Head Charles

F. Lanzieri,

MD

#{149} Manjul

Shah,

MD

#{149} Dennis

Krauss,

MD

Use of Gadolinium-enhanced for Differentiating Mucoceles In the Paranasal Sinuses’ The purpose of this study was to determine whether additional clinically useful information could be obtamed from gadolinium-enhanced magnetic resonance (MR) imaging compared with the information obtamed from nonenhanced MR imaging and computed tomography (CT). Therefore, the authors selected 41 patients, whose results at CT examinations demonstrated a variety of pathologic conditions of the paranasal sinuses, to undergo MR imaging both with and without the use of gadopentetate dimeglumine for contrast enhancement. In 22 of 35 cases of neoplasms and mucoceles occurring separately, the lesions were correctly differentiated by established MR signal criteria alone. However, 32 of these 35 cases were accurately differentiated when gadolinium-enhanced MR images were obtained. Six cases demonstrated coexisting neoplasm and mucocele: Gadolinium-enhanced MR enabled correct differentiation of five of these lesions, while unenhanced MR enabled correct differentiation of three. On the basis of these results, the authors conclude that the use of gadopentetate dimeglumine for contrast enhancement at MR imaging is useful for differentiating mucoceles from neoplasms in the sinonasal tract. Index nance

terms: (MR),

23.255

Gadolinium #{149} Magnetic contrast enhancement

#{149} Paranasal

23.1214

sinuses,

#{149} Paranasal

MR

sinuses,

I From Cleveland,

(MS.)

the

Department 2074 Abington

and

Otolaryngology (D.K.),

From c

the

1990

received RSNA,

RSNA September

1991

map-

the

potential

celes

from

signal

ages

and

the images

neoplasms

intensity

ty on

include

to differentiate on

MR signal intensity In most instances, ease is demonstrated by increased

increased

the

content.

tetate

dirneglummne

hance

MR

sal

of

intensi-

Because

find-

extent

from

that

can

changes

be relied

nor

resoMucocele,

and

23.35,

in the

signal

depending

on

their

The

served from

use

(P.L.), the

scientific

New

Cleveland York

Western OH 44106

Clinic

Hospital,

assembly. Received 7; accepted September

Reserve (C.F.L.);

of tucellularity

University, University Hospitals the Departments of Radiology

Foundation,

Cornell

June 24.

Cleveland;

University

26, 1990;

Address

and

Medical

revision

reprint

the

Center,

requested

requests

Department New

August

to C.F.L.

used

to en-

the

sinona-

hope

that

the

MR studies in differentiating inflammatory

gad-

would pro-

diseases

than

unenhanced

stud-

of gadopentetate

no significant MR enhancement

dime-

of

of

York.

6; re-

contribution with gado-

dimeglumine.

MATERIALS

AND

a 14-month

referred

period,

because

sinus

162 patients

All

ning these

as their patients

the

of

as mucocele

patients

under-

CT scan-

imaging step. In 41 of results of CT scanning

possibility

or both;

referred

suspicion

such

of these

non-contrast-enhanced initial the

of the parana-

of clinical

mass,

went

neoplasm,

METHODS

for imaging

of mucocele,

therefore,

the

for MR imaging.

patients

Patients

were studied in the sagittal, axial, and coronal planes, with Ti- and T2-weighted imaging sequences. Pulse sequences were as follows: repetition time = 500 msec, echo time = 32 msec (TR/TE 500/32); 3,000/32;

and

underwent

3,000/60.

Each

patient

also

MR imag-

contrast-enhanced

ing (500/32), with a dose of gadopentetate dimeglumine equal to approximately 0.1 mmol/kg. MR imaging was performed as soon after ways within

of Radiology, Case Road, Cleveland,

was our

of the

pentetate

were

consis-

characteristics

either

indicated

in

tently identifiable characteristic signals (3-5). For example, wide vaniation has been seen in the signal intensities of mucoceles, depending on their state of hydration and their age,

#{149}

to

within

neoplastic

or neoplasm.

tumors

on to produce

were

sal sinuses a paranasal

of

associated inflammatory changes. However, theme has been evidence the literature to suggest that neither inflammatory

from

During

of differ-

separately

It was

cesses

were

ences in the signal intensities of each lesion, it is possible to recognize the presence of neoplasms and to map their

study

glumine to enhance MR images of the sinonasal tract has been limited. In three cases, Som et al (13) ob-

ages. In many instances, inflammatory changes are superimposed on neoprocesses.

tract.

ies.

ings are in contradistinction to the signal intensity of neoplastic tissue, which may be slightly lower than the signal intensity of inflammatory tissue on both Ti- and T2-weighted im-

plastic

of this

images

olinium-enhanced be more helpful paranasal sinus

reflecting

These

purpose

(1-

basis

signal

characteristics.

the

abiland

im-

images,

water

MD

compare the results obtained at CT and unenhanced MR imaging with the results obtained when gadopen-

muco-

Ti-weighted

T2-weighted

Radiology

histologic It was

characteristics. inflammatory disby intermediate on

Lavertu,

and

and maghas

178:425-428

Otolaryngology vision

that

tumors with imaging

3). These advantages ity to obtain multiplanar

mors, 1991;

indicate

several advantages oven mapping with computed tomography (CT)

23.37

Radiology

#{149} Pierre

Neck

MR Imaging from Neoplasms

ping of sinonasal inflammatory processes netic resonance (MR)

studies,

neoplasms,

reports

ECENT

and

the injection 15 minutes.

as possible, The presence

alof

a paranasal sinus mass was confirmed at surgery and at pathologic examination in all 41 patients. The CT, unenhanced-MR, and gadolinium-enhanced

MR

dependently

reviewed

studies

were

then

by three

at different times. The observers blinded to the results of the other but

were

told

that

in

each

case

in-

observers were studies

there

was

425

Table

1 of 41 Paranasal

Summary Lesions

Sinus

Diagnosis

No.

Mucoceles

15

Neoplasms Squamous Squamous mucocele

cell carcinoma

only

cell

plus

carcinoma

13 5

Adenocarcinoma

only

Adenocarcinoma

plus

6 mucocele

1

Schwannoma

_2_.

Total

41

a. Figure

b.

reveals praorbital

mixed

The reviewers use

the

bone

presence

as the

between

CT

scans. difference

of the

masses

the

hanced MR image (500/32) shows ages of the nonenhancing central ment within the left frontal sinus evidence

Table

left

those

frontal

for

sinus

and

mucocele

orbit

and/or

that the mass portion (black (white arrows),

has

(arrow).

tumor.

low

signal

The

(c) The

sinuses the left sua mass of

mixed

signal

in-

gadolinium-en-

intensity

on

arrow). There is peripheral, suggesting a mucocele.

Ti-weighted

im-

linear enhanceThere is no

of a tumor.

2 Diagnoses

Correct

of Paranasal

Sinus

Lesions Diagnosis

Coexisting Mucocele Mucocele (n 15)

Modality

differ-

Neoplasm (n 20)

and

Neoplasm (n 6)

mucocele

attempted in

the

with

and

for

and

within

compatible

axial CT scan through the frontal left frontal sinus and extending into axial MR image (2,000/60) reveals

to

expansion criteria

tumor

They

trinsic

in advance

of sinus

destruction

entiating on

decided

intensity

are

positive

c.

Mucocele. (a) Non-contrast-enhanced expansive mass originating in the region (arrow). (b) The T2-weighted

signal

tensities

pathologic proof of either a mucocele, a neoplasm, or a combination of mucocele and neoplasm. The observers were asked to place each study within one of these categories. When at least two of the three observers correctly diagnosed the lesion, the study was considered true-positive. When only one of the observers made the correct diagnosis, the study was considered true-negative. There were no significant differences in the number of correct diagnoses made by each observer.

1. an

to use signal

an

CT

in-

Unenhanced MR Gadolinium-enhanced

intensities

13 12 14

imaging MR imaging

13 10 18

2 3 5

on non-contrast-enhanced

MR images to differentiate between mucocele and tumor. The observers further agreed patterns

to use characteristic on contrast-enhanced

to differentiate ring-enhancing

mucocele pattern

enhancement MR images from within

tumor: A the para-

nasal sinus was considered a positive sign of mucocele formation, whereas homogeneous or solid enhancement within the

affected the

sinus

was considered

presence

pathologically Table 1 . There

carcinomas, isting noma.

proved cases appears were 1 5 mucoceles,

six carcinomas

mucoceles, The correct

modality

and

with

and one diagnoses

relative

to each

19

tate

schwanfor each of

in di-

agnosis are shown in Table 2. The sensitivity and specificity for each modality relative to each diagnosis are shown in

Table

3. On

MR images

non-contrast-enhanced

the tumors

of intermediate both

Ti-

and

whereas

most

increased

signal

T2-weighted

images.

en, the signal intensities mucoceles varied widely.

there

were

ment

patterns

hanced

also

MR images,

that

a thin,

was

suggestive

426

#{149}

regular

Radiology

often intensiHowev-

of tumors Although

variations on

often on

images,

the mucoceles the

most

intensity

T2-weighted

demonstrated ty on

were

signal

and

in enhance-

gadolinium-en-

it became rim

apparent

of enhancement

of a mucocele,

whereas

sinuses MR

a knowledge following

coex-

category

obstructed

trast-enhanced in

of

and

and

of the 41 surgically

and Specificity by Modality

Modality

DISCUSSION The results of our study there is significant overlap nal intensities of tumors,

RESULTS

3

Sensitivity Diagnosis Diagnosis

to indicate

of a tumor.

A summary

Table solid or homogeneous enhancement was suggestive of a neoplastic process (Figs 1-3).

indicate in the mucoceles,

on non-con-

images.

Therefore,

of enhancement the

injection

dimeglummne

may

differentiating

tumors

that sig-

of

aid

patterns gadopente-

the from

imager muco-

celes. There was a significant increase (P < .001) in the number of connect diagnoses if gadolinium-enhanced MR images

the

were

addition

available.

of nonenhanced

Interestingly,

Ti-

weighted MR images did not improve the average number of correct diagnoses made with nonenhanced T2weighted MR images. Van Tassel et al (4) concluded that the appearance of mucoceles on MR images is quite variable, having found that any signal characteristic, including signal void, may represent a mucocele on T2-weighted MR images. This vanability limits the utility of MR imaging in the evaluation of sinus disease, since a sinus that appears to be normal may, in fact, be grossly abnormal, containing

Mucocele (n CT

Sensitivity 15)

Unenhanced MR Gadoliniumenhanced MR Tumor (ii = 20)

CT Unenhanced MR Gadoliniumenhanced MR Coexisting tumor and mucocele (0

Unenhanced Gadoliniumenhanced

inspissated

.87

.65

.79

.65

.93

.95

.65 .65

.87 .79

.95

.93

.33

.87

.50

.83

.83

.86

6)

CT

Note-P

Specificity

MR MR < .001.

mucus

or

fungal

infection

(6). The wide variation in signal intensities depends on the duration of obstruction and is due to the presence of water and the lack of protein content as the mucocele matures, or to the presence of infection by fungus. In their neview, Som et al (7) evaluated the factors that contribute to the varied signal intensities that emanate from mucoceles.

February

1991

b. Figure

2.

Mucocele.

Figure

3.

Schwannoma.

(a) Nonenhanced

axial

d.

C.

CT scan

through

the left frontal

sinus

reveals

an expansive

mass extending toward the roof of the left orbit (arrow). This patient had previously undergone resection of an ethmoidal adenocarcinoma, and CT was performed to rule out recurrent tumor. (b) The sagittal nonenhanced MR image (500/32) reveals an isointense mass in the left frontal region (arrow) superior to the previous resection. (c) The T2-weighted axial MR image (2,000/60) shows an area of mixed signal intensity in the left frontal region (arrow), compatible with the signal intensity characteristic of recurrent tumor or mucocele. (d) The sagittal gadolinium-enhanced MR image (500/32) reveals minimal peripheral enhancement (arrows) of this left frontal mucocele. The absence of solid enhancement eliminates the diagnosis of tumor recurrence. No recurrent tumor was found at surgical exploration.

b. that

extend

gion

into

(arrow).

ment

of

row).

There

posed

layers

They tion

the

These

this

intranasal

was

determining weighted T2 signal result

of

sinuses

findings

are

Ti-weighted

(arrow).

(c)

suggestive

schwannoma

no extension

(black

arrows).

of the neuroma

that

viscosity

protein

play

concentra-

a major

role

in

signal intensities on Tiimages, while the changes in intensities can be largely the dehydration.

lignancies,

Despite

their

in-

and low

to

that

these

tumors

intermediate

dem-

instances,

the observations plastic processes

Volume

axial

(d)

reveal

MR

an

image

isointense

(3,000/60)

noted

the frontal

MR

is homogeneous

sinuses,

and

mass reveals

A gadolinium-enhanced

Incidentally

into

178

our data

support

of these authors. Neocan often be differen-

Number

#{149}

tiated

from

nuses

on

2

mucoceles the

weighted However, of

in the

region

of the

ethmoid

signal

intensity

in

increased

image

(500/32)

enhancement

the homogeneously

cases

in

either

cal,

meaning

made

on obstructed

basis

of Ti-

MR imaging there is still

tics

reveals

within

enhanced

which are

with

signal

typical

that

si-

T2-

characteristics. a large minority

the not

and

or

homogeneous

the

area

air the

frontal

re-

enhance-

sinus

is thought

cells

same

(white

ar-

to represent

the op-

charactenis-

is often

extent

a diagnosis

confidence.

equivo-

cannot

In

dimeglumine

are

useful

these

and mapping neoplasms sinonasal tract. There

is probably

MR imaging

be

cases,

whom

the

need

diagnosis

of

to perform

tumor

fossa.

This

treatment

the

overall

a polyp

be

tissue (9).

gests

that

until

that should

proved

otherwise.

an isolated be

sinus

studied

sug-

attenuation

be considered We

on

tumor

further

MR

im-

aging, both with and without the injection of gadopentetate dimeglumine. There are disadvantages to the use of

tion

tumors

by

ostia,

sinus with

sinonasal nied

believe

high-attenuation

of

from

the

pa-

a tumor,

mucocele, or cyst. Differentiating a tumor inflammatory disease and problematic

isolated

the

when

granulation

should

and

na-

made

of the turbinates to be a serious flaw

is a clinical

CT scans

the

is easily

examination,

there

that

inferior

marked enhancement is not considered

made on the basis of results with CT scanning alone. On the other hand, paradigm

the

it enters

judgment

to distinguish

in can

to judge

when

at it

tient is considered. In addition, on the basis of the strict criteria we have employed in this study, it may be difficult

the

for patients

sal

difficult

of a mass

at physical

identifying

within no

studies

for

gadopentetate dimeglummne in this group of patients. Due to the marked enhancement of the nasal turbinates gadolinium-enhanced MR imaging,

are

and for the inexperienced observer, we believe that MR images obtained following the injection of gadopentetate

signal

intensity on both Ti- and T2-weighted images. Minor salivary gland tumors, because of their seromucinous secretions, may demonstrate high signal intensity on T2-weighted images, rendering them similar in signal intensity to obstructed sinuses or mucoceles. Similarly, the signal intensity of tumors of neural origin, such as schwannomas, may also vary.

In most

(500/32)

T2-weighted

of a mucocele.

depth analysis of signal differences, it is still difficult to differentiate mucocele from tumor by means of nonenhanced MR images. Som et al (8) also showed that cellulan tumors, such as squamous cell carcinoma, make up 90% of all sinonasal maonstrate

The

images

of mucoperiosteum.

concluded and

frontal

d.

C.

(a, b) Nonenhanced

the

are

results sinus

scarring

Primary

obstruction

which of

or

on

from from

is also

and

often

accompa-

adjacent

sinus

in

the

attenua-

CT

high

scans.

to distinguish tumor from tory and obstructive changes preoperative surgical planning Further

study

of

recurrent

in

ty

5).

active mature

the

The

abili-

inflammais vital for (Figs 4,

results

Radiology

of

MR

#{149} 427

b.

c.

Figure

4. Carcinoma and mucocele. (a) Nonenhanced axial Ti-weighted MR image (500/32) reveals an area of mixed signal intensity within the left maxillary sinus and both sphenoid sinuses (arrows). (b) The T2-weighted image (3,000/32) shows decreased or mixed signal intensity within the left maxillary sinuses (large arrow), increased signal intensity within the right sphenoid sinus (double arrows), and mixed signal intensity within the left sphenoid sinus (curved arrow). These findings suggest that tumor is present within the left maxillary sinus and that inflammatory changes are present in the right sphenoid sinus. The findings in the left sphenoid sinus are compatible with both tumor and mucocele. (c) The gadolinium-enhanced MR image (500/32) shows homogeneous, patchy enhancement in the left maxillary sinus

and

left nasal

cavity,

nus, the posterior ed is thickened

and

in

a portion

portion of the mucoperiosteum

of

the

left sphenoid within the

left

sphenoid

sinus

Sinus, and right maxillary

the

(small

arrows).

There

anterior portion sinus, with linear

is linear

of the left enhancement

enhancement

maxillary along

b. Figure

5.

CT scan (arrow), MR image mucocele.

Carcinoma superimposed revealed an opacified left with mixed internal signal (500/32)

There

(arrow).

that

is regular

linium-enhanced sinus

shows MR

image

Surgical

signal

linear obtained

(a)

most

hance

in

time

that

MR imaging is necessary. In conclusion, although use

of gadopentetate

re-

help

differentiate

ing tumor and not be justified tion, there are

be distinguished

428

Radiology

#{149}

the judicious

dimeglumine

some

cases

may

with

the use of a

with

an inflammatory is low.

The

which

is also compatible

solid

process. signal

enhancement

a coexisting

agent. on

evaluated for isolated axial MR image (3,000/60)

sinus

studies

right

sphenoid

si-

Also incidentally arrow).

are

with

not-

a mucocele.

and

medially

medial

sinus.

need

to

3.

the for

tumor

from

While

the

routine

MR

all patients

it has

an

use

imaging

with

gadolinium-

role

examination

or

findings at

equivoa sus-

References 1.

Shapiro sinuses

2.

Crawford

Am

MD, Som and nasal

1989;

Hanafee evaluation lesions. 242.

PM. cavity.

27:447-475. SC, Hasnberger WN. The role of extracranial Radiol Clin North

6.

7.

head Am

RB, in the

and neck mass 1989; 27:219-

W.

WP,

the

Sze

Benign

left

C, Lidov

and

a

gado-

maxillary

M,

Biller

malignant

si-

PM,

Dillon

WP, Curtin

HD,

Fullerton

CD,

Lidov M. Hypointense paranasal sinus foci: differential diagnosis with MRI and relation to CT findings. Radiology 1990; 176:777-781. Zinreich SJ, Kennedy DW, Malat J, Curtin HD, Epstein JI, Huff LC, et al. Fungal sinusitis: diagnosis with CT and MRI. Radiology 1988; 169:439-444. Som PM, Dillon WP, Fullerton CD, Zimmerman RA, Rajagopalan B, Maron Z. Chronically obstructed sinonasal secretions: observations on Ti and T2 shortening. Radiology 1989; 172:515-520.

8.

Som

9.

Lawson W. Sinonasal tumors and inflammatory tissues: differentiation with MR imaging. Radiology 1988; 167:803-808. Ross JS, Blaser 5, Masaryk TJ, Emory S. Bolesta M, Carter J, et al. Cd-DTPA enhancement of posterior epidural scar. AJNR 1989; 10:10831088.

MRI of the paranasal Radiol Clin North

HR. Lufkin of gadolinium

within

Dillon

with

Som

unen-

hanced CT or MR imaging are cal, when it may help confirm pected diagnosis. U

compatible a Ti-weighted,

5.

diseases,

when

On

4.

warranted

sinonasal

important

at physical

of is not

PM,

sinus. The sinus

nonasal lesions with intracranial extension: differentiation with MR imaging. Radiology 1989; 172:763-766. Van Tassel P. Lee YY, Jing BS. DePena CA. Mucoceles of the paranasal sinuses: MR imaging with CT correlation. AJNR 1989; 10:607612.

scarring. enhanced

Som

HF, Lawson

of

most (c)

anterior

maxillary left maxillary

gadolinium-enhanced

therefore

anteriorly

the

left

(b) The Ti-weighted,

in

examinations

recurrent

of the an opacified

is seen

cases to determine imaging modality

differentiating

opacification reveals

characteristics

carcinoma

Further

follow-up

postoperative most effective

of coexist-

infection, its use may in every case. In addisome cases that cannot

even

level with

focus

being

(arrows),

higher

contrast

the same

this

enhancement

imaging in these patients is required to ensure that recurrent tumor can be connectly differentiated from scarring. On the basis of enhancement patterns of granulation tissue and scarring as demonstrated within the postoperative lumbar spine (9), Ross et al found that mature scar tissue may cease to en-

at about

The

a mucocele

current tumor begins to enhance. Because of the potential for incorrect diagnosis of a neoplasm, further study of the normal pattern of postoperative enhancement within the sinonasal tract at

was T2-weighted

within

at a slightly revealed

patient

compatible

intensity

peripheral

exploration

This

sinus.

intensity,

the

the

arrows). (curved

C.

on mucocele. maxillary

within

sinus (large its surfaces

PM, Shapiro

MD,

Biller

HF, Sasaki

February

C,

1991