Key points

Epidemiology

Primary parotid tumors are rare and account for approximately 1% to 3% of all head and neck tumors. Fortunately, most (75%–85%) are benign. The annual age-adjusted incidence of benign parotid tumors in the United States is approximately 3.8 per 100,000 per year, with roughly 1300 to 1600 diagnosed cases each year. Worldwide, incidence varies by geography, with reports of 5.3 to 6.2 per 100,000 in the United Kingdom and 1.35 per 100,000 in Poland. Japan and Malay report an incidence of 1.3 and 1.1 for all benign salivary neoplasms. Unlike their malignant counterparts, no national registries exist for benign diseases, making true incidence difficult to ascertain.

Salivary tumor classification schemes include benign versus malignant, major (parotid, submandibular, sublingual) and minor salivary glands, and by individual histopathology. The 2005 World Health Organization’s classification contains 24 malignant salivary histopathologies and 11 benign, excluding hematolymphoid and secondary tumors. Of parotid tumors, the most common benign and malignant tumors are pleomorphic adenoma (PA) and mucoepidermoid carcinoma, respectively. This article is limited to benign parotid neoplasms and does not address malignancies or tumors that primarily occur in the submandibular, sublingual, or minor salivary glands. Many clinicians use the 80/20 rule for salivary gland neoplasms: 80% benign, 80% occur in the parotid, and 80% are PA. However, variations in certain proportions and relative incidences exist. A Ugandan study reported only 29% of benign tumors occur in the parotid gland, with 53.8% of parotid tumors being malignant. Zero cases of Warthin tumors (WTs) were found, a paucity also described in other African studies. Underreporting may be an issue in resource-poor areas, as adequate therapy may not be sought for non–life-threatening diseases. These issues highlight the variations in incidence reporting for benign parotid tumors.

By age, incidence of benign parotid tumors steadily increases starting at 15 to 25 years of age, with a peak in 65 to 74 years of age. There is a female sex preference overall (1.46:1.0) for benign parotid tumors and a racial difference favoring Caucasian patients over African American. Males (ratio 2.31:1) are more affected in WTs (presumably because of historically higher rates of smoking) and parotid malignancies (ratio 3.47:1) overall. PA and WTs combine to make up 83% to 93% of benign parotid tumors. More detailed epidemiologic variables will be forthcoming for individual histopathologies ( Table 1 ).

Epidemiology

Primary parotid tumors are rare and account for approximately 1% to 3% of all head and neck tumors. Fortunately, most (75%–85%) are benign. The annual age-adjusted incidence of benign parotid tumors in the United States is approximately 3.8 per 100,000 per year, with roughly 1300 to 1600 diagnosed cases each year. Worldwide, incidence varies by geography, with reports of 5.3 to 6.2 per 100,000 in the United Kingdom and 1.35 per 100,000 in Poland. Japan and Malay report an incidence of 1.3 and 1.1 for all benign salivary neoplasms. Unlike their malignant counterparts, no national registries exist for benign diseases, making true incidence difficult to ascertain.

Salivary tumor classification schemes include benign versus malignant, major (parotid, submandibular, sublingual) and minor salivary glands, and by individual histopathology. The 2005 World Health Organization’s classification contains 24 malignant salivary histopathologies and 11 benign, excluding hematolymphoid and secondary tumors. Of parotid tumors, the most common benign and malignant tumors are pleomorphic adenoma (PA) and mucoepidermoid carcinoma, respectively. This article is limited to benign parotid neoplasms and does not address malignancies or tumors that primarily occur in the submandibular, sublingual, or minor salivary glands. Many clinicians use the 80/20 rule for salivary gland neoplasms: 80% benign, 80% occur in the parotid, and 80% are PA. However, variations in certain proportions and relative incidences exist. A Ugandan study reported only 29% of benign tumors occur in the parotid gland, with 53.8% of parotid tumors being malignant. Zero cases of Warthin tumors (WTs) were found, a paucity also described in other African studies. Underreporting may be an issue in resource-poor areas, as adequate therapy may not be sought for non–life-threatening diseases. These issues highlight the variations in incidence reporting for benign parotid tumors.

By age, incidence of benign parotid tumors steadily increases starting at 15 to 25 years of age, with a peak in 65 to 74 years of age. There is a female sex preference overall (1.46:1.0) for benign parotid tumors and a racial difference favoring Caucasian patients over African American. Males (ratio 2.31:1) are more affected in WTs (presumably because of historically higher rates of smoking) and parotid malignancies (ratio 3.47:1) overall. PA and WTs combine to make up 83% to 93% of benign parotid tumors. More detailed epidemiologic variables will be forthcoming for individual histopathologies ( Table 1 ).

Embryology and histogenesis

Understanding parotid embryology is important as several theories propose an etiopathogenesis based on salivary cell types and tumor cell origin. All salivary glands derive from ingrowths of oral epithelium, with parotid anlage starting to appear at 4 to 6 weeks of development. Lymphoid tissue develops before the encapsulation of the parotid gland begins but after the encapsulation of both submandibular and sublingual glands, thus explaining the lack of lymphoid tissue in the latter. Because of concomitant parotid encapsulation and lymphatic development, salivary cells sometimes appear within the intraparotid and periparotid lymph nodes.

The most basic secretory unit, the acinus, is formed from acinar cells and is surrounded by contractile myoepithelial cells. Acini drain into intercalated ducts, followed by striated intralobular ducts, and finally intralobular and main excretory ducts ( Figs. 1–3 ). Acinar cells are exclusively of the serous type in the parotid gland. Noncontractile basal cells line the most distant ducts. Intraparotid and periparotid lymph nodes that may give rise to lymphoma are a part of the lymphatic drainage pathway of cancers from the scalp and face or other distant sites.

The normal parotid is made up of lobules that contain acini (a), ducts (d), and adipose tissue (f) (hematoxylin-eosin, original magnification ×20).

Serous-type acinar cells (a) and ducts (d) make up the functional unit of the parotid (hematoxylin-eosin, original magnification ×200).

The flat, contractile myoepithelial cells surrounding acini and ducts are difficult to appreciate with standard hematoxylin and eosin stains, better appreciated with calponin staining ( brown ) (calponin, original magnification ×200).

Controversies exist regarding salivary gland histogenesis. For decades, a bicellular theory of histogenesis was dominant, suggesting that the varied salivary cell types were products of differentiation from 2 stem cell progenitors: excretory duct reserve cells and intercalated reserve cells. Therefore, mucoepidermoid, ductal, and adenocarcinomas arose from proximal excretory cells, whereas the remainder arose from more terminal, intercalated cells. However, animal and human models have demonstrated that even differentiated cells, such as acinar cells, can cycle between one another and any cell could be a potential target of neoplastic processes.

The multicellular theory of histogenesis assumes numerous cell types can contribute to tumorigenesis, even highly differentiated acinar cells. Such a varied input of histogenesis may account for the wide spectrum of pathologies that exist. More recent work has been in classification of tumors by morphologic features regardless of cell origins: tumor organization, types of cell differentiation, materials produced by tumor cells, and so forth. All of these issues place credence to the complex and relatively unknown pathogenesis of parotid tumors.

Cause and risk factors

Despite the poorly understood histogenesis of parotid tumors, several environmental variables have been associated with parotid gland tumors. However, conclusive evidence demonstrating causal relationships is lacking. A Japanese study of atomic bomb survivors found a 3.5% and 11.0% relative risk increase for radiation and salivary gland tumors, with WT being most associated with radiation risk. Another study found a 2.6-fold increased incidence of parotid tumors with head and neck scalp irradiation after a 2-decade latency period. Studies looking at cell phone radiation and tumor risk have found mixed results, though a 2008 Israeli study found a 1.58 increased odds ratio for tumors only in a specific subgroup of the very highest cell phone users.

Cigarette smoking is strongly associated with WTs ; certain occupational exposures, such as heavy metals, and hormonal factors (eg, early menarche) have also been associated with an increased risk of salivary tumors.

Advancements in genetic sequencing technologies have allowed for easier identification and discovery of distinct genetic translocations, of which we now know of 6. These translocations include PLAG1 and HMGA2 rearrangements in PA, CRCT1-MAML2 found in mucoepidermoid carcinoma and WT, and 4 others. Some may serve as prognostic markers, but more investigation is needed. Recent work on single nucleotide polymorphism analysis on a genome-wide level shows promise in finding new genetic markers. However, the genetic basis to parotid tumorigenesis is still largely unknown.

Clinical anatomy for benign neoplasms

For more detailed discussion of parotid gland and facial nerve anatomy, see Chapter 1 of this text .

The paired parotid glands are the largest salivary glands and weigh, on average, 15 to 30 g. The parotid gland is bounded superiorly by the zygomatic arch and inferiorly by the anteromedial margin of the sternocleidomastoid muscle. The posterior extent can reach the external auditory meatus and mastoid tip. The superficial portion may cover a small portion of the masseter muscle anteriorly. The deep portion of the gland wraps around the mandibular ramus and rests on the surface of the posterior digastric muscle, styloid process, and stylohyoid muscle. Any mass in these areas should be considered neoplastic until proven otherwise.

The facial nerve divides the parotid gland into 2 lobes, superficial (lateral) and deep (medial). This distinction is critical for surgical planning as PAs represent most deep lobe and parapharyngeal space tumors. Surgery of deep lobe tumors has a markedly increased rate of facial nerve injury and operative time. This increased rate is due to the frequent need to expose, mobilize, and retract the facial nerve in order to access these tumors.

Ichihara and colleagues recently reviewed 425 cases of benign parotid tumors and proposed instead a 3-category clinical classification scheme: superficial tumors, deep tumors, and lower pole tumors, defined as the region inferior to marginal mandibular nerve. They found distinct characteristics for lower pole tumors. Compared with superficial tumors, lower pole tumors had older patients on average (57.4 vs 52.2 years), more males (ratio 1.6:1), and more WT compared with PA (ratio 2.5:1). More importantly, the investigators concluded that nerve dissection for deep, lower pole tumors could safely be limited to the marginal mandibular nerve in place of a true deep dissection, highlighting the importance of preoperative localization. This method allowed for a faster, safer, and easier dissection relative to a regular deep parotid dissection.

The accessory parotid gland is a tiny 0.5- to 1.0-cm gland found in 21% to 56% of people based on cadaveric studies. It is usually located 6 mm anterior to the main gland, adjacent to Stenson’s duct as it passes over the masseter muscle. Accessory gland lesions make up 1% to 8% of parotid lesions and up to half are malignant. With accessory gland tumors, patients may present with a midcheek mass.

Clinical presentation

Benign parotid tumors classically present with a painless, slow-growing, preauricular or upper-neck swelling. The differential diagnosis of parotid swelling is broad, ranging from mumps, to sialadenitis, to neoplasms. The initial diagnosis will relate to presenting symptoms. Acute-onset fever, redness, parotid swelling, and elevated white blood cell count (WBC) usually signifies an infectious process (ie, sialadenitis) or obstructive process (ie, salivary stones). Although rare, benign and malignant tumors may present acutely via tumor obstruction of a drainage duct, which may get infected and expand rapidly.

Benign tumors may be asymptomatic for months to even decades. However, parotid cancers also lack symptoms 50% to 70% of the time. With the increasing utilization of imaging (eg, computed tomography [CT], ultrasound, MRI, PET) for unrelated indications, there has been an increasing incidence of parotid incidentalomas, tumors that would not have been found otherwise. This increasing incidence results in needle biopsy and/or surgery due to the risk of malignancy or growth of the tumor. A recent review of parotidectomies at the University of Wisconsin found an increase in the number of incidentalomas, up to 10.2% in the 2004 to 2013 period, from 4% in 1994 to 2003. The investigators also noted a significant decrease in the rate of malignancy found within incidentalomas versus clinically apparent masses.

Rapid growth or pain in a parotid mass may herald a malignant transformation (eg, in a known PA). Red flags, such as pain, facial paresis, soft tissue fixation, trismus, skin ulceration, lymphadenopathy, numbness, and weight loss, should heighten suspicion for malignancy. Facial paresis from Bell palsy has a fast onset and eventual resolution. A slow and worsening progression of facial nerve involvement with facial tics or spasms should raise concern for malignancy, although benign cases with facial paresis have been reported. Bilateral parotid involvement is more likely an inflammatory process (eg, mumps, Sjögren disease) than a synchronous neoplastic process. Because of the intraparotid and periparotid lymph nodes, a parotid mass may be metastatic from malignancies of the face, scalp, or even a distant site.

Grossly, benign superficial tumors are typically solid, mobile, and well circumscribed within the parotid gland. These tumors should be easily palpable. Deep tumors within the parapharyngeal space that medialize the oropharynx may be obvious with a deviated tonsil or soft palate.

Diagnostic work-up

This section focuses on the noninfected parotid mass or swelling.

Clinical evaluation should begin with obtaining a full history and physical examination. The location, size, extent, and features of the mass should be characterized. Visual intraoral examination should be performed in every suspected parotid mass. Fiberoptic examination and assessment for mucosal lesions is necessary if malignancy is suspected or deep lobe extension is evident. Additional imaging studies, such as CT, MRI and/or ultrasound, are often warranted. Ruling out malignancy—ideally in the preoperative setting—in each case of parotid mass is paramount.

Historically, surgery was considered for every parotid tumor. Advances in fine-needle aspiration biopsy (FNAB), radiographic imaging, and the well-described nature of tumors have trended evaluation toward preoperative imaging and needle biopsy. Diagnoses of cysts, stones, and lymphoma may preclude surgery; thus, accurate diagnosis and appropriate planning are essential. Usually, any parotid mass can be considered for CT or MRI. CT provides excellent resolution for evaluating tumor location, size, extent, and lymph nodes if malignancy is suspected. CT and MRI provide excellent anatomic information because of the gland’s high fat content. However, they are poor on their own for determining individual histologies and benign versus malignant features. Vascular lesions may necessitate imaging before biopsy or surgery. Ultrasound is another cheap and effective tool that is useful for delineating cystic versus solid masses and characterizing anatomy of superficial lobe tumors. Its primary shortcomings include poor visualization of the deep lobe (obscured by mandibular ramus) and poor visualization of the facial nerve. For tumors with benign clinical and cytologic features, no obvious deep lobe involvement, and benign features on ultrasound, additional imaging with MRI or CT may not be necessary. Otherwise, diagnosis is typically confirmed with surgery and histologic analysis.

FNAB is an accurate and inexpensive method with low complication rates for differentiating benign versus malignant lesions. It has a sensitivity and specificity of 80% and 97% and a positive predictive value (PPV) and negative predictive value (NPV) of 90% and 94%, respectively. For benign disease, Tryggvason and colleagues found a PPV and NPV for FNAB of 94.3% and 98.6%. Carcinoma ex-pleomorphic adenoma caused the most false negatives (n = 6). FNA’s pitfall is its nondiagnostic rate, estimated at 8%. However, given its ease and accuracy in benign disease, FNAB should be considered for every parotid mass before resection. Contraindications to FNAB include bleeding disorders and acute sialadenitis.

Image-guided FNAB may be necessary for deep, clinically nonpalpable tumors or those close to vital structures. Ultrasound or CT-guided core-needle biopsy (CNB) is a proposed alternative for tumors with a prior FNA that was nondiagnostic or could not be reached with other techniques. Compared to FNAB, it involves a larger needle and can be more painful. Because it extracts more tissue, it may be diagnostic for lesions such as lymphoma. Previously, clinicians avoided it out of fears for safety and tumor seeding. Recent studies suggest otherwise, citing it as a safe and more accurate preoperative alternative to the FNAB, especially for malignant tumors. A recent meta-analysis reported a sensitivity of 96% and specificity of 100% when comparing benign versus malignant lesions, with a nondiagnostic rate of 1.6% (versus 8.0% for FNAB). There were no reports of tumor seeding; the most significant adverse event was hematoma formation, at 1.6% per procedure. The investigators concluded CNB may be a reasonable alternative, especially at institutions with a high rate of FNAB inadequacy (up to 30%).

Sonoelastography is a newer technique using ultrasound that measures deformation of tissues after a mechanical force is applied (with the probe). It is under investigation as a diagnostic tool to distinguish between varying histologies of parotid tumors and even to look at malignancy versus benign tissue. PAs and malignant tumors tend to be stiffer than other benign pathologies, but larger studies are needed to validate this as a useful clinical tool. Nevertheless, a thorough history, examination, and work-up are necessary for all parotid masses.

A more detailed discussion of head and neck imaging for parotid lesions and the evaluation of parotid masses is discussed in chapters 2 and 3.