Int. J. Oral MaxilloJktc. Surg.

1994; 2 3 : 2 2 ~ 7 C o p y r i g h t © Munksgaard 1994

Printed in Denmark. All rights reserved
hJtemational]oumal of
Oral8c
MaxillofacialSurgery
ISSN 0901-5027

0ncology
E. Heissler ~, H. J. Steinkamp 2,
Value of magnetic resonance 1". Heim 2, C. Zwicker 2, R. Felix 2,
J, Bier ~
Clinics for 1Maxillofacial Surgery and
imaging in staging carcinomas of 2Radiology, University Clinic Rudolf Virchow,
Free University of Berlin, Augustenburger
Platz 1, D-13344 Berlin, Germany

the oral cavity and oropharynx

E. Heissler, H. J. Steinkamp, T. Heim, C. Zwicker, R. Felix, J. Bier. Value of
magnetic resonance imaging in staging carcinomas of the oral cavity and oropharynx.
Int. J. Oral Maxillofac. Surg. 1994; 23: 2~27. © Munksgaard, 1994

Abstract. The value of magnetic resonance imaging (MRI) in preoperative tumor

(T) staging was assessed prospectively in 46 patients with malignant carcinomas
of the oropharynx and oral cavity. In each case, the MRI findings were compared
with the preoperative clinical and postoperative histopathologic findings. M R I
showed an accuracy of 89% in identifying the T stage, as compared with the
histopathologic result. The accuracy of the clinical examination was 78%.
Compared with the clinical examination, M R I is particularly suited for differen-
tiating T3/T4 and the larger T2 tumors. M R I tends to overinterpret the T stage
when the neighboring tissues are inflamed or edematous, as it allows no demar-
cation between inflamed or edematous tissue and tumor. Moreover, normal oropha-
ryngeal mucous membrane and intrinsic tongue muscle both absorb contrast
medium. In the clinical examination, the T stage is often underestimated, as
the depth of infiltration and bone involvement cannot always be determined with
certainty. The ability of MRI to give an axial, coronal, and sagittal image allows
an exact preoperative view of tumor spread and assessment of infiltration into
adjacent structures.
Zusammenfassung. Bei 46 Patienten mit maligen Tumoren der Mundh6hle und
des Oropharynx wurde der Wert der Kernspintomographie (MR[) fiir das praeope-
rative Tumorstaging prospektiv untersucht. Fiir jeden Patienten wurden die Er-
gebnisse der M R I mit der klinischen Untersuchung und dem histologischen Befund
verglichen. Die M R I wies mit einer Treffsicherheit von 89% das pathologisch
gesicherte Ergebnis nach. Die klinische Untersuchung erreichte eine Treffsicherheit
yon 78%. Verglichen zur klinischen Untersuchung eignet sich die MRI besonders
zur Differenzierung der T3/T4 sowie der gr613eren T2 Tumoren. Die M R I neigt
zu Uberinterpretation des Tumorstadiums, da sic nicht zwischen Tumor Beglei-
tentzfindung und ()dem unterscheidet. Weiterhin nehmen die normale Oropha-
rynxmukosa und die intrinsische Zungenmuskulatur Kontrastmittel auf. Die
klinische Untersuchung auf der anderen Seite untersch/itzt hfiufig das Tumorsta-
dium, da die Tiefenausdehnung und die Knocheninfiltration oft nicht sicher
Key words: magnetic resonance imaging
bestimmt werden k6nnen. Die axiale, coronare und sagittale Schichtung in der
(MRI); tumor staging; oropharyngeal tumor.
MRI erm6glicht eine genaue Darstellung des Tumors und eine Beurteilung seiner
Infiltration in Nachbarstrukturen. Accepted for publication 16 September 1993

For determination of the most appropri- propriate evaluation and assessment of infiltration of vascular structures,
ate therapy for a squamous cell carci- the stage of the tumor is mandatory. It is nerves, and bone and to assess the ex-
noma (SCC) of the oral cavity, an ap- particularly important to assess possible tent of the tumor. It is also necessary to

know whether a tumor in the oral cavity
crosses the midline, as this can signifi-
cantly influence the choice of treatment
and the prognosis s. Clinical examin-
ation alone is often inadequate. Accord-
ing to LENZ et al. s, a correct appraisal
of the t u m o r (T) stage attained by clin-
ical and endoscopic examination alone
was m a d e in only 82/174 cases. This is
an accuracy of only 47%.
With a claimed accuracy of
80°/o-90% 5'9'11, computerized tomo-
graphy (CT) has gained recognition as
a noninvasive T-staging procedure for
diagnosing orofacial tumors. However,
errors in j u d g m e n t with C T frequently
result when metal artifacts are present;
for example, those caused by tooth fill-
ings, or when no distinction in radio-
density can be made between tumor and
surrounding tissue 4'8'12.
Magnetic resonance imaging ( M R I )
is a noninvasive tomographic method
which, on the one hand, provides axial,
sagittal, and coronal images, and, on
the other hand, is not affected by arti-
facts caused by tooth fillings 1'6'11 13 The
purpose o f this study is to determine
whether M R I meets these requirements.
Patients and methods
In 46 patients with proven SCC of the oral
cavity or oropharynx, the T-stage findings
obtained with MRI were compared prospect-
ively with those obtained through clinical and
endoscopic examination. To ensure objec-
tivity, we compared these findings with the
postoperative pathohistologic findings,
which classified the size and extent of the
tumor, according to the tumor-node-meta-
stasis (TNM) system, as well as the number
and size (maximum diameter to transverse
diameter) of lymph nodes. The histology of
the primary SCC and lymph nodes was also
described. Forty-two patients were surgically
treated, including local resection along with
neck dissection. The postmortem results for
four patients with extensive T4 tumors were
incorporated into the study. In the 46 pa-
tients, SCC of varying degrees could be de-
Table 1. Primary tumor location in 46 oral cavity and oropharynx carcinomas examined by
MRI
Tumor Primary location
Oral cavity Body of tongue
(n = 24) Floor of mouth
Alveolar process
Hard palate
Oropharynx Ton sils
(n = 22) Base of tongue
Soft palate
Posterior wall of oropharynx
tected. Tumor distribution and the T classifi-
cation can be seen in Tables 1 and 2.
MRI examinations
lmages were obtained from 46 patients at a
full-body MRI unit (1.5 Tesla Magnetom,
Siemens). A spin-echo image was made for
each patient. Tl-weighted images were done
in the 1.5 Tesla Magnetom with a repetition
time (TR) of 500 ms and an echo time (TE)
of 15 ms. Proton density-weighted (TR: 2500
and TE: 22) and T2-weighted (TR: 2500 and
TE: 90) images were also made. The T1-
weighted scans (TR: 500 and TE: 15) were
repeated after the application of contrast me-
dium with 0.1 mmol/1 gadolinium (Gd)-
dTPA/kg body weight (Schering AG, Berlin,
Germany). The imaging was performed in
the axial, sagittal, and coronal planes with a
layer thickness of 6 mm. A 256 x 256 matrix
was used, with 1.2 x 1.2 mm picture points.
Muscle, particularly the sternocleidomas-
toid muscle, and fat were used as a basis of
comparison for interpreting the MRI images
in regard to the signal density of the tumor.
Only spin-echo sequences were used. The
scans were assessed according to the follow-
ing scale:
I) hypointense relative to muscle
2) isointense relative to muscle
3) slightly hyperintense (+) relative to
muscle
4) hyperintense (+ +) relative to muscle
5) very hyperintense ( + + + ) relative to
muscle.
For evaluation, the MR images were as-
sessed by three clinicians. These results were
compared with the results of the clinical
examinations and the final histopathologic
reports. The tumors were classified by the
TNM system acording to the U I C C 7.
Results
Topography
In the region o f the oral Cavity and the
oropharynx, muscle appeared with me-
dium signal intensity, with individual
muscle tracts separately distinguishable
because of the fat and fascial tissue sur-
rounding each tract. In the tongue, a
diverse signal intensity increase o c -
Number of patients
11
1l
1
1 T1
10
9
2
1 Totals
M R I in oral c a r c i n o m a s 23
curred upon administration of Gd-
dTPA. The intrinsic tongue muscles (in-
ferior and superior longitudinal
muscles, etc.) absorbed the contrast me-
dium well ( + + ) ; the extrinsic muscles
of the tongue and floor of the m o u t h
(genioglossus muscle, hypoglossus
muscle, digastric muscle, stylohyoid
muscle) showed only low enhancement
( + ) (Fig. 4A). Healthy tonsil tissue,
with a high blood flow, showed good
contrast medium absorption ( + + ) .
Therefore, small tumors could not be
distinguished from healthy tissue after
G d - d T P A administration.
The blood vessels of the oral cavity
appeared only with weak signals. In the
body of the tongue, the lingual artery
was usually distinguishable.
The bony structures of the mandible
and maxilla presented with varying sig-
nal intensity. Structures containing mar-
row appeared as signal-intensive zones;
the cortex showed only faint signals
(Fig. 1).
The SCCs of the oral cavity com-
prised carcinomas of the tongue ( n =
11), floor of the m o u t h ( n = 11), hard
palate (n = 1) and alveolar process (n =
1). Tumors of the oropharynx orig-
inated in the base of the tongue ( n =
9), tonsils (n = 10), posterior pharyngeal
wall ( n = 1) and anterior surface of the
soft palate including the uvula ( n = 1).
Table 3 shows the frequency of the cor-
rect T-stage identification by clinical
examination (including endoscopy), as
compared with M R I and the histopath-
ologic results. Altogether, the correct T
stage was identified by M R I in 40/46
Table 2. T categories: oral cavity and oropha-
rynx
TI: tumor measures at its maximum diam-
eter 2 cm or less
T2: tumor measures at its maximum diam-
eter more than 2 cm but not more than
4 cm
T3: tumor measures at its maximum diam-
eter more than 4 cm
T4: tumor infiltrating neighboring structures
Table 3. Comparison of correct identification
of T stage achieved by clinical examination
(CE) and MRI with histopathologic results
T stage CE MRI Histology
3 2 4
T2 10 9 12
T3 15 19 20
T4 8 10 10
36 (78%) 40 (87%) 46
24 Heissler et al.
Table 4. T stages wrongly identified by MRI
Patient T stage CE T stage MRI T stage pathology Deviation of MRI
L.S. T1 TO
L.G. T1 T2
H.M. T2 T3
L.J. T2 T3
RG. T3 T3
R.S. T3 T4
CE = clinical examination.
+ = additional tumor infiltration found by MRI.
= structures not picked up by MRI but shown by pathologic findings.
cases (87%), but by clinical examination
and endoscopy in only 36/46 cases
(78%). By analysis according to the indi-
vidual T1 T4 stages, it can be seen that
clinical examination gives poorer results
in the T3 and T4 stages (23/30) than
MRI (29/30). On the other hand, clin-
ical examination (13/16) showed better
results in the stages T1 and T2 than
MRI (11/16).
Table 4 shows the T stages wrongly
classified by MRI. MRI correctly
graded only two T1 tumors. One T1
tumor was diagnosed as a T2 tumor
because the tumor size was described as
gr~eater than 2 cm. However, the patho-
logic and endoscopic findings revealed
a tumor with a maximum size less than
2 cm.
A superficial T1 tonsil tumor was not
recognized on MRI because the tumor
could not be distinguished from the
healthy lymphatic tonsil tissue sur-
rounding it, either after Gd-dTPA ad-
ministration or in the T2-weighted se-
quence.
With all three incorrectly graded T2
tumors, the T stage was graded too high
(Table 4). Particularly in the T2-
weighted sequence, tumor volume in
one case was estimated too high because
of the nondistinction between tumor
and accompanying inflammation. After
the administration of Gd-dTPA, how-
ever, the edema was seen as a hypointen-
sive structure which could then be dis-
tinguished from the hyperintensive tu-
mor area. Nevertheless, after Gd-dTPA
administration, the tumor was insuf-
ficiently distinguishable from the neigh-
boring tissue in 8/46 cases; thus, only
by the T2-weighted examination could
the tumor be unambiguously dis-
tinguished from adjacent structures.
In the case of the wrongly classified
T3 tongue tumor, further infiltration of
Tl - Superficial tumor (tonsil)
T1 + Tumor spread > 2 cm
T2 + Tumor spread > 4 cm
T2 + Depth of infiltration >4 cm
T2 + Carcinoma in base of tongue
with probable infiltration of
posterior wall of pharynx
T3 + Tongue carcinoma with
massive depth of
infiltration
the oropharyngeal tongue base ap-
peared to be present on the same side
on MRI. This was not confirmed, either
intraoperatively or in the pathologic
findings of the tissue removed during
operation.
All T4 tumors were correctly iden-
tified by MRI. Whereas clinical examin-
ation was more effective in identifying
the stages T1 and T2, 7/30 T3/T4 tu-
mors were incorrectly graded clinically.
In all cases, the stage of the tumor was
underestimated.
MRI can present the depth of infil-
tration of the tumor clearly because of
its good soft-tissue contrast, particu-
larly after Gd-dTPA administration.
The MRI multiplane imaging presents
a relatively accurate picture of tumor
extension (Figs. I ~ D ) .
Value of examination sequences
Native Tl-weighted examination pro-
duced good tumor-fat contrast (Fig.
1A), but it did not allow sufficient dis-
tinction of the tumor from surrounding
muscle and blood vessels. Relative to
muscle, SCCs appeared to vary from
slightly hypointense to isointense.
In only 36/46 cases did Gd-dTPA-
aided, Tl-weighted sequences provide
adequate tumor identification and clear
demarcation from muscle, blood ves-
sels, and bone structures. In 34/46
cases, the tumors showed heterogeneous
contrast medium uptake. These were all
hyperintense ( + - + + ) in relation to the
muscle (Fig. 1B). Within the richly vas-
cularized, contrast-medium-absorbent
tumor tissue, tumor necrosis was seen
as hypointensive structures (Fig. 1B). In
8/46 cases, the tumor showed no signifi-
cantly increased concentration over the
surrounding tissue. The absorption of
contrast medium, particularly in the re-
gions of the intrinsic tongue muscles
and the tonsils, led, in the case of seven
patients, to an overestimation of tumor
size in the axial, sagittal, and coronal
planes. Thus, there was more than a l-
cm difference in size from the tumor size
found in the histopathologic specimen.
In six cases, the T2-weighted se-
quences did not show the high signal
(Fig. 2) intensity ( + + ) expected in the
tumor (Fig. 4B). However, the tumor
still appeared in these cases as clearly
hyperintensive ( + ) relative to the
musculature. Particularly with the tonsil
tumors (n = 7), the tumor in four cases
could not be clearly identified because
of the increased signal intensity of the
healthy tonsil tissue. In this situation,
the tumor had to be diagnosed from
the indirect evidence mentioned earlier.
Clear demarcation of the tumor area
was visible in the T2-weighted sequence
in 38/46 cases - more so than in the
Gd-dTPA-supported, Tl-weighted se-
quence (Figs. 2, 4B). In the T2-weighted
sequence, considerable overestimation
of tumor size occurred because of in-
flammation and adjacent swelling i n
neighboring tissuel
The proton-density images show a
slight hyperintensity ( + ) of the tumor
in relation to muscle (Fig. 1C). Al-
though ~he proton-density images
yielded no additional information, they
did give sati~fact.ory contrast with the
tumor, fat, muscle, and other structures.
Anatomic relationships were easily dis-
cernible. The good presentation of bone
structures in this sequence is noteworthy
(Fig. 1C).
Image planes
MRI allows the presentation of the oral
cavity and oropharynx in the axial, cor-
onal, and sagittal planes. For carci-
nomas of the tongue and the floor of
the mouth, images in both the axial and
the coronal planes are indispensable.
These make it possible to detect or dis-
count tumor extension across the mid-
line, in order to select the appropriate
therapy (Figs. 1D, 3, 4). The coronal
scan is particularly suited for judging
the craniocaudal extent of tumors of the
tonsils.
In our series, sagittal and axial scans
have proved to be very useful for the
detection of oropharyngeal carcinomas,
as they reveal any infiltration of the epi-
glottis, the preepiglottal soft tissue, and
the hypopharynx. Axial imaging is es-
sential, as it allows assessment of poss-
 M R I in oral carcinomas 25

possibility of using MRI to image tumor

of the oropharynx and the oral cavity
continues to be the subject of clinical
studies6,8,t°,~6.The anterior two-thirds of
the tongue is accessible to clinical exam-
ination by inspection and palpation.
Here, CT is less effective than clinical
examination because it does not give
adequate contrast between tongue
muscle and small tumors 8'9. More accu-
racy can be obtained with MRI ~°. With
the use of T2-weighted sequences, tu-
mors in the floor of the mouth and in
the tongue can be identified in spite of
poor enhancement (Fig. 2).
Another advantage of MRI is that
tomograms are not influenced by tooth
fillings or dense bone4'6'11'13'14.However,
for small tumors, MRI is less reliable
than clinical examination, as main-
tained by DmLON & HARNSBERGER,who
recommend MRI for tumors that are
stage T2 or larger 2.
The axial image should be retained
as the standard for orientation examin-
ations in the oropharynx and oral cav-
ity5. Coronal images allow an exact as-
sessment of tumor size in relationship
to the midline in the floor of the mouth
and tongue (Figs. 3 and 4).
The T2-weighted sequence is, in the
first instance, the sequence most sensi-
tive to the presence of tumor. In particu-
lar, contrast against neighboring muscle
is clearest in this sequence because of
its longer T2 repetition time. However,
other researchers such as LENZ et al. s
and VOGLet al. ~5have found, as we did,
that the T2-weighted sequence overesti-
mates tumor volume, since peritumor-
ous inflammation and edema cannot be
distinguished from actual tumor. Tu-
mor, peritumorous edema, and inflamed
Fig. 1A. Forty-three-year-old patient with increasing difficulty in swallowing for 6 months. tissue show the same T2 repetition
Tonsil carcinoma T3 N2b. Axial native Tl-weighted examination: tumor mass (T) in right times3,S,9,16.
tonsil; tumor slightly hypointense relative to muscle (M); tumor varies from isointense to
hypointense relative to blood vessels (thin arrows). Good contrast with fat tissue (F). The administration of Gd-dTPA en-
Fig. lB. Axial Gd-dTPA-supported Tl-weighted examination. Tumor (T) presents as clearly ables demarcation from edematous
hyperintensive relative to muscle (M) and as hypointensive relative to fat tissue (F). Within tissue. However, peritumorous, inter-
tumor, small structures with reduced contrast medium absorption (necroses) (arrows) can be stitial overflow of contrast medium into
distinguished. Visible variations in contrast medium enhancement in area of floor of mouth. edematous tissue has been reported on
Fig. 1C. Axial proton-density (2500/22) examination. Tumor (T) slightly hyperintensive (+) images taken late after the administra-
relative to muscle (arrows). Clear presentation of all anatomic structures. tion of contrast medium. The use of
Fig. 1D. Gd-dTPA-supported Tl-weighted coronal examination. Presentation of craniocaudal native Tl-weighted sequences is indis-
extent of tumor (thin arrows). No infiltration of opposite side (broad arrowhead).
pensable if the relationship of the tumor
to tissue compartments with fat content
is to be ascertained. T2-weighted se-
quences indicate only very little tissue
ible bone infiltration. Infiltration of the contrast. After Gd-dTPA has been ad-
Discussion
bone must be assumed when strong sig- ministered, demarcation is also more
nals are emitted in the cortex. Proof An exact description of tumor size is difficult. Gd-dTPA contrast-medium
can be established with the Gd-dTPS necessary for treatment planning. At absorption corresponds to the intensity
supported, Tl-weighted sequence (Fig. present, CT is recognized as an integral of blood flow, and oropharyngeal tu-
5). part of pretherapeutic staging4'5's. The mors show, with the exception of tonsil
26 Heissler et al.

Fig. 2. Forty-two-year-old patient with cervical swelling noticed 3 weeks previously. MRI: T2-weighted native examination. Clear hyperintensity
is seen in tumorous mass in right side of floor of oral cavity (arrows). However, there is no infiltration or restriction of oropharynx. T2 stage.
T stage identified by pathologic findings after operation: T2.
Fig. 3. Demonstration of T2 tumor of right tongue in enhanced Tl-weighted MRI: after administration of Gd-dTPA, in coronal scan there
is clear demarcation of tumor (broad arrowheads) as hyperintensive structure in posterior third of tongue without exceeding midline (thin
arrow).

tumors, a stronger increase o f signals necrotic tumor, because of their nonper- which do not absorb contrast medium,
than the surrounding area 16. fusion, appear as hypointense (Figs. 1B and tumor recurrence, which shows
However, tumor absorption of con- and 5). Contrast medium also allows marked enhancement 8.
trast medium is not uniform. Areas o f distinction between scarred structures, Tumors of the tonsils often show no

Fig. 5. MRI of tumor on left side of floor of

Fig. 4A. Forty-seven-year-old patient with large submandibular and submental swelling. Cor- mouth with infiltration of left mandible. Gd-
onal Gd-dTPA supported Tl-weighted image. Presentation of large T4 tumor of floor of dTPA Tl-weighted scan. Demonstration of
mouth/tongue (thin arrows) which has clearly crossed midline. M. mylohyoideus (broad extended, partly necrotic (thin arrow) tumor
arrowheads), which has absorbed no contrast medium, not yet infiltrated. Clearly visible which has infiltrated compact substance,
contrast medium absorption by intrinsic tongue muscle. spongiosa, and opposite compact substance
Fig. 4B. In T2-weighted sequence, distinct hyperintensity (+) of tumor (arrows) against (broad arrowheads). Lymph node metastases
nonfiltrated tongue and mouth floor muscle. (M) in left angle of mandible.

increased contrast-medium enhance-
ment against normal tonsil tissue, be-
cause tonsil and gland tissue also show
high enhancement 8'9,~5. In this region,
both M R I and CT depend on indirect
signs of t u m o r presence, such as dis-
placement of normal anatomic contours
and structures.
Infiltration of bone can be deter-
mined with the help of Gd-dTPA-sup-
ported, Tl-weighted and native T2-
weighted sequences (Fig. 5). The conti-
nuity of faint signals in compact sub-
stance can be demonstrated unambigu-
ously. P r o o f of an accompanying high-
signal fat lining excludes the possibility
of infiltration of the bone tissue 4'9. Ac-
cording to GLAZER et al., microinvasion
cannot be ruled out with any certainty
in the case of inadequate fat planes and
apparently intact cortices 4.
Along with the axial image, the cor-
onal image is important for assessing
craniocaudal tumor extension, particu-
larly in the oropharynx. In the oral cav-
ity, the coronal scan is important for
judging mediolatera[ tumor size and ex-
tension over the midline (Fig. 4). Here,
M R I is superior to CT. Sagittal imaging
scans are important in estimating the
encroachment of the t u m o r onto the
supraglottis and the hypopharynx,
As a result of the study, the following
conclusions can be stated:
1. M R I is superior to endoscopy in
identifying higher tumor stages (T3/
T4) because it allows the assessment
of bone involvement, soft-tissue infil-
tration, and depth.
2. M R I tends to overestimate the T
stage because of the inadequate dis-
tinctions between tumor and edema,
and inflammation and normal mu-
cous membrane. Clinical examin-
ation, on the other hand, often
underestimates the T stage of T 3 / T 4
tumors.
3. The T2-weighted sequence is the se-
quence m o s t sensitive for tumor pres-
entation. As opposed to the T2-
weighted sequence, the G d - d T P A -
supported, Tl-weighted sequence
allows distinction between edema
and tumor.
4. Sagittal imaging can provide import-
ant supplementary information on
the spread of oropharyngeal tumors.
For oral cavity tumors, coronal ima-
ging is particularly suited to provide
this information.
References
1. BRAUN IF. MRI of the nasopharynx.
Radiol Clin North Am 1989: 27: 315-.30.
2. DILLONWP, HARNSEERGERHR. The im-
pact of radiologic imaging on staging of
cancer of the head and neck. Semin Oncol
1991: 18: 64-79.
3. F/,)RST G, ZAMBOGLOUN, GREVEN C,
KAHN T, MODDER K. Kontrastmittelein-
satz in der Kernspintomographie von
Kopf- und Halstumoren, Fortschr Ront-
genstr 1986: 149: 489-95.
4. GLAZER H, NIEMEYER Jl-i, BALFES. Neck
neoplasms: MR1 imaging. I. Initial evalu-
ation. Radiology 1986: 160: 343-8.
5. HARNSBERGERHR, DILLONWP. Imaging
tumors of the central nervous system and
extracranial head and neck. CA Cancer
J Clin 1987: 37: 225-38.
6. HELD P, RUPP N. Tumoren des Oropha-
rynx, Hypopharynx and Larynx-Diag-
nose durch MR. Bildgebung 1991: 58:
132-40.
7. HERMANEKP, SCHEIBEO, SPIESSLB, WAG-
NER G, eds. UICC TNM Klassifikation
maligner Tumoren. 4th ed. Berlin: Spring-
er, 1987.
8. LENZ M, SKALEJM, OZDOBA C, BONGERS
H. Kernspintomographie der Mundh6hle
und des Mundbodens: Vergleich mit
M R I in oral carcinomas 27
der Computertomographie. Fortschr
Rontgenstr 1989: 150: 425-33.
9. LUFKIN R, WORTHAMD, D1ETRICHR, et
al. Tongue and oropharynx: findings
on MR imaging. Radiology 1986: 161:
69 75.
10. LYDIATT DD, MARKIN RS, WILLIAMS
SM, DAVIS LF, YONKEkSAJ. Computed
tomography and magnetic resonance im-
aging of cervical metastasis. Otolaryngol
Head Neck Surg 1989: 101: 422-5.
11. MCKENNAKM, JABOURBA, LUFKINRB,
HANAEESWN. Magnetic resonance imag-
ing of the tongue and oropharynx. Top
Magn Reson Imaging 1990: 2: 49-59.
12. M(SDDER U, STEINnmCHW, HEINDELW,
LINDEMANN J, BRUSIS'1~ Indikation zur
Kernspintomographie bei Tumoren des
Gesichtssch/idels und des Halsbereichs.
Digital Bilddiagnostik 1985: 5: 55-60.
13. N1TSC~IEH, IRO H. Prf, operative Diagno-
stik von Halstumoren mit magnetic res-
onance imaging. HNO 1989: 37:373 8.
14. VOGLT, MEESK, GREVERSG. Die diagno-
stische Wertigkeit der Kernspintomo-
graphie bei Raumforderungen des Phar-
ynx. Laryngol Rhinol Otol 1987: 66:
543-6.
15. VOGL T, BR/3NING R, G~EVEkS G, NEES
K, BAUERM, LlSSNERJ. MR imaging of
the oropharynx and tongue: comparison
of plain and Gd-dTPA studies. J Comput
Assist Tomogr 1988: 12: 427-33.
16. WAGNER-MANSLAUC, BAUERR, VAN D~
FLIERDTE, et al. Wertigkeit yon SE-, IE-
und Ge-Sequenzen bei Tumoren im
Kopf-Hals-Bereich. Fortschr Rontgenstr
1990: 153: 154-60.
Address:
Dr Dr Ernst Heissler
Clinic for Maxillofacial Surgery
University Clinic Rudolf Virchow
Free University of Berlin
Augustenburger Platz 1
D-13344 Berlin
Germany