Parapharyngeal space benign tumours: Our experience

Journal of Cranio-Maxillo-Facial Surgery 42 (2014) 101e105

Contents lists available at SciVerse ScienceDirect

Journal of Cranio-Maxillo-Facial Surgery journal homepage:

Parapharyngeal space benign tumours: Our experience Andrea Cassoni, Valentina Terenzi*, Marco Della Monaca, Davina Bartoli, Andrea Battisti, Oriana Rajabtork Zadeh, Valentino Valentini Maxillo-Facial Surg. Dept, Policlinico Umberto I, “Sapienza” University of Rome, Viale del Policlinico 155, 00100 Rome, Italy

a r t i c l e i n f o

a b s t r a c t

Article history: Paper received 19 March 2012 Accepted 21 March 2013

Only about 0.5% of all head and neck neoplasms occur in the parapharyngeal space (PPS) and approximately 80% of these tumours are benign lesions. Various surgical approaches some of which are associated with mandibulotomy to increase exposure have been described. This article describes our 16years’ experience in treating 60 PPS benign tumours with special focus on our surgical techniques intended to ensure adequate mass exposure and structure safety. On the basis of our experience we assert that mandibulotomy is currently not advocated in the surgical management of benign PPS tumours i.e. not even in very select cases. The transparotid approach is the treatment of choice for parotid gland lesions involving PPS and in cases of multinodular or uninodular pleomorphic adenoma relapse involving the PPS. The transcervical approach is suitable for the safe removal of even large PPS masses in most cases. Ó 2013 European Association for Cranio-Maxillo-Facial Surgery. Published by Elsevier Ltd. All rights reserved.

Keywords: Parapharyngeal space benign tumour Parotid gland Transmandibular approach Transcervical approach Transparotid approach

1. Introduction

2. Materials and methods

The parapharyngeal space (PPS) is a well-defined upper neck region that is located immediately lateral to the pharynx and deep to the mandibular ramus, the pterygoid muscles and the parotid gland; it extends from the skull base down to the hyoid bone. Its ventral (prestyloid) compartment is filled with fat and connective tissue, while the dorsal (retrostyloid) compartment corresponds to the carotid sheaths and the associated enclosed structures (e.g. the lymph nodes and cranial nerves from IX to XII) (Som and Curtin, 2003). This space harbours approximately 0.5% of all head and neck neoplasms and approximately 80% of these tumours are benign lesions (Stell et al., 1985; Hughes et al., 1995; Carrau et al., 1990; Servadei et al., 2012). Various surgical approaches, some of which are associated with mandibulotomy to increase exposure have been described (Cohen et al., 2005; Al Otieschan and Velagapudi, 2008). More specifically these procedures include the transcervical, transparotid and transoral approaches and a combination of the three. The final choice is dependent on the histological diagnosis. This article describes our experience of treating PPS benign tumours over a 16-year period, with special focus on our surgical techniques intended to ensure adequate mass exposure and structure safety.

2.1. Patient population

* Corresponding author. Via Eugenio IV n 28, 00167 Rome, Italy. Tel.: þ39 3931301546 (mobile); fax: þ39 (0) 649979107. E-mail address: [email protected] (V. Terenzi).

From January 1996 to September 2011 88 patients with a PPS mass were referred to the Maxillofacial Surgery Department of Policlinico Umberto I. All these patients underwent surgery and 60 of these patients all of whom had benign tumours were enrolled in this retrospective study. Pre-operative imaging consisted of a CT scan and/or an MRI study; in 18 cases (30%) fine-needle aspiration cytology (FNAC) was performed. The follow-up period ranged from 6 months to 16 years with an average of 7.6 years. Thirty-two patients (53.3%) were female and the remaining 28 (46.7%) were male. The average age of the patients was 38.3 years with a range of 16e 80 years. 2.2. Surgical approach 2.2.1. Transcervical approach In this surgical approach a curvilinear transverse skin incision is made approximately 2e3 cm below the lower border of the mandible in a horizontal skin crease. The incision is carried through the subcutaneous tissue and platysma. The marginal branch of the facial nerve is identified, elevated and preserved. After identification and downward retraction of the posterior belly of the digastric muscle the submandibular gland is freed from the surrounding tissues and reflected anteriorly. This manoeuvre allows identification

1010-5182/$ e see front matter Ó 2013 European Association for Cranio-Maxillo-Facial Surgery. Published by Elsevier Ltd. All rights reserved.


A. Cassoni et al. / Journal of Cranio-Maxillo-Facial Surgery 42 (2014) 101e105

of the facial artery which is subsequently clamped and transected. Subsequently blunt dissection is performed in order to identify and free the mass from the surrounding soft tissue so that it can be delivered through the wound and removed. Further manoeuvres may be performed in order to increase exposure and facilitate tumour removal. For example stylomandibular release and anterior dislocation of the mandible can increase exposure up to 50%. Additionally the styloid process, styloid musculature and posterior belly of digastric muscle can be removed further enhancing exposure.

after surgery. If a mandibulotomy has been performed a soft diet is recommended for 30 days.

2.2.2. Transparotid approach In the transparotid approach a Blair incision is performed in the preauricular skin crease, carried around the earlobe and then extended into the neck. The skin and the superficial muscular aponeurotic system (SMAS) are raised. After identification of the facial nerve and all of its branches a superficial parotidectomy is performed. Eventually the superficial lobe of the parotid gland may be dissected, pedicled anteriorly and preserved. Tumours of the deep lobe of the parotid gland often lie underneath the branches of the facial nerve. In such cases the branches of the facial nerve are dissected and freed from the capsule of the tumour and the nerve is gently mobilised. The tumour is then delivered through this opening in a three-dimensional manner with a combination of blunt and sharp dissection. Blunt finger dissection is often sufficient for extraparotid salivary tumours. In addition a combined transcervicaletransparotideal approach can be performed through an extended parotidesubmandibular incision.

3. Results

2.2.3. Transoral approach A mucosal incision is performed in the soft palate. Blunt dissection is performed; in fact, blunt finger dissection can be used to remove the tumour. 2.2.4. Mandibulotomy In the case of mandibulotomy, nasotracheal intubation is used in order to permit evaluation of the occlusion. A cutaneous incision extending from the mastoid to the submandibular region in the midline is performed. The mandibular ramus is exposed and the Spitz foramen identified. Damage to the marginal mandibular branch of the facial nerve is avoided by identifying it 1 cm anteriorly and inferiorly to the mandibular angle or ligating and elevating the facial vessels to protect it. At this point all manoeuvres described in the “transcervical approach” are performed. A vertical osteotomy is performed posterior to the Spitz foramen from the sigmoid incisures to the mandibular angle. Pre-plating using two titanium 2 mm miniplates is performed prior to completion of the osteotomy. This approach allows the surgeon to open the superior portion of the osteotomy without completely separating the inferior bone stumps thus preserving the condylar position. After tumour removal the mandible is repositioned to the pre-plated position. 2.3. Perioperative complications

2.5. Follow-up Post-operative radiological assessment was performed using an MRI or CT scan with contrast. Depending on the histological finding (e.g. pleomorphic adenoma, nodular oncocytosis, schwannoma, neuroma or vascular tumour) imaging was performed every year in order to monitor the patient for local relapse.

The diagnoses of benign masses in the PPS are described in Table 1. In 40 cases (64%) the masses originated from the deep lobe of the parotid or from the oropharyngeal minor salivary glands. The most common histopathological diagnosis was pleomorphic adenoma (67.5%, N ¼ 27), followed by Warthin tumour (N ¼ 4). Ten patients had undergone previous surgery involving the same anatomic region or parotid gland space. The recurrent tumour was a pleomorphic adenoma in all cases. In 41 cases (68.3%) the lesions presented as an oropharyngeal swelling or cervical painless lump with a median diameter of 4.7 cm (range 2.7e6.5 cm in diameter). Three patients were asymptomatic and their tumour was found incidentally. In contrast, in the 10 patients with pleomorphic adenoma relapse who had previously undergone surgery the tumours were detected during follow-up examinations. Facial nerve palsy was the physical indicator of the tumour in one case of a patient with paraganglioma (Table 2). A PPS mass (pleomorphic adenoma of the deep parotid gland) was diagnosed incidentally during a thyroid ultrasound exam in only one case, while the initial manifestation of symptoms leading Table 1 Diagnosis of benign PPS space masses. Diagnosis

No. masses (%)

Salivary gland total Pleomorphic adenoma Warthin tumour Lymphangioma Lymphoepithelial cyst Canalicular adenoma Basal cell adenoma Epidermoidal cyst Nodular oncocytosis Neurogenic total Paraganglioma Chemodectoma Glomic tumour Schwannoma (XI c.n.) Neuroma (V3) Others total Branchial cyst Lipoma Angioma Total

40 (66.6%) 27 4 2 2 1 2 1 1 9 (15%) 4 2 1 1 1 11 (18.4%) 8 2 1 60

V3: mandibular nerve.

Perioperative complications related to these surgical approaches include facial nerve palsy, damage to the mandibular and IX, X, XI, XII cranial nerves, malocclusion and non-union of the mandibular segments. 2.4. Perioperative management Antibiotic therapy (2.2 g of amoxicillin and clavulanic acid every 12 h) is administered intraoperatively and for 5e7 days after surgery. A drainage tube is placed in position and removed 2e3 days

Table 2 Symptoms at presentation. Symptom

No. patients (%)

Fullness in pharynx/neck Follow-up (relapse) Asymptomatic Dysphonia Total

41 10 3 1 60

(68.3%) (16.6%) (5%) (10.1%)

A. Cassoni et al. / Journal of Cranio-Maxillo-Facial Surgery 42 (2014) 101e105

to the diagnoses was the presence of a mass in the upper neck in other cases. Among the 18 patients who underwent pre-operative fine-needle aspiration (FNA) an accurate pre-operative diagnosis was accomplished in 9 cases (50%). 3.1. Surgical techniques The following surgical approaches associated with mandibulotomy were used: a) b) c) d) e)

transcervical approach (N ¼ 21), transcervical approach with mandibulotomy (N ¼ 3), transparotid approach (N ¼ 33), combined approach (N ¼ 1), combined approach with transoral approach (N ¼ 2).

Thirty-three PPS masses (55%) were successfully excised using the transparotid technique and no major post-operative complications (permanent facial nerve palsy, Frey syndrome or salivary fistula) were observed. In 6 patients (18%) a temporary slight-moderate facial nerve palsy (HouseeBrackmann score: IIeIII) was observed but resolved with corticosteroid therapy. One case of moderatesevere temporary palsy required physiotherapy treatment. The transcervical approach was used alone in 21 cases while in another three cases it was combined with a mandibulotomy (Tables 3 and 4). A mandibular nerve deficit was observed in 2 of Table 3 Surgical approaches depending to histological findings. Surgical approach

No. patients (%)

Transparotid approach Pleomorphic adenoma Warthin tumour Lymphangioma Branchial cyst Canalicular adenoma Lipoma Lymphoepithelial cyst Epidermoidal cyst Transcervical approach Branchial cyst Pleomorphic adenoma Basal cell adenoma Chemodectoma Schwannoma (XI c.n.) Lymphoepithelial cyst Lymphangioma Lipoma Paraganglioma Glomic tumour Nodular oncocytosis Combined approach Pleomorphic adenoma (relapse) Chemodectoma Paraganglioma Angioma Neuroma (V3) Total

33 (55%) 23 4 1 1 1 1 1 1 21 (35%) 7 2 2 1 1 1 1 1 3 1 1 6 (10%) 2 1 1 1 1 60

V3: mandibular nerve.


the cases that included a mandibulotomy. More specifically, in one case a V3 sacrifice was required in order to remove a V3 neuroma. A permanent palsy of the marginal mandibular branch of the facial nerve was observed in one case (Tables 4 and 5). As previously mentioned, the diagnosis was parotid gland pleomorphic adenoma relapse in ten cases (Table 5). In all cases but one (case 2) a total parotidectomy incorporating the previous surgical scar via the transparotid approach was performed. In one case a transoral approach was used because it was used in the previous treatment. 4. Discussion Benign tumours are the most common lesions found in the PPS. Approximately 80e90% of lesions in the PPS are benign tumours and most of them originate from the parotid gland (Hughes et al., 1995; Carrau et al., 1990; Cohen et al., 2005, Luksic et al., 2012). Tumours often grow to at least 3 cm in diameter before they are detected so patients may have these masses for long periods of time before symptoms arise (Som et al., 1984). Manifestations include a mass in the oropharynx and/or upper neck, pain, trismus, change in voice and obstruction of the Eustachian tubes and in rare instances obstructive sleep apnoea (Morfit, 1955; Khafif et al., 2005). Physical intraoral examination of patients with tumours that are larger than 3 cm in diameter usually reveals a smooth submucosal mass that anteromedially displaces the lateral pharyngeal wall, tonsil, and soft palate; often, these masses are misdiagnosed as an infection or a tonsil tumour (Olsen, 1994; Bozza et al., 2009). Pre-operative assessment typically includes an MRI and/or CT scan. Note that a neck ultrasound is not useful for this diagnosis. Angiography is required in cases of vascular lesions such as haemangioma or tumours of the neuroectodermal origin. The displacement pattern of the PPS fat indicates the possible origin of the lesions in most cases, thereby limiting the differential diagnosis (Bradley et al., 2011). Peroral or transcervical FNA is accurate in 90e95% of cases and can be performed under CT or ultrasound-guidance (Oliai et al., 2004; Bozza et al., 2009). Nevertheless the collection a nondiagnostic aspirate has been reported in 25e60% of cases due to lack of cellular material, excessive bleeding or other technical problems (Bradley et al., 2011). Core needle biopsy (CNB) has been introduced as an alternative approach for the pre-operative assessment of salivary gland lesions, however direct comparisons between FNAC and CNB can be complicated because both are clearly operator dependent. The main advantage of CNB over FNAC is the collection of a core sample of tissue that can be used for formal histologic and immunohistochemical analysis. Some authors have suggested that the CNB is a reliable technique for the assessment of salivary gland pathologies despite remaining limitations for the subclassification of some neoplastic lesions (Pfeiffer and Ridder, 2011). Other contraindications such as the potential for tumour spillage, capsule rupture and tumour recurrence are a function of the needle diameter and sample volume. Consequently ultrasound-guided CNB potentially has an increased risk profile as compared to that of the accepted FNAC technique despite the fact that studies investigating rates of

Table 4 Cases with mandibulotomy (N ¼ 3). No. patients



Years of surgery

Surgical approach




1 2 3


33 37 41

1996 2000 2001

Transcervical Transcervical Transcervical þ transoral

V3 neuroma Chemodectoma Parotid gland pleomorphic adenoma (multiple relapse)

NED at 16 yrs NED at 12 yrs Relapse 1 yrs after surgery

V3 sacrifice None Marginalis nerve and V3 damage

V3: mandibular nerve; NED: not evidence of disease.


A. Cassoni et al. / Journal of Cranio-Maxillo-Facial Surgery 42 (2014) 101e105

Table 5 Parotid gland pleomorphic adenoma relapses (N ¼ 10). No. patients



Previous surgery


Surgical approach



1 2


27 41

Enucleation Enucleation (2)

Multinodular Multinodular

3 4 5 6 7


34 58 40 52 41

Uninodular (PPS) Multinodular Multinodular Multinodular Uninodular (PPS)

NED at 16 yrs Relapse 1 yrs after surgery NED at 9 yrs NED at 8 yrs NED at 7 yrs NED at 6 yrs NED at 5 yrs

None Marginalis nerve sacrifice, V3 damage None None None None None

8 9 10


24 31 82

N.A. Enucleation Enucleation Enucleation Enucleation (transoral approach) N.A. Enucleation Enucleation

Transparotid Transcervical þ Transoral þ Mandibulotomy Transparotid Transparotid Transparotid Transparotid Transparotid þ Transoral

Multinodular Uninodular (PPS) Multinodular

Transparotid Transparotid Transparotid

NED at 4 yrs NED at 4 yrs NED at 4 yrs

None None None

V3: mandibular nerve; NED: not evidence of disease.

tumour recurrence as a function of needle diameter and biopsy technique are lacking (Douville and Bradford, 2012). In our practice, FNAC was performed in only 18% of the total number of cases and proved to be diagnostic in only 50% of the cases in which it was used. We only consider using this examination in cases of suspected malignancy. Incisional biopsy is not recommended due to the increased risk of potential complications such as haemorrhage, infection and tumour seeding. Frozen section biopsy may be used at the time of surgery if the suspicion of malignancy remains (Oliai et al., 2004; Bradley et al., 2011). We have not used CNB for PPS tumours in our practice to date. In cases of suspected malignancy i.e. when a precise subclassification of the lesion is required we prefer to perform incisional biopsy or biopsy intraoperative frozen sections in order to obtain a more accurate diagnosis. Given the benign nature of these lesions the surgeon must carefully consider the surgical approach. Various surgical approaches, some of which are eventually associated with mandibulotomy to increase exposure have been described. The more commonly used approaches include the transcervical, transparotid, transoral and a combination of these approaches (Cohen et al., 2005; Khafif et al., 2005; Bradley et al., 2011). Typical complications include palsy of cranial nerves VII, IX, X, and XI and vascular rupture. The choice of surgical approach is dependent on the histological diagnosis. The transcervical approach for the removal of PPS tumours was first described in 1955 (Morfit, 1955). It leads to a good exposure of the surgical field, allowing adequate control of neck vessels. Eventually mandibulotomy was added to enhance exposure in case of tumours that were larger than 5 mm in diameter (Cohen et al., 2005; Douville and Bradford, 2012). To date an extra-oral approach is considered as the treatment of choice and a mandibulotomy is only rarely required (Presutti et al., 2011; Bradley et al., 2011). The transparotid approach is considered for those cases in which total parotidectomy or facial nerve isolation is required. In our series this approach was used in 55% of cases since parotid gland tumours involving the PPS were the most frequent lesions. This approach was used in all but one of the pleomorphic adenoma relapses. Our choice is due to the fact that in most cases (77e90%) of pleomorphic adenoma relapse manifests as multinodular lesions, reinforcing the pathogenic hypothesis of incomplete excision, the violation of the pseudocapsule of the tumour or obvious or underestimated tumour spillage during the first surgery related to the impossibility of obtaining healthy free margins when the lesion is adjacent to the facial nerve (Redaelli de Zinis et al., 2008; Stennert et al., 2004). Depending on the location of the recurrence superficial or total parotidectomy has been suggested when the previous operation was a simple local excision. Facial nerve

resection has been considered for patients with a history of multiple recurrences or failed radiotherapy (Redaelli de Zinis et al., 2008). In our cases (Table 5) a total parotidectomy was performed after considering the intrinsic features of the tumour in order to prepare for a radical resection. 70% of the cases involved a multinodular lesion. Marginal mandibular nerve damage was observed postoperatively in only one case of multinodular multiple relapse and this complication was related to the nerve’s adherence to the tumour. Ultimately we used the transcervical approach to remove pleomorphic adenomas of the PPS only for primary treatment and when the lesion was close to the deep portion of the parotid gland without actually involving it. Various types of mandibulotomy have previously been described and in most cases two osteotomies are required (Kolòkythas et al., 2009). Typical complications include difficulty with mandibular healing, malocclusion, TMJ dysfunction and trigeminal numbness. To minimize loss of mental nerve sensation osteotomies distal to the mental foramen have been described (Spiro et al., 1981). In order to further conserve mandibular nerve integrity, we avoid performing the anterior osteotomy since surgical exposure is sufficient for the accurate removal of tumours. Mandibulotomy was necessary in only 3 cases in our practice and the last one was performed 11 years ago. In particular, as explained in Table 4, this approach was used in one case of V3 neuroma in which the sacrifice of the mandibular nerve was absolutely required. In addition it was used in the case of a large chemodectoma in order to reduce haemorrhage risk and in a case of multiple multinodular pleomorphic adenoma relapse in order to permit the most radical removal possible with the intention of reducing the possibility of future relapses. From our perspective and according to other authors an expert surgeon can remove even large PPS masses without mandibular osteotomies (Presutti et al., 2011). The transoral approach should be considered only in very select cases because of the risk of haemorrhage and/or cranial nerve damage and tumour spillage (Ducic et al., 2006). In our practice we used this approach only in association with the transcervical approach for two cases of pleomorphic adenoma relapse (Table 5). These consisted of one case of multiple multinodular relapse to improve the surgical field (case 2) and one case to remove a scar from a previous surgery (case 7). 5. Conclusions Based on our experience we can assess that in the surgical management of benign PPS tumours mandibulotomy is not recommended even in very select cases. The transparotid approach is the treatment of choice in parotid gland lesions involving the PPS

A. Cassoni et al. / Journal of Cranio-Maxillo-Facial Surgery 42 (2014) 101e105

and in cases of multinodular or uninodular pleomorphic adenoma relapse in the PPS. In most cases the transcervical approach is suitable for the safe removal of even large PPS masses. These approaches typically lead to a good surgical result with an acceptable number of complications. References Al Otieschan A, Velagapudi S: Transcervical, transmandibular, transparotid approaches to parapharyngeal tumours. J Craniomaxillofac Surg 36(Suppl. 1): S150eS151, 2008 Bozza F, Vigili MG, Ruscito P, Marzetti A, Marzetti F: Surgical management of parapharyngeal space tumours: results of 10-years follow-up. Acta Otorhinolaryngol Ital 29: 10e15, 2009 Bradley PJ, Bradley PT, Olsen KD: Update on management of parapharyngeal tumors. Curr Opin Otolaryngol Head Neck Surg 19: 92e98, 2011 Carrau RL, Meyers EN, Johnson JT: Management of tumors arising in the parapharyngeal space. Laryngoscope 100: 583e589, 1990 Cohen SM, Burkey BB, Netterville JL: Surgical management of parapharyngeal space masses. Head Neck 27(8): 669e675, 2005 Douville NJ, Bradford CR: Comparison of ultrasound-guided core biopsy versus fineneedle aspiration biopsy in the evaluation of salivary gland lesions. Head Neck, 2012 [Epub Ahead of print] Ducic Y, Oxford L, Pontius AT: Transoral approach to the superomedial parapharyngeal space. Otolaryngol Head Neck Surg 134: 466e470, 2006 Hughes KV, Olsen KD, McCaffrey TV: Parapharyngeal space neoplasms. Head Neck 17: 124e130, 1995 Khafif A, Segev Y, Kaplan DM: Surgical management of parapharyngeal space tumors: a 10 years follow-up. Otolaryngol Head Neck Surg 132(3): 401e406, 2005 Kolòkythas A, Eisel DW, El-Sayed I, Schmidt BL: Mandibular osteotomies for access to select parapharyngeal space neoplasms. Head Neck 31: 102e110, 2009


Luksic I, Virag M, Manoijlovic S, Macan D: Salivary gland tumours: 25 years of experience from a single institution in Croatia. J Craniomaxillofac Surg 40(3): e75ee81, 2012 Morfit HM: Retromandibular parotid surgery. Arch Surg 70: 906, 1955 Oliai BR, Sheth S, Burroughs FH, Ali SZ: “Parapharyngeal Space” tumors: a cytopathological study of 24 cases on fine-needle aspiration. Diagn Cytophatol 32(1): 11e15, 2004 Olsen KD: Tumors and surgery of the parapharyngeal space. Laryngoscope 104: 1e 28, 1994 Pfeiffer J, Ridder GJ: Diagnostic value of ultrasound-guided core needle biopsy in patients with salivary gland masses. Int J Oral Maxillofac Surg, 2011 Dec 26 [Epub ahead of print] Presutti L, Molteni G, Malvè L, Marchioni D, Ghidini A, Tassi S, et al: Parapharyngeal space tumors without mandibulotomy: our experience. Eur Arch Otorhinolaryngol 269(1): 265e273, 2011 Redaelli de Zinis LO, Piccioni M, Antonelli AR, Nicolai P: Management and prognostic factors of recurrent pleomorphic adenoma of the parotid gland: personal experience and review of the literature. Eur Arch Otorhinolaryngol 265: 447e 452, 2008 Servadei F, Romano A, Ferri A, Magri AS, Sesenna E: Giant trigeminal schwannoma with parapharyngeal extension: report of a case. J Craniomaxillofac Surg 40(1): e15ee18, 2012 Som PM, Curtin HD: Head and neck imaging, 4th edn. St Louis: Mosby Year Book, 2003 Stell PM, Mansfield AO, Stoney PG: Surgical approaches to tumors of the parapharyngeal space. Am J Otolaryngol 6: 92e97, 1985 Som PM, Biller HF, Lawson W: Parapharyngeal space masses: an update protocol based upon 104 cases. Radiology 153: 149e156, 1984 Spiro RH, Gerold FP, Strong EW: Mandibular “swing” approach for oral and oropharyngeal tumors. Head Neck Surg 3: 371e378, 1981 Stennert E, Wittekindt C, Klussmann JP, Arnold G, Guntinas-Lichius O: Recurrent pleomorphic adenoma of the parotid gland: a prospective histopathological and immunohistochemical study. Laryngoscope 114: 158e163, 2004