 1.66 0.007
Total 10.77 2.07 10.73 1.85 0.807 11.55 1.79 10.20 1.93 0.000

The t test was used for comparisons of the maxillary central incisors and the total. The Wilcoxon test was applied for comparisons of the maxillary
lateral incisors and the mandibular central and lateral incisors.
LRL, Longest root length; SRL, shortest root length.

Table III. Distribution of the incisors according to the length of the labial and lingual surfaces and their relationships
with the longest root length
L and Li \ LRL

L and Li 5 LRL Li \ LRL L \ LRL L . Li L \ Li L 5 Li

Incisors n (%) n (%) n (%) n (%) n (%) n (%) Total

Maxillary
Central incisors 3 (8.1) 9 (24.3) 12 (32.4) 6 (16.2) 5 (13.5) 2 (5.4) 37
Lateral incisors 2 (7.6) 10 (38.4) 7 (26.9) 3 (11.5) 0 (0.0) 4 (15.3) 26
Mandibular
Central incisors 0 (0.0) 2 (22.2) 7 (77.7) 0 (0.0) 0 (0.0) 0 (0.0) 9
Lateral incisors 1 (10.0) 2 (20.0) 6 (60.0) 0 (0.0) 0 (0.0) 1 (10.0) 10
Total 6 (7.3) 23 (28.0) 32 (39.0) 9 (10.9) 5 (6.1) 7 (8.5) 82

L and Li 5 LRL, corresponding to type 1 of the visual scale. Li \ LRL, corresponding to types 2 and 5 of the visual scale. L \ LRL, corresponding to
types 3 and 6 of the visual scale. L and Li \ LRL, corresponding to type 4 of the visual scale.
L, Labial surface; Li, lingual surface; LRL, longest root length.

Table IV. Frequency of intraobserver agreement at the 2 evaluations

Type 1 Type 2 Type 3 Type 4 Type 5 Type 6 Total

Observer n (%) n (%) n (%) n (%) n (%) n (%) n (%)

1 7 (8.5) 5 (6.1) 5 (6.1) 28 (34.1) 11 (13.4) 3 (3.7) 59 (72.0)
2 15 (18.3) 10 (12.2) 7 (8.5) 12 (14.6) 6 (7.3) 1 (1.2) 51 (62.2)
3 12 (14.6) 8 (9.8) 4 (4.9) 15 (18.3) 15 (18.3) 5 (6.1) 59 (72.0)
4 19 (23.2) 6 (7.3) 3 (3.7) 8 (9.8) 10 (12.2) 8 (9.8) 54 (65.9)
5 11 (13.4) 9 (11.0) 2 (2.4) 26 (31.7) 8 (9.7) 6 (7.3) 62 (75.6)
Total 64 (15.6) 38 (9.3) 21 (5.1) 89 (21.7) 50 (12.2) 23 (5.6) 285 (69.5)

length of the labial and lingual surfaces and their rela- Table V. Kappa coefficient for the intraobserver
tionships with the longest root length. agreement at the 2 evaluations
As for the evaluation of the reproducibility of the vi-
Observer Kappa coefficient 95% CI Agreement degree20
sual scale for classification of apical root resorption, the
1 0.628 0.503-0.753 Substantial
frequencies of agreement between the first and second 2 0.533 0.408-0.659 Moderate
evaluations are shown in Table IV. Of the 410 evalua- 3 0.656 0.540-0.772 Substantial
tions, 285 were coincident, corresponding to 69.5% of 4 0.570 0.440-0.699 Moderate
agreement. Additionally, the intraobserver reproducibil- 5 0.688 0.572-0.803 Substantial
ity was calculated with the kappa coefficient (Table V), Mean 0.615 0.492-0.737 Substantial
showing an average reproducibility of 0.615. CI, Confidence interval.

American Journal of Orthodontics and Dentofacial Orthopedics April 2013
Vol 143
Issue 4
496 da Silva Campos et al

Table VI. Frequency of interobserver agreement for type of apical root resorption at the 2 evaluations
Type 1 Type 2 Type 3 Type 4 Type 5 Type 6 Total

Agreement/type n (%) n (%) n (%) n (%) n (%) n (%) n (%)

T1 3 observers 8 (9.8) 2 (2.4) 3 (3.7) 9 (11.0) 5 (6.1) 0 (0.0) 27 (32.9)
4 observers 2 (2.4) 4 (4.9) 2 (2.4) 10 (12.2) 5 (6.1) 2 (2.4) 25 (30.5)
All observers 8 (9.8) 3 (3.7) 0 (0.0) 2 (2.4) 3 (3.7) 3 (3.7) 19 (23.2)
Total 18 (22.0) 9 (11.0) 5 (6.1) 21 (25.6) 13 (15.9) 5 (6.1) 71 (86.6)
T2 3 observers 3 (3.7) 5 (6.1) 2 (2.4) 6 (7.3) 5 (6.1) 0 (0.0) 21 (25.6)
4 observers 8 (9.8) 1 (1.2) 2 (2.4) 9 (11.0) 4 (4.9) 4 (4.9) 28 (34.1)
All observers 7 (8.5) 3 (3.7) 2 (2.4) 8 (9.8) 4 (4.9) 1 (1.2) 25 (30.5)
Total 18 (22.0) 9 (11.0) 6 (7.3) 23 (28.0) 13 (15.9) 5 (6.1) 74 (90.2)
T1, First evaluation; T2, second evaluation.

For the analysis of interobserver agreement, the fre-

quencies of coincident evaluations at the 2 times were
evaluated as shown in Table VI. At the first evaluation,
there was an agreement frequency of 3 or more observers
in 86.6% of the cases, whereas at the second evaluation,
this value was 90.2%. In addition, the types with higher
frequencies of agreement between the observers were
type 4 (first and second observations, 25.6% and
28.0%, respectively) and type 1 (both observations, 22%).
The interobserver reproducibility was also calculated
by determination of the kappa coefficient (Table VII), re-
sulting in average values of 0.474 and 0.520 at the 2
evaluations, respectively.
DISCUSSION
Apical root resorption is an undesirable effect com-
mon in orthodontic treatment,2 with a higher prevalence
in the maxillary incisors.11,12 Therefore, periapical
radiographs of these teeth are used for monitoring
resorption lesions during orthodontic treatment.2 As
soon as moderate apical root resorption lesions are diag-
nosed, CBCT scans should be taken to evaluate the real
extent of the lesions.2
All root resorption occurs tridimensionally, and its bi-
dimensional radiographic evaluation is not geometri-
cally accurate, yielding questionable extent values of
the lesions.13-15 Thus, apical root resorption should be
assessed by a radiographic method that allows its
differentiation on each root surface, since these
surfaces might be affected in different ways.6,7,16,17
When apical root resorption is evaluated with a peri-
apical or panoramic radiograph, the image shows super-
imposition of all root structures in a labiolingual
direction; this causes the extent of the lesion to be deter-
mined by the longest surface—ie, by the shortest surface
of the apical root resorption—thus hiding the most re-
sorbed areas of the root. Therefore, CBCT yields the
best image quality for an adequate diagnosis of root re-
sorption because it allows visualization in a distinct way
Table VII. Kappa coefficients for interobserver agree-
ment at the 2 evaluations
Kappa Agreement
coefficient 95% CI degree20
Mean at T1 0.474 0.346-0.601 Moderate
Mean at T2 0.520 0.395-0.645 Moderate
T1, First evaluation; T2, second evaluation.
the resorption lesions on the labial and lingual surfaces
of the root.2,4
In addition to superimposition of images, incisor pro-
clination varies during treatment, and the angulation
between the incisors and the radiographic film can affect
the images obtained, shortening the teeth and jeopard-
izing the diagnostic precision of this technique.18
The possibility of underestimating the extent of api-
cal root resorption lesions in periapical radiographs re-
ported by other studies has been confirmed by the
statistically significant difference found in all groups
of incisors between the longest root length, which has
the least resorption, and the shortest root length, which
has the most resorption.2,5
The average difference between the longest and
shortest root lengths was 1.35 mm, indicating that about
11.6% of the root length is affected by resorption not
identifiable by conventional radiographic methods; this
can jeopardize treatment outcome, since it is planned
based on the visual extent of the apical root resorption
lesions.19
As for the compromising of the labial and lingual sur-
faces by apical root resorption lesions, no statistically
significant difference was found between the lengths
of these 2 surfaces, demonstrating that root resorption
associated with orthodontic treatment has indistinctly
affected both surfaces.
The different ways through which root surfaces were
affected by resorption were also determined. Most inci-
sors evaluated (92.7%) showed the labial surface (39%),
the lingual surface (28%), or both surfaces (25.7%)

April 2013
Vol 143
Issue 4 American Journal of Orthodontics and Dentofacial Orthopedics
da Silva Campos et al
shorter than the longest root length, indicating that in
most incisors, flat root resorption does not occur at
the apical region. The labial and lingual surfaces are af-
fected in different degrees; this cannot be seen in con-
ventional periapical or panoramic radiographs.
Differently from the index created by Malmgren
et al,9 which classified quantitatively the apical root re-
sorption lesions in periapical radiographic images, the
qualitative scale shown in this study represents visually
how root resorption has affected the labial and lingual
surfaces of the periapical region of the 82 incisors eval-
uated, and aimed at facilitating communication among
dental professionals when apical root resorption is diag-
nosed through sagittal images of the incisors.
By relating the types of apical root resorption in the vi-
sual scale with the results obtained in the measurements
of the root surfaces (Table III), the occurrence of type 4 in
the intraoberver (21%) and interobserver (25%) agree-
ments was similar to the percentage of incisors with
both surfaces shorter than the longest root length
(25.5%). Likewise, the occurrence of types 2 and 5 in
the intraobserver (21%) and interobserver (27%) agree-
ments was similar to the frequency of incisors with the lin-
gual surface shorter than the longest root length (28%).
The occurrence of types 3 and 6 in the intraobserver
(10%) and interobserver (12%) agreements was less than
the frequency of incisors with the labial surface shorter
than the longest root length (39%). Adversely, the occur-
rence of type 1 in the intraobserver (15%) and interob-
server (22%) agreements was greater than the
frequency of incisors with the same length as the longest
root length (7.3%). This can be explained by a possible
difficulty found by the observers in identifying irregular-
ities in the contours of the labial surface, decreasing the
occurrence of cases where the labial surface shows a di-
agonal type of resorption (type 3) and increasing the
cases with horizontal apical root resorption (type 1).
Overall, the average intraobserver frequency was
69.5% between the 2 evaluation times, and agreement
was higher than 60% for all observers, varying from
62.2% to 75.6%.
The average agreement was substantial, with the
kappa coefficient of 0.61520 varying from moderate (2
observers) to substantial (3 observers). Also, of all agree-
ments between the 2 evaluations, in 221 (77%), the ob-
servers classified the apical root resorption lesions
between levels 2 and 6, where at least 1 root surface
was shorter than the longest root length, confirming
the results obtained with the statistical analysis.
As for the interobserver reproducibility, we obtained
average kappa coefficients of 0.474 and 0.520 at the 2
evaluations, respectively, both suggesting moderate
agreement.20 We also found agreement by 3 or more
American Journal of Orthodontics and Dentofacial Orthopedics
497
observers of 86.6% at the first evaluation and 90.2%
at the second evaluation, indicating that in most of
the incisors classified, more than half of the observers
agreed with the classification according to the method
proposed.
CONCLUSIONS
There was a statistically significant difference be-
tween the longest and shortest root lengths when eval-
uated in the sagittal cut of the incisors, demonstrating
that radiographic superimposition underestimates the
extent of apical root resorption lesions. The visual scale
created and used for classifying apical root resorption le-
sions showed a frequency of agreement above 65% and
a reproducibility coefficient varying from moderate to
substantial.
REFERENCES
1. McNab S, Battistutta D, Taverne A, Symons Al. External apical root
resorption following orthodontic treatment. Angle Orthod 2000;
70:227-32.
2. Dudic A, Giannopoulou C, Leuzinger M, Kiliaridis S. Detection of api-
cal root resorption after orthodontic treatment by using panoramic
radiography and cone-beam computed tomography of super-high
resolution. Am J Orthod Dentofacial Orthop 2009;135:434-7.
3. Sameshima GT, Asgarifar KO. Assessment of root resorption and
root shape: periapical vs panoramic films. Angle Orthod 2001;
71:185-9.
4. Lermen CA, Liedke GS, Silveira HED, Silveira HLD, Mazzola AA,
Figueiredo JAP. Comparison between two tomographic sections
in the diagnosis of external root resorption. J Appl Oral Sci
2010;18:303-7.
5. Dudic A, Giannopoulou C, Martinez M, Montet X, Kiliaridis S. Di-
agnostic accuracy of digitized periapical radiographs validated
against micro-computed tomography scanning in evaluating or-
thodontically induced apical root resorption. Eur J Oral Sci
2008;116:467-72.
6. Hsu JT, Chang HW, Huang HL, Yu JH, Li YF, Tu MG. Bone density
changes around teeth during orthodontic treatment. Clin Oral In-
vestig 2011;15:511-9.
7. Lund H, Gr€ ondahl K, Gr€
ondahl H. Apical root resorption during or-
thodontic treatment. A prospective study using cone beam CT. An-
gle Orthod 2010;80:466-73.
8. El-Beialy AR, Fayed MS, El-Bialy AM, Mostafa Y. Accuracy and re-
liability of cone-beam computed tomography measurements: in-
fluence of head orientation. Am J Orthod Dentofacial Orthop
2011;140:157-65.
9. Malmgren O, Goldson L, Hill C, Orwin A, Petrini L, Lundberg M.
Root resorption after orthodontic treatment of traumatized teeth.
Am J Orthod 1982;82:158-64.
10. Conover WJ. Practice nonparametric statistics. New York: Wiley;
1999.
11. Bresniak N, Wasserstein A. Orthodontically induced inflammatory
root resorption. Part I: the basic science aspects. Angle Orthod
2002;72:175-9.
12. Weltman B, Vig KWL, Fields HW, Shanker S, Kaizar EE. Root
resorption associated with orthodontic tooth movement: a sys-
tematic review. Am J Orthod Dentofacial Orthop 2010;137:
462-76.
April 2013
Vol 143
Issue 4
498
13. Sameshima GT, Sinclair PM. Predicting and preventing root re-
sorption: part I. Diagnostic factors. Am J Orthod Dentofacial Or-
thop 2001;119:505-10.
14. Chan EKM, Darendeliler MA. Exploring the third dimension in root
resorption. Orthod Craniofac Res 2004;7:64-70.
15. Katona TR. Flaws in root resorption assessment algorithms: role of
tooth shape. Am J Orthod Dentofacial Orthop 2006;130:
698.e19-27.
16. Schwarz MS, Rothman SL, Rhodes ML, Chafetz N. Computed to-
mography: part I. Preoperative assessment of the mandible for en-
dosseous implant surgery. Int J Oral Maxillofac Implants 1987;2:
137-41.
April 2013
Vol 143
Issue 4 American Journal of Orthodontics and Dentofacial Orthopedics
da Silva Campos et al
17. Velvart P, Hecker H, Tillinger G. Detection of the apical lesion and
the mandibular canal in conventional radiography and computed
tomography. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
2001;92:682-8.
18. Apajalahti S, Peltola JS. Apical root resorption after orthodon-
tic treatment—a retrospective study. Eur J Orthod 2007;29:
408-12.
19. Cohenca N, Simon JH, Mathur A, Malfaz JM. Clinical indications
for digital imaging in dento-alveolar trauma. Part 2: root resorp-
tion. Dent Traumatol 2007;23:105-13.
20. Landis JR, Koch GG. The measurement of observer agreement for
categorical data. Biometrics 1977;33:159-74.