David M. Yousem, MD #{149}Douglas W. Fellows, MD #{149}David W. Kennedy, MDWilliam E. Bolger, MD #{149}Haskins Kashima, MD #{149}S. James Zinreich, MD

Inverted Papilloma: Evaluation.#{149} #{149}

with MR Imaging

cni

Head and Neck Radiology

The authors examined the magneticresonance (MR) appearance of in-verted papillomas to determine if thishistologically benign lesion could bedistinguished from malignancies ofthe sinonasal cavity. MR images in 10patients with histologically provedinverted papilloma were retrospec-lively reviewed. The signal intensityof inverted papilbomas on short repe-tition time (TR) images was iso- toslightly hyperintense to muscle in all10 patients. Inverted papillomas hadintermediate signal intensity on thelong TR/echo time (TE) images. Thetumors were iso- or slightly hy-pointense to fat on long TR/short TEimages. In the seven patients whoreceived gadopentetate dimeglumine,all inverted papillomas showed solidinhomogeneous enhancement. A re-view of eight sinonasal malignanciesshowed no distinctive signal inten-sity or enhancement characteristics tohelp differentiate inverted papillo-mas from various malignant tumors.The authors conclude that there is nosignature MR appearance for the be-nign inverted papilboma. The mainutility of MR imaging is in definingthe extent of the lesion.

Index terms: Nose, MR. 23.1214 #{149}Nose, neo-

plasms, 26.369 #{149}Papilloma, 26.369 #{149}Paranasal

sinuses, neoplasms, 23.369

Radiology 1992; 185:501-505

I From the Departments of Radiology, Neuro-

radiology Section (D.M.Y.) and Otorhinolaryn-gology, Head and Neck Surgery (D.W.K.,

WEB.), Hospital of the University of Pennsyl-vania, 3400 Spruce St, Philadelphia, PA 19104;and Departments of Radiology, NeuroradiologySection (D.W.F., S.J.Z.) and Otorhinolaryngol-ogy, Head and Neck Surgery (H.K.), The JohnsHopkins Hospital, Baltimore. Received April 15,1992; revision requested May 21; revision re-ceived June 2; accepted June 22. Address reprintrequests to D.M.Y.

e RSNA, 1992

S EVERAL recent studies have at-tempted to differentiate the be-

sions that occur in the sinonasal cay-ity by means of magnetic resonance

(MR) imaging signal intensity charac-teristics and enhancement patterns

(1-3). In general, inflammatorymasses can be distinguished from ma-

lignancies by their very high signalintensity on T2-weighted images. Inthe sinonasab cavity, a peripheral rim-enhancement pattern has been de-scribed in inflammatory mucosal dis-ease, polyps, and mucoceles (3). A

solid enhancement pattern is sugges-tive of malignancy. Elsewhere in thehead and neck, it has been suggestedthat less cellular neoplasms of the pa-rotid gland can be distinguished frommore cellular, poorly differentiatedcounterparts on the basis of signalintensity; the less differentiated be-

sions tend to have a bower signal in-tensity on T2-weighted images (4-6).

The inverted papibboma (IP) is a reb-ativeby common neoplasm of the na-sal cavity, accounting for 4% of allnasal neoplasms (7). IPs generallyarise along the lateral aspect of thenasal cavity but often grow into theparanasab sinuses and can be locallydestructive. The importance of diag-nosing IPs lies in the fact that squa-mous cell carcinoma may be coexist-

ent in 5.5%-27% of cases (7-11). It isnot clearly understood whether squa-mous cell carcinomas arise within bPs

or in separate sites. It would be veryuseful to distinguish the two neo-plasms because a more aggressive

therapeutic intervention at the outsetwould be indicated for squamous cellcarcinoma. The prognostic implica-tions and the necessity of postopera-tive radiation therapy are different forthe two entities.

To determine whether MR imagingis useful in differentiating between IP

and sinonasal malignancies, we retro-speetiveby reviewed the MR imagingexperience with IP at two institutions.Our goal was to describe the appear-

ance of IP at MR imaging and com-pare these characteristics with thoseof sinonasab malignancies reported inthe literature.

MATERIALS AND METHODS

The MR images of all patients with his-tologically proved bPs from two institu-

tions were retrospectively reviewed by

three neuroradiobogists (D.M.Y., D.W.F.,

S.J.Z.). The study population consisted of10 patients (seven men and three women)aged 35-70 years (mean, 54 years). Thenasal cavity was involved in all 10 pa-

tients; extension into the paranasal sinuses

was seen in eight. Two patients had bilat-erab sinonasal extension.

All images were obtained at 1.5 T with a

Signa imager (GE Medical Systems, Mib-

waukee) with a quadrature head coil. TI-weighted images were obtained with rep-etition times (TRs) of 400-850 msee andecho times (TEs) of 11-30 msee. Long TRimages were obtained with TRs of 2,500-

3,600 msec and TEs of 18-36 msee (protondensity) and 80-108 msec (T2 weighted).Fast spin-echo (rapid acquisition with re-

laxation enhancement [RARE]) T2-weighted pulse sequences were used in

three patients (12). All images were ob-

tamed with a 256 x 128-256 matrix, infe-nor saturation pulses, and one or twosignals averaged. Gadopentetate dimeglu-

mine was administered at a dose of 0.1

mmob/kg in seven of the 10 patients. TI-weighted images were obtained immedi-

ately after administration of the gadopen-tetate dimeglumine, with TRs and TEs of600-833 and 17-26 msee, respectively.

Images were evaluated for signal inten-

sity and contrast enhancement character-

istics. The signal intensity of the IP was

compared with that of muscle, fat, mu-

cosa, and cerebrospinal fluid (CSF) for the

Ti-weighted, proton-density, and T2-weighted images. Because fat has high

signal intensity on fast spin-echo T2-

weighted images and intermediate signalintensity on routine spin-echo images, im-

Abbreviations: CSF = cerebrospinal fluid,IP = inverted papilloma, RARE = rapid acquisi-tion with relaxation enhancement, TE = echotime, TR = repetition time.

a. b. c.

502 #{149}Radiology November 1992

Figure 1. Bilateral IP in a 35-year-old man with nasal congestion. (a) Axial T2-weighted MR image (TR msec/TE msee = 3,500/90) obtainedthrough the nasal cavity shows a mass of heterogeneous mixed (predominantly intermediate) signal intensity (arrows) in both nasal cavities.

Note the high signal intensity of obstructed secretions in the maxillary antrum (M). The patient had previously undergone surgery of the eth-moid sinuses. (b) Coronal Ti-weighted MR image (600/16) shows hypointense mass (arrows) within both nasal cavities, with even more hy-

pointense secretions within the inferior left maxillary antrum (M). The patient had previously undergone partial turbinatectomy on the rightside. (c) Fat-suppressed coronal MR image (800/26) obtained after administration of contrast material shows solid enhancement of the nasal

cavity mass with gradational enhancement between the inferior right turbinate and the region of the left middle turbinate. The solid enhance-

ment is to be contrasted again with the peripheral rim enhancement within the maxillary sinus inflammatory secretions.

ages of the three patients who underwentRARE imaging were evaluated separatelyfor comparison of fat signal intensity.

Contrast enhancement of the IP wasassessed for homogeneity and degree and

compared with adjacent inflammatorymueosal enhancement. Enhancement wasclassified as marked (enhancing as muchas inflamed mucosa), moderate (definite

enhancement but not as sinking as that ofmucosa), or mild (slight increase over

baseline signal intensity after administra-tion of gadopentetate dimegbumine, muchless than that of mueosa). The imageswere evaluated by consensus of the threeneuroradiologists viewing the images to-

gether.

RESULTS

At Ti-weighted imaging, five IPswere isointense and five were hyper-intense to muscle. The masses werehyperintense to CSF and hypointenseto fat in all eases (Table).

The proton-density images re-veabed the masses to be hyperintenseto CSF in all patients and isointense tofat in all but two patients, in whom

the lesion was hypointense to fat. ThebPs were hyperintense to muscle.However, heterogeneity within thetumors was seen in eight of the 10 bPson long TR/short TE images.

On the conventional T2-weightedspin-echo images, the masses wereisointense to fat in three of the sevenpatients and hyperintense to fat infour. On the RARE images, all threetumors were hypointense to fat butwere still of intermediate signal inten-sity (similar to that of gray matter)(Fig 1). On T2-weighted images, be-sions were hypointense to CSF andhyperintense to muscle in all cases.On the long TR/TE images, septafions/striations within the IP were seen infive of the 10 patients.

Solid enhancement of the IP wasseen in all 10 cases; the enhancementwas heterogeneous in seven. AllIPs enhanced bess than the adjacentinflammatory mucosa, which hadmarked enhancement, and enhance-ment was considered mild in twopatients and moderate in five(Fig 2).

The extent of disease was con-firmed with review of surgical notes.In two patients, the masses extendedthrough the cribriform plate to theintracranial compartment. No intra-parenchymal cerebral invasion waspresent; however, meningeal infiltra-tion was seen at surgery. The intracra-nial extent of the tumor was not ap-preciated at computed tomography(CT) performed in these two cases atthe same time as MR imaging. In onecase, the mass grew into the mastica-tor space via the pterygopabatinefossa.

One patient underwent extensivepreoperative biopsies, which led to adiagnosis of IP. The postoperativesurgical specimen revealed multifoeabmicroscopic areas of squamous cellcarcinoma within the IP (Fig 3). Evenin retrospect, none of the three neuro-radiologists could identify the onepatient with both squamous cell earci-noma and IP.

In 1 year, eight patients with mabig-nancies of the sinonasal cavity (threeesthesioneurobbastomas, two melano-

b. d.

Figure 3. Left nasal IP with microscopic foci of squamous cell carcinoma in a 26-year-oldman. (a) Axial MR image (3,500/18) through the nasal mass reveals a mass of heterogeneousintermediate signal intensity within the left nasal cavity, with extension into the left maxillaryantrum. High signal intensity is seen in the obstructed secretions in the left maxillary antrum.0’) Axial MR image (3,500/90) again shows a mass of intermediate signal intensity within theleft nasal cavity, with nonaggressive expansion and deformity of the left medial maxillary si-nus wall. This signal intensity would be identical to that seen with squamous cell carcinoma.(c) Axial Ti-weighted image (600/Il) obtained after administration of gadopentetate dimeglu-

mine shows a solid mass extending from the left nasal cavity into the left maxillary antrum.Note the peripheral rim enhancement of the maxillary antrum, denoting inflammatory dis-ease as opposed to a neoplastic process. (d) Fat-suppressed coronal MR image (700/26) obtained

after administration of gadopentetate dimeglumine reveals left nasal cavity enhancementwith a small nodule (arrow) extending through the left maxillary sinus ostium. Opacificationof the left maxillary antrum with obstructed secretions is noted. Note also the extension of theneoplasm through the enbnform plate (arrowhead) on the left side, where the low signal in-tensity of the cortical bone is lost. Compare this with the appearance of the contralateral side.

Multiple foci of microscopic squamous cell carcinoma were found scattered within the pre-dominant IP.

a.

mas, two squamous cell carcinomas,

and one small cell carcinoma) under-went MR imaging with similar param-eters as those performed for the IPs.An informal review of these cases byone neuroradiobogist revealed thatthree olfactory neurobbastomas (Fig4), one melanoma, one squamous cellcarcinoma, and a small cell carcinoma(Fig 5) had some of the same signalintensity characteristics and enhance-ment patterns as those of the bPs re-ported herein. The septated/striatedappearance noted in five IPs wasfound in one squamous cell carci-noma and one olfactory neuroblas-toma. The signal intensity patterns ofIP and sinonasab malignancies are alsocompared with those reported in theliterature in the following discussion.

DISCUSSION

bPs are masses that typically arisealong the lateral nasal wall or in theparanasab sinuses (most frequently inthe maxillary antrum). Patientspresent with nasal congestion,epistaxis, nasal discharge, and recur-rent sinusitis. bPs originate from thenasal septum in only 5.5%-18% ofcases (9,i3). Bilateral IPs occur in lessthan 5% of cases (8). Men are three tofive times more commonly affectedthan women, and whites are morecommonly affected than blacks(7,8,11,13). The typical age range is40-60 years, although IPs in the nasalseptum often occur at a younger age(iO,i4-16). The etiobogic characteris-tics of IPs are unknown, although an

b.

Figure 2. Right nasoantral IP in a 48-year-

old woman. (a) Axial 12-weighted MR image(2,700/108) through the maxillary antrum

reveals a mass of heterogeneous signal inten-sity within the right maxillary antrum and

right nasal cavity, with well-defined borders.

Intermediate signal intensity on 12-weightedimages is not always indicative of high-grade

malignancy. (b) Fat-suppressed coronal MR

image (750/26) obtained after administration

of contrast material reveals a solidly enhanc-ing mass occupying the right nasal cavity

and right maxillary antrum.

association with human papibbomavi-rus-li has been proposed (ii).

The IP is seen pathologically as avascular mass with prominent mu-cous cyst inclusions interspersed

throughout the epithelium and a highintracellular glycogen content (8,17).The IP gets its name because of itsunusual growth pattern into the un-derlining stroma rather than in anexophytic direction. bPs have beencabled by many names, including en-dophytie papilboma (to suggest itsgrowth pattern), squamous cell papib-boma, transitional cell papibboma, cy-bindricab epithebioma, inverting papil-boma, and sehneiderian papilboma.The terms used for the exophytic pap-ilbomas are fungiform (most common)or cylindrical cell (rare) papilbomas(8,17). Fungiform papibbomas tend tooccur on the nasal septum. Cybindri-cab cell papilbomas favor the maxillarysinus (8).

Squamous cell carcinomas may beassociated with bPs in 5.5%-27% ofpatients (7-11,18). Other neoplastic

a. b.Figure 4. Olfactory neuroblastoma in a 66-year-old woman. (a) Axial RARE 12-weighted im-

age (4,300/90) shows a mass with heterogeneous signal intensity and what appear to be stria-tions/septations within the left nasal cavity component. Obstructed secretions (S) have highsignal intensity. (b) Axial MR image (750/26) shows heterogeneous enhancement of the neu-roblastoma. In a blinded reading, two of the three neuroradiobogists thought that this imageshowed IP.

a. b.Figure 5. Small cell carcinoma of the nasal cavity in a 72-year-old woman. (a) Axial 12-

weighted MR image (3,500/90) reveals a low-signal-intensity mass within the nasal cavity,

with moderate heterogeneity of signal intensity. The mass expands the lamina papyracea on

the right side more than the left. (b) Fat-suppressed coronal MR image (700/26) obtained afteradministration of contrast material reveals that the carcinoma has a homogeneous pattern of

enhancement as well as extension through the eribriform plate (arrowheads). The signal in-

tensity of this mass and the enhancement characteristics are no different than those of IP; the

two cannot be distinguished.

504 #{149}Radiology November 1992

forms, such as mucoepidermoid carci-noma, verrucous carcinoma, and ade-noid cystic carcinomas, have alsobeen reported to coexist with IP. Ma-lignant change of an IP has a weakassociation with human papibbomavi-rus-16 (ii). Because of the associationof IP and squamous cell carcinomaand the very high recurrence rate ofIP (15%-78%) (7,8,10,11,15,18), com-plete surgical extirpation is the treat-ment of choice. This is usually accom-plished by means of lateral rhinotomyand en bloc excision of the lateral na-sal wall (7,15,19). Midfaciab degbovingprocedures may be an option that hasimproved cosmetic results; however,recurrence rates are still reported tobe 22%-29%, even with lateral rhinot-omy and midfaciab degboving (18).

When the IP is far advanced andentering the intracranial space, cran-iofaciab resection may be required,with a combined neurosurgicab-otorhinobaryngobogic team (18,19).The morbidity of craniofaciab resec-tion must be weighed against the ben-efit of completely removing the be-sion. If an imaging modality couldenable accurate prediction of thepresence and location of a coexistentmalignancy, surgical planning wouldbe greatly aided.

Unfortunately, we have found thatthe signal intensity characteristics andenhancement pattern of IP are identi-cal to those reported with squamouscell carcinoma (1-3,19). We could notdistinguish IP from olfactory neuro-blastomas, mebanomas, or small cellcarcinomas. Nor could we identify the

IP with squamous cell carcinomawithin it. Although calcification mayoccasionally be seen in IPs and wouldbe a useful finding in an intranasabmass (usually limiting the diagnosis to

an olfactory neurobbastoma and IP),in the one patient in whom this waspresent at CT, we were unable toidentify the calcification with MR im-

aging. (The presence of increased at-tenuation on CT scans may not be aresult of dystrophic calcification butmay be reactive change along residualethmoid sinus septa.)

Som et al (20) have previously re-ported that the pattern of bone de-struction of IP may be identical to thatof squamous cell carcinoma, particu-larly in the sphenoid and frontal si-nuses and along the anterior cranialfossa. Bone destruction has been re-ported in up to 30% of IPs and is pre-sumed to be due to pressure necrosis(7,20,21). The presence of bone de-struction is better evaluated with CTthan with MR imaging; in the regionof the eribnform plate, where the

bone is sievelike, tumor can extendintracraniably with only minimal os-seous changes.

Differentiation of IP from inflam-matory disease may be more success-ful in routine cases in which the in-flamed mucosa has low signalintensity on Ti-weighted images andvery high signal intensity on T2-weighted images. The signal intensityof IPs was never as high as that of be-nign inflamed mucosa or obstructedsecretions. Nonetheless, it has beenshown that hyperproteinaceous si-nonasal secretions may have a vari-able signal intensity that may overlapthe intensity pattern of bPs (22). Thesolid enhancing pattern of the IP maydistinguish it from inflamed mucosaand/or mucoceles, which have pe-ripheral rim enhancement (3).

Minor salivary gland tumors andsarcomas of the sinonasal cavity havebeen reported to have high and low

signal intensity characteristics withlong TR pulse sequences (2,23,24).The premise that a mass with bow sig-nal intensity at T2-weighted imagingrepresents a poorly differentiated tu-mor (and, therefore, a poor prognosis)is fraught with exceptions, just as

high signal intensity on T2-weightedimages is not always indicative of be-nign tumors or inflammatory lesions.

Nonetheless, MR imaging is veryuseful in determining the locationand extent of sinonasab neoplasms. Intwo eases, intracranial extent was notdefinitely suggested with CT. In oneof these instances, the incorrect CTinterpretation was a result of severemetal artifact from tooth amalgam. Inthe other case, infiltration was verysubtle and was not visualized with CTbecause of its bower soft-tissue resobu-tion.

There is no signature pattern of MRsignal intensity characteristics or en-

hancement suggestive of a specificdiagnosis of IP. One cannot differenti-ate between patients with coexistent

squamous cell carcinoma and thosewith IP. U

ReferencesI. Som PM, Shapiro MD, Biller HF, Sasaki C,

Lawson W. Sinonasal tumors and inflam-matory tissues: differentiation with MRimaging. Radiology 1988; 167:803-808.

2. Sigal R, Monnet 0, de Baere 1, et al. MRimaging of head and neck adenoid cysticcarcinomas: a study of twenty-seven caseswith elinicopathological correlation. Radi-obogy 1992; 184:95-101.

3. Lanzieri CF. Shah M, Krauss D, Lavertu P.Use of gadolinium-enhanced MR imagingfor differentiating mueoceles from neo-plasms in the paranasal sinuses. Radiology1991; 178:425-428.

4. Som PM, Dillon WI’, Sze C, et al. Benignand malignant sinonasal lesions with in-traeranial extension: differentiation withMR imaging. Radiology 1989; 172:763-766.

5. Som PM, Biller HF. High-grade malig-nancies of the parotid gland: identificationwith MR imaging. Radiology 1989; 173:823-826.

6. yogI TJ, Dresel SHJ, Spath M, et al. Pa-rotid gland: plain and gadolinium-en-haneed MR imaging. Radiology 1990; 177:667-674.

7. Myers EN, Fernau JL, Johnson JT, TabetJC,Barnes L. Management of inverted papil-loma. Laryngoscope 1990; 100:481-490.

8. Hyams DJ. Papillomas of the nasal cavityand paranasal sinuses. Ann Otob RhinolLaryngol 1971; 80:192-206.

9. KellyJH, Joseph M, Carroll E, et al. In-verted papilloma of the nasal septum. Arch

Otolaryngol 1980; 106:767-771.10. Lawson W, LathangerJ, Som P. Bernard PJ,

Billar HF. Inverted papilloma: analysis of87 eases. Laryngoscope 1989; 99:117-124.

11. Furuta Y, Shinohara 1, Sano K. Molecularpathologic study of human papillomavirusinfection in inverted papilloma and squa-

mous cell carcinoma of the nasal cavitiesand paranasal sinuses. Laryngoscope 1991;101:79-85.

12. HennigJ, Friedburg H. Clinical applica-tions and methodological developments ofthe RARE technique. Magn Reson Imaging1988; 6:391-395.

13. Outzen JE, Grontved A, Jorgensen K,Clausen PP. Inverted papilloma of thenose and paranasal sinuses: a study of 67patients. Clin Otolaryngol 1991; 16:309-312.

14. Sim DW. Coexistent inverted papilbomaand squamous carcinoma of the nasal rep-tum. J Laryngol Otol 1989; 103:774-775.

15. Buehwald C, Nielsen LH, Nielsen PL, et al.

Inverted papilloma: a follow-up study in-eluding primarily unacknowledged eases.Am J Otolaryngol 1989; 10:273-281.

16. Buchwald C, Nielsen LH, Ahlgren P. et al.Radiologie aspects of inverted papilloma.EurJ Radiol 1990; 10:134-139.

17. Barnes L. Surgical pathology of head and

neck lesions. New York: Dekker, 1985; 403-451.

18. Dolgin SR, Zaveri VD, Casiano RR, Man-iglia AJ. Different options for treatment ofinverting papilloma of the nose and para-nasal sinuses: a report of 41 cases. Laryngo-scope i992; 102:231-236.

19. van Olphen AF, Lubsen H, van’t VerlaatJW. An inverted papilloma with intracra-nial extension. J Laryngol Otol 1988; 102:534-537.

20. Som PM, Lawson W, Lidov MW. Simu-lated aggressive skull base erosion in re-sponse to benign sinonasal disease. Radiol-ogy 1991; 180:755-759.

21. Weissler WW, Montgomery WW, TurnerPA, et al. Inverted papilloma. Ann OtolRhinol Laryngol 1986; 95:215-221.

22. Som PM, Dillon WP, Fullerton GD, et al.Chronically obstructed sinonasal secre-tions: observations on Ti and T2 shorten-ing. Radiology 1989; 172:515-520.

23. Meyers SP, Hirsch WL, Curtin HD, et al.

Magnetic resonance imaging features ofehondrosareomas of the skull base. Radiol-ogy 1992; 184:103-108.

24. Yousem DM, Lexa FJ, Bilaniuk LT, Zimmer-man RI. Rhabdomyosareomas in the headand neck: MR imaging evaluation. Radiol-ogy, 1990; 177:683-686.