Annals of Surgical Oncology

, Volume 20, Issue 7, pp 2236–2241

Rare Breast Cancer: 933 Adenoid Cystic Carcinomas from the National Cancer Data Base


  • Nandini Kulkarni
    • Department of SurgeryAbington Memorial Hospital
    • Department of SurgeryAbington Memorial Hospital
  • Jon M. Greif
    • Bay Area Breast Surgeons, Inc.
    • Alta Bates Summit Medical Center
  • V. Suzanne Klimberg
    • University of Arkansas for Medical Sciences
  • Lisa Bailey
    • Bay Area Breast Surgeons, Inc.
    • Alta Bates Summit Medical Center
  • Soheila Korourian
    • University of Arkansas for Medical Sciences
  • Marlene Zuraek
    • Stanford University Medical Center
Breast Oncology

DOI: 10.1245/s10434-013-2911-z

Cite this article as:
Kulkarni, N., Pezzi, C.M., Greif, J.M. et al. Ann Surg Oncol (2013) 20: 2236. doi:10.1245/s10434-013-2911-z



Adenoid cystic carcinoma (ACC) is a rare subtype of breast malignancy.


Patients with ACC and infiltrating ductal carcinoma (IDC) reported to the National Cancer Data Base from 1998 to 2008 were reviewed for patient age, ethnicity, tumor size, nodal status, American Joint Committee on Cancer TNM Stage, tumor grade, initial treatment, hormone receptor status (for patients from 2004 to 2008), and survival (for patients from 1998 to 2003).


A total of 933 patients with ACC and 729,938 with IDC were identified. No differences were found for incidence by race/ethnicity (p = 0.97). The group with ACC was older (median 60 vs. 58 years), had larger tumors (median 18 vs. 16 mm), had more grade 1 tumors (46 vs. 18 %), was less likely to undergo axillary lymph node evaluation (75.9 vs. 96.3 %), had fewer node-positive patients (5.1 vs. 35.5 %), had fewer estrogen receptor–positive tumors (15.4 vs. 75.6 %), had fewer progesterone receptor–positive tumors (13.3 vs. 65.2 %), and underwent breast-conserving surgery more often (69.8 vs. 59.8 %). Chemotherapy was provided less often for ACC (11.3 vs. 46.4 %), as was hormone therapy (9.1 vs. 42.3 %). All of these differences were statistically significant (p < 0.0001). With a median follow-up of 65.7 months (ACC) and 64.9 months (IDC), 5-year overall survival (OS) was 88 % for ACC vs. 84 % for IDC (p = 0.02). Grade 1 OS (ACC, 91 % vs. IDC, 92 %; p = 0.50) and stage I OS (ACC, 90 % vs. IDC, 91 %; p = 0.93) were equal.


Compared with IDC, ACC has different characteristics (lower grade, hormone receptor negative, node negative), is treated differently (less axillary surgery, fewer mastectomies, less chemotherapy, less hormone therapy), and has an improved prognosis, with 88 % 5-year survival.

Heterogeneity in breast malignancies has been increasingly appreciated. Many factors, including hormone receptor expression, variable patterns of gene expression, and varying histologic appearance, define biologically different breast tumors. Women are often treated differently and have different patterns of failure, times to failure, and overall prognosis on the basis of these biologic differences. To treat a patient with breast cancer, it is critical to understand not only the patient’s characteristics, but also the specific biologic characteristics associated with her malignancy, and how differences in those characteristics will affect the response to various treatments. The study of rare subtypes of breast cancer helps clinicians design appropriate treatment plans for patients diagnosed with these tumors.

Adenoid cystic carcinoma (ACC) of the breast is a particularly rare subtype of breast cancer. ACC constitutes 0.1–1 % of all breast malignancies.1 Originally called cylindroma by Billroth in 1856, the name ACC was coined by Geschickter in 1945.2,3 The diagnostic criterion for ACC includes malignant cells showing both epithelial and myoepithelial differentiation.4 These cells express myoepithelial markers which are also expressed by other more aggressive basal like carcinomas of the breast. In addition this tumor shows strong positivity with CD117 (c-KIT).59 Similar to basal-like breast cancer, ACC is most often hormone receptor negative and does not show Her2-neu amplification.6,911 Despite some similarities, microarray-based comparative genomic evaluation has shown ACC of the breast to be a distinct entity when compared to other triple-negative, grade-matched, basal-like, infiltrating ductal carcinomas (IDC).12 Unlike the poor prognosis associated with other triple-negative breast cancer, ACC has been reported to have an excellent survival and is not associated with other malignancies.4

The literature on ACC of the breast is limited by the rarity of the disease, consisting of case reports, small case series, and small population-based reports, but it suggests a favorable prognosis.1,1219


The National Cancer Data Base (NCDB) is a joint program of the Commission on Cancer (CoC) of the American College of Surgeons (ACoS) and the American Cancer Society (ACS). The NCDB, established in 1989, is a nationwide, facility-based, comprehensive clinical surveillance resource oncology data set that captures 70 % of all newly diagnosed malignancies in the United States annually. These cases come from approximately 1500 CoC-accredited cancer programs. It is the world’s largest oncology outcomes database and contains over 28 million cancer cases, a number that grows daily. Access to this Health Insurance Portability and Accountability (HIPAA)-compliant data was provided to the authors as part of the NCDB’s alpha test of the Participant Use File (PUF) program.

Women with ACC (ICD-O-3 histology code 8200/3) of the breast (primary site codes C500 to C509) entered in the NCDB from 1998 through 2008 were the subjects of this study. A control group consisted of women diagnosed with IDC (ICD-O-3 code 8500/3) of the breast during the same time period. In-situ cases are not included. Only patients for whom breast cancer was the first cancer diagnosis (Sequence No. 00 or 01) receiving the entire first course of their treatment at or through the reporting institution (Class of Case 1 or 2) were included.

Comparisons included patient age, ethnicity, tumor size, and nodal status, American Joint Commission on Cancer (AJCC) stage, grade, and initial treatment. Initial treatment included the surgery of the primary site, the use of radiotherapy and/or chemotherapy and/or hormone therapy. Survival analysis was performed for patients diagnosed from 1998 to 2003 to allow for a minimum of 5-year follow-up for all patients in the survival analysis. Hormone receptor status was examined for patients diagnosed from 2004 to 2008. Demographic, treatment, and tumor characteristic data were analyzed by the Chi square test and by the nonparametric Wilcoxon rank-sum test. Survival rates were calculated by the Kaplan–Meier method and compared by the log-rank test.


Nine hundred thirty-three patients with a new diagnosis of ACC of the breast and 729,938 patients with IDC were reported to the NCDB between 1998 and 2008. The characteristics of the patients and their tumors, initial treatment, and overall survival (OS) rates are summarized in Table 1. There were no differences in race/ethnicity, with 87 % of both groups identified as white. Patients with ACC were 2 years older on average than those with IDC (60 vs. 58 years, p < 0.0001).
Table 1

Characteristics, treatment, and survival of patients by histologic type of breast cancer









Age, median (years)




Reported race/ethnicity


Median tumor size (mm)




Grade 1 disease, n




Axillary node evaluation (%)




Node positive (%)




ER-positive tumors (%)




PR-positive tumors (%)




Breast conservation (%)




Chemotherapy (%)




Hormone therapy (%)




5-year OS (%)




Grade 1 OS (%)




Stage 1 OS (%)




IDC infiltrating ductal carcinoma, ACC adenoid cystic carcinoma, ER estrogen receptor, PR progesterone receptor, OS overall survival

The histologic grade of the two tumor types was markedly different, with the most common grade for ACC being grade 1 (well differentiated), accounting for 46 % of all ACCs, compared to only 18 % of IDCs. Only 18 % of ACC were reported to be grade 3 (poorly differentiated) or grade 4 (undifferentiated), compared to 40 % of IDC. Patients with ACC had a median tumor size that was larger than patients with IDC (18 vs. 16 mm, p < 0.0001).

Surgery was performed as part of the initial treatment of 908 of 933 (97.3 %) of patients with ACC, and 720,923 of 729,938 (98.8 %) of IDC patients. Breast-conserving surgery was performed more commonly in patients with ACC than in patients with IDC (69.8 vs. 59.8 %, p < 0.0001). Fewer patients with ACC underwent surgical evaluation of axillary lymph nodes compared to IDC (75.9 vs. 96.3 %, p < 0.0001).

Of those patients who underwent surgical evaluation of axillary lymph nodes, only 36 of 703 (5.1 %) of patients with ACC had axillary lymph node metastasis compared with 247,942 of 697,759 (35.5 %) of patients with IDC (p < 0.0001). Twenty-nine (4.1 %) patients with ACC had one to three nodes positive, 5 (0.7 %) had four to nine nodes positive, and 2 (0.3 %) had 10 or more nodes positive. In patients with IDC, 23.8 % had one to three nodes positive, 7.8 % had 4–9 nodes positive, and 3.6 % had more than 10 nodes positive.

Despite the higher use of breast-conserving surgery for ACC, adjuvant radiation treatment was administered to only 438 (47.1 %) patients with ACC, compared to 54.5 % of patients with IDC.

Systemic chemotherapy was administered as part of the initial treatment for 104 patients (11.3 %) with ACC compared to 331,262 (45.4 %) of IDC patients (p < 0.0001). For another 5.1 % of patients of each histologic type (48 ACC, 37,337 IDC), chemotherapy was recommended but either was not administered or was unknown whether it was administered. Specific details concerning the chemotherapy drugs delivered, cycles and duration of treatment, and dosages are not reported in the NCDB.

Estrogen receptor (ER) status and progesterone receptor (PR) status were evaluated for a subgroup of 391 ACC and 289,112 patients diagnosed from 2004 to 2008. Hormone receptor testing was not performed in 20 (5.1 %) of these ACC patients and 1.1 % of IDC patients, and it was borderline or unknown in 27 (6.9 %) of ACC patients and 4.3 % of IDC patients. Hormone receptor status was dramatically different for the two histologic types, with only 53 (15.4 %) of ACC tumors being ER positive compared with 75.6 % of IDC tumors (p < 0.0001). Similarly, only 46 (13.3 %) of ACC tumors were PR positive compared to 65.2 % of IDC tumors (p < 0.0001).

The use of hormone therapy paralleled the significant difference in hormone receptor status between the tumor types, with only 83 (8.9 %) patients with ACC receiving hormone therapy, compared to 290,397 (39.8 %) patients with IDC (p < 0.0001). Another 24 (2.6 %) ACC patients and 43,687 (6.0 %) IDC patients had contraindications to hormone therapy, hormone therapy was recommended but not provided, or it was unknown whether hormone therapy was provided. Hormone therapy was not provided or offered to the remaining 826 (88.5 %) of ACC patients or 395,546 (57.7 %) of IDC patients.

The 5-year OS of the 448 patients with ACC (Fig. 1) treated from 1998 to 2003 was significantly better than for the 370,816 patients with IDC (88 vs. 84 %, p = 0.02). The OS benefit for ACC compared to IDC persisted when the subgroup of 136 ACC patients treated by mastectomy was compared to the 155,318 patients with IDC treated by mastectomy (88 vs. 78 %, p = 0.007) (Fig. 2). However, when grade 1 patients with ACC were compared to grade 1 patients with IDC (Fig. 3), no difference in survival was found (91 vs. 92 %, p = 0.50). In addition, when stage 1 patients with ACC were compared to stage 1 patients with IDC (Fig. 4), no difference in OS was found (90 vs. 91 %, p = 0.93).
Fig. 1

Kaplan–Meier OS curve, ACC vs. IDC 1998–2003
Fig. 2

Kaplan–Meier OS curve, mastectomy group only, ACC vs. IDC, 1998–2003
Fig. 3

Kaplan–Meier OS curve for grade 1, ACC vs. IDC, 1998–2003
Fig. 4

Kaplan–Meier OS curve for stage 1, ACC vs. IDC, 1998–2003


The NCDB, established in 1989, contains data from the records of more than 28 million patients treated at more than 1500 hospitals in the United States. Data from approximately 70 % of all new patients diagnosed with cancer in the United States are entered into the NCDB.20 The volume of data in the NCDB allows a unique opportunity to examine rare tumors. In this study, we utilized the NCDB to examine the characteristics, demographics, and patterns of treatment for patients diagnosed with ACC and IDC between 1998 and 2008 and compared the two groups.

ACC of the breast is a rare subtype of breast malignancy that accounts for approximately 0.1 % of all breast cancers reported to the NCDB. ACC is most commonly seen as a tumor of the salivary glands, but it also arises in the breast, lacrimal glands, auditory canal, digestive tract, skin, prostate, tracheobronchial tree, uterine cervix, larynx, and Bartholin gland.5

ACC of the breast morphologically resembles ACC of the salivary glands, with both having similar architectural patterns, most commonly the cribriform type (and less commonly solid, tubular, or trabecular types). The solid growth pattern is associated with higher grade and worse prognosis. Both also are reported to share the molecular feature of a recurrent chromosomal translocation, t(6,9)(q22–23:p23–24), and both express c-KIT in a high percentage of cases. However, although ACC of the breast is associated with a more benign course, with only rare nodal or distant metastasis, ACC of the salivary gland is a more aggressive malignancy that frequently recurs locally and/or metastasizes to distant sites, resulting in 5-year survival rates of 60–75 %, which drop to 30–54 % at 10 years.5 ACC of salivary glands is treated by radical surgery and often also with systemic chemotherapy.

After the description by Geschickter, multiple case reports subsequently appeared in the literature.1,1219 Hopkins and Tullis performed a retrospective analysis of cases of breast cancer in a single institution from 1958 to 1971 and found three cases of ACC.17 All three patients underwent radical mastectomy in their study, although the authors suggest that in view of the benign, slow-growing nature of the disease, it would not be unreasonable to consider a partial mastectomy in the properly selected patient. In the current study, 69.8 % of patients diagnosed with ACC were treated with breast-conserving surgery.

More recently, Sumpio et al. examined the Connecticut Tumor Registry from 1952 to 1982 and found 40,350 cases of invasive breast cancer.21 They reported 37 cases of ACC and reviewed available histologic slides of 27 of these cases. They found that only 14 of the 27 cases (52 %) were truly ACC; the others proved to be breast cancer of other histologic types. They also found that the diagnosis of ACC improved from 3 % accuracy before 1974 to 79 % after 1974.

A potential weakness of this study is that histologic slides were not reviewed, and neither the diagnosis nor the tumor grade was confirmed independently. It is possible that some cases—specifically some of the hormone receptor–positive and/or higher-grade cases—included in this series might not be found to be true ACC if a complete histologic review had been undertaken. We are also unable to report the percentage of our cases with the various architectural growth patterns of ACC (cribriform, solid, tubular, trabecular) because these data are not reported to the NCDB. It is not known how the percentage of solid architectural growth pattern would correlate with survival from our data.

Sumpio et al. found ACC to have a favorable biologic behavior and good prognosis.21 They advocated simple mastectomy as primary local treatment; they also thought that no axillary dissection was needed as a result of the low incidence of axillary metastases, and a yearly chest x-ray was sufficient for follow-up because most of the distant metastases were in the lungs.

Mammographically, a study by Santamaria et al. suggested that ACC of the breast usually presents as a circumscribed, lobulated nodule most commonly in the outer quadrants or periareolar region.22 The NCDB does not capture the mode of presentation (e.g., palpable mass, screening mammograms), so the current study has not examined presenting signs and symptoms for ACC of the breast.

Major findings in this study of ACC of the breast, when compared to patients with IDC, include larger tumor size for ACC, a high percentage of patients with low-grade/well-differentiated tumors, low incidence of axillary lymph node metastasis, greater use of breast-conserving surgery, lower use of axillary lymph node surgery, and lower use of chemotherapy despite very low rates of hormone receptor positivity.

Overall 5-year survival was better for patients with ACC of the breast compared to IDC. However, examining the survival of only patients with grade 1 ACC of the breast compared to those with grade 1 IDC of the breast revealed no significant difference in survival. Similarly, the survival of patients with stage 1 ACC of the breast was not significantly different than those with stage 1 IDC of the breast. The improved OS of ACC compared to IDC appears to be largely explained by the lower grade and earlier stage of patients presenting with ACC compared with IDC. Data on survival for patients with ACC beyond 5 years of follow-up were not available in this data set.

Immunohistochemical studies have shown that adenoid cystic cancer of the breast has a characteristic phenotype that is ER, PR and Her2-neu negative and c-KIT positive.511 In this study, only 15.4 % of patients with ACC were ER positive and only 13.3 % were PR positive. In contrast to IDC, where triple negativity in breast tumors is associated with a poor prognosis, ACC is a triple-negative tumor that has good OS and a very low incidence of axillary metastases. Wetterskog et al. microdissected 14 ACCs and 14 grade-matched IDCs and used high-resolution microarray comparative genomic hybridization to show that ACC of the breast is a genomically distinct tumor from other triple-negative and basal-like breast cancer.12

In our study, only 76 % of patients with ACC had axillary nodal evaluation, and only 5 % of these were found to have positive axillary lymph nodes. Many previous reports have concluded that axillary node dissection is not a required part of management of ACC of the breast. This study of nearly 1000 women with ACC confirms the very low incidence of axillary lymph node involvement in ACC and an even lower incidence of involvement of more than three axillary lymph nodes, allowing surgeons and their patients to make a more informed decision regarding the need for axillary node surgery.

Systemic adjuvant therapy was used infrequently to treat ACC patients reported to the NCDB. A total of 11.3 % of patients received adjuvant chemotherapy and 8.9 % hormone therapy. This pattern of care is supported by the low incidence of hormone receptor positivity, the low grade of ACC of the breast, and the high OS achieved in these 933 patients.

In summary, it is important to recognize ACC of the breast as a distinct clinical entity from IDC, as the implications for management and outcome are significantly different.


The authors declare no conflict of interest.

Copyright information

© Society of Surgical Oncology 2013