i 273

Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 06/01/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

Oral-Maxillary Sinus Fistula

(Oroantral Fistula): Clinical Features
and Findings on Multiplanar CT

James J. Abrahams1 OBJECTIVE. Oroantral fistula, an abnormal communication between the oral cavity
Scott B. Berger and the maxillary sinus, is infrequently diagnosed radiologically. The purpose of this
study was to describe the CT findings and clinical features of oroantral fistula and to
show that dental CT multiplanar reformatting programs can be instrumental in diag-
nosing this condition.
SUBJECTS AND METHODS. The study included eight patients with clinically con-
firmed oroantral fistula or with radiologic evidence of oroantral fistula. Fistula size,
degree of alveolar atrophy, nature of maxillary sinus disease, and related dental dis-
ease were assessed along with the clinical presentations.
RESULTS. The most frequent CT findings were bony discontinuity of the maxillary
sinus floor, communication between the oral cavity and the sinus, soft-tissue opacifi-
cation of the ipsilateral maxillary sinus, focal areas of alveolar atrophy, and associ-
ated periodontal disease. In 6 of 8 patients there was 100% opacification of the
ipsilateral maxillary sinus, and in 5 of 8 patients the contralateral sinus appeared
completely normal. The fistula size ranged from 13.5 mm2 to 189 mm2.
CONCLUSION. The appearance of oroantral fistula on multiplanar CT imaging is
disruption of the bony floor of the maxillary sinus with soft-tissue opacification of the
ipsilateral sinus. Dental reformatted CT can be useful for evaluating patients sus-
pected of having oroantral fistula, and this condition may be found incidentally in
patients referred for evaluation for osseointegrated root-form dental implants.

AJR i995;i65:1273-i276

Oroantral fistula is an abnormal communication between the maxillary sinus

and the oral cavity and may be the result of several different pathologic processes.
Most commonly, it occurs after dental extraction; however, other causes include
infection [i-3], inflammatory conditions [4], neoplasm [5, 6], Pagets disease [7],
iatrogenic injury [8], and trauma [9]. Imaging of oroantral fistulas has been limited
because of technical limitations of CT scanning; however, more recently, dental
reformatted CT scanning has permitted improved visualization of structures of the
jaw and oral cavity.
Oroantral fistula has been well described in articles about clinical dentistry and
oral surgery [9-i i ] but is relatively unfamiliar to radiologists. Our aim is to describe
the radiologic findings of oroantral fistula, correlate the findings with clinical pre-
sentation and history, and raise awareness among radiologists of oroantral fistula
Received December 16, 1994; accepted after
as a possible factor related to unilateral maxillary sinusitis.
revision June 7, 1995.
Presented at the annual meeting of the American
Roentgen Ray Society, Washington, DC, May 1995. Subjects and Methods
1 Both authors: Section of Neuroradiology, De- Selection and Clinical Review
partment of Diagnostic Imaging, Yale-New Haven
Medical Center, 20 York St., New Haven, CT 06830. We identified eight patients either with clinically confirmed (four patients) oroantral fistula
Address correspondence to J. J. Abrahams. or with no clinical suspicion but radiologic evidence of oroantral fistula (four patients). For

O361-8O3X/95/1 655-1 273 patients without a definite clinical diagnosis of oroantral fistula, the radiologic criterion for
American Roentgen Ray Society inclusion was disruption of the floor of the maxillary sinus. Clinical information was obtained
 1274 ABRAHAMS AND BERGER AJR:165, November 1995

for each patient by obtaining clinical records, either verbally from the motion. A lateral digital scout image is obtained, and the position of
referring clinician or by review of the patients record. the maxilla is confirmed.
Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 06/01/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

The patients were from 29 to 62 years old (mean age, 48 years). After obtaining the images, a curved line is first superimposed on
One patient was immunocompromised, with a i-year history of one of the axial images by depositing several fiducial markers along
AIDS. Five patients were referred by dental practitioners, including the curve of the jaw. Cross-sectional images are obtained perpen-
four referred by general dentists and one by a periodontist. Two dicular to the curved line, at 1- to 2-mm intervals. The final images
patients were referred by an otolaryngologist, and one directly from are axial, cross sectional, and panoramic. All images were inter-
the outpatient emergency department. Of the five patients referred preted by one neuroradiologist.
by dental practitioners, oroantral fistula was suspected in only one.
The remaining four patients were referred for routine evaluation for
dental implants. Results
Two patients had recently undergone maxillary tooth extractions.
A third patient with extremely poor dental hygiene had self extracted Dental reformatted CT scans showed disruption of the
several maxillary teeth. In one patient, oroantral fistula developed bony floor of the maxillary sinus with communication to the
after use of a blade type implant. The implant had failed and was oral cavity in all patients. The margins of the fistulas were
found floating in an infected maxillary sinus. irregular, and all were associated with opacification of the
adjacent maxillary sinus. Representative examples are
Imaging Methods shown in Figures 1-3. The fistula tract itself was always of
All CT scans were obtained using GE 9800 or GE Advantage soft-tissue density. Five of the fistulas were on the right, and
machines. Six patients were evaluated with DentaScan software three were on the left.
(GE Medical Systems, Milwaukee, WI), one with ISG Systems soft- Asymmetrically greater opacification of the ipsilateral max-
ware (Toronto, Ontario), and one with standard coronal CT protocol. illary sinus was noted in every patient studied. In six of eight
Images were assessed for fistula size (caliper measurements in patients, there was iOO% opacification of the ipsilateral
three planes), associated maxillary sinus disease, degree of alveo- sinus, and in five of eight patients, the contralateral side was
lar atrophy, associated periodontal disease, and degree of edentia. completely clear.
The technique for dental reformatting is described in detail else-
The fistula size ranged from 13.5 to 189 mm2. Five of eight
where [12-15]. Briefly, axial thin-slice images are obtained parallel
patients were partially edentulous, two were totally edentu-
to the alveolar ridge using a bone algorithm, dynamic mode, 15-cm
field of view, 512 x 5i2 matrix, and slice thickness of either 1 mm
bus, and one had no tooth loss except for the extracted molar
(GE 9800) or 1 .5 mm (GE Advantage). Patients are placed supine in
that led to the fistula. All patients had at least minimal pen-
a head holder to reduce motion. Patients are instructed not to swal- odontal disease, and two patients had severe periodontal dis-
low or move during the scan, and sponges may be used to reduce ease with extensive bony destruction.

Fig. 1 -45-year-old man with left oroantral

fistula.
A, Standard axial CT scan at level of maxil-
lary sinuses shows unilateral opacificatlon of
ipsilateral sinus.
B, Panoramic reconstructed CT scan from a
dental reformatting program shows disruption
of bony floor of maxillary sinus, with fistula tract
of soft-tissue density (arrow). Note associated
soft-tissue opacification of ipsilateral maxillary
sinus (S), whereas contralateral sinus Is aerated.
C, Three cross-sectional CT scans are in true
parallel planes to alveolar ridge and show fistu-
Ia tract (arrow) and sinusitis.
 AJR:165, November 1995 CT OF OROANTRAL FISTULA i275
Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 06/01/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

Fig. 2.-29-year-old woman with large oroantral fistula.

A, Standard coronal CT scan shows extensive beam-hardening artifacts projected from dental prostheses. Fistula of 189 mm3 was detected.
B, Dental reformatted (panoramic) CT scan projects extensive beam-hardening artifacts into plane orthogonal to plane of fistula (arrow).

Fig. 3.-46-year-old woman with persistent

oroantral fistula.
A, Dental reformatted CT scan shows promi-
nent fistula tract (arrow) with unilateral maxil-
lary sinusitis.
B, Hydroxyapatite plug (arrow) was surgical-
ly placed into defect to repair fistula, and dental
reformatted CT scan shows closure of fistula
with normal aeration of sinus cavity (S).
B

Alveolar bone atrophy was seen in the segment adjacent odontal disease with bone resorption may, however, be a
to the fistula in all but one patient. Alveolar bone height predisposing factor for development of oroantral fistula.
around the margins of the fistula ranged from 8 to 20 mm
(average 9.7 mm), and the buccal lingual bone width ranged
Discussion
from 0 to 8 mm (average 3 mm). Because most patients
were referred originallyfor evaluation for implants, moderate Imaging of oroantral fistula has been limited for three main
to severe bone atrophy was not surprising. However, this reasons: (1 ) plain radiographs suffer from superimposition of
finding is not necessarily related to oroantral fistula and structures obscuring small defects in the bony floor of the
instead most likely reflects selection (referral) bias. Pen- sinus; (2) coronal CT scans obtained directly in planes paral-
 1276 ABRAHAMS AND BERGER AJR:165, November 1995

lel to the fistula are often degraded by dental artifacts that vation of dental implant technology and the need for CT eval-
project along the plane of the fistula, thereby interfering with uation of potential implant sites have led to the widespread
Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 06/01/15 from IP address 128.122.253.212. Copyright ARRS. For personal use only; all rights reserved

visualization; and (3) standard 3- to 5-mm-thick axial CT use of dental reformatted CT scans and hence to the involve-
slices often lack resolution sufficient for visualization of small ment of radiologists for assessing this region of the body.
defects. Unlike traditional CT, dental reformatting CT pro- This development has created the need for radiologists to
grams use thin axial CT slices to reformat multiple cross-sec- become familiar with oroantral fistula and a wide spectrum of
tional and panoramic views and have the major advantage of related diseases.
projecting any artifacts from dental structures into a plane In conclusion, oroantral fistula appears as discontinuity of
orthogonal to the plane of the fistula. The images, therefore, the bony floor of the maxillary sinus with associated soft tis-
are free of such artifacts,improving interpretation. sue opacification of the ipsilateral maxillary sinus. Dental
Interestingly, of the five patients referred by dentists, only reformatted CT can be useful for evaluating patients sus-
one was thought by the referring clinician to have oroantral pected of having oroantral fistula, and this condition may be
fistula. The remainder were sent for routine evaluation for found incidentally in patients referred for evaluation for osseoin-
dental implants, and the finding of oroantral fistula was unex- tegrated root-form dental implants. Radiologists should be
pected. We did not formally evaluate an association between familiar with its appearance, as it may be related to implant fail-
oroantral fistula and implant failure. Of note, however, one ure or secondary complications.
patient with failed implants was found to have an oroantral
fistula. A second patient was diagnosed with an oroantral fis-
tula prior to implant placement, and the oroantral fistula was REFERENCES
not repaired; this patients implant failed soon thereafter. 1 . Kurien M, Anandi V, Raman J, et al. Maxillary sinus fusarosis in immuno-
Finally, one patient with an untreated oroantral fistula went competent hosts. J Laryngo/ Otol I 992:106:733-736
on to develop a severe inflammatory reaction in the maxillary 2. Spoor TC, Harding SA. Orbital tuberculosis. Am J Ophtha!mo/ 1981 91:
644-647
sinus, and eventually required sinus surgery that compro-
3. Felix DH, Wray D, Smith GL, et al. Oro-antral fistula: an unusual complica-
mised her nasolacnimal apparatus. The possibility that oroan- tion of HIV-associated periodontal disease. Br Dent J 1991 171:61-62
tral fistula predisposes patients to implant failure and 4. Edwards MB, Buckerfield JP. Wegeners granulomatosis: a case with pri-
secondary complications warrants further investigation. mary mucocutaneous fistula. Oral Pathol 1978:46:53-63
5. Babajews A. Multiple myeloma presenting as oro-antral fistula. Br J Oral
In the most overt presentation, the patient with oroantral fis-
Maxiiofac Sung 1986;24:52-57
tula had a history of recent tooth extraction and complained of 6. Chaudhuri P. Ameloblastoma of the upper jaw. J Lanyngo! Oto! 1975;
a salty taste and the escape of air into the mouth on blowing 89:457-465
the nose. If infection is present, there is a foul-smelling post- 7. McMahon P, Tzadok A. Pagets disease of the bone presenting as appar-
nasal discharge associated with pain in the cheek and general ent oral antralfistula.Arch Oto!aryngo!1986:112:668-670
8. Abe K, Beppu K, Shinohara M, et al. An iatrogenic foreign body in the
symptoms of infection. Only one of our patients presented as
maxillary antrum: a report of two cases. Br Dent J I 992:173:63-65
such. A previous clinical review [9] revealed the occurrence of 9. Norman JE. Oro-antral fistula. Aust Dent J 1977;22:248-258
oroantral fistula in patients from i 5 to 72 years old. Our group 10. von Wowern N. Clinical and radiographic findings in maxillary sinus with
was older, on average, with only a single patient less than 45 oro-antral fistula. Int J Ona/Sung 1981 :1O[suppl 1]:1 38-1 42
1 1 . Norman JE, Craig G. Oro-antral fistula: an analysis of 100 cases. Oral
years old. This age bias most likelyreflectsa referralbias on
Sung Oral Med Oral Pathol 1971:31:734-744
the basis of dental implant planning, as implants are planned 12. Abrahams JA. The role of diagnostic imaging in dental implantology.
most commonly in older patients. It is possible that younger Radio! C/in North Am 1993:31 :163-1 80
patients with fistula go untreated or that the diagnosis is com- 13. Abrahams JA. Anatomy of the jaw revisited with a dental CT software pro-
monly made on the basis of history, clinical examination, and gram. AJNR 1993:14:979-990
14. Abrahams JA. CT assessment of dental implant planning. Oral Maxillofac
findings on plain radiographs.
C/in NonthAm 1992:4:1-18
Traditionally, most plain radiography of dental structures 1 5. DelBalso AM, Greiner FG, Licata M. Role of diagnostic imaging in evalua-
has been done and interpreted in the dental office. The inno- tion of the dental implant patient. RadioGraphics 1 994; 14:699-719