SEARCH

SEARCH BY CITATION

Keywords:

  • adenoid cystic carcinoma;
  • Surveillance, Epidemiology, and End Results;
  • epidemiology;
  • site;
  • survival

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

BACKGROUND:

The scientific literature to date lacks population-based studies on the demographics, clinical features, and survival of patients with adenoid cystic carcinoma (ACC) of different anatomic sites.

METHODS:

The authors identified 5349 patients who had ACC of the major salivary glands (N = 1850), minor salivary glands (N = 2077), breast (N = 696), skin (N = 291), lung and bronchus (N = 203), female genital system (N = 132), and eye and orbit (N = 100) from the Surveillance, Epidemiology, and End Results (SEER) registry. Differences in demographics, clinical features, and survival of patients were assessed.

RESULTS:

ACC of the eye and orbit was associated with younger age at presentation (mean age,49.9 years). ACC of the skin or breast tended to present with less aggressive prognostic features, whereas ACC of the lung and bronchus or eye and orbit tended to present with more aggressive prognostic features. In a multivariate survival analysis of patients who presented with localized disease, patients with ACC of the breast (hazard ratio [HR], 0.40) or skin (HR, 0.40) had a significantly lower risk death than patients with ACC of the major salivary glands; whereas patients with ACC of the lung and bronchus (HR, 3.72) or the eye and orbit (HR, 3.67) had a significantly higher risk. For patients who presented with regional disease, the only clear prognostic difference in multivariate analysis was that patients with ACC of skin fared significantly better.

CONCLUSIONS:

The demographics and clinical features of patients with ACC differ by disease site. The current results indicated that site may be an important predictor of survival for patients who present with localized disease but is less important for patients who present with regional disease. Cancer 2012. © 2011 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Adenoid cystic carcinoma (ACC) is a relatively rare malignancy that arises mostly in major and minor salivary glands. ACC accounts for only 1% of all head and neck malignancies1 and for approximately 25% of all salivary gland carcinomas.2 ACC also occurs in other glandular areas, including breast, lung, cervix, skin, and lacrimal glands; and various other sites also have been reported. ACC is a slow-growing but highly invasive malignancy with a propensity for recurrence and delayed onset of distant metastasis. The prolonged clinical course of ACC is well known, and patients may survive for years even after recurrence. However, little is understood about the etiology of ACC.3

It is noteworthy that, although ACCs at different anatomic sites share common histologic features and expression of the proto-oncogene c-KIT (v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog),4 the clinical features and outcomes of ACC appear to vary by site. For example, in large ACC studies,5, 6 96% of ACCs arising in the breast were localized, whereas only 57% of ACCs arising in the head and neck were localized. Distant metastases and perineural invasion are rare for ACCs of the breast, whereas metastases and perineural invasion are common for ACCs of the salivary glands.4 The reported 10-year survival rates for patients with ACCs of the breast, major salivary glands, and lung are >90%,4 71%,6 and 39%,7 respectively.

Because ACC is an uncommon malignancy, most published studies of ACC have been case reports or clinical series with sample sizes of fewer than 70 patients.8-10 There are a few published population-based studies of ACC5, 11, 12; however, those studies investigated only the clinical behaviors of ACCs at a single anatomic site and did not include comparisons between different sites. Therefore, we conducted a population-based study using data from the Surveillance, Epidemiology, and End Results (SEER) Program to determine the demographics, clinical features, and survival of patients with ACC at different anatomic sites.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Patients

The SEER Program databases contain information on cancer incidence and survival from specific geographic areas across the United States. We used the SEER 17 registries, which includes 28% of the US population and covers Atlanta, Connecticut, Detroit, Hawaii, Iowa, New Mexico, San Francisco-Oakland, Seattle-Puget Sound, Utah, Los Angeles, San Jose-Monterey, Rural Georgia, Alaska, Greater California, Kentucky, Louisiana, and New Jersey. Because the SEER Program developed from 9 registries in 1973 to 17 registries in 2000, not all registries have contributed cases since 1973. Atlanta, Connecticut, Detroit, Hawaii, Iowa, New Mexico, and San Francisco-Oakland were the original registries and contributed cases for the entire period from 1973 through 2008; whereas Seattle-Puget Sound and Atlanta contributed cases beginning in 1974 and 1975, respectively. Los Angeles, San Jose-Monterey, Rural Georgia, and the Alaska Native Tumor Registry contributed cases beginning in 1992; and Greater California, Kentucky, Louisiana, and New Jersey contributed cases beginning in 2000. The data were extracted using SEER*Stat software, version 7.0.4 (National Cancer Institute, Bethesda, Md) from the 1973 to 2008 databases, which were released in April 2011.

On the basis of the International Classification of Diseases for Oncology, 3rd edition (ICD-O-3) codes provided in the SEER database, we identified all cases of the following 4 histologic tumor types, which are the most common types occurring in salivary glands: 1) ACC, “adenoid cystic carcinoma 8200/3”; 2) mucoepidermoid carcinoma, “mucoepidermoid carcinoma 8430/3”; 3) acinic cell carcinoma, “acinar cell carcinoma 8550/3” and “acinar cell cystadenocarcinoma 8551/3”; and 4) carcinoma-expleomorphic adenoma/malignant mixed tumor, “carcinoma in pleomorphic adenoma 8941/3” and “mixed tumor, malignant, not otherwise specified (NOS), 8940/3.”

In total, 5464 cases of ACC diagnosed from 1973 to 2008 with known patient age and malignant tumor behavior were identified. We categorized sites on the basis of the variable “site recode with Kaposi and mesothelioma” in the SEER database. Among all cases, 5349 cases arose in 7 sites with ≥100 cases and were included in the study. Sixty-three cases of ACC of unclassified site and 52 cases of ACC of unusual sites (23 digestive system, 18 male genital system, 7 soft tissue, 3 urinary system, and 1 endocrine system) were excluded. The sites with ≥100 cases each were grouped into 7 groups as follows: 1) major salivary glands; 2) minor salivary glands, including “lip,” “tongue,” “floor of mouth,” “ gum and other mouth,” “nasopharynx,” “oropharynx,” “hypopharynx,” “other oral cavity and pharynx,” “nose, nasal cavity and middle ear,” “larynx,” and “trachea”; 3) breast; 4) skin; 5) lung and bronchus; 6) female genital system; and 7) eye and orbit. We used “SEER Historic Stage A” to group patients according to tumor stage. Stage was defined according to the SEER Program as follows: 1) localized with malignancy limited to the organ of origin, no spread beyond the organ of origin, infiltration past the basement membrane of epithelium into stroma of organ; 2) regional with tumor extension beyond the limits of the organ of origin but no distant disease; and 3) distant with tumor spread to areas of the body distant or remote from the primary tumor. The SEER staging system, which we used in this study, is somewhat independent on the lymph node metastasis status of the tumor. For instance, a tumor at SEER regional stage may or may not have lymph node metastasis, because tumor extension beyond the local organ may or may not be caused by regional metastases. Lymph node metastasis data and tumor size information were not available until 1983. Information on tumor grade was not available for most cases. Specific data about perineural invasion are not available in the SEER database.

Statistical Analysis

The distributions according to histologic type, demographics, and clinical features were calculated using the frequency session of SEER*Stat 7.0.4. Differences in demographics and clinical features were assessed with chi-square tests or t tests. Patients who were diagnosed with second or later primary ACC, patients who were not actively followed, and patients who were diagnosed by death certificate only or autopsy only were excluded during survival analysis. Patients were stratified by tumor stage during the survival analysis. The 5-year, 10-year, 15-year, 20-year, and 25-year disease-specific survival rates were estimated using the survival session of SEER*Stat 7.0.4 using the Kaplan-Meier method. Survival was calculated from the date of diagnosis to the date of death caused by ACC or December 31, 2008, which was the follow-up cutoff date. Patients who died from causes other than ACC and patients who were still alive at the time of the follow-up cutoff were censored. Log-rank tests were used to assess differences between survival curves. Hazard ratios (HRs) and their 95% confident intervals (CIs) were determined using univariate and multivariate Cox proportional hazards models and SAS statistical software, version 9.2 (SAS Institute Inc., Cary, NC).13 Year of diagnosis, age, sex, race, marital status, and treatment were adjusted when the relative risk was estimated. All factors were treated as categorical variables except age, which was treated as a continuous variable. Tumor grade, tumor size, and lymph node metastasis were not included in multivariate analysis; because the large amount of missing information on these variables would have resulted in a drastic decrease in our sample size. For all statistical tests, significance was considered to be achieved when P < .05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Table 1 lists the distribution by anatomic site of the 4 most common histologic tumor types occurring in salivary glands: ACC, mucoepidermoid carcinoma, acinar cell carcinoma, and carcinoma expleomorphic adenoma/malignant mixed tumor. With the exception of acinar cell carcinoma, which also is a histologic variant of prostate carcinoma, these cancers occurred most commonly in major and minor salivary glands. ACC demonstrated a relatively widespread distribution: 13% of tumors occurred in the breast, 5% occurred in the skin, 4% occurred in the lung or bronchus, 2% occurred in the female genital system, and 2% occurred in the eye or orbit. ACC occurred with almost equal frequency within the parotid gland and submandibular gland, whereas other salivary gland carcinomas occurred with much greater frequency in the parotid gland.

Table 1. Site Distribution of Adenoid Cystic Carcinomas, Mucoepidermoid Carcinomas, Acinar Cell Carcinomas, and Carcinoma Expleomorphic Adenoma/Malignant Mixed Tumors
 No. of Patients (%)
Tumor SiteAdenoid Cystic Carcinoma, N = 5464Mucoepidermoid Carcinoma, N = 6751Acinar Cell Carcinoma, N = 17,066Carcinoma-Ex/ Malignant Mixed, N = 912
  1. Abbreviations: Ex, expleomorphic adenoma; NOS, not otherwise specified.

Oral cavity and pharynx3208 (58.7)5837 (86.5)1822 (10.7)755 (82.8)
 Major salivary gland1850 (33.9)3652 (54.1)1681 (9.8)673 (73.8)
  Parotid gland907 (16.6)3107 (46)1611 (9.4)492 (53.9)
  Submandibular gland783 (14.3)369 (5.5)39 (0.2)123 (13.5)
  Sublingual gland57 (1)66 (1)1 (0)1 (0.1)
  NOS or overlapping103 (1.9)110 (1.6)30 (0.2)54 (5.9)
 Lip90 (1.7)139 (2.1)21 (0.1)7 (0.8)
 Tongue217 (4)225 (3.3)2 (0)2 (0.2)
 Floor of mouth94 (1.7)182 (2.7)7 (0)1 (0.1)
 Gum and other mouth790 (14.5)1506 (22.3)103 (0.6)59 (6.5)
 Nasopharynx109 (2)35 (0.5)2 (0)5 (0.5)
 Oropharynx26 (0.5)58 (0.9)3 (0)2 (0.2)
 Hypopharynx12 (0.2)21 (0.3)1 (0)0 (0)
 Other oral cavity and pharynx20 (0.4)19 (0.3)2 (0)6 (0.7)
Digestive system23 (0.4)62 (0.9)324 (1.9)9 (1)
 Esophagus16 (0.3)16 (0.2)2 (0)1 (0.1)
 Pancreas1 (0)15 (0.2)280 (1.6)0 (0)
 Anus, anal canal, and anorectum6 (0.1)16 (0.2)0 (0)0 (0)
Respiratory system922 (16.9)579 (8.6)1159 (6.8)26 (2.9)
 Nose, nasal cavity, and middle ear525 (9.6)131 (1.9)17 (0.1)17 (1.9)
 Larynx62 (1.1)43 (0.6)0 (0)2 (0.2)
 Trachea132 (2.4)21 (0.3)1 (0)0 (0)
 Lung and bronchus203 (3.7)384 (5.7)1141 (6.7)7 (0.8)
Bones and joints0 (0)0 (0)0 (0)0 (0)
Soft tissue including heart7 (0.1)4 (0.1)1 (0)4 (0.4)
Skin291 (5.3)49 (0.7)5 (0)53 (5.8)
Breast696 (12.7)5 (0.1)35 (0.2)8 (0.9)
Female genital system132 (2.4)15 (0.2)2 (0)17 (1.9)
 Cervix uteri74 (1.4)13 (0.2)0 (0)0 (0)
 Corpus and uterus, NOS1 (0)0 (0)1 (0)5 (0.5)
 Ovary2 (0)2 (0)1 (0)12 (1.3)
 Vagina5 (0.1)0 (0)0 (0)0 (0)
 Vulva50 (0.9)0 (0)0 (0)0 (0)
Male genital system18 (0.3)0 (0)13659 (80)5 (0.5)
 Prostate16 (0.3)0 (0)13659 (80)2 (0.2)
Urinary system3 (0.1)0 (0)27 (0.2)15 (1.6)
Eye and orbit100 (1.8)46 (0.7)0 (0)10 (1.1)
 Lacrimal gland81 (1.5)21 (0.3)0 (0)9 (1)
Brain and other nervous system0 (0)0 (0)0 (0)0 (0)
Endocrine system1 (0)35 (0.5)6 (0)1 (0.1)
 Thyroid gland1 (0)26 (0.4)6 (0)1 (0.1)
Miscellaneous63 (1.2)119 (1.8)26 (0.2)9 (1)

Table 2 lists the distribution of demographics and clinical features in patients with ACC across the 7 sites that were included in this study. The majority of ACCs occurred in patients ages 40 to 69 years. However, almost 33% of patients with eye and orbit ACC presented before age 40 years, and almost 50% of women with genital system ACC presented at age ≥70 years. Patients with ACC of the eye and orbit were significantly younger at presentation than patients with ACC of the major salivary glands (mean age, 49.9 years vs 55.5 years; P = .001), whereas patients with tumors of the female genital system (mean age, 65.0 years), skin (mean age, 57.5 years), and breast (mean age, 61.5 years) were significantly older (P ≤ .001, P < .001, and P < .001, respectively). For all sites except lung and bronchus, which included a significantly lower proportion of women (P < .001), more women than men were affected. The proportion of African Americans was significantly greater among patients with ACC of the female genital system or eye and orbit than among patients with ACC of major salivary glands (P < .001 and P = .001, respectively), whereas the percentage of African Americans was significantly lower among patients with ACC of the breast (P = .038).

Table 2. Demographics and Clinic Features of Patients With Adenoid Cystic Carcinoma by Tumor Site
 No. of Patients (%)
VariableaMajor Salivary Glands, N = 1850Minor Salivary Glands, N = 2077Breast, N = 696Skin, N = 291Lung and Bronchus, N = 203Female Genital System, N = 132Eye and Orbit, N = 100
  • Abbreviations: Sep/Div/Wid, separated, divorced, or widowed; XRT, radiation therapy.

  • a

    The unknown group for each variable was not indicated.

  • b

    The Surveillance, Epidemiology, and End Results database only provides data from 1983 to 2008.

Year of diagnosis       
 1973-1981311 (16.8)330 (15.9)148 (21.3)33 (11.3)45 (22.2)25 (18.9)22 (22)
 1982-1990300 (16.2)374 (18)86 (12.4)52 (17.9)30 (14.8)15 (11.4)16 (16)
 1991-1999398 (21.5)478 (23)136 (19.5)64 (22)35 (17.2)41 (31.1)13 (13)
 2000-2008841 (45.5)895 (43.1)326 (46.8)142 (48.8)93 (45.8)51 (38.6)49 (49)
Age, y       
 ≤39337 (18.2)260 (12.5)28 (4)26 (8.9)13 (6.4)9 (6.8)29 (29)
 40-691084 (58.6)1237 (59.5)466 (67)155 (53.3)128 (63.1)58 (43.9)53 (53)
 ≥70429 (23.2)580 (28)202 (29)110 (37.8)62 (30.5)65 (49.3)18 (18)
Sex       
 Men754 (40.8)879 (42.3)11 (1.6)123 (42.3)110 (54.2)0 (0)43 (43)
 Women1096 (59.2)1198 (57.7)685 (98.4)168 (57.7)93 (45.8)132 (100)57 (57)
Race       
 White1508 (81.8)1663 (80.7)606 (87.7)251 (89)171 (84.2)92 (69.7)70 (70)
 Black164 (8.9)214 (10.4)46 (6.7)18 (6.4)17 (8.4)32 (24.2)18 (18)
 Other172 (9.3)184 (8.9)39 (5.6)13 (4.6)15 (7.4)8 (6.1)12 (12)
Marital status       
 Single269 (15)268 (13.6)72 (10.6)37 (15.9)31 (15.6)24 (18.6)21 (22.1)
 Married1178 (66)1256 (63.8)411 (60.6)137 (58.8)122 (61.3)44 (34.1)54 (56.8)
 Sep/Div/Wid339 (19)446 (22.6)195 (28.8)59 (25.3)46 (23.1)61 (47.3)20 (21.1)
Tumor grade       
 Low: 1 or 2210 (62.3)243 (61.8)272 (80.2)20 (87)33 (75)12 (42.9)7 (41.2)
 High: 3 or 4127 (37.7)150 (38.2)67 (19.8)3 (13)11 (25)16 (57.1)10 (58.8)
Tumor size, cmb       
 ≤2528 (42.8)422 (38.3)318 (63.1)91 (72.2)28 (27.4)17 (27.4)10 (22.2)
 >2705 (57.2)679 (61.7)186 (36.9)35 (27.8)74 (72.6)45 (72.6)35 (77.8)
Lymph node metastasisb       
 None1049 (82.7)1111 (90.2)466 (95.5)146 (98)74 (61.2)79 (96.3)56 (100)
 Yes219 (17.3)121 (9.8)22 (4.5)3 (2)47 (38.8)3 (3.7)0 (0)
Disease stage       
 Localized910 (51.8)635 (35.5)622 (90.5)173 (69.2)34 (26.8)77 (63.1)41 (44.5)
 Regional638 (36.3)901 (50.4)55 (8)67 (26.8)59 (46.4)37 (30.3)41 (44.5)
 Distant208 (11.9)252 (14.1)10 (1.5)10 (4)34 (26.8)8 (6.6)10 (11)
Treatment       
 Surgery and XRT1125 (62.5)968 (48.4)207 (30.4)37 (13)32 (16.7)35 (27.6)52 (54.7)
 Neither60 (3.4)124 (6.2)11 (1.6)17 (5.9)39 (20.3)4 (3.1)3 (3.2)
 XRT only51 (2.8)269 (13.4)2 (0.3)6 (2.1)41 (21.3)21 (16.5)2 (2.1)
 Surgery only563 (31.3)640 (32)461 (67.7)226 (79)80 (41.7)67 (52.8)38 (40)

Information on tumor grade was available for very few cases. Greater than 80% of tumors in the breast and skin were low grade, whereas tumors of the female genital system and eye and orbit typically were high grade. Greater than 60% of ACCs of the skin or breast measured ≤2 cm in greatest dimension; whereas greater than 70% of ACCs of the lung and bronchus, female genital system, and eye and orbit measured ≥2 cm. The mean tumor size at presentation was significantly smaller for ACC of the breast (2.2 cm) than for ACC of the major salivary glands (2.8 cm) (P < .001), whereas the mean tumor size was significantly larger for ACC of the female genital system (3.7 cm) and lung and bronchus (3.7 cm; P = .012 and P < .001, respectively).

Lymph node metastasis was not common for any site except lung and bronchus. Similarly, only ACC of the lung and bronchus was commonly associated with distant metastases at presentation (27%). Ninety-one percent of breast ACCs were localized at diagnosis compared with only 36% of minor salivary gland ACCs and only 27% of lung and bronchus ACCs. Patients who had ACC in the breast, skin, lung and bronchus, and female genital system were more likely to undergo surgery without radiation; whereas patients who had ACC of the major salivary glands, minor salivary glands, and eye and orbit were more likely to undergo surgery and receive radiation.

Because disease stage is such an important predictor of prognosis,6, 11 we stratified tumors using the SEER Historic Stage A system (localized, regional, and distant stage) when we conducted the survival analysis. Table 3 provides the disease-specific survival rates at 5 years, 10 years, 15 years, 20 years, and 25 years for patients with ACC. Figures 1 and 2 are Kaplan-Meier disease-specific survival curves for patients with localized disease and regional disease, respectively. Patients who presented with localized disease had favorable survival; and, for all sites except eye and orbit, the disease-specific survival rates for patients with localized disease at presentation were >50% even 25 years after diagnosis. Patients who presented with distant disease had low survival rates, and their disease-specific survival rate was only approximately 30% 10 years after diagnosis. In Kaplan-Meier disease-specific survival analysis of patients with localized disease, ACCs of the skin and breast were associated with favorable survival compared with ACCs of the major salivary glands (log-rank P = .017 and P < .001, respectively); whereas ACCs of the lung and bronchus and the eye and orbit were associated with worse survival (both log-rank P < .001) (Fig. 1). Among the patients who had regional disease at diagnosis, the disease-specific survival rates were significantly better for patients with ACC of the skin than for those with ACC of the major salivary glands (log-rank P = .003), whereas disease-specific survival rates were significantly worse for patients with ACC of the minor salivary glands or the lung and bronchus (log-rank P = .008 and P = .017, respectively) (Fig. 2).

thumbnail image

Figure 1. These are Kaplan-Meier survival curves for patients who presented with localized disease by anatomic site. SG indicates salivary glands.

Download figure to PowerPoint

thumbnail image

Figure 2. These are Kaplan-Meier survival curves for patients who presented with regional disease by anatomic site. SG indicates salivary glands.

Download figure to PowerPoint

Table 3. Disease-Specific Survival for Patients With Adenoid Cystic Carcinoma by Anatomic Site and Stage at Presentation
 Anatomic Site
Disease StageMajor Salivary GlandsMinor Salivary GlandsBreastSkinLung and BronchusFemale Genital SystemEye and Orbit
  1. Abbreviations: DSS, disease-specific survival; NA, not enough intervals to produce statistics.

Localized stage       
 No. of patients809552542144226338
 DSS, %       
  5-y93.992.496.896.471.587.288.7
  10-y88.185.59493.771.576.858.7
  15-y79.178.39093.7NA66.652.2
  20-y71.77485.788.5NA66.644.7
  25-y69.665.284.588.5NA66.644.7
Regional stage       
 No. of patients5758044859473635
 DSS, %       
  5-y78.976.278.597.87262.684.9
  10-y64.756.775.494.745.556.936.8
  15-y55.447.37081.115.256.9NA
  20-y49.137.564.654.1NA56.9NA
  25-y42.133.164.654.1NANANA
Distant stage       
 No. of patients1872269828610
 DSS, %       
  5-y43.355.415.954.734.78040
  10-y30.733.415.90NA26.726.7
  15-y21.120.900NA26.726.7
  20-y21.117.100NANA26.7
  25-y21.18.600NANANA

Tables 4 and 5 provide results from the univariate and multivariate Cox proportional hazards model analyses of ACCs diagnosed at localized and regional stage, respectively. Model 1 indicates the HR of each site after adjustment for factors that remained in the model after stepwise selection. Model 2 is the full model, which provides the HR for each site after adjustment for year of diagnosis, age, sex, race, marital status, and treatment. Unfortunately, information on tumor size, tumor grade, and lymph node metastases was available for only a minority of patients. The crude HRs and their P values (Table 4) indicate that localized tumors in the breast and skin were associated with approximately half the risk of death of localized tumors in the major salivary glands, whereas tumors in the lung and bronchus and the eye and orbit were associated with 5 times and almost 3 times, respectively, the risk of death. No significant survival differences were observed between ACCs of the minor salivary glands or female genital system and ACCs of the major salivary glands. The HRs changed slightly after adjustment for other factors; however, the results were consistent with the results in the unadjusted model.

Table 4. Survival Analysis for Patients Presenting With Localized Disease
 Unadjusted ModelModel 1aModel 2b
SiteCrude HR (95% CI)PHR (95% CI)PHR (95% CI)P
  • Abbreviations: CI, confidence interval; HR, hazard ratio; Ref, reference group.

  • a

    Adjusted by age (continuous variable) and treatment.

  • b

    Adjusted by year of diagnosis, age (continuous variable), sex, race, marital status, and treatment.

Major salivary glands1.00Ref1.00Ref1.00Ref
Minor salivary glands1.12 (0.85-1.49).4150.90 (0.67-1.21).4950.87 (0.64-1.18).360
Breast0.47 (0.32-0.69)<.0010.40 (0.26-0.60)<.0010.43 (0.28-0.66)<.001
Skin0.46 (0.22-0.93).0320.40 (0.19-0.82).0130.47 (0.22-0.98).043
Lung and bronchus4.89 (1.98-12.05).0013.72 (1.50-9.19).0053.85 (1.51-9.82).005
Female genital system1.49 (0.84-2.65).1721.05 (0.56-1.94).8891.03 (0.54-1.96).925
Eye and orbit2.60 (1.44-4.72).0023.67 (2.02-6.68)<.0013.79 (2.06-6.95).001
Table 5. Survival Analysis for Patients Presenting With Regional Disease
 Unadjusted ModelModel 1aModel 2b
SiteCrude HR (95% CI)PHR (95% CI)PHR (95% CI)P
  • Abbreviations: CI, confidence interval; HR, hazard ratio; Ref, reference group.

  • a

    Adjusted by age (continuous variable), sex, marital status, and treatment.

  • b

    Adjusted by year of diagnosis, age (continuous variable), sex, race, marital status, and treatment.

Major salivary glands1.00Ref1.00Ref1.00Ref
Minor salivary glands1.30 (1.08-1.57).0061.08 (0.89-1.32).4241.10 (0.90-1.34).344
Breast0.85 (0.46-1.56).6021.04 (0.54-1.99).9151.06 (0.55-2.04).856
Skin0.29 (0.12-0.70).0060.26 (0.09-0.69).0070.26 (0.10-0.70).008
Lung and bronchus1.84 (1.12-3.04).0161.49 (0.88-2.52).1351.54 (0.91-2.62).110
Female genital system1.76 (0.99-3.09).0511.35 (0.72-2.55).3521.34 (0.71-2.56).369
Eye and orbit1.38 (0.75-2.54).3011.50 (0.79-2.85).2141.65 (0.86-3.14).130

Table 5 indicates that, among the patients who presented with regional disease, patients who had tumors of the skin had only approximately 33% the risk of death of patients who had ACCs of the major salivary glands; whereas patients who had tumors of the minor salivary glands had a 30% increased risk of death, and patients who had tumors of the lung and bronchus had an 84% increased risk of death. After adjustment for other factors, only patients who had ACC of the skin had a significantly different risk (75% reduced) compared with the risk for patients who had ACC of the major salivary glands. Because of the small sample size of patients with ACC who presented with distant stage disease, we were not able to conduct multivariate analysis of such patients.

In an attempt to also adjust for tumor size, tumor grade, and lymph node metastases (variables that were unavailable for most patients), we also conducted survival analysis among the patients who had ACCs at the 3 most common sites: major salivary glands (reference group), minor salivary glands, and breast. In a Cox proportional hazards model analysis, we took into consideration all demographics and clinical features, including tumor grade, tumor size, and lymph node metastases (lymph node metastases data were excluded when we analyzed the survival of patients who presented with localized stage, ie, all with negative lymph nodes). The effect model after stepwise selection revealed that, for patients with ACC who presented with localized disease, the HRs for minor salivary glands and breast were 0.94 (95% CI, 0.49-1.82; P = .864) and 0.21 (95% CI, 0.09-0.51; P < .001), respectively. For patients with ACCs who presented with regional disease, the HRs for minor salivary glands and breast were 0.87 (95% CI, 0.54-1.41; P = .584) and 1.29 (95% CI, 0.53-3.14; P = .574), respectively.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

Our current findings demonstrate that, compared with the other 3 common histologic types of salivary gland carcinoma, ACCs arise at more widespread sites across the body. Greater proportions of ACCs occur at sites outside the major and minor salivary glands (especially the breast) compared with other histologic types (with the exception of many prostate carcinomas classified as acinar cell carcinomas). The demographics, clinical features, and survival of patients with ACC differ by primary tumor site. Disease stage (as classified by the SEER local/regional/distant staging system), as expected, appeared to be an important predictor of survival, although we were unable to assess the TNM staging system within this cohort. Furthermore, tumor site appeared to be an important predictor of survival for patients who presented with localized ACC but was less important for patients who presented with more advanced disease.

Our analysis by site indicated that ACC occurs predominantly among individuals aged >55 years, except for ACC of the eye and orbit, for which the mean age at diagnosis was 49.9 years. An even younger mean age at diagnosis, 39.5 years, was reported in a retrospective case series of 20 patients with ACC of the lacrimal gland who received treatment at The University of Texas MD Anderson Cancer Center between 1952 and 2002,14 consistent with the results from other small retrospective case series.15-17 We also observed that the proportion of patients who were African American was significantly greater among women with ACC of the genital system than among patients with ACC of other sites. It is noteworthy that 90% of ACCs of the female genital system among African American women occurred in the cervix compared with only 46% of those in white women. A high proportion of African American women was reported previously in a case series of cervical ACC.18 With the exception of lung and bronchus, ACC was more common in women than in men for all sites. A slight predominance of men among patients with ACC of the lung and bronchus has been reported by others19; and, in most studies, a slight predominance of women among patients with ACCs of other sites also has been reported.4, 6, 14 Because of the rarity of ACC, sample size may be an issue in the evaluation of age, race, or sex distributions; however, because our current study was based on a relatively large population, these demographics may be more reliable.

A European population-based study of ACCs of major and minor salivary glands also indicated that primary site was an important predictor of prognosis: ACCs in the nasal cavity, pharynx, and larynx were associated with an increased risk of death compared with ACCs in major salivary glands.6 In the same study, localized ACCs in the oral cavity were associated with a better outcome than localized ACCs in the major salivary glands; whereas, for ACCs of more advanced stages, differences in survival were not apparent.6 These findings were somewhat consistent with our finding that survival differences were obvious among localized ACCs but less apparent among more advanced ACCs. Among patients with ACCs who presented with regional disease, only those with tumors of the skin had a significantly reduced risk of death. This may be an artifact of the SEER staging system, in which regional disease includes some cases with local extension without lymph node metastases. ACCs of the skin and of the eye and orbit had the lowest incidence of lymph node metastases, as indicated in Table 2; consequently, it is likely that most patients with “regional” ACC of the skin have their disease staged as regional because of local extension, which may be easier to treat surgically than regional ACC of other sites. Patients with localized ACCs at presentation generally had better survival than patients with regional or distant disease. However, the TNM staging system would permit more precise stratification of ACCs to allow comparisons. Furthermore, although the TNM staging system is used more commonly among clinicians and health care professionals, the SEER database only provides the applicable TNM staging information for patients who were diagnosed after 2004, but the local/regional/distant staging information is provided for most of the patients who were diagnosed during the years from 1973 through 2008. Because ACC is a rare malignancy, we chose to use the local/regional/distant staging system so that our current analyses would have adequate sample size and power.

Regardless, our current results suggest that ACCs of the breast and skin have excellent clinical behavior. At presentation, the vast majority of them were localized, smaller than 2 cm, of lower grade, and without lymph node metastasis. The excellent clinical behavior of breast ACC also has been demonstrated in a rare cancer network study from Switzerland,20 in which the median tumor size was 2.0 cm, and approximately 90% of patients presented with T1 or T2 disease. Our subgroup multivariate survival analysis (including tumor size and tumor grade) suggested that patients with localized ACC of the breast have better survival than patients with localized ACC of major salivary glands even after adjustment for clinical features. It is possible that factors like variations in anatomy, pathology, and molecular characteristics contribute to these survival differences among sites.21

Unfortunately, we could not include tumor size, tumor grade, or lymph node metastasis in our multivariate analysis for other sites, because the missing information decreased our sample size and power, especially in less common sites, to such a degree that analysis was meaningless. However, the results we obtained for the 3 most common sites when we included size, grade, and lymph node metastasis in the analysis were concordant with the results from our more limited multivariate analysis with the full sample set, and, to a large extent, this confirmed the validity of our survival analysis of all 7 sites. Although information on surgical margins and perineural invasion is not available in the SEER database, future SEER studies will have more information on the clinical tumor features for each case; thus, adjustment for tumor size, tumor grade, and lymph node metastases will become feasible in such site-comparative analyses.

Our survival analysis was also limited by the rather small number of ACCs of the skin, lung and bronchus, female genital system, and eye and orbit. In addition, site and histology misclassification are possible. However, the SEER registry is considered the largest population-based cancer registry in the United States; and, to our knowledge, our current study is the largest to date of ACCs of these less common sites. Although our results should be interpreted with caution, our study is important because it presents findings for ACCs of less common sites within a population database and is the first to specifically provide insights into differences in demographics, clinical features, and survival for ACCs of different anatomic sites. Therefore, our findings can provide guidance in ACC detection, therapy, and prognosis and may serve as inspiration for future studies to explore the determinants of the distinct evolution and clinical behavior of ACCs at each site.

In conclusion, the demographics and clinical features of ACC differ by anatomic site. Patients who present with localized ACCs generally have better survival than those who present with regional or distant ACCs; thus, early detection may enhance survival and possibly quality of life for patients with ACC. Anatomic site may be an important predictor of survival for patients who present with localized disease but is less important for patients who present with regional disease. However, limitations in the SEER database prevent adequate control for major prognostic confounders, such as grade, size, lymph node metastases, perineural invasion, and surgical margins.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES

This work was supported by the National Institutes of Health through grant U01 DE019765-01 (to Dr. El-Naggar; Dr. Sturgis is Project 2 Leader), by MD Anderson's Cancer Center Support Grant (CA016672), and by Halliburton Employees Fellow in Cancer Prevention funds (to Dr. Xu, fellow).

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. FUNDING SOURCES
  9. REFERENCES