Computed Tomography of Malignant Tumors of the Nasal Cavity and Paranasal Sinuses

MAKOTO KONDO, MD, MASATOSHI HORIUCHI, MD,' HAYAO SHIGA, MD,f YUKIO INUYAMA, MD,' TAKUSHI DOKIYA, MD, YASUNORI TAKATA, MD,* SHOJl YAMASHITA. MD, KUNIO IDO, MD,t

YUTAKA ANDO. MD, YOSHIRO IWATA. MD, AND SHOZO HASHIMOTO, MD

Staging of malignant tumors of the nasal cavity and paranasal sinuses by computed tomography (CT) was studied in a total of 49 patients, 33 with squamous cell carcinoma and 16 with tumors of other histologic types. Involved sites by the tumor were studied, and clinical staging was made using CT findings alone according to AJC classification for maxillary sinus tumors. Surgical findings for com- parison were available for most cases. Of 33 squamous cell carcinomas and of 16 tumors with other histologic types, the maxillary sinus was the site of origin in 29 and eight, respectively. Of these 37 maxillary sinus tumors, 11 were staged T3,26 T4, and none was staged T1 or T2. None of these tumors were down staged, and one T3 was upstaged after surgical procedures, although all sinuses were not explored in some cases. Sinusitis due to obstruction was indistinguishable from the tumor without bone destruction. And the determination of the site of origin was difficult in some cases. Despite these, CT should be used for pretreatment evaluation of the tumors of these sites.

Cancer 50226-231, 1982.

N APPROPRIATE classification system is needed for A any tumor to aid oncologists in determination of treatment modalities and to help exchange of infor- mation among institutions. Of the nasal cavity and paranasal sinuses, cancer staging systems have been proposed only for the maxillary sinus, yet come to agreement.'-3 The reasons for this may be the inherent difficulties of accurate evaluation, since the sites are relatively difficult to assess noninvasively and most pa- tients present with advanced disease. Moreover, tools for staging may vary among institutions. Noninvasively, it has long been the practice to perform plain radiog- raphy and polytomography for this purpose, although the site of origin and the extent of the tumor is often difficult to define with accuracy.

Computed tomography (CT) has proven to be a valu- able diagnostic tool in the diagnosis of lesions of the nasal cavity and paranasal sinuses.4-' 'However, the role of CT in staging of these tumors has not been well established, since they are rather rare tumors, only 0.2%

From the Department of Radiology, *Department of Otorhinolar- yngology, ?Division of Diagnostic Radiology, Keio University School of Medicine, Tokyo, Japan.

$ Present address: Department of Radiology, Saiseikai Utsunomiya Hospital, Tochigi-ken, Japan.

Address for reprints: Makoto Kondo, MD, Department of Radiol- ogy, Keio University School of Medicine, 35 Shinanomachi, Shinjuku- ku, Tokyo 160, Japan.

Accepted for publication May 6, 1981.

of all cancers and 3% of head and neck cancers.I2 At our institution, not only biopsy but also surgical explo- ration or resection are performed for most cases, making it possible to compare CT findings with surgical find- ings. The purpose of this study is to provide the data concerning the incidences and pattens of involvement of these sites observed by CT, and to examine the value and limitations of CT in the classification of the tumors by the current staging system.

Materials and Methods

During the period 1978-1980, CT was performed in 49 patients with malignant tumors of the nasal cavity and paranasal sinuses. Those patients were referred to the Department of Radiology for preoperative or de- finitive ratiotherapy, and represented most of the cases seen at our hospital in the same period. Those cases consisted of 33 with squamous cell carcinoma and 16 with other histologic types including six with adenoid cystic carcinoma, three with adenocarcinoma, two with metastasis from hypernephroma, two with rhabdomyo- sarcoma, and single cases with mucoepidermoid car- cinoma, malignant lymphoma, and meningioma. The diagnosis of all cases was confirmed histologically, ei- ther by biopsy or following surgical excision. The tu- mors that originated at the alveolar ridge and extended into the maxillary sinus were excluded from the study. Although not malignant microscopically, meningioma

0008-543X/82/07 15/0226 $0.80 0 American Cancer Society

226

No. 2 CT OF MALIGNANT NASAL/PARANASAL TUMORS - Kondo et al. 227

TABLE I . Incidence of Bone Destructions and Soft Tissue Density Detected by CT in the Anatomical Sites According to Histologic Type

Squamous cell Adenoid cystic Adeno- Other carcinoma carcinoma carcinoma histologies

No. of cases (33) (6) (3) (7)

Maxillary wall(s) 30 (91%) 5 I 3 Anterior wall 23 (70%) 4 1 1 Medial wall 27 (82%) 5 1 3 Posterior wall 23 (70%) 3 1 2

Pterygoid plate 9 (27%) 3 1 1 Orbital wall(s) 1 1 (33%) 0 0 5 Skull base 4 (12%) 1 0 3

Maxillary sinus 31 (94%) 5 2 5 Ethmoid sinus 23 (79%) 2 3 7 Sphenoid sinus 9 (27%) 2 1 3 Frontal sinus 3 (9%) I 0 2 Nasal cavity 25 (76%) 3 3 4 Orbital cavity 14 (42%) 2 1 6 Nasopharynx 3 (9%) 2 1 0 Infratemporal fossa 23 (70%) 3 I 2 Pterygoid or temporal muscle 9 (27%) 2 1 1 Intracranial cavity 2 (6%) 2 0 1

Bone destruction

Soft tissue density

was included because of its invasiveness and fatality. Metastasis from hypernephroma was also included since the paranasal sinus tumors were found before the pri- mary was suspected in both cases.

Forty-two cases were examined with an EM1 CT 1010 scanner, and the last seven cases with a GE CT/ T X2 scanner. All examinations were performed with the patient positioned to produce standard transverse

FIGS. I A AND IB. (A, left) The maxillary sinus was involved with destruction of the medial and posterior walls. The pterygoid plate was displaced (arrow).( B, right) Squamous cell carcinoma involved the ethmoid sinus and orbit, and destructed the skull base. Intracranial extension became evident after contrast enhancement (arrow).

228 CANCER July 15 1982 Vol. 50

FIG. 2. Adenocarcinoma of the nasal cavity confirmed by surgical resection. The anterior and posterior ethmoids, and sphenoid sinus were explored, and sinusitis due to obstruction was found without tumor extension.

FIG. 3. Mucoepidermoid carcinoma involved the maxillary sinus, skin, nasal cavity, ethmoid and sphenoid sinuses, infratemporal fossa and skull base. It was difficult to estimate the site of origin, even after wide tumor resection in the sinuses and nasal cavity.

axial views, with the gantry directed parallel to the or- bitomeatal line. Coronal scans were obtained in only 18 cases due to a heavy case load and difficulties to maintain the position in older-age group. Contrast in- jection was done only when intracranial extension was suspected.

Tumors considered to arise within the maxillary sinus were staged according to American Joint Committee classification system:’ T1 tumors are confined to the mucosa of the infrastructure; T2 tumors are confined to the mucosa of the suprastructure, or produce medial or inferior bony wall destruction; T3 tumors invade the skin, orbit, anterior ethmoid sinus or pterygoid muscle; and T4 tumors extend to the cribriform plate, posterior ethmoids, sphenoid, nasopharynx, pterygoid plate or base of skull. The staging was based on CT findings alone.

CT findings about the maxillary sinus were confirmed surgically in most cases of which maxillary sinus was involved. At our institution, the treatment of choice has been surgical excision of the tumor with preoperative irradiation; even in advanced cases, surgical procedures such as maxillectomy or tumorectomy were performed for debulking the tumor. This policy permitted us to compare CT findings with surgical findings, at least some extent, for most cases.

The site of origin was estimated for all cases. The site having most significant soft tissue mass and/or bone destructions was considered the site of origin. If the medial wall of the maxillary sinus was involved; the maxillary sinus was thought to be the site of origin when either the anterior or posterior wall of the maxillary sinus was also involved; otherwise, the decision was made taking into account surgical findings or by con- jecture. I t is generally accepted that if there is doubt concerning the corect T, N, or M category to which a particular case should be alloted, the lower, i.e., less advanced category should be used. Along with this line, soft tissue density in the paranasal sinuses was not con- sidered a sign for involvement by the tumor unless frank bone destruction of that sinus was shown.

Results

The incidence of tumor involvement according to the sites and histologies is shown in Table 1. The maxillary sinus was most frequently involved. In 28 of 33 squa- mous cell carcinomas and six of 16 tumors of other histologic types, more than one maxillary walls were de- structed. Destruction of the posterior maxillary wall always accompanied soft tissue mass density extending into the infratemporal fossa (Fig. IA.). Soft tissue den- sity in the maxillary sinus usually coexisted with bone

No. 2 CT OF MALIGNANT NASAL/PARANASAL TUMORS - Kondo et al. 229

FIGS. 4A A N D 4B. (A, left) Adenocarcinoma of the maxillary sinus extended into the sinus on the opposite side. Air was trapped in the soft tissue density (arrow). The posterior wall of the maxillary sinus was destructed. (B, right) In this slice, extension into the nasopharynx (white arrow) and loss of the tissue plane in front of the pterygoid muscle (small arrow) were evident.

destruction of the maxillary sinus (39 out of 43); of the four cases without bone destruction, one had tumor in the maxillary sinus and three had sinusitis.

The presence of soft tissue density in the nasal cavity always accompanied soft tissue density either in the maxillary or ethmoid sinus (Fig. 2). The incidence of orbital wall destruction observed by CT was lower than that detected by conventional radiological methods, since coronal scans were not obtained in two-thirds of the patients, making observation of the inferior orbital wall difficult. The incidence of soft tissue density in the orbit was higher than that of orbital bone destruction

demonstrated by CT. Involvement of other structures such as the skull base (Fig. 1B and 3), nasopharynx (Figs. 4A and 4B) pterygoid plate (Fig. 1A) and in- tracranial cavity (Fig. 1 B) was readily delineated. In- volvement of the opposite side was observed and proved in four maxillary sinus tumors (Figs. 4 A and 4B).

There was no characteristic pattern of involvement specific to any histology. However, the tumors arising in the sphenoid sinus and frontal sinus were seen in the tumors of histologic types other than squamous cell carcinoma. The primary site estimated by CT is shown in Table 2. Of the tumors arising within the maxillary

TABLE 2. The Site of Origin Estimated by CT According to Histologic Type

Maxillary Nasal Ethmoid Sphenoid Frontal sinus cavity sinus sinus sinus

- - Squamous cell carcinoma 29 2 2 Adenoid cystic carcinoma 4 1 1 Adenocarcinoma 1 2 Mucoepidermoid carcinoma I Malignant lymphoma 1 Rhabdomyosarcoma 1 Meningioma - - -

1 Metastasis from hypernephroma - -

Total 3 1 5 5 1 1

- -

- - -

- - - -

- - - -

- - 1

I

-

- 1 -

230 CANCER July 15 1982 Vol. 50

sinus, ten squamous cell carcinomas and one adenoid cystic carcinoma were staged T3, and 19 squamous cell carcinoma and seven tumors with other histologic types were staged T4. Of these, no downstaging was done, and upstaging one T3 was done after surgical proce- dures including at least maxillotomy, tumorectomy and necrotomy. The tumors considered to arise outside the maxillary sinus were surgically explored in all but one case. Again, without bone destruction observed by CT, it was impossible to differentiate sinusitis from tumor extension; especially in the ethmoid and sphenoid sinus, in which sinusitis was frequently encountered.

Discussion

From the definition, clinical staging is preferable to surgical staging in any tumor of any site. For tumors of the nasal cavity and paranasal sinuses, however, sur- gical staging is often used since clinical evaluation has limitations to define the extent of the tumors.' Never- theless, reliable methods to delineate the tumor extent must be obtained, since not all tumors of these sites are treated surgically, making comparison of the treatment results among institutions difficult, and TNM classifi- cation should provide useful information about decision of treatment modalties and prognosis of the patients. Conventional plain radiography and polytomography are capable of showing fine bone structures and soft tissue density in the air filled cavities. However, the extension into the orbit, infratemporal fossa and intra- cranial cavity is only inferred by adjacent bone destruc- tions.

CT demonstrates both bone destructions and soft tis- sue mass density on the same image. I t readily shows the extension into the orbit, infratemporal fossa, pter- ygopalatine fossa, nasopharynx and intracranial cavity. These sites are very important both for staging and treatment planning. Our results reconfirmed the ability of CT to delineate the extent of the tumors. In a large number of patients, involvement of these sites was dem- onstrated by CT with good mapping quality. The in- terpretation of the images was easy and constant, and seemed to necessitate a little experience for most cli- nicians, an important factor from a clinical point of view. In addition, we felt that although the number was few, imaging quality of the newer GE scanner was far better than the EM1 CT 1010 scanner, and the in- creased capability of delineating fine bone structures might find the newer scanners superior to conventional radiographic met hods.

There were, however, two major problems, i.e., the site of origin and sinusitis. The site of origin of the tumor was often difficult to determine. If the medial wall of the maxillary sinus is involved, the site of origin

may be either the nasal cavity, ethmoid sinus or max- illary sinus." Even with surgical exploration, the site of origin may be obscured. Therefore, the estimation of the site of origin is often a matter of conjecture rather than science, yet it constitutes the basis for staging of the tumors. In our series, most of the patients with soft tissue density in the maxillary sinus accompanied bone destructions, which is a good sign for the presence of the tumor. A majority of them had destruction of more than one maxillary walls, which was considered a sign for the tumor arising within the maxillary sinus, al- though extensive tumors posed a question (Figs. 1A and IB, 3, and 4A and 4B).

Sinusitis due to obstruction is common in the tumors of the nasal cavity and paranasal sinuses. Sinusitis can- not be differentiated from the tumor based on CT num- bers,' although changes after contrast injection might be a clue.I3 In our study, soft tissue density in the para- nasal sinuses without bone destruction of those sinuses, there were about equal number of tumor involvement and sinusitis as far as surgically explored (Fig. 2). Since there was rather a large number of patients with soft tissue density in the ethmoid and/or sphenoid sinuses, these findings without bone destructions were not con- sidered a sign for tumor involvement for the purpose of T-staging. Although one T3 case of the maxillary sinus tumors was upstaged, most of them remained at the same stage after surgical exploration. The reason was that most of them were staged T4 by other findings such as extension into the nasopharynx, pterygopalatine fossa, etc, and only two T3 cases showed soft tissue density in the sphenoid and posterior ethmoid sinuses, one of which prove to be due to tumor extension.

There was no characteristic pattern concerning squa- mous cell carcinoma versus other histologic types. How- ever, bone destruction of the maxillary sinus was more frequent with the squamous cell carcinoma than with the other histologic types (P < 0.05). This reflected the preponderance of squamous cell carcinoma arising in the maxillary sinus. Squamous cell carcinoma arises in the nasal cavity and ethmoid sinus less frequently, and very rarely arises in the sphenoid and frontal s i n ~ s e s . ' ~

There are conflictions as to which classification sys- tem is preferable to others.' In the current study, based on the American Joint Committee classification system, 11 maxillary sinus tumors were staged T3 and 26 T4. These figures were concurred with other studies.'.'' However, it may be more important to provide reliable information concerning the extent of the tumors rather than to simply stage the patients, since such information should constitute the basis for developing an appropri- ate staging system.

The authors believed that CT was a reliable, re- peatable, noninvasive method for pretreatment evalu-

No. 2 CT O F MALIGNANT N A S A L I P A R A N A S A L TUMORS * Kondo et d. 23 1

ation, which provided good mapping information lead- ing to quick and easy understanding of the extent of the tumors.

REFERENCES

1. Harrison DFN. Critical look at the classification of maxillary sinus carcinomata. Ann Otol 1978; 87:3-9.

2. Sakai S, Shigematsu Y, Fuchitaka H. Diagnosis and T N M clas- sification of maxillary sinus carcinoma. Acta Otolaryngol (Stockh)

3. American Joint Committee for Cancer Staging and End-Results Reporting; Manual for Staging of Cancer. American Joint Commit- tee, Chicago, 1978; 45-48.

4. Caille J M , Constant PH, Renaud-Salis JL, Dop A. C T studies of tumors of the skull base, facial skeleton and nasopharynx. Comput Tomogr 1977; 1:217-224.

5. Takahashi M, Tamakawa Y, Shindo M, Konno A. Computed tomography of the paranasal sinuses and their adjacent structures. Comput Tomogr 1977; 1 :295-3 1 1.

6. Hesselink JR, New PFJ, Davis KR, Weber AL, Roberson G H , Taveras JM. Computed tomography of the paranasal sinuses and face. Part I : Normal anatomy. J Comput Assist Tomogr 1978; 2~559-567.

1972; 741123-129.

7. Hesselink JR, New PFJ, Davis KR. Weber AL, Roberson G H , Taveras JM. Computed tomography of the paranasal sinuses and face. Part 11: Pathological anatomy. J Comput Assist Tomogr 1978; 2568- 576.

8. Forbes WSTC, Fawcitt RA, lsherwood I , Webb R, Farrington T. Computed tomography in the diagnosis of diseases of the paranasal sinuses. C/in Radio/ 1978; 29:501-51 I .

9. Jing BS, Goepfert H, Close LG. Computerized tomography of paranasal sinus neoplasms. Laryngoscope 1978; 88: 1485- 1503.

10. Parsons C, Hodson N. Computed tomography of paranasal sinus tumors. Radiology 1979; 132:641-645.

1 1 . Guilbert-Tranier F, Piton J , Calabet A, Caille JM. Orbital syndrome-CT analysis of 100 cases. Comput Toniogr 1979; 3:241- 265.

12. Goldstein JC, Sisson GA, Tumors of the nose, paranasal si- nuses, and nasopharynx. In Otolaryngology. Philadelphia, W B Saun- ders, 1980; 2078-21 14.

13. Manelfe C, Bonafe A, Fabre P, Pessey JJ. Computed tomog- raphy in olfactory neuroblastoma: One case of esthesioneuroepithe- lioma and four cases of esthesioneuroblastoma. J Comput Assist Tom- ogr 1978; 2:412-420.

14. Frazell EL, Lewis JS. Cancer of the nasal cavity and accessory sinuses: A report of the management of 416 patients. Cancer1963; l6:l293-l301,