Vol. 98 No.

5 November 2004

ORAL AND MAXILLOFACIAL RADIOLOGY Editor: Stephen R. Matteson

Anatomical considerations for the spread of odontogenic infection originating

from the pericoronitis of impacted mandibular third molar: Computed
tomographic analyses
Aya Ohshima, DDS,a Yoshiko Ariji, DDS, PhD,b Masakazu Goto, DDS,c Masahiro Izumi, DDS, PhD,d
Munetaka Naitoh, DDS, PhD,d Kenichi Kurita, DDS, PhD,e Kazuo Shimozato, DDS, PhD,f and Eiichiro
Ariji, DDS, PhD,g Nagoya, Japan
AICHI-GAKUIN UNIVERSITY

Objectives. The aims of the present study were to clarify the anatomy of impacted mandibular third molars in relation to
surrounding structures and to investigate the pathway of infection originating from pericoronitis of this tooth.
Study design. Computed tomography (CT) images were evaluated in 87 patients with uninfected mandibular third molar
impaction and in 12 patients with infection originating from an impacted mandibular third molar. In uninfected patients, bony
features around the impacted crown were investigated together with the relationship between the crown and surrounding
muscles. In infected patients, involvements of bony and soft tissue structures were evaluated according to the disappearance of
cortices and lateral asymmetry of density and shape in the spaces and muscles.
Results. In uninfected patients, the disappearance of the lingual cortical plate was observed in 48 (35.3%) impacted molars,
while only in 11 (8.1%) teeth for buccal cortices. The cortical thickness was thinner on the lingual side than the buccal side.
Sixty-five percent of the masseter muscle horizontally overlapped the crown, while almost all of the medial pterygoid muscle
was posteriorly situated apart from the crown. The mylohyoid muscle horizontally overlapped the crown at below or
intermediate vertical positions. In infected patients, the involvement of lingual structures was more frequently observed than
that of buccal structures. The mylohyoid muscle was involved in 10 (83.3%) of 12 patients. Among them, 8 showed
submandibular space involvement.
Conclusion. CT findings supported the clinical observations of infection spread in patients with pericoronitis of the impacted
mandibular third molar. CT appeared to be an effective tool for investigating the pathway of infection originating from the
pericoronitis of impacted mandibular third molars.
(Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2004;98:589-97)

a
Research Student, Department of Oral and Maxillofacial Surgery II, The anatomy of maxillofacial structures can influence the
Aichi-Gakuin University School of Dentistry. spread of odontogenic infection. Since an impacted
b
Assistant Professor, Department of Oral and Maxillofacial Radiol-
mandibular third molar is one of the most frequent causes
ogy, Aichi-Gakuin University School of Dentistry.
c
Instructor, Department of Oral and Maxillofacial Radiology, Aichi- of odontogenic infection,1,2 it is important to understand
Gakuin University School of Dentistry. the anatomy of this tooth impaction. To accomplish this,
d
Associate Professor, Department of Oral and Maxillofacial Radiol- many studies have been carried out using conventional
ogy, Aichi-Gakuin University School of Dentistry.
e
radiography.3,4 The prevalence of infection spread
Professor and Chairman, Department of Oral and Maxillofacial depends on the vertical position on panoramic radio-
Surgery I, Aichi-Gakuin University School of Dentistry.
f
Professor and Chairman, Department of Oral and Maxillofacial graph.3-5 An incompletely impacted third molar results in
Surgery II, Aichi-Gakuin University School of Dentistry. odontogenic infection more frequently than fully erupted
g
Professor and Chairman, Department of Oral and Maxillofacial or completely impacted molars do.3-6
Radiology, Aichi-Gakuin University School of Dentistry. Based on the anatomical observations, mandibular
Received for publication Apr 3, 2004; returned for revision May 23, molars are situated lingually in the mandible.7 This
2004; accepted for publication Jul 15, 2004.
1079-2104/$ - see front matter
supports the clinical observation that the spread of
Ó 2004 Elsevier Inc. All rights reserved. odontogenic infection arising from mandibular teeth is
doi:10.1016/j.tripleo.2004.07.012 more likely to occur on the lingual side than the buccal

589

590 Ohshima et al
side. The bucco-lingual characteristics of impacted third
molars can be assessed by computed tomography (CT),
and many authors have investigated the relationship
between the tooth apex and the mandibular canal to
provide guidance on how to avoid accidental injury of
the mandibular nerve during removal of this tooth.8,9
However, there are no CT studies on bucco-lingual bony
anatomy that take the spread of infection into account,
and the anatomical relationship between the crown of an
impacted mandibular third molar and surrounding soft
tissue structures has not been completely elucidated.
CT examinations are frequently used and are effective
for examining various maxillofacial lesions, especially
in patients with trismus caused by the spread of infection
to the masticatory muscles. If the infection spreads to
deep spaces such as the parapharyngeal space, rapid
airway obstruction may occur, resulting in a life-
threatening condition. Therefore, the spread of infection
should be evaluated by imaging at the early stage.
Although the spread of infection originating in the
mandibular third molar frequently causes deep fascial
space infection, most of imaging studies also focus on
periapical infection of the first and second molars.
Infection spread from the pericoronitis of mandibular
third molar should be separately investigated from those
originating from other teeth and lesions.
The first aim of the present study was to clarify the
anatomy of impacted mandibular third molars on CT
image with emphasis on the bucco-lingual aspects and
their relation to surrounding muscles. We investigated
the CT images of uninfected patients who were referred
to our hospital for extraction of impacted mandibular
third molars. The second aim was to investigate the
infection pathway on CT images in patients with
odontogenic infection originating from the pericoronitis
of the mandibular third molar.
MATERIALS AND METHODS
Uninfected patients
Eighty-eight patients underwent CT examination
before extraction of their mandibular third molars
between 1996 and 2003 at the Department of Radiology
and Diagnostic Imaging, Aichi-Gakuin University Dental
Hospital. For all patients, CT was performed to examine
the relationship between the root and mandibular canal.
With the exception of 2 normally erupted teeth, 136
impacted third molars without definitive infection in 87
patients (43 females and 44 males, mean age of 32.8
years) were enrolled in this study.
The state of impaction was classified into 3 types on the
basis of physical and panoramic x-ray examination as
suggested by Hugoson and Kugelberg.10 When a part of the
crown could be observed in the oral cavity, it was defined as
‘‘partial impaction.’’ Third molars that were embedded
OOOOE
November 2004
under the gingival mucosa but not completely impacted in
the mandibular bone on panoramic radiograph were
classified as ‘‘submucosal impaction.’’ A tooth completely
embedded in the mandibular bone on panoramic radio-
graph was defined as ‘‘complete impaction.’’ The vertical
position of the crown was also assessed on panoramic
radiograph and was classified into 3 levels (positions A, B,
and C) according to a report by Pell and Gregory.11 In
position A, the occlusal plane of the impacted third molar
was situated at the same level as that of the second molar.
Position B showed the occlusal plane of the impacted tooth
between the occlusal plane and the cervical line of the
second molar. In position C, the impacted tooth was below
the cervical line of the second molar.
CT examination was performed by Somatom ART
(Siemens AG; Erlangen, Germany) or Hi speed NX-I
Pro (GE Yokogawa Medical Systems; Tokyo, Japan)
with patients in the supine position. Axial scans were
contiguously obtained, with a section thickness of 2 to 3
mm and the scan plane parallel to the inferior margin of
the mandible.
Analyses of CT images
The minimum length between the surface of the
impacted third molar and the outer surface of the
cortical plate was measured and defined as the bony
width (Fig 1). This measurement was performed on
a line perpendicular to a line tangential to the outer
surface of the cortex passing through the nearest point to
the cortex on the surface of the crown. Cortical thick-
ness was also measured on the same line. These measure-
ments were performed by an observer (A.O.) on the
buccal and lingual sides at the slice where the maximal
crown area was observed. To verify the reproducibility
of the measurement, 10 impacted third molars were
randomly selected and the bony width and cortical
thickness were measured twice. The measurement error
(Se) was calculated by Dahlberg’s formula:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
Se ¼ Sðd1
d2Þ =2n
ðd1; d2 : measured values; n : sample sizeÞ
where d1 and d2 are measured values and n is the sample
size.
The values for the bony width were 0.26 mm and 0.16
mm for the buccal and lingual sides, respectively. For
cortical thickness, values were 0.24 mm and 0.11 mm
for the buccal and lingual sides, respectively. These
were sufficiently small compared to the variance of the
measured values.
The positions of the masseter muscle (MM), medial
pterygoid muscle (MPM), and mylohyoid muscle (MhM)
were determined on CT image in relation to the crowns
of impacted third molars (Fig 2). The horizontal
OOOOE
Volume 98, Number 5
Fig 1. Bony width and cortical thickness were measured on
the buccal and lingual sides at the slice where the maximal
crown area was observed. Bony width was measured as the
minimum length between the outer surfaces of the crown and
the cortical plates. The measurement was performed on the
line perpendicular to the line tangential to the outer surface of
the cortex passing through the point nearest to the cortex on
the surface of the crown. The cortical thickness was also
measured on the same line as the length between the inner and
outer surfaces of the cortical plate on the line on which the
bony width was measured.
relationships were classified into 3 groups. As for the
MM, they were classified into the anterior group when the
most anterior points of the muscles were situated more
anterior to a line passing through the most medial position
of the impacted crowns parallel to the line on which the
buccal bony width was measured. The anterior muscle
point for the intermediate group was positioned between
the lines passing the most medial and distal positions of
the crown. For the posterior group, the anterior points
were situated posterior to a line through the most distal
position of the crown. The classification of MPM position
was performed with reference to the line parallel to the
line on which the lingual cortical thickness was measured.
For the MhM, the horizontal positions were determined
based on the position of the posterior border of the muscle
to the crown of the impacted third molar. The vertical
positions of the MhM were also evaluated and classified
into 3 groups. When the posterior border of the muscle
could be observed on an axial image where a part of the
crown was simultaneously visible, the position was
classified into the intermediate position. The posterior
borders above or below the crown were classified as above
or below positions, respectively.
Ohshima et al 591
Fig 2. The horizontal relationships between the masseter and
medial pterygoid muscles was classified as belonging to the
anterior group when the most anterior points of the muscles
were situated more anterior to the most medial position of
the impacted crown. The anterior muscle points of the in-
termediate group were situated between the most medial and
distal positions of the crown. In the posterior group, the
anterior points were situated posterior to the most distal
position of the crown. The horizontal position of the
mylohyoid muscle was determined based on the position of
the posterior border of the muscle to the crown of the
impacted third molar.
Infected patients
Between 1996 and 2003, 90 patients were examined
with CT at our department under the diagnosis of
odontogenic infection likely originating from the
mandibular teeth, because the spread could not be
evaluated by physical and conventional radiographic
examinations. Among them, 12 patients (3 females and
9 males) had an infection originating from impacted
mandibular third molar, which had not been extracted at
CT examination. CT images were obtained with the
same scan conditions as were used on uninfected
patients. The causal tooth was identified with reference
to the clinical findings and clinical course together with
radiographic appearances. Patient age ranged from 15 to
85 years old (mean age of 46.4 6 12.2 years). The
duration from the onset of infection to CT examination
ranged from 5 days to 90 days. Infection appeared
within 7 days in 8 patients, within 30 days in 2 patients,
and after 30 days in 2 patients. The clinical data are
summarized in Table I.
 OOOOE
592 Ohshima et al November 2004

Table I. Clinical and CT features of infected patients

Clinical sign and symptom CT appearance
Bony change Muscle and space involvement
State of Vertical Buccal Lingual Periosteal
No. Age Sex impaction position Trismus Dysphagia Dyspnea Pyrexia plate plate reaction BS MM TM MPM MhM SMS PPS SLS PS
1 18 M Partial A + + + + +
2 40 M Partial A + + + +
3 47 M Submucosal B + + + + +
4 55 F Submucosal B + + + + + + + +
5 25 M Submucosal B + + + + + + + +
6 39 M Submucosal B + + + + + + + + +
7 22 M Partial A + + + + + + + +
8 24 F Partial A + + + +
9 24 M Partial B + + + + + + +
10 28 M Partial B + + + + + +
11 24 F Partial A + + + + + + + + +
12 27 M Partial A + + + + + + + + + + +

BS, buccal space; MM: masseter muscle; TM, temporal muscle; MPM, medial pterygoid muscle; MhM, mylohyoid muscle; SMS. submandibular space; PPS,
parapharyngeal space; SLS, sublingual space; PS, parotid space.

Table II. State of impaction and vertical position of
the mandibular third molar in uninfected patients (n=136
teeth)
State of impaction Partial impaction
Submucosal impaction
Complete impaction
Vertical position on Position A
panoramic radiography
Position B
Position C
Position A, the occlusal plane of the impacted third molar was at the same level
to those of the second molar; Position B, the occlusal plane of impacted third
molar situated between the occlusal plane and the cervical line of the second
molar; Position C, the occlusal plane of impacted third molar was below the
cervical line of the second molar.
All images were evaluated by 2 observers (A.O. and
Y.A.). Bony involvement was assessed for the disappear-
ance or interruption of continuity of cortical plates and the
presence of periosteal reaction. The spread of infection to
the soft tissues was determined with reference to bilateral
asymmetry of shape and density. A final assessment was
reached by consensus after discussion when the evalua-
tion was different between the observers.
RESULTS
Uninfected patients
The states of impaction and vertical positions are
summarized in Table II. Approximately one half of the
impacted molars were classified as ‘‘submucosal impac-
tion,’’ while complete impaction was seen in only 14
(10.3%) patients. For the vertical position, 74 teeth
belonged to position B; that is, the occlusal plane of the
impacted molar was situated between the occlusal plane
and cervical line of the second molar. The lowest position
(position C) was observed in 20% of the patients.
In 48 (35.3%) of 136 impacted third molars, the
lingual cortical plate could not be observed in the slice
where the crown area was maximal (Fig 3), while only
52 (38.2%) 11 were observed on the buccal side (Table III). The
70 (51.5%) bony width was thicker on the buccal side than the
14 (10.3%) lingual side. This difference was statistically significant
35 (25.7%)
with a P value less than .001. The thickness of the buccal
74 (54.4%) cortical plate was more than twice that of the lingual
27 (19.9%) side, and this difference was significant (P \.001).
The MM and crown horizontally overlapped (anterior
or intermediate group) in 89 (65.4%) of 136 impacted
third molar areas, while 129 (94.8%) of the MPM were
situated posteriorly apart from the impacted crowns
(posterior group) (Table IV). Almost all posterior
borders of the MhM were situated at the intermediate
or posterior position of the impacted crowns. Moreover,
the muscle was vertically positioned equivalent or
inferior to the crowns.
There were no relationships between the bucco-
lingual aspects on CT image and the state of impaction
or the vertical positions determined by panoramic x-ray
examination.
Infected patients
There were no complete impactions of the mandibular
third molars in the infected patients (Table I). For the
vertical position, no impacted third molars were found
in position C.
Eight (66.7%) of 12 patients showed the absence of
the lingual cortical plate around the crown of the
infected mandibular third molar (Fig 4, A), while only 4
patients exhibited this on the buccal side (Fig 5, A). In 3
of these 4 patients, the MM was involved (Fig 5, B). Two
patients (Patient Nos. 11 and 12) showed widespread
mandibular bony changes with periosteal reaction,
OOOOE
Volume 98, Number 5 Ohshima et al 593

Table III. Bony features of 136 uninfected mandibular

third molars
Disappearance of the Buccal side 11/136 (8.1%)
cortical plate
Lingual side 48/136 (35.3%)
Bony width* Buccal side (n=125) 3.6 6 1.8mm
Lingual side (n=88) 1.9 6 1.5mm
Cortical thickness* Buccal side (n=125) 2.9 6 1.2mm
Lingual side (n=88) 1.3 6 0.7mm

*A significant difference was found between the values of buccal and

lingual sides with P\ .001

Table IV. Relationship between the crown and the

muscles in uninfected patients (n=136)
Masseter muscle Anterior 9 (6.6%)
(horizontal position)
Intermediate 80 (58.8%)
Posterior 47 (34.6%)
Medial pterygoid muscle Anterior 0 (0%)
(horizontal position)
Intermediate 7 (5.2%)
Posterior 129 (94.8%)
Mylohyoid muscle Anterior 1 (0.7%)
Fig 3. In an uninfected patient, the lingual cortical plate (horizontal position)
cannot be observed on the left side (arrow) at the slice where Intermediate 50 (36.8%)
the maximal crown area is shown. Posterior 85 (62.5%)
Mylohyoid muscle Above 0 (0%)
(vertical position)
suggesting the appearance of osteomyelitis of the
Intermediate 16 (11.8%)
mandible (Fig 6, A). Below 120 (88.2%)
All patients showed soft tissue changes caused by
mandibular third molar infection. Among the muscles
evaluated, the MhM was most frequently involved in 10
(83.3%) of 12 patients (Fig 4, C, and Fig 7). Of them, 8 DISCUSSION
patients showed submandibular space involvement (Fig Partial or submucosal impaction is known to be a
4, C). Masticator space involvement was found in 10 significant risk factor for acute pericoronitis in impacted
patients. Among the masticatory muscles, which were mandibular third molars.3-6 As for the vertical position
included in the masticator space, the MPM was most on panoramic radiographs, pericoronitis frequently
frequently involved in 7 patients (Fig 6, A, and Fig 7) occurs in positions A and B, in which the occlusal plane
followed by the MM. Structures situated on the lingual of an impacted third molar is situated at the same level of
side such as the MhM and MPM were more frequently the occlusal plane or between the plane and cervical line
involved than those on the buccal side (the MM). The of the second molar, respectively.2,4 The present study
parapharyngeal space was involved in 5 patients (Fig 4, supports these observations because no infected patients
B), and 4 of them were accompanied by submandibular showed complete impaction or position C, in which the
space involvement. Two patients with sublingual occlusal plane of the third molar was positioned below
space changes had concomitant involvement of the the cervical line of the second molar. Hugoson and
submandibular and parapharyngeal spaces. Two Kugelberg10 carried out a randomized panoramic survey
patients with osteomyelitis exhibited involvement of to determine the frequency of each state of impaction.
the parotid or temporal space. According to their results, submucosal, partial, and
Nine of 11 patients with trismus showed changes in complete impactions occur at frequencies of 47%, 37%,
the MM and/or MPM. In 5 of 7 patients with dysphagia, and 18%, respectively. The results of the present study
the parapharyngeal spaces were involved. Three were roughly equivalent to theirs, although all un-
patients with dyspnea were accompanied by pa- infected patients in the present study were scanned with
rapharyngeal space involvement. All 4 patients with CT for preextraction examination.
pyrexia had multistructure involvement, especially on On the basis of anatomical observations, the lingual
the lingual side. wall is thinner than the buccal wall in the mandibular
 OOOOE
594 Ohshima et al November 2004

Fig 4. A CT image expressed by bony windows (A) shows the disappearance of the left lingual cortical plate (arrow) adjacent to the
crown of the impacted mandibular third molar in a 25-year-old male (Patient No. 5). Parapharyngeal space involvement can be
clearly observed (*), but the fat line adjacent to the medial pterygoid muscle is preserved (arrow heads) (B). At the inferior slice,
displacement of the fatty midline of the tongue is visible (arrow) and swelling of the mylohyoid muscle (arrow heads) and
submandibular space involvement (*) can also be observed (C).

third molar region.7 The present study supports this
observation through cortical thickness measurement.
The uninfected cortical thickness was more than twice
that found on the buccal side compared with the lingual
side. A difference between the bony width and cortical
thickness indicates the thickness of cancerous bone
around the crown. The present results show that
cancerous bone thickness was approximately equal
between the buccal and lingual sides. Therefore, the
cortical thickness mainly contributes to the bony width
and may influence the bucco-lingual spread of infection
originating from an impacted mandibular third molar. In
uninfected patients, 48 (35.3%) cortices disappeared on
the lingual side, whereas only 11 (8.1%) disappeared on
the buccal side. This observation supports the hypoth-
esis that pericoronitis of the mandibular third molar is
more likely to spread to the lingual side than the buccal
side. In infected patients, involvement was observed
more frequently in lingual structures (lingual cortical
plate, MPM, and MhM) than in buccal structures
(buccal cortical plate, buccal space, and MM). For the
cortical disappearance, the frequencies in infected
patients were higher than those in uninfected patients
for the lingual and buccal sides. Although some patients
with preexisting cortical disappearance were included,
this result suggests that the disappearance was caused by
infection. Taken together, the anatomical features of
bony structures support the hypothesis that pericoronitis
of the third mandibular molars is more likely to spread to
the lingual side than the buccal side.
The MM is frequently involved in odontogenic
infection originating from mandibular teeth,12-15 and
various incidences of this muscle involvement have
been reported.14,15 Most of reports include infections
originating from areas other than the third molar or
infections occurring after the extraction of an impacted
OOOOE
Volume 98, Number 5 Ohshima et al 595

Fig 5. In an 18-year-old male (Patient No. 1), a CT image shows a disappearance of the right buccal cortical plate (arrow) of the
mandibular third molar (A). This appearance can be recognized when compared to that of the left side where the third molar was
extracted. A CT image expressed by soft tissue windows depicts an enlarged MM (white arrow heads) together with buccal space
involvement (*) (B).

Fig 6. A soft tissue CT image shows swelling of the left masseter muscle (*) and medial pterygoid muscle (arrow) in a 27-year-pld
male (Patient No. 12) (A). The inferior slice expressed by bony windows clearly depicts cortical resorption (arrows) and periosteal
reactions (arrow heads) of the left ramus (B).

third molar. Therefore, these incidences are not specific
to infection originating in the pericoronitis of an
impacted mandibular third molar. In the present study,
6 (50%) of 12 infected patients showed involvement of
the MM. In 65.4% of uninfected patients, the MM
horizontally overlapped the crown of impacted third
molars. This indicates that infection frequently involves
the MM when it spreads to the buccal side. Indeed, all 4
infected patients that exhibited changes in their buccal
cortices, which was a proof of buccal spread, showed
MM involvement.
An uninfected MPM was situated posterior to and apart
from the crown. In infected patients, however, involve-
ment occurred at a relatively high rate (58.3%) in the
MPM. This discrepancy indicates the existences of
pathways of infection to the MPM from the pericoronitis

596 Ohshima et al
Fig 7. On a CT image of a 55-year-old female (Patient No. 4),
enlargement of the medial pterygoid and mylohyoid muscles
and the disappearance of fatty plane between them can be
observed on the right side (arrow).
of the impacted third molar. Two pathways may exist. In
the first, infection spreads posteriorly, involving the
retromolar triangle. Thereafter, it spreads into the MPM
through the fat space between the mandible and the MPM,
which is generally known as the pterygo-mandibular
space. The same route has been reported for the spread of
cancer of the mandibular gingiva into the masticator
space.16 The second pathway includes that infection
reaching the MPM through the mandibular bone or
periostium. It may be accompanied by bony or periosteal
changes on images. In 2 (Patient Nos. 11 and 12) of 7
infected patients with involvement of the MPM, a wide-
spread appearance of osteomyelitis and interruptions of
cortices at the attached portion of the MPM were
observed. In these patients, infection may extend via the
mandibular bone (the second pathway). In the remaining
5 infected patients with MPM involvement, infection may
have spread through the first pathway.
The relationship between the MhM and the root apices
of mandibular molars is essential for infection spread.14
Periapical infection perforates the lingual cortices and
involves the sublingual space and the MhM when the
root is situated superiorly to the attached portion of the
MhM in the mandible. Conversely, infections originat-
ing in the root apices, which are positioned beyond the
attached portion of the MhM, directly extend into the
submandibular space (SMS). Despite the position of
the apices, most infections consequently involve the
SMS.2,14,17 The present study allowed for observations
of the relationships between the MhM and an impacted
crown. Most uninfected MhMs horizontally overlapped
and were situated inferior to the crown of the impacted
third molar. These anatomical relationships indicate that
OOOOE
November 2004
pericoronitis of the third molar frequently involves the
MhM and subsequently spreads into the SMS.
Moreover, the free posterior edge of the MhM serves
a direct pathway to the SMS from the crown. Our
observations of the infected patients support these
hypotheses. Ten (83.3%) of 12 infected patients showed
involvement of the MhM, and 8 (80%) of them showed
SMS involvement. As previously reported, infection
involving the SMS frequently spreads into the para-
pharyngeal space (PPS) because there are no distinct
fascial barriers between these spaces.14 Therefore, SMS
involvement should be carefully observed.
Since the buccinator muscle is attached bucco-
lingually to the mandible near the impacted third molar,
it probably plays a role in the spread of infection.
However, we could not investigate the muscle in the
present study because the uninfected buccinator muscle
could not be clearly identified on axial CT images.
In 2 patients (Patient Nos. 11 and 12) with extensive
osteomyelitis, involvement was observed in the tempo-
ral muscle or parotid gland, which are situated apart
from the crown of the third molar. In these patients,
infection may spread through the marrow space and
perforate the cortical plate near the involved structures.
As previously reported, trismus is a significant
symptom in the majority of patients with masticator
space involvement.13,17,18 Nine of 11 patients with
trismus had MM and/or MPM involvement. The PPS is
an important space because patients with this space
involvement often have accompanying dysphasia or
dyspnea.19,20 Moreover, the PPS is a space through
which infection can spread into the retropharyngeal and
carotid spaces, resulting in a life-threatening condition.
In the present study, 5 of 7 patients with dysphasia and
all 3 patients with dyspnea had accompanying PPS
involvement. When PPS involvement is suggested on
images, early and appropriate treatment should be
performed.
In conclusion, anatomical relationships supported the
clinical observations of infection spread in patients with
pericoronitis of the impacted mandibular third molar.
CT was effective in the evaluation of spread of infection
originating from this tooth.
REFERENCES
1. Haug RH, Hoffman MJ, Indresano AT. An epidemiologic and
anatomic survey of odontogenic infections. J Oral Maxillofac
Surg 1991;49:976-80.
2. Indresano AT, Haug RH, Hoffman MJ. The third molar as
a cause of deep space infections. J Oral Maxillofac Surg 1992;
50:33-5; discussion 35-6.
3. Leone SA, Edenfield MJ, Cohen ME. Correlation of acute
pericoronitis and the position of the mandibular third molar. Oral
Surg Oral Med Oral Pathol 1986;62:245-50.
OOOOE
Volume 98, Number 5
4. Venta I, Turtola L, Murtomaa H, Ylipaavalniemi P. Third molars
as an acute problem in Finnish university students. Oral Surg
Oral Med Oral Pathol 1993;76:135-40.
5. Halverson BA, Anderson WH 3rd. The mandibular third molar
position as a predictive criteria for risk for pericoronitis: a retro-
spective study. Mil Med 1992;157:142-5.
6. Knutsson K, Brehmer B, Lysell L, Rohlin M. Pathoses associated
with mandibular third molars subjected to removal. Oral Surg
Oral Med Oral Pathol Oral Radiol Endod 1996;82:10-7.
7. Shicher H, Dubrul EL. Oral anatomy. 6th ed. St Louis: C.V.
Mosby; 1975. p. 426.
8. Feifel H, Riediger D, Gustorf-Aeckerle R. High resolution
computed tomography of the inferior alveolar and lingual nerves.
Neuroradiology 1994;36:236-8.
9. Pawelzik J, Cohnen M, Willers R, Becker J. A comparison of
conventional panoramic radiographs with volumetric computed
tomography images in the preoperative assessment of impacted
mandibular third molars. J Oral Maxillofac Surg 2002;60:
979-84.
10. Hugoson A, Kugelberg CF. The prevalence of third molars in
a Swedish population. An epidemiological study. Community
Dent Health 1988;5:121-38.
11. Pell GJ, Gregory TG. Report on a ten-year study of a tooth
division technique for the removal of impacted teeth. Am J
Orthod Dentofacial Orthop 1942;28:660-6.
12. Garcia AG, Sampedro FG, Rey JG, Vila PG, Martin MS.
Pell-Gregory classification is unreliable as a predictor of difficulty
in extracting impacted lower third molars. Br J Oral Maxillofac
Surg 2000;38:585-7.
13. Ariji E, Moriguchi S, Kuroki T, Kanda S. Computed tomography
of maxillofacial infection. Dentomaxillofac Radiol 1991;20:
147-51.
Ohshima et al 597
14. Kim HJ, Park ED, Kim JH, Hwang EG, Chung SH. Odontogenic
versus nonodontogenic deep neck space infections: CT mani-
festations. J Comput Assist Tomogr 1997;21:202-8.
15. Yonetsu K, Izumi M, Nakamura T. Deep facial infections of
odontogenic origin: CT assessment of pathways of space
involvement. AJNR Am J Neuroradiol 1998;19:123-8.
16. Kimura Y, Sumi M, Sumi T, Ariji Y, Ariji E, Nakamura T. Deep
extension from carcinoma arising from the gingiva: CT and MR
imaging features. AJNR Am J Neuroradiol 2002;23:468-72.
17. Ariji Y, Gotoh M, Kimura Y, Naitoh M, Kurita K, Natsume N,
et al. Odontogenic infection pathway to the submandibular space:
imaging assessment. Int J Oral Maxillofac Surg 2002;31:165-9.
18. Hardin CW, Harnsberger HR, Osborn AG, Doxey GP, Davis RK,
Nyberg DA. Infection and tumor of the masticator space: CT
evaluation. Radiology 1985;157:413-7.
19. Schuman NJ, Owens BM. Ludwig’s angina following dental
treatment of a five-year-old male patient: report of a case. J Clin
Pediatr Dent 1992;16:263-5.
20. Peterson LJ. Contemporary management of deep infections of
the neck. J Oral Maxillofac Surg 1993;51:226-31.
Reprint requests:
Eiichiro Ariji, DDS, PhD
Department of Oral and Maxillofacial Radiology
Aichi-Gakuin University School of Dentistry
2-11 Suemori-dori, Chikusa-ku
Nagoya 464-8651, Japan
ariji@dpc.aichi-gakuin.ac.jp