Imaging tumors of the minor salivary glands

Takashi Kaneda, DDS, PhD,a Manabu Minami, MD,b Kaoru Ozawa, DDS, PhD,C Yoshiaki Akimoto, DDS, PhD,d Manabu Okada, DDS,” Hirotsugu Yamamoto, DDS, PhD,f Hiromi Suzuki, DDS, PhD,s and Yasuhito Sasaki, MD, PhD,h Matsudo and Tokyo, Japan NIHON UNIVERSITY SCHOOL OF DENTISTRY AT MATSUDO AND UNIVERSITY OF TOKYO

HOSPITAL

Magnetic resonance imaging evaluations of nine histopathologically confirmed minor salivary gland tumors were made retrospectively and compared with evaluations of images obtained by computed tomography. All tumors had low-to-intermediate Tl signal intensities and intermediate-to-high T2 signal intensities. Malignant tumors had an irregular margin in all but one case. Benign tumors invariably had well-defined margins. In terms of tumor margination, the magnetic resonance imaging findings correlated well with the histopathologic findings. Magnetic resonance imaging demonstrated the internal architecture of the minor salivary gland tumors multidirectionally and was superior to computed tomography in this respect and in the ability to locate the tumors. (ORAL SURC ORAL MED ORAL PATHOL 1994;77~385-90)

The minor salivary glands are located within the sub- mucos,a of the ‘oral cavity and oropharynx. A wide range of neoplas,ms that arise from salivary glands can

occur in these regions.‘? 2 In imaging tumors of the major salivary glands,

sialography, scintigraphy, magnetic resonance imag- ing (MRI), computed tomography (CT), and sonog- raphy (are frequently used.3-5 However, these modal- ities have not been widely used for the evaluation of tumors of the minor salivary glands. Recently MRI has been used for the diagnosis of various pathologic conditions of the parotid gland,6-s although there is a dearth of reports concerning MRI of the minor sali- vary glands.

The purposes of this study were (1) to evaluate the efficacy of MRI in the diagnosis of tumors of the mi- nor salivary glands, (2) to compare the MRI findings with those of CT, and (3) to correlate the MRI find- ings with histopathologic observations.

aAssistant Professor, Department of Radiology, Nihon University School of Dentistry at Matsudo. bInstructor, Department of Radiology, University of Tokyo Hospi- tal. CResearch Assistant, Department of Radiology, Nihon University School of Dentistry at Matsudo. dAssociate Professor, Department of Oral Surgery, Nihon Univer- sity School of Dentistry at Matsudo. eGraduate student, Department of Radiology, Nihon University School of Dentistry at Matsudo. ‘Professor and Director, Department of Pathology, Nihon Univer- sity Schalol of Dentistry at Matsudo. sprofessor and Director, Department of Radiology, Nihon Univer- sity School of Dentistry at Matsudo. hProfessa8r and Director, Department of Radiology, University of Tokyo Hospital. Copyright @ 1994 by Mosby-Year Book, Inc. 0030-4220/94/$3.00+0 7/16/55023

MATERIAL AND METHODS Diagnostic images from nine tumors of the minor

salivary glands treated at the Dental Hospital of Ni- hon University School of Dentistry at Matsudo between 1990 and April 1993 were used for this study. The patients comprised two men and seven women with a mean age of 47 years (range, 32 to 65 years). MRI, CT, and hnstopathologic examinations were done in all cases.

In three patients, MRI examinations were done with a 0.064 Tesla permanent magnet system (AC- CESS, Toshiba America MRI Inc., San Francisco, Calif.) with a belt-type neck coil. Axial Tl-weighted images were obtained with the use of a three-dimen- sional Fourier transformation method with the gradi- ent echo technique. The imaging parameters were repetition time (TR) 68 msec, echo time (TE) 24 msec, flip angle 45 degrees, one acquisition, 256 X 256 matrix, 1.1 X 1.1 mm pixel size, 3.5 mm section thickness, and 3;! sections. Contiguous axial T2- weighted images were obtained by the spin-echo techaique. The scianning parameters were TR 1500 msec, TE 105 msec, three acquisitions, 192 X 256 matrix, 1.1 X 1.1 mm pixel size, 5 mm section thick- ness, and 10 sections. In addition, coronal or sagittal Tl-weighted images were made with the same pulse sequence as that for axial T 1 -weighted images. In the other six patients, MRI examinations were done with a 0.2 Tesla permanent magnet system (Magnetom P8 Plus, Siemens-Asahi Meditech, Tokyo, Japan) with the use of a head coil. Axial Tl -weighted images were obtained by the spin-echo technique. The scanning parameters were TR 363 msec, TE 20 msec, 2 acqui- sitions, 256 X 256 matrix, 300 X 300 mm field of view, 7 mm section thickness, and 8.4 mm section in-

385

386 Kaneda et al. ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY September 1994

Table I. Summary of Histopathologic and CT and MR imaging data in nine minor salivary gland tumors

CTfindings

Location Margins Density Margins

MRI &dings

TI-weighted T2-weighted signal signal

Homogeneity intensity intensity;

Patient number/ i age/sex I Diagnosis

1/4l/male Pleomorphic Hard Ill defined Soft tissue Well defined adenoma palate density mass

2/50/female Pleomorphic adenoma

3/49/female Pleomorphic adenoma

4/ 5 1 /female Myoepithelioma

with bone defect

Hard Undetected palate

Hard Undetected palate

Soft Ill defined Soft tissue palate density

mass Hard Ill defined Soft tissue

palate density mass

Floor Undetected of the mouth

Hard I11 defined Soft tissue palate density

mass Floor Undetected

of the mouth

Hard Well Soft tissue palate defined density

mass

Well defined Inhomogeneous

Well defined Inhomogeneous

Well defined Inhomogeneous

5/65/female Mucoepidermoid carcinoma

Ill defined

6/42/female Mucoepidermoid carcinoma

Ill defined

7/32/male Adenoid cystic carcinoma

Ill defined

8/32/female Adenoid cystic carcinoma

Ill defined

9/60/female Carcinoma ex pleomorphic adenoma

Well defined Inhomogeneous

Inhomogeneous Intermediate to high

High

Intermediate to high

Intermediate

Intermediate

High

High

High

Homogeneous Intermediate High

Inhomogeneous Low IIigh

Inhomogeneous Low to intermediate

High

Inhomogeneous Low to intermediate

Intermediate to high

High

High

terval. TZweighted images were acquired by the spin-echo technique. The scanning parameters were TR 3000 msec and TE 120 msec, and other param- eters were identical to those for obtaining T 1 -weighted images. In addition, coronal or sagittal Tl-weighted images were taken with the same pulse sequence as that for axial Tl-weighted images. CT contiguous axial and coronal images 2 or 3 mm thick were obtained without administration of contrast media.

The images were evaluated by two oral and max- illofacial radiologists (T. K. and M. M.) to determine the tumor locations, and the CT (margination, den- sity) and MRI (margination, internal architechure, Tl and T2 signal intensities) findings of the tumors. The radiologists were blind to histopathologic diag- noses. Tumor density on CT was compared with that of the surrounding normal soft tissues. The MRI sig- nal intensities of the lesions were compared with those of fat (high) and muscle (low) on Tl-weighted images and those of cerebrospinal fluid (high) and muscle (low) on TZ-weighted images.

All tumors were evaluated histopathologically by an oral pathologist (H.Y.). The pathologic classifica-

tion used was that of the World Health Organiza- tion .9

RESULTS A summary of the pathologic diagnoses, tumor lo-

cations, CT findings, and MRI findings is provided in Table I. Seven tumors occurred in the palate (all ex- cept one in the hard palate) and two occurred in the floor of the mouth. Four tumors were diagnosed his- topathologically as benign (three pleomorphic ade- nomas and one myoepithelioma) and five as malig- nant (two mucoepidermoid carcinomas, two adenoid cystic carcinomas, and one carcinoma ex pleomorphic adenoma) .

MRI On MRI with Tl weighting, pleomorphic adenoma

appeared as a well-defined mass of inhomogeneous intensity between that of muscle and that of fat. On T2-weighted images, all the tumors had an inhomo- geneously high intensity. The myoepithelioma had a similar appearance to that of pleomorphic adenoma (Fig. 1). Mucoepidermoid carcinoma in two cases was

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volun ye 78, Number 3

Kaneda et al.

Fig. 1. Myoepithelioma in 5 1 -year-old woman. A, Axial CT image reveals soft tissue density mass that ex- tends from soft palate to retropharyngeal space. Lesion is ill defined (arrow). B, Axial Tl-weighted image shows well-defined mass of inhomogeneously intermediate signal intensity (arrow). C, Axial T2-weighted im- age shows well-defined mass of inhomogeneously high signal intensity. D, Histopathologically, myoepithe- lioma is surrounded by thin layer of fibrou.s connective tissue (arrow). (Hematoxylin and eosin stain; original magnification X 10.)

revealed as an ill-defined mass of inhomogeneously low-to-intermediate Tl -weighted signal intensity sur- rounded by a rim of intermediate-to-high intensity and of inhomogeneously high T2-weighted signal in- tensity (Fig. 2). The two cases of adenoid cystic car- cinoma were ill-defined masses of inhomogeneously intermediate intensity with surrounding high inten- sity on Tl-weighted images; the T2-weighted images showed masses of inhomogeneously high intensity. The carcinoma ex pleomorphic adenoma was similar to pleomorphic adenoma with regard to margination on MRI (Fig. 3).

CT CT showed masses of soft tissue density in five

cases, four of whdch were ill-defined and one of which was well-defined. In the other four cases (two in the palate and two in the floor of the mouth), the lesions were not detected on CT. One case of pleomorphic adenoma caused bony erosions of the palate, which could be detected on CT.

Histopathology Kstopathologically the five cases of malignant tu-

mor showed an inhomogeneous architecture that

388 Kaneda et al. ORAL SURGERY ORAL MEDlClNE ORAL PATHOLOGY September 1994

Fig. 2. Mucoepidermoid carcinoma in 42-year-old woman. A, Axial Tl-weighted image shows ill-defined mass of inhomogeneously low signal intensity surrounding high intensity rim in floor of mouth (arrows). B, Axial T2-weighted image reveals ill-defined mass of homogeneously high signal intensity. C, Histopatholog- ically, lesion reveals ill-defined tumor margination with hemorrhage in peripheral region (arrows). (Hema- toxylin and eosin stain; original magnification X IO.)

contained necrotic tissue and hemorrhage. This in- homogeneity correlated well with the inhomogeneous signal intensities found on MRI. Four malignant tu- mors with a peripheral rim of high Tl-weighted in- tensity were found pathologically to have associated hemorrhage peripherally (Fig. 2). With respect to margins, MRI findings were in close agreement with the histopathologic results.

DISCUSSION Among all salivary gland tumors 9% originate in

the minor salivary glands. to More than 50% of tumors in the minor salivary glands are localized in the pal- ate.‘O The remaining 50% occur in the lips (15%), cheeks (12%) tongue (5%), floor of the mouth (5%), and other regions. lo In this study, seven of nine cases arose in the palate. Seifert et al.” reported that 45% of tumors of the minor salivary glands are malignant,

and this reflects the proportions in the present study. Concerning histopathologic diagnosis, Chaudhry et al.2 reported that among 94 intraoral minor salivary gland tumors, 43 (46%) were benign and 51 (54%) were malignant. The benign tumors included 40 pleomorphic adenomas (43%) and 3 papillary cysta- denomas (3%). The malignant group comprised 19 adenoid cystic carcinomas (20%), 17 adenocarcino- mas (38%), 10 mucoepidermoid carcinomas (1 I%), 3 malignant pleomorphic adenomas (3%), 1 undiffer- entiated carcinoma (I%), and 1 unclassified carci- noma (1%). The clinical appearances of benign and malignant tumors of minor salivary glands are so similar that in most cases a tumor cannot be ranked as benign or malignant without histopathologic anal- ysis.”

Radiographically, benign salivary gland tumors may produce a well-defined saucerlike depression in

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 78, Number 3

Kaneda et al. 389

Fig. 3. Carcinoma ex pleomorphic adenoma in 60-year-old woman. A, Coronal CT image reveals inhomo- geneous: soft tissue density mass in hard palate (arrows). There is little information about inner structure of tumor. :No bony erosion is apparent. B, Axial Tl-weighted image shows well-defined mass of inhomogeneous mainly high intensity on palate (arrow). C, Sagittal T2-weighted image reveals well-defined mass of inhomo- geneously high signal intensity (arrow). D, Histopathologically, carcinoma ex pleomorphic adenoma shows capsule of fibrous connective tissue. There are several hemorrhagic foci in tumor (arrows). (Hematoxylin and eosin stain; original magnification X 10.)

the underlying bone. In contrast, malignant tumors ate may destroy the alveolar bone and invade the may invade the bone and produce a ragged radiolu- maxillary sinus. l l cent defect with poorly defined borders. Such tumors On CT, the normal parotid gland has low radio- locat,ed on the lateral aspect of the posterior hard pal- pacity12 because of its high fat content and is usually

390 Kaneda et al. ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY September 1994

well demarcated from the neighboring muscle. There- fore soft tissue tumors that occur in the parotid gland can also be detected with high contrast by CT. How- ever, the minor salivary glands are very small and have x-ray absorption characteristics similar to those of the surrounding soft tissue. In addition, tumors of the minor salivary glands are considered to have ra- diopacity similar to that of the surrounding tissue. Therefore minor salivary gland tumors are not always clearly detected by CT, even though histopathologic examination showed well-defined tissue margins in five tumors.

By way of contrast, all minor salivary gland tumors were clearly detected by MRI, because this provides high-contrast resolution. l3 In particular, T2-weighted images showed tumors to be more hyperintense than the surrounding soft tissue and provided better infor- mation about the extent of the tumors. The tumor margins tended to be ill-defined on MRI in the ma- lignant cases.

Histopathologically, the tumors cells infiltrated the surrounding tissues. Mandelblatt et a1.14 reported that the most reliable MRI appearance in malignant tumors of the parotid gland is the margin. The present study showed the same to be the case for tumors of the minor salivary gland. Seifert et al.‘O reported that the histopathologic criteria for malignancy of salivary gland tumors include infiltration, vascular invasion, perineural spread, and metastases. In the present study, however, the case of carcinoma ex pleomorphic adenoma had similar margins on MRI to those of the benign tumors.

MRI can demonstrate the internal architecture of a tumor more clearly than CT.15 Although this was true in our study, the internal architecture of benign tumors did not differ from that of malignant tumors. We observed that four of the five malignant tumors had peripheral areas of high Tl-weighted signal intensity, which represented evidence of hemorrhage on histopathologic examination. We considered that the surrounding high-intensity rim on Tl -weighted images might be suggestive of malignancy, although a confident differentiation between benign and ma- lignant tumors could not be made on the basis of hemorrhagic foci alone.

In conclusion, MRI is able to demonstrate the presence, extent, margins, and internal architecture of minor salivary gland tumors and is superior to CT as a diagnostic tool for the detection of these tumors. Differentiation between benign and malignant tu- mors, however, requires histopathologic analysis.

REFERENCES

1.

2.

3.

4.

5.

6.

I.

8.

9.

10.

11.

12.

13.

14.

15.

Fine G, Marshall RB, Horn RC Jr. Tumors of tbe minor sai- ivary glands. Cancer 1960;13:653-69. Chaudhry AP, Vickers RA, Gorlin RJ. Intraoral minor salivary gland tumors: an analysis of 1414 cases. ORAL SURG ORAL MED ORAL PATH~L 1961;14:1194-226. Wittich GR, Scheible WF, Hajek PC. Ultrasonography of the salivary glands. Radio1 Clin North Am 1985;23:29-37. Schmelzeisen R, Milbrat II, Reimer P, Gratz P, Wittekind C. Sonography and scintigraphy in the diagnosis of disease of major salivary glands. J Oral Maxillofac Surg 1991;49:798- 803. Stone DN, Mancuso AA, Rice D, Hanafee WN. Parotid CT sialography. Radiology 1981;138:393-7. Schaefer SD. Maravilla KR. Burns DK. Close LG. Merkel MA, Suss RA. Evaluation of NMR versus CT for parotid masses: a preliminary report. Laryngoscope 1985;95:945-50. Teresi LM, Lufkin RB, Wortham DG, Abemayor E, Hanafee WN. Parotid masses: MR imaging. Radiology 1987;163:405-9. Casselman JW, Mancuso AA. Major salivary gland masses: comparison of MR imaging and CT. Radiology 1987; 165:183-9. Seifert G, ed. Histological typing of salivary gland tumors, 2nd ed. Berlin: Springer-Verlag, 1991;8-38. Seifert 6, Miehlke A, Haubrich J, Chilla R, ed. Disease of the salivary glands. In:Pathology-diagnosis-treatment-facialnerve surgery. First series, Stuttgart: Georg Thieme Verlag, 1986: 171-2. Wood NK, Goaz PW, ed. Differential diagnosis of oral lesions. 3rd ed. St. Louis: CV Mosby, 1985;162-6. McGahan JP, Walter JP, Bernstein L. Evaluation of the pa- rotid gland: comparison of sialography, non-contrast computed tomography, and CT sialography. Radiology 1934;152:453-8. Mirich DR, McArdle CB, Kulkarni MV. Benign pleomorphic adenomas of the salivary glands: surface coil MR imaging versus CT. J Comput Assist Tomogr 1987;11:620-3. Mandelblatt SM, Braun IF, Davis PC, Fry SM, Jacobs LH, Hoffman JC Jr. Parotid masses: MR imaging. Radiology 1987;163:411-4. Rice DH, Becker T. Magnetic resonance imaging of the sal- ivary glands: a comparison with computed tomographic scan- ning. Arch Otolaryngol Head Neck Surg 1987;113:78-80.

Reprint requests. Takashi Kaneda, DDS, PhD Department of Radiology Nihon University School of Dentistry at Matsudo 2-870-l Sakaecho Nishi Matsudo, Chiba 271; Japan