squamous cell carcinoma of the parotid gland: a population-based analysis of 2545 cases

A M E R I C A N J O U R N A L O F O T O L A R Y N G O L O G Y – H E A D A N D N E C K M E D I C I N E A N D S U R G E R Y X X ( 2 0 1 4 ) X X X – X X X

Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

ScienceDirectwww.e l sev i e r . com/ loca te /amjo to

Squamous cell carcinoma of the parotid gland: Apopulation-based analysis of 2545 cases☆,☆☆,★

Michael J. Pfisterer, MDa,1, Alejandro Vazquez, MDa,1, Leila J. Mady, PhDa,Mohemmed N. Khan, MDb, Soly Baredes, MD, FACSa, c, Jean Anderson Eloy, MD, FACSa, c, d,⁎a Department of Otolaryngology—Head and Neck Surgery, Rutgers New Jersey Medical School, Newark, NJ, USAb Department of Otolaryngology—Head and Neck Surgery, Mount Sinai School of Medicine, New York, NY, USAc Center for Skull Base and Pituitary Surgery, Neurological Institute of New Jersey, Rutgers New Jersey Medical School, Newark, NJ, USAd Department of Neurological Surgery, Rutgers New Jersey Medical School, Newark, NJ, USA

A R T I C L E I N F O

☆ Financial disclosures: None.☆☆ Conflicts of interest: None.★ Presented at the 2014 Combined Sections⁎ Corresponding author at: Department of Oto

8100, Newark, NJ 07103, USA. Tel.: +1 973 972E-mail address: Jean.anderson.eloy@gma

1 These authors contributed equally to this

http://dx.doi.org/10.1016/j.amjoto.2014.03.0030196-0709/© 2014 Elsevier Inc. All rights rese

Please cite this article as: Pfisterer MJ, e2545 cases, Am J Otolaryngol–Head and

A B S T R A C T

Article history:Received 23 January 2014

Purpose: Squamous cell carcinoma (SCC) of the parotid gland is an uncommon tumor, whichgenerally affects older patients. In this study, we explore various aspects of this entity usinga national population-based database.Methods: The Surveillance, Epidemiology, and End Results (SEER) registry was used toextract data on frequency, incidence, and disease-specific survival (DSS) from 1973 to 2009.Variables analyzed included age, gender, race, histologic grade, stage and treatment. Coxproportional hazards analysis was conducted.Results: A total of 2545 cases were identified. Parotid SCCwasmost common inmales (79.8%),whites (92.9%), and patients aged ≥75 years (51.4%). Incidence increased slightly over the pastthree decades (annual percent change 1.90%, p < 0.05). Overall 5-year DSS was 54.4%.Statistically significant poor prognostic factors included black race, age ≥75 years, tumor T3or greater, and higher clinical stage at diagnosis. Elective neck dissection (END) in patientsstagedN0wasassociatedwithhigherDSS (78.3%versus 51.1%,p < 0.0001). Theomissionof ENDwas associated with a three-fold greater hazard of death (hazard ratio 3.19, 95% confidenceinterval 1.53–7.26, p = 0.0016), regardless of whether or not radiotherapy was given.Conclusion: Parotid SCC is uncommon, and data on treatment decisions are limited. Ourstudy profiles the demographic, clinicopathologic, incidence, and survival features of thisentity. Perhaps most notably, our results support the practice of END of the N0 neck.

© 2014 Elsevier Inc. All rights reserved.

1. Introduction

Salivary gland neoplasms are relatively uncommon, with anoverall incidence of 2.5 to 3 cases per 100,000 persons in the

Meeting of the Triologiclaryngology–Head and Ne4588; fax: +1 973 972 376

il.com (J.A. Eloy).work.

rved.

t al, Squamous cell carcNeck Med and Surg (20

Western world [1]. The majority of these lesions (approxi-mately 70%–80%) originate in the parotid gland [1–3]. Malig-nant salivary gland tumors are considerably less common;they account for approximately 3% to 5% of all head and neck

al Society, Miami Beach, FL, on January 10–12, 2014.ck Surgery, Rutgers New Jersey Medical School, 90 Bergen St., Suite7.

inoma of the parotid gland: A population-based analysis of14), http://dx.doi.org/10.1016/j.amjoto.2014.03.003

http://dx.doi.org/10.1016/j.amjoto.2014.03.003

mailto:[email protected]



http://www.sciencedirect.com/science/journal/01960709


2 A M E R I C A N J O U R N A L O F O T O L A R Y N G O L O G Y – H E A D A N D N E C K M E D I C I N E A N D S U R G E R Y X X ( 2 0 1 4 ) X X X – X X X

cancers [1,2] and have a tendency to present in the sixth orseventh decades of life [4–6]. Squamous cell carcinoma (SCC)originating in the parotid gland is a rare, often aggressivemalignancy which makes up 0.1% to 3.4% of all parotidmalignancies [7–9]. Given its rarity, epidemiologic data anddata on the management of primary parotid SCC are limited.

Although surgical resection is the principal treatmentmodality for primary parotid gland malignancies, severalfactors influence management decisions, among themtumor size, extent, facial nerve function, and the presenceof metastatic disease [10,11]. One particular area of interestconcerns the management of the regional lymphatics. Theindications for therapeutic neck dissection in the presenceof regional metastasis are relatively well defined; however,although most experts recommend elective neck dissection(END) in the management of the clinically negative (N0)neck, evidence regarding its impact on survival is limited[12–14]. Occult metastases are reported in 12% to 48% ofpatients with all forms of parotid gland carcinoma; never-theless, data describing the impact of elective neck treat-ment (neck dissection or radiotherapy) on survival are ratherlimited [11,12,14–17].

In the present study, we use a national population-baseddatabase to investigate the demographic, clinicopathologic,incidence, and survival and prognostic features of primarySCC of the parotid gland. In addition, we aim to examine therole of neck dissection in the treatment of patients withoutclinically evident neck disease. To our knowledge, our studyrepresents the largest population-based series of primaryparotid SCC to date.

2. Materials and methods

Frequency, incidence and survival data were obtained fromthe SEER Program database using SEER*Stat 8.0.4 software(National Cancer Institute, Bethesda, MD). Frequency andsurvival data were obtained from the SEER18 dataset(consisting of 18 registries, covering the years 1973–2009).Incidence data were obtained from the SEER9 dataset (9registries), as it provided the greatest number of data points(1973–2009). IRB approval was not required, as SEER does notuncover sensitive patient information. Using the InternationalClassification of Diseases for Oncology, Third Edition (ICD-O-3)codes, we extracted data for the histologic subcategories of“SCC NOS [not otherwise specified]” (8070/3) and “SCCkeratinizing NOS” (8071/3) and considered them collectivelyas SCC. Next, we restricted our results to the parotid gland(C07.9).

Frequency and survival data were analyzed on the basis ofage, gender, race, histologic grade, stage at diagnosis, andtreatment (including surgery, radiotherapy, or both). Race wasdivided into white, black, and “other” (including AmericanIndian, Alaskan Native, Asian/Pacific Islander, unspecified, orunknown). Treatment was further analyzed by stage atdiagnosis. Contemporary TNM staging information (i.e., AJCC6th edition) [3] was available only for patients diagnosed in2004 and after. Information on treatment (surgery, radiation,or both) was obtained from SEER site-specific treatment codes.Information on neck dissection was derived using SEER's

Please cite this article as: Pfisterer MJ, et al, Squamous cell carc2545 cases, Am J Otolaryngol–Head and Neck Med and Surg (201

regional lymph node surgery codes and correlated with Nclassification. Incidence rate data were adjusted to the year2000 U.S. standard population and reported as incidence per100,000 persons. Net survival of the cohort was estimated byKaplan–Meier analysis (yielding disease-specific survival,DSS). Joinpoint Regression Program 4.0.1 (National CancerInstitute, Bethesda, MD) was used to calculate annualpercentage change (APC) in incidence and assess the statis-tical significance of the trends observed. Survival and hazardanalysis was carried out with JMP Statistical Discovery 10 (SASInstitute, Cary, NC). Kaplan–Meier survival curves werecompared using the log-rank test. Hazard analysis wasconducted using the multivariate Cox proportional hazardsmodel. Microsoft Office Excel 2007 (Microsoft Corporation,Redmond, WA) was used for additional data processing. Aprobability value (p-value) of <0.05 was considered statisti-cally significant for all tests.

3. Results

3.1. Analysis of demographic and clinicopathologic features

SEER18 was queried; 2545 cases were identified. The resultsare summarized in Tables 1 and 2. Males made up 79.8% ofcases; whites made up 92.9% of the cohort. A majority ofpatients were aged ≥75 years (51.4%), with the mean age atdiagnosis being 72.9 (±13.1) years. Most tumors were histo-logic grades II (27.0%) or III (36.4%). AJCC staging informationwas available for the years 2004–2009 (1081 patients; Table 2).Approximately 70% of lesions were staged T3 or greater atdiagnosis, whereas only 13.8% of cases were staged T1.Regional metastases (N1–3) were present in 40.4% of cases,but only 5% had distant metastases (M1).

3.2. Trends in incidence

Joinpoint regression analysis revealed a statistically signifi-cant increase in overall incidence from 1972 to 2009, with anAPC of 1.90% (p < 0.05) (Fig. 1). Average incidence for this timeperiod was 0.154 cases per 100,000 persons.

3.3. Survival and hazard analysis

Overall 5-year DSS was 54.4%. Statistically significant resultswere observed when race and age were examined (Table 1).DSS at 5 years was 54.8% for whites, 40.5% for blacks, and61.9% for “others” (p = 0.0048). Among the standardized agegroups, DSS was worst among those aged ≥75 years (48.4%)and best among those aged ≤44 years (68.0%) (p < 0.0001). Forthe years 2004–2009, AJCC stage group, T stage, and M stageshowed statistically significant differences in DSS (Table 2).

Hazard analysis revealed statistically significant worstprognosis for black race (HR 1.56, 95% CI 1.13–2.10, p = 0.0080relative to whites), age ≥75 years (HR 2.08, 95% CI 1.34–3.47, p =0.0007 relative toage≤44 years), andhistologic grade III (HR1.44,95% CI 1.04–2.04, p = 0.0278 relative to grade I) (Table 1).Differences were also identified for stage III and IV disease(relative to stage I); stages T3, T4a and T4b (relative to T1); andM1 stage (relative to M0) (Table 2).



Table 1 – Frequency, survival and hazard analysis according to various demographic and clinicopathologic factors forparotid squamous cell carcinoma.

Frequency 5-year survival Hazard analysis (Cox PH)

n % DSS p (log-rank) HR 95% CI p

Total cohort/overall 2545 100% 54.4% – – – –GenderMale 2030 79.8% 55.2% 0.0738 Ref. – –Female 515 20.2% 51.4% 1.17 0.98–1.40 0.0849

RaceWhite 2364 92.9% 54.8% 0.0048 ⁎ Ref. – –Black 98 3.9% 40.5% 1.56 1.13–2.10 0.0080 ⁎Other 56 2.2% 61.9% 0.83 0.52–1.25 0.3914

Agea

15–44 years 67 2.6% 68.0% <0.0001 ⁎ Ref. – –45–54 years 184 7.2% 65.7% 1.14 0.68–2.00 0.638055–64 years 372 14.6% 55.8% 1.47 0.92–2.50 0.113065–74 years 612 24.0% 57.8% 1.45 0.92–2.44 0.118175+ years 1,309 51.4% 48.4% 2.08 1.34–3.47 0.0007 ⁎

Age at diagnosis, mean (±SD) 72.9 (±13.1) yearsGradeI 197 7.7% 67.8% 0.1114 Ref. – –II 686 27.0% 56.2% 1.33 0.95–1.91 0.1032III 926 36.4% 54.5% 1.44 1.04–2.04 0.0278 ⁎IV 71 2.8% 66.3% 1.01 0.54–1.79 0.9811Unknown 665 26.1% – – – –

a A single patient in the 0–14-years age group was excluded.⁎ Indicates statistical significance (p < 0.05).

3A M E R I C A N J O U R N A L O F O T O L A R Y N G O L O G Y – H E A D A N D N E C K M E D I C I N E A N D S U R G E R Y X X ( 2 0 1 4 ) X X X – X X X

Survival and hazard analysis of treatment by stage was alsoconducted. For stages I–III, DSS was 75.9% with surgery alone,53.2%with radiotherapy (RT) alone, and 80.2%with surgery and

Table 2 – Frequency, survival and hazard analysis by AJCC 6carcinoma.

Frequency 5-year Sur

n % DSS p

AJCC stageI 91 8.4% 80.1%II 99 9.2% 72.5%III 232 21.5% 71.3%IV 463 42.8% 50.3%

Unknown 196 18.1% –T stagea

T1 149 13.8% 75.1%T2 177 16.4% 77.0%T3 243 22.5% 57.8%T4a 314 29.0% 54.0%T4b 197 18.2% 32.6%TX 149 13.8% –

N stageN0 505 46.7% 66.7%N1 247 22.8% 60.7%N2 170 15.7% 58.2%N3 20 1.9% 46.3%NX 132 12.2% –

M stageM0 897 83.0% 65.3%M1 54 5.0% 16.1%MX 130 12.0% –

a A single patient staged T0 was excluded.⁎ Indicates statistical significance (p < 0.05).


RT in combination (p = 0.0013). For stage IV disease, DSS was41.3% with surgery, 37.5% with RT, and 61.8% with bothcombined (p < 0.0001) (Table 3). The use of RT alone in earlier

th edition staging (2004–2009) for parotid squamous cell

vival Hazard analysis (Cox PH)

(log-rank) HR 95% CI p

<0.0001 ⁎ Ref. – –1.94 0.75–5.57 0.17362.66 1.20–7.07 0.0146 ⁎4.91 2.36–12.6 <0.0001 ⁎– – –

<0.0001 ⁎ Ref. – –1.44 0.71–3.03 0.31343.08 1.70–6.08 0.0001 ⁎3.54 1.98–6.92 <0.0001 ⁎7.79 4.03–16.0 <0.0001 ⁎– – –

0.1310 Ref. – –1.29 0.91–1.82 0.15571.18 0.78–1.73 0.42992.15 0.59–1.39 0.0623– – –

<0.0001 ⁎ Ref. – –4.71 3.03–7.05 <0.0001 ⁎– – –



Fig. 1 – Incidence trends from 1973-2009. APC, annualpercentage change.

Table 4 – Hazard analysis by treatment according to stagefor parotid squamous cell carcinoma.

Stages I–III (n = 255) Stage IV (n = 296)

HR 95% CI p HR 95% CI p

S Ref. – – Ref. – –RT 2.72 1.02–6.65 0.0462 ⁎ 1.29 0.73–2.28 0.3771S + RT 0.60 0.29–1.24 0.1637 0.49 0.31–0.81 0.0058 ⁎

S, surgery; RT, radiotherapy; S + RT, surgery and radiotherapy.⁎ Indicates statistical significance (p < 0.05).


stage disease (i.e., I–III) was a poor prognosticator (HR 2.72, 95%CI 1.02–6.65, p = 0.0462), whereas in stage IV disease combinedsurgery and RT were associated with a lower hazard of death(HR 0.49, 95% CI 0.31–0.81, p = 0.0058) (Table 4). These findingsare illustrated in Fig. 2.

In order to evaluate the roles of therapeutic neck dissection(TND) and elective neck dissection (END), the cohort wasdivided into patients with (N+) or without (N0) neck disease atdiagnosis (Table 5). Patients with N+ disease who underwentTND had DSS of 63.1%, whereas those that did not had DSS of41.0% (p < 0.0001). When patients treated with any form of RTwere excluded from the cohort, DSS changed to 60.6% (ND)and 31.7% (no ND) (p = 0.0009) (Fig. 3). The omission of TNDwas associated with a roughly three-fold greater hazard ofdeath in this group (Table 5).

Patients with N0 disease who underwent END had a DSS of78.2%, whereas those that did not have END had a DSS of51.1% (p < 0.0001). After excluding RT from the cohort, DSSchanged to 81.3% (END) and 50.7% (no END) (p = 0.0016) (Table 5)(Fig. 3). Similar to TND, the omission of END was associatedwith a roughly three-fold greater hazard of death in thisgroup (Table 5).

4. Discussion

Primary SCC of the parotid gland is a rare entity. Indeed, mostSCC affecting the parotid gland represents metastatic diseaserather than a primary salivary gland process [6], and it is

Table 3 – Survival analysis by treatment according tostage for parotid squamous cell carcinoma.

Stages I–III (n = 255) Stage IV (n = 296)

DSS p (log-rank) DSS p

S 75.9% 0.0013 ⁎ 41.3% <0.0001 ⁎RT 53.2%a 37.5%S + RT 80.2% 61.8%

S, surgery; RT, radiotherapy; S + RT, surgery and radiotherapy.a Insufficient data to calculate DSS beyond 49 months.⁎ Indicates statistical significance (p < 0.05).


estimated that 40% of these lesions are attributable to aprimary cutaneousmalignancy [18]. The rarity of true primaryparotid SCC, which comprises 0.1% to 3.4% of all parotidmalignancies, accounts for the limited data on the subject. Todate, there has only been a small collection of retrospectivereviews that address management and outcomes in primaryparotid SCC [7,8,19–23]. The sample sizes of these studies haveranged from 14 to 57 patients and account for a total of 230cases of primary SCC. By using data available from thepopulation-based SEER registry, we were able to compile alarge sample size, totaling 2545 cases of primary parotid SCC.To our knowledge, this study represents the largest series ofprimary parotid SCC to date.

The demographics of our cohort resemble those previouslyreported for parotid SCC. The vast majority of patients wereidentified as white, comprising 92.9% of the cohort. Thesignificance of this finding is not clear, but it is on par withpreviously reported demographics. A retrospective review by

Fig. 2 – Treatment of parotid squamous cell carcinoma bystage. DSS for treatment for stages I–III combined (A) andstage IV (B). S, surgery; RT, radiotherapy; S + RT, surgery andradiotherapy combined.



Table 5 – Survival and hazard analysis by whether or not neck dissection was performed.

N0 neck N+ neck

Survival Hazard analysis (Cox PH) Survival Hazard analysis (Cox PH)

DSS p (log-rank) HR 95% CI p DSS p HR 95% CI p

All patientsND 78.2% <0.0001 ⁎ Ref. – – 63.1% <0.0001 ⁎ Ref. – –No ND 51.1% 3.15 1.98-5.16 <0.0001 ⁎ 41.0% 2.82 1.72-4.48 <0.0001

RT excludedND 81.3% 0.0016 ⁎ Ref. – – 60.6% 0.0009 ⁎ Ref. – –No ND 50.7% 3.19 1.53-7.26 0.0016 ⁎ 31.7% 3.21 1.52-6.74 0.0026 ⁎

S, surgery; RT, radiotherapy; S + RT, surgery and radiotherapy⁎ Indicates statistical significance (p < 0.05).


Ying et al. [6], compiled 66 patients with parotid SCC during a21-year period, of which 65 were white. Males also composedthe majority of cases in this study, making up 79.8% of thecohort. Additionally, patients with parotid SCC tended to beolder than 75 years of age and with a mean age at diagnosis of72.9. The mean age previously reported has ranged between55 and 71.3 [6,23]. AJCC staging information was available for1081 patients reported to the SEER registry. Staging data werenot available for all 2545 patients, because SEER did not begincompiling data for staging until 2004. The majority of lesionswere stage T3 or greater at diagnosis (70%). Regional metas-tases were present in 40.4% of cases. Previously reportedregional lymph nodemetastases have ranged from 50% to 70%[12,15]. Distantmetastases were only present in 5% of patients

Fig. 3 – Elective and therapeutic neck dissection for parotidsquamous cell carcinoma. DSS for patients staged N0undergoing elective neck dissection (A) and patients withN+ disease undergoing therapeutic neck dissection (B). Allforms of radiotherapy are excluded. ND, neck dissection.


in the present study. In our analysis, we detected a slight (APC1.90%) but statistically significant increase in incidence overthe past three decades. The reason for this increasedincidence is not apparent.

The prognosis for parotid SCC is poor, with previouslyreported 5-year survival being less than 50% [8,10,11,20]. In aretrospective review of 12 cases, Lee et al. [23] have reported a5-year DSS of 31.3% for primary parotid SCC. Our results showa 5-year DSS of 54.4%. The survival analysis identified severalfactors that influenced survival rates among subgroups ofpatients. Black race had a statistically significant worse 5-yearDSS (40.5%, p = 0.0048) when compared to white (54.8%) andother (61.9%). The reason for this disparity is unclear, but maybe associated with socioeconomic factors known to be linkedto race in the US; many of these factors have been shown toplay a role in the survival of cancer patients [24–26]. Furtherstudies are required in order to evaluate this speculation.Survival was also negatively affected by age at diagnosisgreater than 75 (p < 0.0001), a finding consistent with the priorreport by Lee et al. [23], who demonstrated that age greaterthan 60 years was associated with a lower 5-year DSS. Inour study, an age greater than75was associatedwith a twofoldgreater hazard of death relative to patient age younger than 44.

A comparison of histologic grade found a 1.4-fold greaterhazard of death for histologic grade III relative to grade I (p =0.0278). As may be expected, survival was remarkably poor inthe setting of distant metastases (16% at 5 years). Moreover, Tstage appeared to have a major effect on survival, with T1 andT2 lesions having essentially equivalent DSS (75%–77%); T3and T4a having survival in the 54%–58% range; and T4b lesionshaving the lowest observed survival (32.6%). Although in-creasing N stage appeared to show a trend toward worsesurvival, this trend did not reach statistical significance.

In assessing the effect of primary treatment modality onsurvival, we separated our cohort into two groups: those withstage IV disease (n = 296), and those with stage I–III disease(n = 255). Stage IV includes not only patients with advanced T-stage disease, but also patients with distant metastases oradvanced nodal disease, all of which can confound ouranalysis if all stages were considered collectively. By separat-ing stage IV from the remainder of the cohort, we sought toattain a clearer understanding of the effects of therapy on lessadvanced disease. In stages I–III, we found that the use of RTalone was associated with significantly lower survival (53.2%)than surgery alone (75.9%) or surgery with RT (80.2%). Hazard




analysis, moreover, showed a greater than 2.5-fold hazard ofdeath in patients treated with RT only. Although DSS wasslightly greater in those treated with surgery and RT (80.2%,versus 75.9% for surgery alone), hazard analysis failed toreveal a statistically significant difference between the twomodalities. It should be noted that, although some statisti-cally significant results were observed in the stage IV group(e.g., an apparently protective effect for surgery with RT), theheterogeneity inherent to it significantly limits the conclu-sions that can be drawn from these data.

Several prior studies have sought to address the manage-ment of the clinically negative neck in patients with known orsuspected primary parotid carcinoma (various histologies)[12–14]. Conclusions drawn by these prior studies havefavored ENDs when the pathology is a high-grade carcinoma.However, the rarity of primary parotid SCC has precluded aprecise assessment of what effect neck dissection has onsurvival. For instance, one of the larger studies is a retrospec-tive analysis of 83 patients with primary parotid carcinomaand a clinically negative neck (N0) [12]. Only 6 of the 83patients had SCC. So although trends in treatments andoutcomes had a tendency to approach statistical significance,the study is underpowered when attempting to draw conclu-sions regarding the treatment of parotid SCC.

In our study, we sought to investigate the role of END bydividing the cohort into patients with clinically positive (N+)or negative (N0) neck disease at diagnosis. As expected,patients who did not receive a neck dissection when therewas positive neck disease had a 2.8-fold increased hazard ofdeath compared to patients that underwent a TND. Whenpatients treated with any form of RT were excluded from thecohort, the hazard ratio increased to 3.1. Regarding patientswith N0 disease, those who underwent END had a DSS of78.2%. In comparison, patient with N0 disease who did notundergo END had a DSS of 51.1%. The hazard of death was 3times greater for patient with a clinically negative neck thatdid not have END. This finding remains unaffected by theexclusion of patients who were treated with RT. Thesefindings may help to delineate and solidify the role of ENDin the treatment of primary parotid SCC.

The current recommendations regarding management ofthe neck offered by Zbären et al. [12] is that routine ENDshould be offered to all patients with primary carcinoma of theparotid gland. Kelley and Spiro [15] temper this opinion byfurther defining high-risk pathology and conclude that anEND is appropriate for patients who present with facial nerveparalysis, high tumor grade, older age, perilymphatic inva-sion, and extraparotid tumor extension. High-grade tumorswere defined as SCC, mucoepidermoid carcinoma, adenoidcystic carcinoma, acinic cell carcinoma, adenocarcinoma andmalignant mixed tumors.

Several limitations should be noted in our study. First,registry-based data are subject to informational biases:without access to the original clinical records, accuracy ofthe data is wholly dependent on the registration/codificationprocess, and it is difficult to control for certain aspects ofthe data. For instance, inherent to the diagnosis of parotidSCC is the determination of whether the disease in questionrepresents a primary neoplasm or metastatic disease (e.g.,from a cutaneous primary malignancy); similarly, it is


impossible to independently confirm the accuracy of thehistologic diagnosis without access to primary patient data.Second, the absence of detailed treatment information (e.g.,classification of neck dissection, use of chemotherapy,radiation dose/regimen) limits our conclusions regardingtreatment. Treatment bias is likely to affect the interpreta-tion of outcome and survival data because patients withworse prognosis at the onset were most likely to havereceived more aggressive treatment. In order to address thisbias, a randomized trial would be needed. Another limitationof the SEER data is the absence of information regarding theuse of chemotherapy, an adjuvant treatment modality oftenused in conjunction with radiotherapy in advanced salivarygland malignancies.

5. Conclusion

Parotid SCC is uncommon, and data on which to basetreatment decisions are limited. This retrospective review of2545 cases of primary parotid SCC represents the largest studyon the subject to date. Our study profiles the demographic,clinicopathologic, incidence, and survival features of thisentity using a national, standardized patient population.Perhaps most notably, our results support the practice ofroutine END in patients with primary parotid SCC and aclinically negative neck.

R E F E R E N C E S

[1] Speight PM, Barrett AW. Salivary gland tumours. Oral Dis2002;8:229–40.

[2] DeVita VT, Lawrence TS, Rosenberg SA. DeVita, Hellman, andRosenberg's cancer: principles & practice of oncology.Philadelphia: Wolters Kluwer Health/Lippincott Williams &Wilkins; 2011.

[3] Edge SB, American Joint Committee on Cancer. AJCC cancerstaging manual. New York: Springer; 2010.

[4] Ellis GL, Auclair P. Tumors of the salivary glands. WashingtonDC: American Registry of Pathology; 1996.

[5] Wahlberg P, Anderson H, Biorklund A, et al. Carcinoma of theparotid and submandibular glands—a study of survival in2465 patients. Oral Oncol 2002;38:706–13.

[6] Ying YL, Johnson JT, Myers EN. Squamous cell carcinoma ofthe parotid gland. Head Neck 2006;28:626–32.

[7] FlynnMB, Maguire S, Martinez S, et al. Primary squamous cellcarcinoma of the parotid gland: the importance of correcthistological diagnosis. Ann Surg Oncol 1999;6:768–70.

[8] Batsakis JG, McClatchey KD, Johns M, et al. Primary squamouscell carcinoma of the parotid gland. Arch Otolaryngol1976;102:355–7.

[9] Taxy JB. Squamous carcinoma in a major salivary gland: areview of the diagnostic considerations. Arch Pathol Lab Med2001;125:740–5.

[10] Woods JE, Chong GC, Beahrs OH. Experience with 1,360primary parotid tumors. Am J Surg 1975;130:460–2.

[11] Armstrong JG, Harrison LB, Thaler HT, et al. The indicationsfor elective treatment of the neck in cancer of the majorsalivary glands. Cancer 1992;69:615–9.

[12] Zbären P, Schupbach J, Nuyens M, et al. Elective neckdissection versus observation in primary parotid carcinoma.Otolaryngol Head Neck Surg 2005;132:387–91.


http://refhub.elsevier.com/S0196-0709(14)00058-1/rf0005


































[13] Medina JE. Neck dissection in the treatment of cancer ofmajor salivary glands. Otolaryngol Clin North Am1998;31:815–22.

[14] Regis De Brito Santos I, Kowalski LP, Cavalcante De Araujo V,et al. Multivariate analysis of risk factors for neck metastasesin surgically treated parotid carcinomas. Arch OtolaryngolHead Neck Surg 2001;127:56–60.

[15] Kelley DJ, Spiro RH. Management of the neck in parotidcarcinoma. Am J Surg 1996;172:695–7.

[16] Frankenthaler RA, Byers RM, Luna MA, et al. Predicting occultlymph node metastasis in parotid cancer. Arch OtolaryngolHead Neck Surg 1993;119:517–20.

[17] Bron LP, Traynor SJ, McNeil EB, et al. Primary and metastaticcancer of the parotid: comparison of clinical behavior in232 cases. Laryngoscope 2003;113:1070–5.

[18] Guillamondegui OM. Salivary gland cancers, surgery, andirradiation therapy. Arch Otolaryngol 1982;108:709.

[19] Gaughan RK, Olsen KD, Lewis JE. Primary squamous cellcarcinoma of the parotid gland. Arch Otolaryngol Head NeckSurg 1992;118:798–801.


[20] Sterman BM, Kraus DH, Sebek BA, et al. Primary squamouscell carcinoma of the parotid gland. Laryngoscope1990;100:146–8.

[21] Marks MW, Ryan RF, Litwin MS, et al. Squamous cellcarcinoma of the parotid gland. Plast Reconstr Surg1987;79:550–4.

[22] Shemen LJ, Huvos AG, Spiro RH. Squamous cell carcinoma ofsalivary gland origin. Head Neck Surg 1987;9:235–40.

[23] Lee S, Kim GE, Park CS, et al. Primary squamous cellcarcinoma of the parotid gland. Am J Otolaryngol2001;22:400–6.

[24] Konski A, Berkey BA, Kian Ang K, et al. Effect of educationlevel on outcome of patients treated on Radiation TherapyOncology Group Protocol 90-03. Cancer 2003;98:1497–503.

[25] Chen AY, Schrag NM, Halpern MT, et al. The impact of healthinsurance status on stage at diagnosis of oropharyngealcancer. Cancer 2007;110:395–402.

[26] Kwok J, Langevin SM, Argiris A, et al. The impact of healthinsurance status on the survival of patients with head andneck cancer. Cancer 2010;116:476–85.











































squamous cell carcinoma of the parotid gland: a population-based analysis of 2545 cases

Documents