Limitations of Magnetic Resonance Imaging in the Evaluation of Perineural Tumor Spread Causing Facial Nerve Paralysis

ORIGINAL ARTICLE Limitations of Magnetic Resonance Imaging in the Evaluation of Perineural Tumor Spread Causing Facial Nerve Paralysis

Markus Jungehuelsing, MD; Christian Sittel, MD; Roman Fischbach, MD; Mathias Wagner, MD; Eberhard Stennert, MD

Objective: To present and discuss the clinical presen- patients, the initially performed MRI did not detect any tation and treatment in patients with long-duration uni- parotid gland lesion causing the paralysis, whereas long lateral facial paralysis and normal magnetic resonance im- duration of the paralysis and electroneurography indi- aging (MRI) findings. cated malignancy.

Design: Case series. Results: Exploration surgery was performed as total parotidectomy in these 8 patients and malignant parotid Setting: Ear, nose, and throat department of the Uni- gland tumors were proved in all 8 patients. versity of Cologne, Cologne, Germany. Conclusions: Individuals with facial nerve paralysis Patients: A total of 486 patients with unilateral facial without any signs of regeneration 6 months after the paralysis who were treated from 1986 to 1998. Besides onset of paralysis and/or persistent electrophysiological the usual diagnostic workup, a complete electrophysi- evidence of ongoing neuronal degeneration should ological evaluation, including investigations such as needle undergo surgical exploration of the parotid gland and electromyography and neuromyography (also known as facial nerve, even if MRI studies show no tumoral electroneurography), of the facial nerve was performed lesion. at repeated intervals. In 19 patients, a malignant tumor was delineated with ultrasonography or MRI. In 8 of these Arch Otolaryngol Head Neck Surg. 2000;126:506-510

ELL PALSY, by definition, is a sion is generally considered an ominous diagnosis of exclusion. sign, almost uniformly heralding malig- Eighty percent of all periph- nant disease. Usually, tumors are pal- eral facial paralyses are la- pable and detectable using ultrasonogra- beled as idiopathic or Bell phy, computed tomography, or magnetic palsy and a complete recovery occurs in resonance imaging (MRI).3 Nevertheless, B 1 70% of all patients. it is known that computed tomography Conversely, approximately 20% of and ultrasonography may fail to depict these facial palsies can be demonstrated to small parotid gland lesions, especially if have a specific cause. All patients exhib- they are located in the deep, retroman- iting facial paralysis thus should undergo dibular portion of the parotid gland or a thorough neurotologic evaluation to close to the stylomastoid foramen.4,5 Mag- identify the underlying abnormality. netic resonance imaging has higher accu- Facial paralysis of neoplastic origin racy in identifying soft tissue lesions. With is uncommon. It is estimated to repre- its high soft tissue definition, even very sent the etiology in approximately 5% of small lesions of the parotid gland and the From the Departments of all cases.2 Neoplastic involvement may be facial nerve become identifiable, as well as Otorhinolaryngology–Head and by neurogenic primary lesions of the sev- the deep portion of the parotid gland and Neck Surgery enth cranial nerve or by secondary, ex- the intramastoid, tympanic, and labyrin- (Drs Jungehuelsing, Sittel, and trinsic neoplasms. thine sections of the facial nerve.6-9 Stennert), Diagnostic and Interventional Radiology Parotid gland disease must be ac- The purpose of this article was to pre- (Dr Fischbach), and Pathology knowledged as a significant clinicopatho- sent a series of cases of facial paralyses (Dr Wagner), University of logic entity in the pathogenesis of facial caused by malignant parotid tumors, which Cologne Medical School, paralysis. The presentation of facial pa- had not been identified by gadolinium- Cologne, Germany. ralysis in the presence of parotid gland le- enhanced MRI.

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, APR 2000 WWW.ARCHOTO.COM 506

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 Table 1. Diagnoses and Nerve Function in Patients PATIENTS AND METHODS With Unilateral Facial Paralysis

No. (%) of From 1986 to 1998, we treated in our outpatient clinic Patients Facial Nerve Function 486 patients with peripheral, unilateral facial paraly- (N = 486) Diagnosis After 6 mo* sis. In addition to the usual neurotologic investiga- 344 (71) Idiopathic (“Bell palsy”) Recovery tions (standard pure tone and speech audiometry, im- 22 (4) Herpes zoster (IgM) Recovery pedance audiometry, modified Short Increment 37 (8) Pontine angle tumor Mixed, degeneration and Sensitivity Index, Carhart tests, brainstem audiom- regeneration etry in cases of equivocal findings, Schirmer tear test, 11 (2) Any central dysfunction† No recovery, denervation stapedial reflex testing, and taste evaluation), we per- potentials formed ultrasound of the parotid gland. A complete 27 (6) Malignant parotid gland No recovery, denervation electrophysiological workup including needle elec- tumor potentials tromyography and neuromyography (also known as 0 Benign parotid gland No paralysis electroneurography) was performed at repeated in- tumor 45 (9) Trauma Depending on extent of tervals. As a routine, all patients were submitted to trauma these investigations at the time of first contact with our department, then follow-up visits were sched- *Electromyographic findings. uled 2, 6, and 12 weeks after onset of paralysis. In †Central dysfunction includes apoplectic insult, hemorrhage, and other cases of persistent paralysis with or without electro- central nervous system disorders proved unequivocally by computed physiological signs of neuronal degeneration (ie, posi- tomography or magnetic resonance imaging. tive sharp waves in needle electromyography) follow-up was continued at 4-week intervals. In 344 and neck coil and high-resolution MRI reconstruction. patients (76%) any specific cause of the paralysis was In 5 patients, the repeated Gd-DTPA–enhanced MRI did ruled out and the paralysis was labeled as idiopathic. The other patients showed specific lesions as not depict any pathologic alteration. described in Table 1. In 27 patients (6%) a unilat- All 8 patients underwent total parotidectomy for di- eral facial paralysis was the first symptom of a ma- agnostic and therapeutic reasons. Histopathological in- lignant tumor of the parotid gland. Eight of these pa- vestigations revealed malignant tumors of the parotid tients had been referred to our service for facial gland in all 8 patients: adenoid cystic carcinoma with peri- reanimation surgery because of abnormally long du- neural spread in 4 patients, basal cell adenocarcinoma ration of complete facial nerve paralysis (Table 2). in 2 patients, mucoepidermoid carcinoma in 1 patient, In all 8 patients, previously performed gadopen- and acinic cell carcinoma in 1 patient. tetate (Gd-DTPA)–enhanced MRI of the parotid gland The MRI study did not detect the lesion in all pa- and of the pontine angle region had not shown any tients with adenoid cystic carcinoma and in 1 patient with pathological change, destruction, or lesion. In these 8 patients we performed, in addition to the neuroto- mucoepidermoid carcinoma. Histopathological studies logic examination and electromyography, another revealed predominant perineural, intraneural, and peri- Gd-DTPA–enhanced MRI and exploration surgery. vascular spread of the cancer without a real circum- script tumor (Figure 2). In 1 patient with basal cell adenocarcinoma, the initially performed MRI did only depict an accompanying RESULTS mastoiditis, but not the tumor itself. A second high- resolution Gd-DTPA–enhanced MRI scan hinted at the Our subgroup of 8 patients showed no electrophysi- parotid gland tumor; perhaps the histological results ologic regeneration after more than 6 months of paralysis proved very low mitotic activity and subsequently the slow (range, 6-36 months; mean, 17.6 months; Table 2). After growth of the tumor may have affected the documented electromyographic diagnosis indicated neoplastic inva- MRI signal intensity similarly (Figure 3). In the re- sion, another Gd-DTPA–enhanced MRI of the cerebello- maining 2 patients, the initial MRI studies simply did not pontine angle and the parotid gland was performed in pa- cover the parotid gland lesion. In these patients the MRI tients in whom MRI studies did not cover the whole parotid studies were performed primarily to exclude a cerebel- gland or in whom MRI studies had low diagnostic qual- lopontine angle tumor, but later served as guarantee for ity. Electromyographic criteria for neoplastic invasion were safe tumor exclusion. pathological spontaneous activities without interference activity pattern or intermediate activity pattern, but some- COMMENT times with transitory single oscillations in patients with clinically complete facial paralysis. Delineation of circumscribed tumoral lesions of the pa- In 2 patients a parotid gland tumor was diagnosed rotid gland today is the domain of ultrasonography, and, immediately (Figure 1). The formerly performed MRI for the deeper portions of the parotid gland, of MRI. Es- studies had simply not covered the parotid gland lesion, pecially for the evaluation of tumors of the stylomastoid but only the cerebellopontine angle region for acoustic foramen and the retromandibular part of the parotid gland, neuroma exclusion and the upper parts of the parotid MRI is the method of choice.6-10 gland. In 1 patient, tumoral invasion of the deep por- Magnetic resonance imaging allows excellent visu- tion of the parotid gland was demonstrated using a head alization of the facial nerve throughout its entire course

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, APR 2000 WWW.ARCHOTO.COM 507

Patient Age, Time Since Paralysis No. of MRI Presumed Cause for Histologic y/Sex Onset, mo Studies False-Negative MRI Findings 40/M 18 1 Perineural spread? size? ACC 52/M 24 3 Perineural spread? size? ACC 67/M 9 1 Tumor not covered Acinic cell carcinoma 42/M 6 1 Similar signal intensity BCC 23/F 30 3 Tumor not covered BCC 57/F 24 2 Perineural spread? size? ACC 55/M 14 1 Perineural spread? size? ACC 63/M 16 1 Size? Mucoperidermoid carcinoma

*MRI indicates magnetic resonance imaging; ACC, adenoid cystic carcinoma; and BCC, basal cell adenocarcinoma.

T1 10

Gd-DTPA

Figure 1. Left, Magnetic resonance imaging (MRI) study of a 23-year-old woman with complete left facial paralysis of 30 months’ duration (September 1998). T1-weighted, T2-weighted, and gadopentetate (Gd-DTPA)–enhanced transversal MRI of the parotid gland depicts a large tumor of the lower deep portion of the left parotid gland (arrow). The tumor is fairly visible in native T1 image, but disappears after Gd-DTPA enhancement (repetition time, 600/2000 milliseconds; echo time, 20/80 milliseconds; field of view, 240 mm; slice thickness, 4 mm). Right, MRI study of the same patient in January 1997. No tumor is visible on the 2 contiguous lowest slices. The scout does not cover the tumor region. The results were interpreted as exclusion of tumor, idiopathic facial nerve paralysis.

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, APR 2000 WWW.ARCHOTO.COM 508

Figure 2. Left, Magnetic resonance imaging study of a 55-year-old man with adenoid cystic carcinoma of the parotid gland and left facial nerve paralysis of 14 months’ duration. T1- and T2-weighted transversal slices of the region where the tumor was proved histologically are circled (echo time, 16/100 milliseconds; repetition time, 600/2500 milliseconds, contiguous slices; field of view, 240 mm; slice thickness, 4 mm). Right, Histological specimen of the same patient showing intraneural spread of the cribriform adenoid cystic carcinoma (hematoxylin-eosin, original magnification ϫ60). 1 indicates neurons; 2, cell clusters of the adenoid cystic carcinoma; 3, Schwann cells; and 4, parotid gland tissue.

from the brainstem to the stylomastoid foramen, and con- T1 T2 trast-enhanced Gd-DTPA MRI has become an essential part of the evaluation of patients with facial paralysis, even if the data remain somewhat inconclusive.11,12 Although Bell palsy has been considered a diagnosis largely made by exclusion, several reports of gadolinium-enhanced MRI for facial paralysis suggest that the diagnosis of Bell palsy in some instances can be confirmed by MRI rather than by deduction from the absence of abnormal findings.13-17 While MRI has become a valuable tool in evalu- ating patients with facial paralysis, its limitations should be recognized and must be interpreted in conjunction with the clinical presentation of facial paralysis.11 We present 8 patients with unilateral facial paraly- Figure 3. Magnetic resonance imaging study of a 42-year-old man with sis, all of them initially diagnosed as having Bell palsy. Ini- basal cell adenocarcinoma of the right parotid gland and incomplete facial tial and subsequent MRI showed no evidence of tumor- paralysis of 6 months’ duration. T1- and T2-weighted transversal sections of ous lesion. Although the history of each patient was not the skull base depict notable tissue surplus located at the deep portion of the consistent with Bell palsy in regard to the long duration right parotid gland (circle), but no tumor delineation. There is no signal Ͼ intensity difference between tumor and normal parotid gland tissue (echo of the unilateral facial paralysis ( 6 months without signs time, 20/100 milliseconds; repetition time, 600/3000 milliseconds; slice of regeneration), once the diagnosis was suggested by the thickness, 5 mm; field of view, 256 mm). negative results of the MRI studies, it was accepted in each case. No electrophysiological evaluation was ordered, and function over a 6- to 12-week course has never been, in the patients were not appropriately followed up clinically our experience, idiopathic in origin.18 A facial paralysis by their physicians. of sudden onset does not necessarily rule out tumor in- Unilateral facial paralysis that progresses beyond 3 volvement of the facial nerve. Fisch19 noted sudden on- weeks strongly suggests neoplastic involvement.1 An in- set of facial paralysis in 20% of patients with malignant complete facial paralysis progressive to complete dys- parotid gland tumors. These cases must be acknowl-

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, APR 2000 WWW.ARCHOTO.COM 509

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 edged as exceptions but should demonstrate the need for Accepted for publication June 18, 1999. neurotologic screening of every case of facial paralysis Reprints: Markus Jungehuelsing, MD, University, HNO- whether immediate or slowly progressive in onset. Klinik Ko¨ln, Joseph-Stelzmann-Strasse 9, 50924 Ko¨ln, Ger- Worse than gradual progression of facial paralysis many. is the remarkable persistence of a facial paralysis.1,20 Ab- sence of facial tone or of any resumption of clinical func- REFERENCES tion 6 months following paralysis suggests neoplasm and the patient must be treated accordingly. 1. May M, Hardin WB. Facial palsy: interpretation of neurologic findings. Trans Pa Our patient group demonstrates the need to per- Acad Ophthalmol Otolaryngol. 1977;84:710-722. form close electrophysiological monitoring in patients with 2. Shambaugh GE, Clemis JD. Facial nerve paralysis. In: Paparella MM, Shumrick atypical facial paralysis. The diagnosis of Bell palsy on the M, eds. Otolaryngology. Vol 2. Philadelphia, Pa: WB Saunders Co; 1973:275. basis of MRI findings in some situations may result in a 3. Spiro RH. Changing trends in the management of salivary tumors. Semin Surg false sense of security in the management of patients with Oncol. 1995;11:240-245. 4. Gosepath K, Hinni G, Mann W. The state of the art of ultrasonography in the head unilateral facial paralysis. Jackson et al wrote in 1980, and neck [editorial]. Ann Otolaryngol Chir Cervicofac. 1994;111:1-5. Despite the technological sophistication of the various diag- 5. Weber AL. Imaging of the salivary glands. Curr Opin Radiol. 1992;4:117-122. 6. Yanagida M. MRI enhancement of the facial nerve with Gd-DTPA—second report— nostic modalities (CAT scan), they remain limited in their pow- investigation of enhanced nerve portions in patients with facial palsy [in Japa- ers of resolution. False-negative data are not uncommon. It is nese]. Nippon Jibiinkoka Gakkai Kaiho. 1993;96:1329-1339. therefore not unusual that diagnosis and therapeutic direction 7. Jonsson L, Hemmingsson A, Thomander L, Bergstrom K, Stalberg E, Thuomas depend on the very direct approaches of surgery. In these cir- KA. Magnetic resonance imaging in patients with Bell’s palsy. Acta Otolaryngol cumstances, surgery serves as a diagnostic and therapeutic Suppl (Stockh). 1989;468:403-405. agent.18(p1585) 8. Murphy TP, Teller DC. Magnetic resonance imaging of the facial nerve during Bell’s palsy. Otolaryngol Head Neck Surg. 1991;105:667-674. Nineteen years later we still suggest surgery as diagnostic 9. Castelijns JA, van den Brekel BM. Magnetic resonance imaging evaluation of ex- and therapeutic agent in patients with peripheral facial pa- tracranial head and neck tumors. Magn Reson Q. 1993;9:113-128. ralysis and without signs of regeneration 6 to 8 months 10. Kane WJ, McCaffrey TV, Olsen KD, Lewis JE. Primary parotid malignancies: a after paralysis onset. Individuals experiencing a facial nerve clinical and pathologic review. Arch Otolaryngol Head Neck Surg. 1991;117: 307-315. paralysis without any signs of regeneration 6 months af- 11. Freije JE, Harvey SA, Haberkamp TJ. False-negative magnetic resonance ter onset of paralysis and/or persistent electrophysiologi- imaging in the evaluation of facial nerve paralysis. Laryngoscope. 1996;106: cal evidence of ongoing neuronal degeneration should un- 239-242. dergo surgical exploration of the parotid gland and facial 12. Gebarski SS, Telian SA, Niparko JD. Enhancement along the normal facial nerve nerve, even if MRI studies show no tumorous lesion. in the facial canal: MR imaging and anatomic correlation. Radiology. 1992;183: 391-394. 13. Murphy TP. MRI of the facial nerve during paralysis. Otolaryngol Head Neck Surg. CONCLUSIONS 1991;104:47-51. 14. Teresi LM, Lufkin RB, Wortham DG, Abemayor E, Hanafee WN. Parotid masses: Magnetic resonance imaging studies may prove but never MR imaging. Radiology. 1987;163:405-409. can exclude neoplastic lesions. Bell palsy is a diagnosis 15. Millen SJ, Daniels DL, Meyer GA. Gadolinium-enhanced magnetic resonance imaging in temporal bone lesions. Laryngoscope. 1989;99:257-260. of exclusion. Electrophysiological findings and regular 16. Engstrom M, Thuomas KA, Naeser P, Stalberg E, Jonsson L. Facial nerve en- follow-up visits continue to be the most important as- hancement in Bell’s palsy demonstrated by different gadolinium-enhanced mag- pects in the management of patients with unilateral, pe- netic resonance imaging techniques. Arch Otolaryngol Head Neck Surg. 1993; ripheral facial paralysis. Although MRI with and with- 119:221-225. 17. Engstrom M, Abdsaleh S, Ahlstrom H, Johansson L, Stalberg E, Jonsson L. out Gd-DTPA enhancement of the parotid gland, the Serial gadolinium-enhanced magnetic resonance imaging and assessment of stylomastoid foramen, and the intratemporal segment of facial nerve function in Bell’s palsy. Otolaryngol Head Neck Surg. 1997;117: the facial nerve provides useful information, negative ra- 559-566. diographic studies should not be taken as the final word 18. Jackson CG, Glasscock ME, Hughes G, Sismanis A. Facial paralysis of neoplas- indicating absence of neoplasms. In patients with long- tic origin: diagnosis and management. Laryngoscope. 1980;90:1581-1595. 19. Fisch U. Diagnosis and therapy of salivary gland tumors [in German]. Schweiz duration paralyses and no signs of regeneration, surgi- Med Wochenschr. 1978;108:927-932. cal exploration of the parotid gland and the facial nerve 20. Neely JG. Neoplastic involvement of the facial nerve. Otolaryngol Clin North Am. is compelling even with normal MRI findings. 1974;7:385-396.

(REPRINTED) ARCH OTOLARYNGOL HEAD NECK SURG/ VOL 126, APR 2000 WWW.ARCHOTO.COM 510