Учебники / Textbook and Color Atlas of Salivary Gland Pathology - DIAGNOSIS AND MANAGEMENT Carlson 2008

186 Tumors of the Parotid Gland

BENIGN TUMORS

Pleomorphic Adenomas

The PA is the commonest benign salivary gland tumor and the commonest salivary tumor overall, although it is comparatively rare in young children. It is slow growing and can reach giant proportions if neglected, and there is a 2–4% malignant change. PAs will recur if the tumor is inadequately removed. Although PAs have a pseudo-capsule of compressed fibrous tissue, the buds and pseudopodia from the tumor involve the capsule so that simple enucleation will leave tumor remnants and lead to multifocal recurrence. The concept of whether the capsule is incomplete and whether pseudopodia of the tumor involves the parotid tissue is currently being questioned, and with it the need for complete superficial parotidectomy. Although parotidectomy is designed to remove PAs with a cuff or margin of normal tissue to prevent recurrence, the tumor’s proximity to the facial nerve frequently means that the dissection at some points leaves no tissue around the capsule. In a recent histologic analysis of the capsular form in PAs, 81% showed capsular exposure following parotidectomy or submandibular gland excision (Webb and Eveson 2001) (Figure 8.4f). This paper also showed 57% bosselations, 33% enveloping of the capsule, with 42% microinvasion and 12% “tumor buds” in the capsule; and large >25 mm hypocellular tumors had thinner capsules, possibly easier to rupture at surgery. The article suggested that a minimum of 1 mm of normal tissue around PAs was required as a margin. However, in an article reviewing 475 PAs of the superficial lobe of the parotid, 380 treated by extracapsular dissection and 95 by superficial parotidectomy, there was no difference in recurrence rate or permanent facial nerve palsy (McGurk et al. 1996). These surgeons postulated that tumor buds or micro-invasion into the capsule had little significance and that extracapsular dissection could be done safely. In 1999 a series of 59 partial parotidectomies with selective nerve dissection for benign and low-grade malignant tumors reported a zero incidence of permanent facial nerve paralysis or paresis and zero recurrence (Witt 1999). Although Witt in a later paper confirmed that capsular exposure occurred in virtually all types of parotid surgery, and could find no difference in recurrence, capsular rupture, tumor-facial nerve interface, and permanent facial palsy between total parotidectomy, superficial

parotidectomy, and extracapsular dissection, he recommended against minimum margin resection in extracapsular dissection (Witt 2002). Further evidence for extracapsular dissection is provided by a series of 83 cases in which the overall recurrence rate was 6%, but 17.6% when the tumor itself was at the margins; however, cases with margins of <1 mm had a recurrence of only 1.8% (Ghosh et al. 2003). Ghosh et al. also reported that intraoperative capsular rupture and microscopic invasion of the capsule had no influence on recurrence, suggesting that a fraction of a millimeter of normal tissue was an adequate margin and that only tumors that actually involved the margin were at risk for recurrence. These authors recommend that preservation of vital structures is a more important consideration than preserving a cuff of normal tissue.

In contrast, Piekarski, Nejc, and Szymczak et al. (2004) found a recurrence rate of 8.2% and an unacceptable rate of complications with extracapsular dissection and did not recommend the technique as too “technically demanding.” In a separate publication with 213 patients who were operated for pleomorphic adenoma of the parotid, 5 of 9 primary tumors (56%) that recurred were found to have pseudopodia extending outside the capsule on histologic review. This was statistically higher than the examined cases that did not recur (8%), and the authors concluded that pseudopodia extending outside the capsule was a significant risk for recurrence (Henriksson et al. 1998). Interestingly, in the same study only 2 of 28 cases that ruptured during surgical removal recurred (7.1%), which was not significantly different than the 4.1% recurrence rate for the tumors that had no rupture. A further cautionary note is raised by the histologic analysis of Zbaren and Stauffer (2007), in which 160 of 218 (73%) of PAs were found to have adverse capsular characteristics, 33% with an incomplete capsule and 13% with satellite nodules. These were most frequently seen in the stroma rich myxoid subtype. Similar findings with stroma rich PAs showing 71% of focal absence of a capsule and 33% of satellite nodules have been reported with recommendations against local dissection (Stennert et al. 2004.)

It does not appear that extracapsular dissection is just a “euphemism for enucleation” as some have claimed, as recurrence rates are comparable to parotidectomy and most papers show lower morbidity. The exact margin required for complete

Tumors of the Parotid Gland 187

removal of PA remains controversial. A criticism of extracapsular dissection has been that even if this technique is suitable for a presumed benign PA, what should the surgeon do if the final histopathologic diagnosis turns out to be malignant? In a review of 662 clinically benign parotid tumors, 503 treated by extracapsular dissection and 159 by superficial parotidectomy, 5% were malignant and there was no difference in 5- or 10-year survival or recurrence rates between the malignant tumors in the two surgical groups, although morbidity was significantly lower in the extracapsular dissection group (McGurk, Thomas, and Renehan 2003).

A superficial lobe parotid tumor clinically benign and diagnosed as a PA on FNAB may also be treated with a limited superficial parotidectomy (without complete dissection of the facial nerve) and may not require a complete superficial parotidectomy for cure (O’Brien 2003). This is probably most commonly undertaken for tumors in the parotid tail (Figure 8.9).

Deep lobe PAs are usually larger and frequently will have less surrounding parotid tissue, especially deeply, where they abut the prevertebral muscles of the neck. However, the inability to obtain a surrounding cuff of parotid does not seem

b

a

c

Figures 8.9a, 8.9b, and 8.9c. MR images of large cystic benign tumor in the parotid tail.

188 Tumors of the Parotid Gland

Figure 8.9d. Following partial parotidectomy the two arrows point to the dissected cervico-mandibular branch of the facial nerve, the parotid tail tumor having been resected inferior to this nerve branch. The upper branches of the facial nerve have not been dissected in this case.

Figure 8.9e. The long arrow points to the resected parotid tail region, while the shorter arrow points to the remaining superfi cial parotid lobe preserved intact.

to lead to increased recurrence. Harney, Murphy, and Hone et al. (2003) found that the capsules of deep lobe tumors were significantly thicker and that there was less extracapsular extension of tumor in the deep lobe tumors (58% vs. 79%), which may explain this phenomenon.

If the capsule of the tumor is ruptured during surgery, then recurrence is not inevitable and perhaps liberal irrigation with sterile water followed by normal saline may be tumoricidal (Webb and Eveson 2001). When recurrence occurs it is frequently multinodular (Figure 8.10) and requires

Figure 8.9f. Surgical specimen.

more radical en bloc surgery with excision of the previous scar, muscle, overlying skin, and facial nerve if they are involved. Maxwell, Hall, and Freeman (2004) in a retrospective study of 35 patients treated with surgery alone found a locoregional control of 77% with a malignant transformation of 5.7%. In a separate study of 42 cases of multinodular recurrence (6 with prior radiation), there were 2 patients with malignant transformation who died of distant metastases. Twelve patients had subtotal parotidectomy, 25 total parotidectomy, 5 subtotal petrosectomy; 14 had facial nerve resection. Seven patients of 36 who were followed developed further recurrences (19.4%), all of whom had only undergone subtotal parotidectomy (Leonetti et al. 2005). In a further series of 33 patients, 73% multifocal, 9% with malignant transformation, treated surgically, 6 (18%) recurred at an average of 9 years, and 23% of patients with initial enucleation and 14% with initial superficial parotidectomy had permanent partial facial nerve injury (Zbaren Tschumi et al. 2006). Renehan, Gleave, and McGurk (1996) reviewed 144 cases of recurrent PAs and suggested a role for radiation in multinodular cases. A recent paper of 34 cases of

Figure 8.10a. Two large nodules of recurrent pleomorphic adenoma exist in the left parotid gland.

Figure 8.10b. Two smaller nodules are seen in the parotid tail.

Tumors of the Parotid Gland 189

recurrent PAs with radiation therapy post–gross resection shows a 20-year actuarial control rate of 94% (Chen et al. 2006)

Warthin’s Tumor

This is the second commonest benign tumor of the parotid. If it is diagnosed when small and asymptomatic it may not require treatment in an old or infirmed patient. There is a 12% incidence of multiple ipsilateral or bilateral tumors. There appears to be a link to heavy smoking and bilateral Warthin’s tumors (Klussman, Wittekindt, and Preuss et al. 2006). Eight percent of these tumors occur in extra-parotid cervical lymph nodes and may be found at the time of parotidectomy or serendipidously in neck dissection specimens. Treatment is as for PA. Warthin’s tumors have a tendency to occur in the parotid tail, where the majority of parotid lymph nodes occur, so partial parotidectomy is often all that is required.

MALIGNANT TUMORS

Principles of Management of Parotid Carcinoma

There is no universally agreed method for managing parotid cancer; however, prognosis and management are related to two variables: the histologic classification/grade of the tumor and the staging. In reviewing 2,465 patients with carcinoma of parotid and submandibular glands, Wahlberg, Anderson, and Bjorklund et al. (2002) found 10-year survival of 88% for acinic cell carcinoma, 80% for mucoepidermoid carcinoma (MEC), and 74% for adenoid cystic carcinoma (ACC), but only 55% for adenocarcinoma unspecified and 44% for undifferentiated carcinoma. It should be noted that 5-year survival figures for ACC will give an artificially high value, as late local recurrence and distant metastasis continue over a 20+ year period (Chen, Garcia, Bucci et al. 2006; Lima, Tavares, Dias et al. 2005; Longuet, Nallet, Guedon et al. 2001). Harbo, Bungaard, and Pederson et al. (2002) also found acinic cell carcinoma to have the best 10-year survival, but in their Cox hazard regression analysis found T stage, N stage, M stage, and histologic differentiation to be significant in predicting prognosis and recommended the use of both staging and histologic diagnosis to assess prognosis. In other series significant factors include extraglandular extension, aggressive histology,

190 Tumors of the Parotid Gland

and nodal disease (Bhattacharyya and Fried 2005), histologic grade, T stage, N stage, and facial nerve dysfunction (Lima et al. 2005), and N stage and perineural involvement (Hocwald, Korkmaz, and Yoo et al. 2001).

The only other predictor of adverse prognosis reported in several series was advancing age (Bhattacharyya and Fried 2005; Lima et al. 2005; Kirkbridge, Liu, and O’Sullivan et al. 2001).

Figure 8.11a. CT scan of large superficial lobe tumor in 22-year-old woman.

The reported survival related to stage varies between authors, which may reflect differences in therapy as well as different patterns of histopathology. Luukkaa, Klemi, and Leivo et al. (2005) found

Figure 8.11b. Clinical appearance; the tumor has no signs of malignancy.

Figure 8.11c. Complete parotidectomy. The facial nerve was not involved.

Figure 8.11d. Histopathology of acinic cell carcinoma. The patient is alive and well 8+ years postoperatively.

5-year survival Stage I–IV 78%, 25%, 21%, and 23%, while Lima et al. (2005) found 10-year disease-specific survival Stage I–IV 97%, 81%, 56%, and 20%.

In considering management, Kaplan and Johns (1986) divide parotid cancers into 4 groups to recommend treatment. Group I T1-2 low-grade tumors are treated by parotidectomy with preservation of the facial nerve (Figures 8.11 and 8.12). Group II T1-2 high-grade are treated with parotidectomy plus first echelon node removal and postoperative radiation therapy (RT) (Figure 8.13). Group III T3 tumors, any positive nodes and recurrent tumor not in Group IV, are treated with radical parotidectomy with sacrifice of the facial nerve if necessary and radical neck dissection plus RT. Group IV includes T4 and tumors with significant local extension; they are treated by radical paroti-

Tumors of the Parotid Gland 191

Figure 8.12b. Operative image shows Blair incision incorporating 1 cm skin margin surrounding the biopsy.

Figure 8.12c. The level II nodes (fi rst echelon nodes) will be taken in continuity in this case.

Figure 8.12a. Patient who had a “skin cyst” biopsied; it was histologically a low-grade mucoepidermoid carcinoma of the parotid. Note preauricular biopsy scar.

Figure 8.12d. Histology shows a focus of mucoepidermoid carcinoma (arrow) in the biopsy scar between the skin and parotid, demonstrating the importance of excising “seeded” skin.

192 Tumors of the Parotid Gland

Figure 8.12e. Mucicarmine stain confirms intracellular mucus.

Figure 8.12f. Surgical specimen.	Figure 8.12g. The patient is alive and disease-free 14
	years postoperatively.

dectomy plus skin, muscle, and bone as indicated with radical neck dissection and postoperative RT (see chapter 11, Figure 11.16).

Controversy exists in the exact indications for RT, neck dissection, and facial nerve sacrifice. The majority of recent papers do show that RT is indicated for advanced parotid carcinoma and confers a survival benefit (Bhattacharyya and Fried 2005) or longer disease-free survival (Hocwald, Korkmaz, and Yoo et al. 2001). However, there is a move toward suggesting RT for earlier stage disease. Zbaren, Nuyens, and Caversaccio et al. (2006) retrospectively analyzed T1-2 carcinomas with and without postoperative RT and found local recurrence rates of 3% and 33%, respectively, and actuarial and disease-free survival of 93% and 92% with and 83% and 70% without RT. In an earlier publication from the same unit RT was suggested

not just for high-grade tumors but low-grade T2-4 (Zbaren, Schupbach, and Nuyens et al. 2003). So perhaps RT is indicated for earlier stage disease than was previously recommended. The latest data regarding fast neutron therapy in the management of advanced salivary cancer with gross residual disease shows a 6-year local-regional control of 59%, and 100% with no evidence of gross residual disease (Douglas et al. 2003.) Benefits of chemotherapy have not been clearly demonstrated for parotid cancer.

Similarly, although lymph node dissection was recommended for positive nodes and highgrade tumors, there is an increasing interest in the N0 neck. Occult metastasis rates of 22–45% led Stennert, Kisner, and Jungehuelsing et al. (2003) and Zbaren, Schupbach, and Nuyens et al. (2003) to recommend an elective neck dissection in the

Tumors of the Parotid Gland 193

Figure 8.13a. MR of parotid tumor that radiologically was diagnosed as a pleomorphic adenoma, but fine needle aspiration biopsy showed adenoid cystic carcinoma.

Figure 8.13b. In view of the cytologic diagnosis, proposed treatment includes resection of overlying skin (which clinically was felt to be tethered) and level II cervical nodes.

N0 neck for parotid cancer (Figure 8.14). Elective neck dissection for high-grade tumors and >T2 low-grade tumors should encompass levels I–III and upper V (Teymoortash and Werner 2002). In comparing elective neck dissection for the N0 neck against observation, Zbaren et al. (2005) found an

Figure 8.13c. Surgical specimen shows parotid with skin. The arrows show where level II nodes and fat are in continuity with the parotid tail. Final diagnosis was cellular pleomorphic adenoma with the FNAB diagnosis being a false positive.

actuarial and disease-free survival of 80% and 86% for the elective neck dissection patients vs. 83% and 69% for the observation group in a retrospective study.

Regarding the facial nerve, Spiro and Spiro (2003) recommend preservation unless the nerve is adherent to/embedded in the tumor. They feel that close margins to the nerve can be treated successfully by RT. This view is supported by the work of Carinci, Farina, and Pelucchi et al. (2001), who found that sacrifice of the nerve was not always able to improve survival rate. In a series of 107 patients with parotid cancer, 91 had normal nerve function preoperatively and facial nerve preservation was possible in 79 patients. The 5- year disease-free rate and 5- and 10-year survival rates were 65%, 83%, and 54% in the preserved nerve cohort and 56%, 62%, and 42% in the patients with nerve sacrifice. The authors felt that

Figure 8.14a. High-grade parotid malignancy with skin fi xation and N0 neck. Incision modified to include skin excision (arrow), and cervical incision extended to allow supraomohyoid neck dissection.

Figure 8.14b. The skin fl aps developed with the skin overlying the tumor left on the parotid gland.

Figure 8.14c. The selective neck dissection is complete with the specimen in continuity with the parotid gland. The facial nerve trunk has been exposed and the superficial parotidectomy is being performed.

Figure 8.14d. The surgical site following parotidectomy and selective neck dissection.

Figure 8.14e. The surgical specimen with the parotid superiorly and level I nodes pinned out with white pins.

194

preservation of the facial nerve by careful dissection gave favorable oncologic results (GuntinasLichius, Klussman, and Schroeder et al. 2004). Finally, a disease-free survival in patients with normal, partially, and completely impaired facial nerve function preoperatively of 69%, 37%, and 13% despite the use of facial nerve sacrifice and postoperative RT indicates what a poor prognosis invasion of the nerve confers (Terhaard, Lubsen, and Tan et al. 2006).

In specific histologic tumor types variable results for different treatments have been reported. Mucoepidermoid carcinoma is the commonest salivary malignancy and most cases are fortunately low or intermediate grade. A series of 89 cases at the Mayo Clinic, 69 T1-2, 85 N0, and 83 low/intermediate grade, were treated by parotidectomy with “appropriate” neck dissection and only 7 had RT. Kaplan-Meier estimated cancer-specific survival rates at 5, 15, and 25 years were 98.9%, 97.4%, and 97.4% (Boahene, Olsen, and Lewis et al. 2004). Using a point grading system for histopathologic features in a series of 234 mucoepidermoid carcinomas of the major salivary glands, cystic component <20%, 4 or more mitotic figures per 10 high power fields, neural involvement, necrosis, and anaplasia were found to have prognostic significance for parotid MEC (Goode, Auclair, and Ellis 1998). Intermediate-grade MEC tends to behave more like low-grade MEC, while high-grade MEC behaves aggressively with local recurrence and regional and distant metastases in the majority of cases. Other low-grade tumors such as acinic cell carcinoma, epimyoepithelial carcinoma, and low-grade adenocarcinoma all can be treated like low-grade MEC. Polymorphous low-grade adenocarcinoma is rare in the major glands, being seen mostly in the minor salivary glands of the oral cavity.

On the other hand, results for high-grade tumors such as primary squamous carcinoma of the parotid are poor; in one published series twothirds were treated with radical surgery and RT and one-third with RT alone, but 5-year actuarial survival and disease-free survival was 31% and 33%, respectively (Lee, Kim, and Parks et al. 2001). Malignant change in PAs is most commonly seen as carcinoma ex-pleomorphic adenoma, and prognosis will depend on the histologic type of malignancy and whether the malignancy has spread outside the capsule. In carcinoma ex-pleo- morphic adenoma the use of postoperative RT improved 5-year local control from 49% to 75%

Tumors of the Parotid Gland 195

and improved survival in patients without cervical metastasis (Chen, Garcia, and Bucci et al. 2007). Two other forms of malignant PA occur, both rare: the “true” malignant mixed tumor or carcinosarcoma where malignant change is seen in both the epithelial and myoepithelial component of the PA, and the benign metastasizing PA, which as its name suggests retains a benign histologic appearance despite the presence of metastases.

It is hard to interpret survival figures in some series, as ACC is very slow growing and 5-year survival is less meaningful in this neoplasm as survival continues to fall on 20-year follow-up. Thus in series with short follow-up ACC will erroneously be thought to have a good prognosis. Typical long-term survival figures are 84.3% 2 year, 75.9% 5 year, 50.49% 10 year, and 20.11% after 20 years (Issing, Hemmanouil, and Wilkens et al. 2002). The type of histologic appearance, solid vs. cylindrical, and the presence of perineural invasion are important prognostic factors. Even with documented lung metastases patients can live 5+ years, the average survival between the appearance of lung metastases and death being 32.3 months in one series (van der Waal et al. 2002). Wide field adjuvant RT post–radical surgery is usually recommended for ACC.

The histologic grade of the tumor must be taken into account as well as TNM staging when interpreting survival results in reported series. Every parotid cancer will be unique and the decision for what is the correct surgery will be made on an individual basis for each patient.

Summary

•80% of parotid tumors are benign and most commonly pleomorphic adenomas.

•Less than one-third of malignant tumors will have obvious clinical signs of malignancy, for example, facial nerve palsy, ulceration, fixation, or lymphadenopathy.

•Routine use of CT or MR imaging does not appear justified and should be used selectively for malignant neoplasms and deep lobe tumors.

•Preoperative open biopsy is contraindicated and FNAB is the modality of choice for preoperative cytologic diagnosis.