Traditionally, all patients with breast cancer have been treated in a similar fashion, and breast cancer was treated as a single disease. Increasingly, the heterogeneity of breast malignancies is being recognized and encompasses many factors, including hormone receptor expression, variable patterns of gene expression, and varying histologic appearance. The optimal management of a patient who receives a diagnosis of a malignant tumor of the breast depends on both a tumor-specific and a patient-specific approach. This requires knowledge about the tumor type, as well as its predicted behavior, and response to various treatments. Data concerning subtypes of breast malignancy, including extremely rare subtypes, will aid in these therapeutic decisions.

Metaplastic breast cancer (MBC) is a rare malignancy characterized by various combinations of adenocarcinoma, mesenchymal, and other epithelial components. Because MBC was not officially recognized as a distinct pathologic diagnosis until 2000,1 knowledge about the patient demographics, presentation, tumor characteristics, and treatment patterns is limited. To date, only small series and case reports have attempted to delineate the factors that make MBC different from more common malignant breast histologies.2–17 We hypothesized that MBC has markedly different characteristics at presentation and is managed differently than infiltrating ductal carcinoma (IDC).

MATERIALS AND METHODS

Malignant tumors are reported to the National Cancer Data Base (NCDB) with the use of diagnosis codes from the 3rd edition of the International Classification of Diseases for Oncology (ICD-O-3), which was implemented in 2001 and included a new histologic code for MBC: 8575/3.1 Data reported to the NCDB for all women with a diagnosis of MBC that used this code from January 2001 through December 2003 were included in this study and analyzed. To obtain a homogenous group of patients with MBC, other overlapping or potentially related histologic diagnoses were neither included nor combined.

Specific data examined included the year of diagnosis, patient age at diagnosis, race/ethnicity, primary tumor size, lymph node status, overall American Joint Committee on Cancer (AJCC) stage, tumor grade, and estrogen and progesterone receptor activity. Primary tumors with diffuse breast involvement or of unknown size, and patients who did not undergo surgery at the site of the primary tumor were not included in the analysis of primary tumor size. Also analyzed were the following: data concerning initial treatment, including surgery of the primary site, and the use of radiotherapy and/or chemotherapy; and information about the primary payer or type of medical insurance to provide insight concerning access to screening examinations, specifically mammography, of the study group. For comparison, the same data for all patients with a diagnosis of infiltrating ductal carcinoma (ICD-O-3 code 8500/3) for the same time period were also retrieved. The Pearson χ2 test was used for statistical analysis. A value of P < .05 was considered statistically significant, and all P values are two-tailed.

RESULTS

Eight hundred ninety-two patients with a new diagnosis of MBC were reported to the NCDB between 2001 and 2003 (Table 1). During this period, data of 255,164 patients with IDC were entered. This represents .24% and 69.8%, respectively, of the total of 365,464 patients with a new diagnosis of breast cancer whose data were entered into the NCDB during the study period. The number of patients with MBC reported to the NCDB increased during each year of the study period, with 243 patients in 2001, 318 patients in 2002, and 331 patients in 2003, whereas the number of patients reported with IDC did not increase (85,559 in 2001, 87,891 in 2002, and 81,714 in 2003). The mean age of the group with MBC was older, 61.1 years, compared with the mean age of the group with IDC, 59.7 years (P = .001).

TABLE 1. Demographics and tumor characteristics
Full size table

Patients with MBC were statistically significantly more likely to be African American or Hispanic, 126 (14.1%) and 49 patients (5.5%), respectively, compared with patients with IDC, 25,900 (10.2%) and 9,947 patients (3.9%), respectively, (odds ratio [OR], 1.455, P < .001 for African American, and OR, 1.817, P < .001 for Hispanic). Both groups had a similar percentage of patients identified as Asian or Pacific Islander (2.7% for MBC vs. 2.6% for IDC) or as other or unknown (2.5% for MBC vs. 2.6% for IDC).

The size of the primary tumor was reported in 807 patients (90.5%) with MBC and 228,501 patients (89.6%) with IDC who underwent surgery at the primary site. The size of the primary tumors was much larger for patients with MBC than for patients with IDC (Fig. 1). Only 5.5% patients with MBC had tumors <10 mm in size (T1a or T1b) compared with 26.9% patients with IDC (OR, 6.3, P < .001). Most patients with MBC (70.5%) presented with tumors larger than 20 mm (T2 and above), whereas most patients with IDC (65.2%) presented with tumors smaller than 20 mm (OR, .7, P < .01). Tumors larger than 5 cm in size accounted for 20.4% of MBC, and only 5.2% of IDC.

FIG. 1.
figure 1

Primary tumor size by histology.

Full size image

Lymph node involvement was reported in 21.9% of patients with MBC for whom nodal status was available. This compares with 34.3% of patients with IDC with positive lymph nodes (Fig. 2). This difference in the percentage of patients with negative lymph nodes (78.1% for MBC vs. 65.7% for IDC) was statistically significant (OR, .538, P < .001). Of the 174 patients with positive lymph nodes in the MBC group, data concerning the specific number of positive nodes were available in 140 patients. One to three positive nodes were reported in 72.1% of patients with MBC, four to nine nodes in 19.3%, and 10 or more positive nodes in 8.6%. In the IDC group, 67.3% of patients had one to three nodes positive, 22.4% of patients had four to nine nodes positive, and 10.2% of patients had 10 or more nodes positive, of the total of 71,255 patients with IDC with positive nodes for whom the specific number of involved nodes was known.

FIG. 2.
figure 2

Lymph node status by histology.

Full size image

The AJCC stage (5th edition) grouping was available for 825 patients with MBC and 246,575 patients with MBC, with the remainder unknown (Table 2). Despite the lower frequency of lymph node involvement, patients with MBC more commonly presented with stage II disease (MBC 55% vs. IDC 37.3%, OR, .4, P < .001) as a result of larger primary tumors. Patients with MBC were also more likely to be diagnosed with stage III (10.5%) or stage IV disease (4.6%) at presentation compared with patients with IDC (8.4% and 3.4%, respectively).

TABLE 2. American Joint Committee on Cancer stage by histology (%)
Full size table

Tumor grade was classified as well differentiated, moderately differentiated, or poorly/undifferentiated in 2.7%, 7.1%, and 67.8% of patients with MBC, respectively. Grade was not recorded or was unknown for 22.5% of patients with MBC. Tumor grade for patients with IDC was well differentiated, moderately differentiated, or poorly/undifferentiated in 17.1%, 37.9%, and 38.8% of patients, respectively, and was not recorded or was unknown in 6.3%.

Estrogen receptor (ER) and progesterone receptor (PR) testing was not performed on 3.0% of MBC specimens. It is unknown whether these tests had been performed or whether the results were unknown in 25.7% and 26.1% of cases, respectively. Of the cases in which ER testing was performed and the results were known, 72 (11.3%) of 636 patients with MBC were ER positive compared with 137,050 (74.1%) of 184,937 patients with IDC (OR, 22.4, P < .001). For PR testing, 66 (10.4%) of 632 patients with MBC were positive compared with 114,061 (62.4%) of 182,936 patients with IDC (OR, 14.2, P < .001).

Surgery was performed in 847 patients (95.0%) with MBC as part of their primary treatment and included a mastectomy in 471 (55.6%). Breast-conserving surgery (BCS) was performed in 373 patients (44.0%) with MBC. Three patients (.4%) underwent an unspecified surgical procedure. Of the 241,192 patients with IDC (94.5%) who underwent surgery as part of their primary treatment, mastectomy was performed in 92,424 patients (38.3%), and BCS was performed in 147,814 patients (61.3%). The more common use of BCS for IDC was statistically significant (OR, 2.0, P < .001). An unspecified surgical procedure was performed in 954 patients (.4%) in the IDC group. Surgery was used as the sole treatment modality for 263 patients (29.5%) in the MBC group and 58,639 patients (23%) in the IDC group, and was otherwise used in combination with radiotherapy, chemotherapy, and/or hormonal therapy. An analysis of the choice of surgical treatment by tumor size shows that for a given tumor size, the rates of BCS and mastectomy were similar for both patients with MBC and IDC (Fig. 3).

FIG. 3.
figure 3

Breast-conserving surgery by primary tumor size and histology (%).

Full size image

Radiotherapy was delivered as part of the primary treatment to 379 (42.5%) of 892 patients with MBC, compared with 132,691 patients (52.0%) in the IDC group. Radiotherapy was most commonly combined with surgical treatment for both histologic types of breast cancer in 368 (97.1%) of 379 patients with MBC, and 129,168 (97.3%) of 132,691 patients with IDC. Eleven patients with MBC (2.9%) underwent radiotherapy without surgery, either alone (n = 2), with chemotherapy (n = 8), or with a combination of chemotherapy and hormonal therapy (n = 1). Similarly, 3523 patients (2.7%) with IDC underwent radiotherapy without surgery, either alone (n = 1072), with chemotherapy (n = 1193), with hormonal therapy (n = 772), or with a combination of chemotherapy and hormonal therapy (n = 486).

Systemic chemotherapy was administered to 53.4% patients with MBC and was provided in addition to surgery (preoperatively or postoperatively) in all but 16 patients (3.4%). In comparison, 42.1% of patients with IDC were treated with systemic chemotherapy, again usually in addition to surgical treatment, in all but 4.1%. The increased use of chemotherapy for patients with MBC compared with those with IDC (53.4% vs. 42.1%, respectively) was statistically significant (OR, 1.6, P = .001). Hormonal therapy was provided in only 57 patients (6.4%) with MBC overall, compared with 83,485 patients (32.7%) with IDC. An analysis of treatment with systemic chemotherapy by AJCC stage (Fig. 4) shows that chemotherapy was almost twice as likely to be administered to stage I patients with MBC compared with identically staged patients with IDC (39.7% vs. 21.7%), but that the percentage of stage II and III patients treated with chemotherapy was identical in both groups.

FIG. 4.
figure 4

Systemic chemotherapy by American Joint Committee on Cancer (AJCC) stage and histology.

Full size image

The most common primary payors or types of medical insurance for the MBC group of patients (in order of decreasing frequency) were managed care, Medicare with supplement, unspecified insurance, Medicare alone, and Medicaid, accounting for 35.5%, 24.3%, 17.2%, 9.6%, and 6.2%, respectively. Only 3.7% of patients with MBC were recorded as not insured or as self-pay, and the insurance status of 2.7% was unknown. Tricare, military, Veterans Administration, and Indian/Public Health Service accounted for the remainder of patients. Similarly, the patients with IDC had the same top five payors or types of medical insurance accounting for 37.4%, 21.9%, 20.1%, 9.3%, and 4.4%, respectively, with 2.5% of patients recorded as not insured or self-pay, and the insurance status unknown in 3.5%.

DISCUSSION

The NCDB, established in 1989, contains data from the records of more than 18 million patients treated at more than 1400 approved cancer programs in the United States. Data from approximately 75% of all new patients diagnosed with cancer in the United States are entered into the NCDB. The size of the NCDB allows the examination of extremely rare tumors in relatively large numbers and can provide valuable information that would otherwise not be available from smaller sources of data about their presentation, characteristics, treatment, and outcomes. We examined the characteristics at presentation, demographics, and patterns of treatment for patients diagnosed with MBC and IDC between 2001 and 2003 and compared the two populations.

MBC is a rare breast malignancy accounting for only one-quarter of 1% of all breast malignancies reported to the NCDB during the study period. The increasing number of patients with MBC reported each year during the study period may represent an actual increase in incidence of the disease in the United States. However, it is likely that the increase seen in the number of cases is a result of increasing recognition of and familiarity with the diagnosis of MBC by both pathologists and the tumor registrars who assign and report codes to the NCDB. Since early reports more than 30 years ago,18,19 MBC has been increasingly recognized and reported as a distinct histologic type of breast cancer, but because of its rarity, only relatively small series have been reported. In the largest previous publications,8–12 Wargotz et al. separately described five subgroups of MBC, including matrix-producing carcinoma, spindle cell carcinoma, carcinosarcoma, squamous cell carcinoma of ductal origin, and metaplastic carcinoma with osteoclastic giant cells, with 26, 100, 70, 22, and 29 patients, respectively. Oberman16 reported 29 cases of MBC and concluded that the lack of a correlation of the microscopic pattern of these tumors with prognosis, as well as the presence of apparent overlapping microscopic findings, supported the concept that the subgroups described by Wargotz et al. are all variants of a single entity. There is still no strong consensus on this issue. Cytokeratin positivity in both the epithelial and mesenchymal components of MBC has led many to conclude that it is metaplasia of the epithelial elements of a carcinoma that gives these lesions their pseudosarcomatous appearance.17

With the recognition by ICD-O-3 of a specific code for MBC, cancer registries began collecting data from across the United States in 2001. The continued use of other overlapping or potentially related diagnoses, some more specific and some less so, is a confounding factor in the effort to study MBC. These additional diagnoses include squamous cell carcinoma, spindle cell carcinoma, carcinosarcoma, sarcoma not otherwise specified, pleomorphic carcinoma, spindle cell sarcoma, spindle cell squamous cell carcinoma, osteosarcoma, and others. Undoubtedly, some patients diagnosed with these alternatives would be considered by many pathologists to have MBC. We chose to limit our study only to the specific diagnosis of MBC for this analysis, to maximize the likelihood that our observations would be applicable to clinicians faced with a diagnosis of MBC in the future.

A difference was seen in the race/ethnicity of patients with MBC compared with IDC, with a higher percentage of patients identified as African American or Hispanic. The reasons for these differences are unknown, but represent a small but increased risk for MBC in these groups. The difference observed in the age of patients diagnosed with MBC compared with IDC, although statistically significant (mean age, 61.1 vs. 59.7 years, P = .001), is not likely of major clinical importance.

Patients with MBC had larger primary tumors than patients with IDC. This difference in size could possibly be explained by a difference in the use of screening mammography between the groups, or by a failure of screening examinations to identify MBC compared with IDC. With similar primary payor and types of medical insurance in both groups, it is unlikely that differences in access to or use of screening mammography alone can explain the large difference in size seen in this study. In addition, several reports13,14 have described the mammographic and sonographic findings in MBC, including a round to ovoid, high-density mass with ill-defined or obscured margins and with architectural distortion. In one study, most patients presented with a rapidly growing palpable mass, which was then visible as a mammographic mass in 15 of 16 patients.13 These studies suggest that MBC is not mammographically occult.

It is most likely that the larger size at presentation for MBC is the result of a more rapid growth rate. Successful early detection of IDC by mammography depends on a relatively long, detectable, preclinical phase resulting from slower tumor growth, which may not exist with MBC. Most patients with MBC had poorly differentiated or undifferentiated tumors compared with IDC, which would also support this theory of more rapid growth for MBC.

The lower incidence of axillary lymph node involvement compared with IDC (21.9% vs. 34.3%) also represents an observed biological difference between these two tumors. The difference is even more dramatic when the larger tumor size of the patients with MBC is considered. If these tumors were biologically similar, we would expect a higher incidence of nodal involvement for patients with MBC on the basis of the larger tumor size. The presence of mesenchymal components and sarcomatous elements seen in MBC may explain this different biologic behavior and pattern of metastasis. Wargotz et al.8–12 reported an incidence of axillary lymph node metastasis ranging from 6% to 26%, depending on the subtype of MBC, and was highest for carcinosarcoma. Additionally, for patients with carcinosarcoma, it was carcinoma that was the most frequent component to metastasize.

The higher AJCC stage seen in the MBC group, despite having more lymph node–negative tumors, is strongly influenced by the larger tumor size. The 29.5% of patients with MBC having tumors <2 cm in size (T1), compared with 65.2% of patients with IDC with T1 tumors, explains the much higher incidence of AJCC stage I disease in patients with IDC, and AJCC stage II disease in patients with MBC.

The very low incidence of hormone receptor positivity in MBC compared with IDC represents another biologic difference in these tumors and has obvious implications for differences in their treatment. Hormonal therapy was rarely provided to patients with MBC, consistent with this low incidence of hormone receptor positivity in these patients. As would be suspected, the use of hormonal therapy in the patients in both groups in this study was directly related to the number of cases with hormone receptor–positive tumors.

The lack of hormonal therapy as a therapeutic option for adjuvant treatment, combined with an increased risk of systemic metastasis that would be predicted given the larger tumor size, higher AJCC stage, higher tumor grade, and hormone receptor negativity, explains the increased frequency of treatment with systemic chemotherapy in the patients with MBC. Data supporting the effectiveness of adjuvant chemotherapy for patients with MBC are lacking, and its use represents extrapolation from data related to more common histologic varieties of breast cancer. Given the apparent differences in tumor biology, the benefit of adjuvant chemotherapy, as well as the most effective drugs if it is to be administered, is unclear. Rayson et al.15 reported 27 patients over a 30-year period from the Mayo Clinic, 13 of whom (50%) developed distant metastasis. Ten different chemotherapy regimens were used to treat metastatic disease, and only one partial response was observed. They concluded that systemic therapy seems to be less effective.

Surgery was used to treat approximately 95% of both patients with MBC and IDC, but mastectomy was most common for patients with MBC (55.6%), whereas BCS was most common for patients with IDC (61.3%). This difference was related to the larger tumor size of the patients with MBC, including 20.4% of the MBC tumors >5 cm in size compared with only 5.2% for IDC tumors. When corrected for tumor size, the rates of BCS and mastectomy were similar between the two groups, suggesting that the principles of breast cancer surgery used for more common histologies are being applied to patients with MBC. The use of radiotherapy parallels the use of BCS, as would be anticipated. In both histologic groups, however, fewer patients received radiation than were treated with BCS. BCS was performed in 44.0% patients with MBC. The results of BCS and radiation in this group of patients with MBC, including ipsilateral breast recurrence rates, are being reported to the NCDB and will require further follow-up.

In summary, MBC is a very rare tumor and is different from IDC, with different presenting characteristics, demographics, and tumor biology. It is increasingly being reported to the NCDB since the ICD-O-3 codes have been implemented. Patients with MBC are more commonly African American or Hispanic and present with larger tumor size, less nodal involvement, higher AJCC stage, higher tumor grade, and hormone receptor negativity. Because of the larger tumor size and higher AJCC stage, patients with MBC are treated more aggressively than IDC, more often with mastectomy and systemic chemotherapy. In the absence of histology-specific data, patients are treated by use of the same principles that govern treatment of more common breast cancers. Follow-up is needed to determine the long-term results of the current treatment.