CLINICAL ARTICLES

J Oral Maxillofac Surg

61:417-421, 2003
Panoramic Radiographic Risk Factors for
Inferior Alveolar Nerve Injury After Third
Molar Extraction
Bart F. Blaeser, DMD, MD,* Meredith A. August, DMD, MD,
R. Bruce Donoff, DMD, MD, Leonard B. Kaban, DMD, MD,
and Thomas B. Dodson, DMD, MPH
Purpose: The purpose of this study was to estimate the association between specic pan-
oramic radiographic signs and inferior alveolar nerve (IAN) injury during mandibular third molar
removal.
Patients and Methods: A case-control study design was used; the sample consisted of patients
who underwent removal of impacted mandibular third molars. Cases were dened as patients with
conrmed IAN injury after third molar extraction, whereas controls were dened as patients without
nerve injury. Five surgeons, who were blinded to injury status, independently assessed the preop-
erative panoramic radiographs for the presence of high-risk radiographic signs. Bivariate analyses
were completed to assess the relationship between radiographic ndings and IAN injury. The
sensitivity, specicity, and positive and negative predictive values were computed for each radio-
graphic sign.
Results: The sample was composed of 8 cases and 17 controls. Positive radiographic signs were
statistically associated with an IAN injury (P .0001). The presence of radiographic sign(s) had positive
predictive values that ranged from 1.4% to 2.7%, representing a 40% or greater increase over the baseline
likelihood of injury (1%) for the individual patient. Absence of these radiographic ndings had a strong
negative (99%) predictive value.
Conclusions: This study conrms previous analyses showing that panoramic ndings of diver-
sion of the inferior alveolar canal, darkening of the third molar root, and interruption of the cortical
white line are statistically associated with IAN injury. Based on the estimated predictive values, the
absence of positive radiographic ndings was associated with a minimal risk of nerve injury,
whereas, the presence of one or more of these ndings was associated with an increased risk for
nerve injury.
2003 American Association of Oral and Maxillofacial Surgeons
J Oral Maxillofac Surg 61:417-421, 2003
Received from the Department of Oral and Maxillofacial Surgery,
Massachusetts General Hospital, Harvard School of Dental Medi-
cine, Boston, MA.
*Clinical Associate.
Assistant Professor.
Professor of Oral and Maxillofacial Surgery and Dean, Harvard
School of Dental Medicine.
Walter C. Guralnick Professor and Chairman.
Associate Professor and Director of Resident Training.
Presented as an abstract at the 78th General Session of the
International Association for Dental Research, Washington, DC,
April 2000.
This project was funded by the Department of Oral and Maxil-
lofacial Surgery Research Fund, Massachusetts General Hospital
(B.F.B., M.A.A., R.B.D., and L.B.K.) and the Mid-Career Investigators
in Patient-Oriented Research award from the NIDCR (K24-
DE00448 (T.B.D.).
Address correspondence and reprint requests to Dr Dodson:
Department of Oral and Maxillofacial Surgery, Massachusetts Gen-
eral Hospital, 55 Fruit St, Warren 1201, Boston, MA 02114; e-mail:
tdodson@partners.org
2003 American Association of Oral and Maxillofacial Surgeons
0278-2391/03/6104-0002$30.00/0
doi:10.1053/joms.2003.50088
417
A well-recognized, but uncommon, serious complica-
tion of mandibular third molar extraction is injury to
the inferior alveolar nerve (IAN).
1-6
The overall risk of
IAN injury associated with third molar extraction
ranges from 0.5% to 5%. In most cases, the injured
nerve recovers spontaneously. The reported rate of
permanent IAN injury is less than 1%.
1
Although the
risk of permanent nerve injury is low, for affected
patients there is considerable dissatisfaction and mor-
bidity.
The risk of IAN injury after third molar extraction is
a function of the anatomic relationship between the
nerve and the tooth. Panoramic imaging is the stan-
dard radiologic examination to evaluate the anatomic
relationship of third molars and the IAN.
2
Several
authors have identied high-risk radiologic signs asso-
ciated with an intimate, anatomic relationship be-
tween third molars and the IAN. These radiologic
signs include 1) darkening of the root where it crosses
the inferior alveolar canal, 2) deected or hooked
roots around the inferior alveolar canal, 3) narrowing
of the root implying perforation or grooving by the
nerve, 4) a bid root apex representing intimacy of
the apical periodontal membrane, 5) interruption or
obliteration of either of the radiopaque white (corti-
cal) lines of the inferior alveolar canal, 6) diversion or
bending of the inferior alveolar canal in the region of
the root apices, and 7) narrowing of the inferior
alveolar canal.
3-6
Of these 7 signs, Rood and
Nooraldeen Shehab
7
identied 3 (radiologic signs 1,
5, and 6) that were signicantly associated with IAN
injury. Although the panoramic radiograph is a useful
screening tool for assessing the anatomic relationship
between third molars and the IAN, it is imperfect. In
Rood and Nooraldeen Shehabs study, the reported
sensitivity for these 3 signs ranged from 24% to 38%,
and the reported specicity ranged from 96% to 98%.
As an oral and maxillofacial surgery group working
in a referral center for the management of IAN inju-
ries, we have the opportunity to document the pres-
ence and nature of such injuries at the time of repar-
ative surgery. In this study, we correlated operative
ndings with preoperative radiographic signs to eval-
uate their predictive value in a sample of patients with
conrmed injury to the IAN. Given the variability of
the sensitivity and specicity estimates in the litera-
ture, the purpose of this study was to measure the
association between radiographic signs, noted on
panoramic imaging, and documented IAN injury by
limiting the sample to patients with conrmed nerve
injury. We believe this to be the rst study to use a
sample composed of patients with operatively con-
rmed IAN injury to assess the relationship between
panoramic radiographic ndings and risk for IAN in-
jury.
Patients and Methods
STUDY DESIGN/SAMPLE
This was a retrospective case-control study com-
posed of patients who had 2 impacted mandibular
third molars removed. Cases were dened as patients
with a documented IAN injury, manifested by com-
plaints of persistent sensory decit or dysesthesia
after third molar removal, who underwent operative
exploration of the nerve. At the time of exploratory
surgery, the injury was conrmed. The case-control
study design is specically targeted to address situa-
tions where the outcome of interest, that is, IAN
injury, is a rare event. Controls were dened as pa-
tients who had third molars extracted but did not
complain of a postoperative IAN injury and the neg-
ative complaint was conrmed by a normal neurosen-
sory examination. The control patients were ran-
domly derived from a population of patients treated in
the Oral and Maxillofacial Surgery Unit, Massachusetts
General Hospital, Boston, MA. Data were collected by
reviewing charts and panoramic radiographs. The In-
stitutional Review Board approved the project.
STUDY VARIABLES
The predictor variable was the presence or absence
of 1 or more preoperative panoramic radiographic
ndings: 1) diversion or bending of the canal, 2)
darkening of the tooth root, and 3) interruption of the
cortical white line of the canal. These 3 radiographic
signs were selected because they have been shown to
be associated with IAN injury.
7
Five surgeons who were blinded to the IAN injury
status independently reviewed the preoperative pan-
oramic radiographs and assessed the images for the
presence of the high-risk radiographic signs. The sur-
geons also evaluated third molar angulation and the
level of third molar impaction. Angulation was de-
ned as mesioangular, distoangular, horizontal or ver-
tical. Level of impaction was classied using the Pell
and Gregory system.
8
The outcome variable was pres-
ence (case) or absence (control) of an IAN injury. In
addition, we abstracted age, gender, and indications
for extraction from the patients record.
DATA ANALYSES
A database was constructed using Microsoft Access
(Redmond, WA). Data analyses were completed using
SPSS version 8.0 (SPSS, Inc, Chicago, IL). Descriptive
statistics of the sample were computed. Bivariate anal-
yses were used to assess the relationship among pan-
oramic signs, IAN injury, and variability of individual
clinicians radiographic interpretations. The sensitivity
and specicity of each radiographic sign were com-
puted. If P .05, the results were considered statis-
tically signicant.
418 PANORAMIC RADIOGRAPH AND IAN INJURY
Results
The study sample was composed of 25 patients, of
whom 8 had documented evidence of IAN injury
(cases). The remaining 17 patients were classied as
controls. The mean ages of the case and control sam-
ples were 34 and 27 years, respectively. Both case and
control samples were predominantly composed of
females: 88% and 65%, respectively. There was no
statistically signicant difference between the cases
and controls in terms of age or gender. In all of the
cases, at operation, there were 1 or more ndings
consistent with IAN injury: nerve compression (n
2), neuroma formation (n 5), and partial (n 2) or
total (n 1) IAN transection.
The 5 surgeons independently evaluated 2 third
molars for each of the 25 patients, resulting in a total
of 250 third molar observations. There were 40 ob-
servations of third molars associated with IAN injury
and 210 observations of third molars not associated
with IAN injury. The clinicians evaluations of the
radiographs were consistent as evidenced by the ab-
sence of a statistically signicant difference in classi-
cation of third molar angulation, level of impaction,
or identication of the high-risk radiographic signs
(P .31). Neither angulation nor level of impaction
was associated with IAN injury (P .72).
Table 1 summarizes the association of high-risk
radiographic signs and IAN injury. Consistent with the
results of Rood and Nooraldeen Shehab,
7
diversion of
the canal, darkening of the root, and interruption of
the white line were each statistically associated with
IAN injury (P .0001). The sensitivity and specicity
of the radiographic signs, however, were variable.
Depending on the radiographic sign, the sensitivity
ranged from 50% to 80%, and the specicity ranged
from 52% to 82% (Table 2). The sensitivity and spec-
icity for one or more positive ndings were 100%
and 33%, respectively.
Discussion
IAN injury after third molar extraction is a well-
recognized complication with a reported prevalence
ranging from 0.5% to 5%.
9,10
Various contributing fac-
tors, including age, history of infection, use of rotary
instruments, use of concomitant general anesthesia,
and level of impaction, have been reported.
11-13
The
most important factor contributing to IAN injury may
be the anatomic proximity of the IAN to the third
molar root. This anatomic intimacy is well known by
oral and maxillofacial surgeons; however, direct con-
tact is only rarely and anecdotally discussed. At the
Massachusetts General Hospital, many patients (1 to 3
per month) undergo operative treatment for postex-
traction nerve injuries. As such, the existence of an
anatomic injury to the inferior alveolar or lingual
nerves can be denitively documented by direct ex-
amination at the time of surgery. Given this unique
opportunity, the authors thought it was an excellent
opportunity to reevaluate the usefulness of panoramic
imaging to establish the relationship between the IAN
canal and third molars and the risk for associated
Table 1. RELATIONSHIP OF RADIOGRAPHIC
FINDINGS TO IAN INJURY
IAN Injury
Yes No Totals
Diversion of the canal
Yes 20 38 58
No 20 172 193
Total 40 210 250
Darkening of root
Yes 26 56 82
No 14 154 192
Total 40 210 250
Interruption of cortical line
Yes 32 96 128
No 8 114 122
Total 40 210 250
Any positive radiographic nding
Yes 40 140 180
No 0 70 70
Total 40 210 250
NOTE. For each category, the difference between presence and
absence of IAN injury was signicant (P .0001).
Abbreviation: IAN, inferior alveolar nerve.
Table 2. ESTIMATES OF SENSITIVITY, SPECIFICITY, AND PPF AND NPV
Radiographic
Finding
Sensitivity
(%)
Specicity
(%)
16% Prevalence of
IAN Injury
1% Prevalence of
IAN Injury
PPV
(%)
NPV
(%)
PPV
(%)
NPV
(%)
Diversion 50 82 34 89 2.7 99
Darkening 65 73 31 93 2.3 99
Interruption 80 54 25 93 1.7 99
Any nding 100 33 22 100 1.4 100
Abbreviations: IAN, inferior alveolar nerve; PPV, positive predictive value; NPV, negative predictive value.
BLAESER ET AL 419
nerve injury. Needless to say, this rigorous denition
for nerve injury limited the number of patients eligi-
ble for inclusion as cases. In the setting where the
outcome variable of interest is rare, such as IAN in-
jury, the case-control study design is indicated.
In this study, third molar angulation was not asso-
ciated with risk for IAN injury. This nding contrasts
with other reports in which third molar angulation
was associated with IAN injury.
6,14
Given this studys
small sample size, we caution against weighting the
negative ndings heavily because there is a signicant
risk for type II errors (reporting no difference when a
difference exists).
Currently, panoramic radiographic imaging is the
standard technique for assessing the risk of IAN injury
after third molar extraction.
15-17
We hypothesized that
by limiting the sample to patients with conrmed IAN
injury, we would identify radiographic signs associ-
ated with nerve injuries with a high level of sensitivity
and specicity. The results of this study showed that
the high-risk radiographic signs had fair to poor sen-
sitivity and specicity. These ndings contrast with
the study reported by Rood and Nooraldeen Shehab,
7
in which the high-risk radiographic signs had weak
sensitivity but high specicity. The discrepancy in the
results may be due to the difference in case selection.
In Rood and Nooraldeen Shehabs
7
study, a case was
dened as any IAN injury after third molar removal. In
the current study, a case was limited to those patients
in whom the surgeon visualized nerve injury at the
time of nerve exploration and repair.
Despite the numerical differences between the cur-
rent study and that of Rood and Nooraldeen Shehab,
7
the clinical implications are similar. In both studies,
the radiographic signs were associated with nerve
injury with high statistical signicance. However,
when the surgeon applies these ndings in a clinical
setting, it is important to estimate the overall predic-
tive value of the diagnostic test. Positive predictive
value (PPV) is the probability that the disease is
present when the test or sign is present. In this case,
PPV is the probability that an IAN injury will occur
when a high-risk radiographic sign is present. Nega-
tive predictive value (NPV) is the probability that the
disease is absent if the sign is absent. Specically for
this study, the NPV is the probability that an IAN
injury will not occur if high-risk radiographic signs are
absent.
The predictive value of any diagnostic test or sign is
a function of the sensitivity and specicity of the test
or sign and the underlying prevalence of the disease
or disorder. In this study, the prevalence of IAN injury
(16%) was articially elevated due to the selection
criteria used. Given the sensitivity and specicity in
this study and an IAN injury prevalence of 16%, the
PPV of the various radiographic signs ranged from
25% to 34%. The NPV ranged from 89% to 93%. If we
decrease the prevalence of IAN injury after third mo-
lar removal to 1%, the PPV and NPV change dramati-
cally. The PPV ranges from 1.4% to 2.7%, and the NPV
ranges from 99% to 100% (Table 2).
Table 3 summarizes the sensitivity and specicity
for the various high-risk signs reported by Rood and
Nooraldeen Shehab
7
and estimates the PPV and NPV
assuming a 1% rate of IAN injury. The PPVs range
from 6% to 12%, and the NPV is 99%. In the setting of
an unlikely event such as IAN injury, the presence of
a positive radiographic nding suggests that there is a
slight increase in the absolute risk for injury. Speci-
cally, given a positive radiographic nding, the risk
increases from 1% (background prevalence) to 1.7%
to 12% based on the results of our data and those of
Rood and Nooraldeen Shehab.
7
Conversely, the ab-
sence of positive radiographic ndings suggests that
the risk for IAN injury is remote (1%).
In summary, a patient with 1 or more high-risk
radiographic ndings has a signicantly increased risk
(70% to 1,200%) for nerve injury, although overall this
is still an uncommon event. Conversely, in the ab-
sence of any positive ndings, the risk for nerve injury
is negligible. In the absence of high-risk radiographic
signs, there is little preventive value in obtaining ad-
ditional imaging. However, when 1 or more high-risk
radiographic signs are present, additional imaging,
such as computed tomographic imaging in the axial,
coronal, and sagittal planes, may be indicated to bet-
ter establish the anatomic relationship between the
IAN canal and third molars.
18-20
More accurate risk
assessment may help the surgeon and patient in mak-
ing the decision for or against removal of the tooth in
question and in discussing alternative management
strategies.
21-23
The results of this study conrm other reports of a
statistically signicant association between specic
panoramic radiographic signs and the risk for IAN
injury during third molar removal. To apply these
ndings in a clinical setting, one should consider the
Table 3. ESTIMATES OF PPV AND NPV BASED ON
ROOD AND NOORALDEEN SHEHAB
7
Radiographic
nding
Sensitivity
(%)
Specicity
(%)
1%
Prevalence
of IAN
Injury
PPV
(%)
NPV
(%)
Diversion 29 98 12 99
Darkening 38 95 7 99
Interruption 24 96 6 99
Abbreviations: IAN, inferior alveolar nerve; PPV, positive predic-
tive value; NPV, negative predictive value.
420 PANORAMIC RADIOGRAPH AND IAN INJURY
PPV and NPV of these radiographic signs. The results
of this study indicate that in the absence of predictive
radiographic signs, the risk of IAN injury is negligible.
In the presence of a high-risk radiographic nding,
however, additional patient assessment may be indi-
cated.
Acknowledgments
We would like to thank Drs Edward Seldin, Walter C. Guralnick,
Kenneth A. MacAfee, II, John A. Buehler, and Carole A. Lorente for
volunteering to review the radiographs.
References
1. Van Gool AV, Ten Bosch JJ, Boering G: Clinical consequences
of complaints and complications after removal of the mandib-
ular third molar. Int J Oral Surg 6:29, 1977
2. Smith AC, Barry SE, Chiong AY, et al: Inferior alveolar nerve
damage following removal of mandibular third molar teeth. A
prospective study using panoramic radiography. Aust Dent
42:149, 1997
3. Howe GL, Poyton HG: Prevention of damage to the inferior
dental nerve during extraction of mandibular third molars. Br
Dent J 109:353, 1960
4. Azaz B, Shteyer A, Piamenta M: Radiographic and clinical man-
ifestations of the impacted mandibular third molar. Int J Oral
Surg 5:153, 1976
5. Rud J: Third molar surgery: Relationship of root to mandibular
canal and injuries to inferior dental nerve. Tandlaegebladet
87:619, 1983
6. Kipp DP, Goldstein BH, Weiss WW: Dysesthesia after mandib-
ular third molar surgery: A retrospective study and analysis of
1377 surgical procedures. J Am Dent Assoc 100:185, 1980
7. Rood JP, Nooraldeen Shehab BAA: The radiological prediction
of inferior alveolar nerve injury during third molar surgery. Br J
Oral Maxillofac Surg 28:20, 1990
8. Pell GJ, Gregory GT: Report on a ten-year study of a tooth
division technique for the removal of impacted teeth. Am J
Orthod 28:660, 1942
9. Von Arx DP, Simpson MT: The effect of dexamethasone on
neuropraxia following third molar surgery. Br J Oral Maxillofac
Surg 27:477, 1989
10. Alling CC: Dysesthesia of the lingual and inferior alveolar
nerves following third molar surgery. J Oral Maxillofac Surg
44:454, 1986
11. Brann CR, Brickley MR, Shepherd JP: Factors inuencing nerve
damage during lower third molar surgery. Br Dent J 186:514,
1999
12. Sisk AL, Hammer WB, Shelton DW, Joy ED Jr: Complications
following removal of impacted third molars: The role of the
experience of the surgeon. J Oral Maxillofac Surg 44:855, 1986
13. Rood JP: Degrees of injury to the inferior alveolar nerve sus-
tained during the removal of impacted mandibular third molars
by the lingual split technique. Br J Oral Surg 21:103, 1983
14. Tulloch JF, Antcazk-Bouckoms AA: Decision analysis in the
evaluation of clinical strategies for the management of mandib-
ular third molars. J Dent Educ 51:652, 1987
15. Nortje CJ, Farman AG, Grotepass FW: Variations in the normal
anatomy of the inferior dental canal: A retrospective study of
panoramic radiographs from 3612 routine dental patients. Br J
Oral Surg 15:55, 1977
16. Westesson PL, Carlsson LE: Anatomy of mandibular third mo-
lars. A comparison between radiographic appearance and clin-
ical observations. Oral Surg Oral Med Oral Pathol 49:90, 1980
17. Lindh C, Petersson A, Klinge B: Measurements of distances
related to the mandibular canal in radiographs. Clin Oral Im-
plants Res 6:96, 1995
18. Feifel H, Riediger D, Gustorf-Aeckerle R: High resolution com-
puted tomography of the inferior alveolar and lingual nerves.
Neuroradiology 36:236, 1994
19. Yang J, Cavalcanti MG, Ruprecht A, et al: 2-D and 3-D recon-
structions of spiral computed tomography in localization of the
inferior alveolar canal for dental implants. Oral Surg Oral Med
Oral Pathol 87:369, 1999
20. Rothman SL, Chaftez N, Rhodes ML, et al: CT in the preoper-
ative assessment of the mandible and maxilla for endosseous
implant surgery. Work in progress. Radiology 168:171, 1988
21. Alantar A, Roisin-Chausson MH, Commissionat Y, et al: Reten-
tion of the third molar roots to prevent damage to the inferior
alveolar nerve. Oral Surg Oral Med Oral Pathol 80:126, 1995
22. Amin M, Haria S, Bounds G: Surgical access to an impacted
lower third molar by sagittal splitting of the mandible: A case
report. Dent Update 22:206, 1995
23. Checchi L, Bonetti GA, Pelliccioni GA: Removing high-risk
impacted mandibular third molars: A surgical-orthodontic ap-
proach. J Am Dent Assoc 127:1214, 1996
BLAESER ET AL 421