Clinical Research

Identification of a C-shaped Canal System in Mandibular

Second MolarsPart II: The Effect of Bone Image
Superimposition and Intraradicular Contrast Medium on
Radiograph Interpretation
Bing Fan, DDS, MSc, PhD,* Yuan Gao, DDS, MSc, PhD,* Wei Fan, DDS, MSc, PhD,*
and James L. Gutmann, DDS, PhD, FACD, FICD, FADI

Abstract
The purpose of this study was to investigate the effect
of image superimposition from surrounding alveolar
bone and intraradicular contrast medium on radio-
graphic interpretation of a C-shaped canal system in
mandibular second molars. Thirty mandibular second
molars with fused roots were collected and stored in
10% formalin solution. After being scanned by micro-
computed tomography and reconstructed, the buccal-
lingual radiographs of each tooth were taken with or
without intraradicular contrast medium and mandibular
bony plates. The number of mesial and distal canals in
the coronal-to-middle canal portion and the canal num-
ber in the apical 4-mm area were recorded. The buccal-
lingual and three-dimensional view of reconstructed
canal images were used to evaluate the accuracy of
radiograph reading. The results were subjected to the
Friedman and chi-square test. The results showed that
without intraradicular contrast medium, the bone su-
perimposition could decrease the accuracy of canal
recognition significantly (p 0.001), especially in the
apical area. When contrast medium was in canals, the
canal reading seemed not to be affected by the bone
superimposition (p 0.05). Based on these findings,
the intraradicular contrast medium could increase the
accuracy of canal recognition of a C-shaped canal sys-
tem in mandibular second molars despite the bone
image superimposition, especially in the detection of
continuous C-shaped canals and the canals in the apical
area. (J Endod 2008;34:160 165)
Key Words
Anatomy, bone image superimposition, C-shaped ca-
nal, contrast medium, mandibular second molar
C leaning, shaping, and filling the root canal system effectively is the basis of suc-
cessful root canal therapy. To achieve this, thorough knowledge of the morphology
of the root canal system is an essential prerequisite (13). Many methods are used to
investigate the root canal anatomy in vitro and in vivo (4 7). However, with the excep-
tion of radiography, these methods either destroy the specimen by grinding or splitting
or preclude further use of the specimen, such as tooth clearing and dye assessments.
Based on these concerns, radiography becomes the most practical and often used
method to predict the root canal anatomy in both laboratory and clinical studies.
Careful assessment of the preoperative radiograph is a key step for subsequent
root canal procedures, such as root canal preparation and obturation. Preoperative
radiographs may aid in visualizing and observing the anatomy of various root canal
systems, especially the complicated ones, such as a C-shaped canal system. A C-shaped
canal system is a canal variant mostly seen in mandibular second molars, although it can
also appear in maxillary and mandibular premolars and molars (6, 8 12). The main
anatomic feature of C-shaped canals is the presence of a fin or web connecting mesial
and distal individual canals, which makes the canal cross-section C-shaped and pre-
sents difficulty in its thorough cleaning, shaping, and obturation. Recognition of a
C-shaped canal configuration before treatment can facilitate a more effective manage-
ment of the root canal system.
Radiographs may be the only practicable way in the clinic to provide clues about
the morphology of the root canal system. Recent research has revealed some radio-
graphic features of the mandibular second molars with a C-shaped canal system (13)
and found there may be certain links existing between these radiographic features and
the canal anatomy. These findings would help in the recognition of a C-shaped canal
system in mandibular second molars through an assessment of the preoperative radio-
graphs.
Although radiography is very helpful in aiding the recognition of canal anatomy, its
interpretation suffers from many factors, the main one of which is the image superim-
position from surrounding hard tissues (bone in most situations). It is often difficult to
characterize the morphology of the root canal system when this image superimposition
exists (14). This is especially true in the area having thick cortical bone plates, such as
the area surrounding mandibular second molars. A method that could enhance the
image of canals may, therefore, be considered of value in root canal treatment.
The idea of introducing water-soluble radiographic contrast medium into the
canal system is not new, but the methods used are different (14, 15). Most of these

From the *Key Laboratory of Oral Biomedical Engineering of Ministry of Education, Endodontic Center, School and Hospital of Stomatology, Wuhan University,
Wuhan, China; and Department of Endodontics, Baylor College of Dentistry, Texas A&M University system Health Science Center, Dallas, Texas.
Address requests for reprints to Dr Bing Fan, Key Laboratory of Oral Biomedical Engineering of Ministry of Education, School and Hospital of Stomatology, Wuhan
University, 237 Luoyu Road, Wuhan 430079, China. E-mail address: bingfan8@hotmail.com.
0099-2399/$0 - see front matter
Copyright 2008 by the American Association of Endodontists.
doi:10.1016/j.joen.2007.10.010

160 Fan et al. JOE Volume 34, Number 2, February 2008


methods could either only be used in a laboratory study or ineffective in
contrast medium distribution (15). In part I of this series studies, The
device developed by Lussi et al (16) for a noninstrumental technique
was modified and proved successful in introducing contrast media into
a C-shaped canal system (17). However, it is unknown whether or not
the intraradicular contrast medium would improve the accuracy of ra-
diograph interpretation on a C-shaped canal system in mandibular sec-
ond molars, especially when the image superimposition from bone
exists.
To obtain the three dimensional (3D) canal anatomy nondestruc-
tively, a microcomputed tomography (CT) scan is introduced to scan
the tooth and produce the root and canal cross-section images, based
on which the 3D canal morphology could be reconstructed by using a
specialized software (5, 18 23). The features of a 3D reconstructed
C-shaped canal system in mandibular second molars has been eluci-
dated in recent research (20), and the reconstructed 3D canal config-
uration could be referred to as the criterion when assessing other ob-
servation methods.
The purpose of this study was to investigate the effect of image
superimposition from surrounding alveolar bone and intraradicular
contrast medium on radiographic interpretation of a C-shaped canal
system in mandibular second molars.
Materials and Methods
Thirty mandibular second molars with fused roots and deep lin-
gual longitudinal grooves on root surfaces were selected and stored in
10% formalin solution. All of the teeth were scanned by a microcom-
puted tomography (CT 20) scan (Scanco Medical AG, Bassersdorf,
Switzerland) from the crown to the apex. The scanning interval was 0.5
mm. Based on these CT-scanned images, the 3D canal configuration
of each tooth was reconstructed by using the 3D-Doctor software (Able
Software Corp, Lexington, MA) and would be referred to as the criterion
for the assessment of subsequent radiograph reading.
After scanning, the pulp chamber of each tooth was accessed by
using an engine-driven fissure bur and unroofed completely with an
Endo Z bur (Dentsply Ltd, Surrey, United Kingdom). Pulp tissue in each
canal was removed by a fine barbed broach (Dentsply Maillefer, Bail-
luges, Switzerland) without exerting any pressure. Subsequently, the
teeth were soaked in 5% sodium hypochlorite (NaOCl) solution for 30
minutes (24, 25). The canals were then irrigated with 20 mL distilled
water, and a jet of water from the apical foramina revealed that the
canals were negotiable. Canals without a visible jet of water from the
apical foramina were negotiated carefully with a no. 20 Lightspeed
(Lightspeed Technology Inc, San Antonio, TX) instrument until the fo-
ramina were patent. Subsequently, all canals were irrigated with 20 mL
5% NaOCl solution and 20 mL distilled water. All the canals were dried
briefly with an air syringe.
All the teeth were put on the same radiographic device described
in part I (17), and buccal-lingual radiographs were taken in the same
way. A right-half human mandible collected from a cadaver was split
into halves longitudinally with the plane of section extending mesial
distally. The teeth in the mandible were removed, and the socket of
mandibular second molar was slightly enlarged without extensively cut-
ting away buccal and lingual cancellous bone. The two halves of man-
dible plates were screwed onto a metal frame with a 2-cm distance
between the plates. The metal frame with two mandible plates could be
mounted onto and demounted from the radiographic device used in
part I (17) with the tooth root standing between and in the exact area of
original mandibular second molar socket (Figs. 1 and 2). The cortical
bone surfaces were vertical to the x-ray beam. The placement of the
mandibular plates could be accurately repeated during the mount-
JOE Volume 34, Number 2, February 2008
Clinical Research
Figure 1. The device used to take radiographs with bone plates. 1, x-ray tube; 2,
aluminum wedge; 3, mandibular bony plates; 4, radiograph.
ing and demounting. Then, the buccal-lingual radiographs of the
teeth with the mandibular plates were taken in the same way as
mentioned previously. When the plates were in position, the expo-
sure time was 0.32 seconds. When the plates were removed, the
exposure time was 0.10 seconds. All the films were developed under
the same conditions. A 4-mm wide stepwedge made of aluminum
1100 was fixed onto the jig 3 mm above the occlusal surface of the
teeth as the reference standard for radiographic density in all films
(26). The stepwedge contained ten 1-mm thick steps with 2 mm as
each step span.
After the roots of all teeth were covered totally with sticky wax to
seal all possible canal exits on the root surface, the vacuum device
described in part I was used to introduce the 76% compound meglu-
mine diatrizoate (Xudong Haipu Pharmaceutical Co Ltd, Shanghai,
China) into the canal system. After the contrast medium was introduced,
the radiographs with or without the mandible plates were taken as
aforementioned (Fig. 2).
The number of mesial and distal main canals in the coronal-to-middle
canal portion and the canal number in the apical 4-mm area were recorded
based on radiograph reading by three precalibrated and standardized end-
odontic postgraduate students according to a set of written criteria. The
completely merged mesial and distal canal in the coronal-to-middle canal
portion (ie, continuous C-shaped canal cross-section described in part I
throughout the whole canal portion) would be considered as a continuous
C-shaped canal when it appears in the 3D reconstructed. Any variances or
disagreements in the recording were discussed among the observers until
an agreement was reached. Before formally viewing the radiographs, all the
evaluators were trained to obtain full understanding about the objectives
and criteria. The kappa value, which was subject to the Mann-Whitney U test
for significance at p 0.05, was calculated to be from 0.59 to 1.00, sug-
gesting good interobserver agreement (15, 27). The buccal-lingual and 3D
view of reconstructed canal images based on CT scanning were used to
evaluate the accuracy of radiograph reading. The accuracy was calculated
based on the correct recognition of canal number in each canal areas
compared with the reconstructed canal images. The results were subjected
to the Friedman and chi-square test. The statistical significance was consid-
ered at p 0.05.
Results
Canal Recognition Under Different Observing Conditions
The distribution of canal number in the coronal-to-middle canal por-
tion and apical 4 mm area under six different observing conditions was
shown in the Table 1. There were significant differences among the six
different conditions (p 0.001). The canal number distribution on the
radiographs taken without bone was similar to that from the radiographs
taken with both contrast medium and bone and reconstructed canal images
(p 0.05), whereas the canal recognition on the radiographs taken with
C-shaped Canal System in Mandibular Second MolarsPart II 161
Clinical Research

Figure 2. Radiographic and reconstructed canal images. (A) Radiographic image without bone and contrast medium showing one mesial and one distal canal,
(B) radiographic image with bone but without contrast medium showing obscure canal images, especially in the apical area, (C) radiographic image with contrast
medium and without bone clearly showing one mesial and one distal canal, (D) radiographic image with bone and contrast medium (canal images are still very clear),
(E) buccal-lingual view of reconstructed canal image showing one mesial and one distal canal; (F) 3D view of reconstructed canal image showing two mesial canals
and one distal canal.

TABLE 1. Canal Recognition under Six Different Observing Conditions

Coronal-to-Middle Canal Portions Apical 4-mm Area
No. of Teeth With No. of Teeth With No. of Teeth With
No. of Teeth With 1,2,3,or X Mesial 1,2,3,or X Distal 1,2,3,or X Apical
Observing
Continuous C-Shaped Canals Canals Canals
Conditions
Canal
1 2 3 X 1 2 3 X 1 2 3 X
L 0 30 0 0 0 30 0 0 0 10 20 0 0
LB 0 26 0 0 4 22 0 0 8 1 2 0 27
CL 3 25 2 0 0 27 0 0 0 11 18 1 0
CLB 3 25 2 0 0 27 0 0 0 9 19 1 1
CT-BL 3 22 5 0 0 27 0 0 0 11 14 5 0
CT-3D 3 15 12 0 0 27 0 0 0 11 11 8 0
X, indiscernible canals; L, radiographs without contrast medium and mandible plates; LB, radiographs without contrast medium but with mandible plates; CL, radiographs with contrast medium but without mandible
plates; CLB, radiographs with contrast medium and mandible plates; CT-BL, buccal-lingual view of reconstructed images based on CT scanning; CT-3D, 3D view of reconstructed images based on CT scanning.
Tall (T value among all conditions, Friedman test) 63.676, p 0.001:T1 (T value among L, CL, CLB, CT-BL and CT-3D) 8.049, p 0.09; p (between LB and other five conditions) 0.05.

162 Fan et al. JOE Volume 34, Number 2, February 2008


TABLE 2. Canal Recognition on Radiographs under Different Conditions Compared with Buccal-Lingual View of Reconstructed Root Canal System
Number of Correctly Read Canals (%) in Coronal-to-Middle Canal
Portion
Conditions Continuous C-Shaped Canal Mesial Canal
L 0 (0.0) 22 (81.5)
LB 0 (0.0) 19 (70.4)
CL 3 (100.0) 20 (74.1)
CLB 3 (100.0) 20 (74.1)
L, radiographs without contrast medium and mandible plates; LB, radiographs without contrast medium but with mandible plates; CL, radiographs with contrast medium but without mandible plates; CLB,
radiographs with contrast medium and mandible plates.
2 all (chi-square value among all conditions) 90.892, p 0.001; 21 (chi-square value between L and LB) 29.000, p 0.001; 22 (chi-square value among L, CL and CLB) 12.000, p 0.002; 23
(chi-square value among LB, CL and CLB) 70.000, p 0.001.
bone but without contrast medium was significantly different from radio-
graphs taken with contrast medium and reconstructed canal images (p
0.05).
The Effect of Bone Image Superimposition and Intraradicualr
Contrast Medium on the Accuracy of Canal Recognition
The differences in accuracy of canal recognition on radiographs un-
der four different observing conditions are shown in Tables 2 and 3 com-
pared with both the buccal-lingual and 3D view of reconstructed canal
images. Without the intraradicular contrast medium, the bone superimpo-
sition could decrease the accuracy of canal recognition significantly (p
0.05), especially in the apical area (Tables 2 and 3). When the contrast
medium was in canals, the canal reading seemed not to be affected by the
bone superimposition (p 0.05) (Tables 2 and 3). The accuracy of canal
recognition on contrasted radiographs was significantly higher than non-
contrasted radiographs (p 0.001), especially in revealing the continuous
C-shaped canals and the canals in the apical area.
Discussion
The radiographic images of normal structures surrounding the tooth
root have been thought to interfere with the determination of root canal
anatomy, especially the image of overlying alveolar bone (27). The mandib-
ular second molar lies in the far posterior area of the mandible arch where
the buccal and lingual cortical bone plates are much thicker than other
areas because primarily of the internal and external oblique ridges. In this
study, a dried human mandible was used to simulate this situation in vivo,
and the results showed that the bone image superimposed over the man-
dibular second molars decreased the accuracy of radiograph interpretation
significantly if no contrast medium was introduced into canals, especially in
apical areas. According to the results, the radiographic images of a C-shaped
canal system in mandibular second molars should be interpreted with cau-
tion under clinical situations.
Water-soluble intraradicular contrast medium has been used to en-
hance the canal image against other surrounding tissue images (14, 28). In
TABLE 3. Canal Recognition on Radiographs under Different Conditions Compared with Three-Dimensional View of Reconstructed Root Canal System
Number of Correctly Read Canals (%) in Coronal-to-Middle Canal
Portion
Conditions Continuous C-Shaped Canal Mesial Canal
L 0 (0.0) 14 (51.9)
LB 0 (0.0) 13 (48.1)
CL 3 (100.0) 16 (59.3)
CLB 3 (100.0) 16 (59.3)
L, radiographs without contrast medium and mandible plates; LB, radiographs without contrast medium but with mandible plates; CL, radiographs with contrast medium but without mandible plates; CLB,
radiographs with contrast medium and mandible plates.
2 all (Chi-square value among all conditions) 77.880, p 0.001; 21 (chi-square value between L and LB) 24.000, p 0.001; 22 (chi-square value among L, CL and CLB) 14.000, p 0.001; 23
(chi-square value among LB, CL and CLB) 62.000, p 0.001.
JOE Volume 34, Number 2, February 2008
Clinical Research
Number of Correctly Read Canals
(%) in Apical 4mm Area
Distal Canal
27 (100.0) 22 (73.3)
20 (74.1) 3 (10.0)
27 (100.0) 24 (80.0)
27 (100.0) 24 (80.0)
view of the bone superimposition on the canals of mandibular second mo-
lars, the intraradicular contrast medium should be of assistance in canal
recognition. Previous studies have found the contrast medium in canals
could increase the amount of canal anatomy information (14, 15). In the
present experiment, the contrast medium increased the accuracy of radio-
graph interpretation significantly despite the bone image superimposition,
especially in the recognition of apical canals. These findings indicate that
introducing the contrast medium into a C-shaped canal system in mandib-
ular second molars can improve canal identification even when a bone
image superimposition exists and provide more information about the canal
anatomy in the apical area.
In this experiment, both the buccal-lingual and 3D view of recon-
structed canal images were used as the criterion to assess the accuracy of
canal recognition. When compared with the buccal-lingual view, the con-
trasted radiograph showed its main advantages over noncontrasted radio-
graphs in the detection of continuous C-shaped canals and the canals in
apical areas. While in comparison with the 3D view of reconstructed canal
images, the reading accuracy of the contrasted radiograph decreased al-
though the contrast radiograph seemed better than the noncontrast radio-
graphs. The discrepancies between the contrasted radiographs and recon-
structed canal images might be caused by the following: (1) the
superimposition of the buccal canal image on the lingual canal image (or
vice versa) influenced the canal reading; (2) the contrast medium failed to
fill the whole canal system because of the incomplete removal of pulp tissue,
canal calcification, or small bubble formation; or (3) in some cases, the
image superimposition from the cementum, dentin, and surrounding bone
may still influence the observation on some fine or thin canal areas, although
this adverse effect has been reduced significantly through the introduction of
contrast medium (Fig. 3). These problems and challenges will be discussed
further in ongoing studies.
Based on the findings of this laboratory experiment, although the
bone image superimposition could significantly decrease the accuracy
of canal recognition, the intraradicular contrast medium could increase
the accuracy of canal recognition of C-shaped canal system in mandib-
Number of Correctly Read-
Canals (%) in Apical 4-mm
Distal Canal Area
27 (100.0) 19 (63.3)
20 (74.1) 3 (10.0)
27 (100.0) 21 (70.0)
27 (100.0) 21 (70.0)
C-shaped Canal System in Mandibular Second MolarsPart II 163
Clinical Research

Figure 3. Incomplete detection of irregular canal areas in contrasted radiographs compared with buccal-lingual view of reconstructed canal images. (a) Small
bubbles in the contrast medium (arrows), (b) obscure fine canal images in the radiograph (arrow), and (c) reconstruction image of the tooth and canal in A, and
(d) reconstruction image of the tooth and canal in B.

ular second molars despite the bone image superimposition. This is
especially valuable in the detection of continuous C-shaped canals and
the canals in the apical area.
Acknowledgment
This study was supported by the National Natural Science
Foundation of China (Grant No. 30572042) and the Natural Science
Foundation of Hubei Province of China (Grant No. 304161200).
References
1. Walid N. The use of two pluggers for the obturation of an uncommon C-shaped canal.
J Endod 2000;26:422 4.
2. Yilmaz Z, Tuncel B, Serper A, Calt S. C-shaped root canal in a maxillary first molar: a
case report. Int Endod J 2006;39:162 6.
3. Sponchiado EC Jr, Ismail HA, Braga MR, de Carvalho FK, Simoes CA. Maxillary central
incisor with two root canals: a case report. J Endod 2006;32:1002 4.
4. Iyer VH, Indira R, Ramachandran S, Srinivasan MR. Anatomical variations
of mandibular premolars in Chennai population. Indian J Dent Res 2006;17:710.

164 Fan et al. JOE Volume 34, Number 2, February 2008


5. Lee JK, Ha BH, Choi JH, Heo SM, Perinpanayagam H. Quantitative three-dimensional
analysis of root canal curvature in maxillary first molars using micro-computed
tomography. J Endod 2006;32:9415.
6. Lu TY, Yang SF, Pai SF. Complicated root canal morphology of mandibular first premolar in a
Chinese population using the cross section method. J Endod 2006;32:9326.
7. Al-Qudah AA, Awawdeh LA. Root canal morphology of mandibular incisors in a
Jordanian population. Int Endod J 2006;39:8737.
8. De Moor RJ. C-shaped root canal configuration in maxillary first molars. Int Endod
J 2002;35:200 8.
9. Bolger WL, Schindler WG. A mandibular first molar with a C-shaped root configura-
tion. J Endod 1988;14:5159.
10. Seo MS, Park DS. C-shaped root canals of mandibular second molars in a Korean
population: clinical observation and in vitro analysis. Int Endod J 2004;37:139 44.
11. Carlsen O, Alexandersen V. Root canals in two-rooted maxillary second molars. Acta
Odontol Scand 1997;55:330 8.
12. Cleghorn BM, Christie WH, Dong CC. Root and root canal morphology of the human
permanent maxillary first molar: a literature review. J Endod 2006;32:81321.
13. Fan B, Cheung GS, Fan M, Gutmann JL, Fan W. C-shaped canal system in mandibular
second molars: part IIradiographic features. J Endod 2004;30:904 8.
14. Shearer AC, Wasti F, Wilson NH. The use of a radiopaque contrast medium in end-
odontic radiography. Int Endod J 1996;29:95 8.
15. Naoum HJ, Love RM, Chandler NP, Herbison P. Effect of X-ray beam angulation and
intraradicular contrast medium on radiographic interpretation of lower first molar
root canal anatomy. Int Endod J 2003;36:129.
16. Lussi A, Suter B, Fritzsche A, Gygax M, Portmann P. In vivo performance of the new
non-instrumentation technology (NIT) for root canal obturation. Int Endod J
2002;35:352 8.
17. Fan W, Fan B, Gutmann JL, Cheung GSP. Identification of C-shaped canal systems in
mandibular second molars. Part I: radiographic and anatomic features revealed by
intraradicular contrast medium. J Endod 2007;33:806 10.
JOE Volume 34, Number 2, February 2008
Clinical Research
18. Peters OA, Laib A, Gohring TN, Barbakow F. Changes in root canal geometry after
preparation assessed by high-resolution computed tomography. J Endod 2001;
27:1 6.
19. Fan B, Cheung GS, Fan M, Gutmann JL, Bian Z. C-shaped canal system
in mandibular second molars: part Ianatomical features. J Endod
2004;30:899 903.
20. Gao Y, Fan B, Cheung GS, Gutmann JL, Fan M. C-shaped canal system in mandibular
second molars part IV: 3-D morphological analysis and transverse measurement. J
Endod 2006;32:10625.
21. Plotino G, Grande NM, Pecci R, Bedini R, Pameijer CH, Somma F. Three-dimensional
imaging using microcomputed tomography for studying tooth macromorphology.
J Am Dent Assoc 2006;137:1555 61.
22. Min Y, Fan B, Cheung GS, Gutmann JL, Fan M. C-shaped canal system in mandibular
second molars. Part III: The morphology of the pulp chamber floor. J Endod
2006;32:11559.
23. Jung M, Lommel D, Klimek J. The imaging of root canal obturation using micro-CT.
Int Endod J 2005;38:61726.
24. Lussi A, Nussbacher U, Grosrey J. A novel noninstrumented technique for cleansing
the root canal system. J Endod 1993;19:549 53.
25. Koskinen KP, Stenvall H, Uitto VJ. Dissolution of bovine pulp tissue by endodontic
solutions. Scand J Dent Res 1980;88:406 11.
26. Tagger M, Katz A. Radiopacity of endodontic sealers: development of a new method
for direct measurement. J Endod 2003;29:7515.
27. Nicopoulou-Karayianni K, Bragger U, Patrikiou A, Stassinakis A, Lang NP. Image
processing for enhanced observer agreement in the evaluation of periapical bone
changes. Int Endod J 2002;35:61522.
28. Naoum HJ, Chandler NP, Love RM. Conventional versus storage phosphor-plate dig-
ital images to visualize the root canal system contrasted with a radiopaque medium.
J Endod 2003;29:349 52.
C-shaped Canal System in Mandibular Second MolarsPart II 165