Metaplastic breast cancer: a comparison between the most common histologies with poor immunohistochemistry factors

Background Metaplastic carcinoma of the breast (MCB) is a rare histological type of breast cancer. This study aimed to determine whether MCB exhibits shorter overall survival (OS) and disease-free survival (DFS) compared with other histologies that are considered unfavorable. Methods We retrospectively analyzed 157 clinical file records of the Mexico City-based National Institute of Cancerology and compared the clinical characteristics and treatment of 24 patients with MCB, 37 patients with triple-negative invasive lobular carcinoma (TN-ILC), 48 patients with high-grade invasive ductal carcinoma (HG-IDC), and 48 patients with triple-negative invasive ductal carcinoma (TN-IDC), paired by clinical stage and age. We performed a comparative analysis and analyzed OS and DFS using a log-rank test. Results In patients with MCB, the 5-year DFS was 52.1% (mean, 48.52 months; 95%: 35.32-61.72), and the 5-year OS was 72.2% (mean, 59.77 months; 95% CI: 48.55-71.00). No differences were observed in the DFS of MCB compared with each of the other histologies (MCB vs. HG-IDC, p = 0.865; MCB vs. TN-IDC, p = 0.966, and MCB vs. TN-ILC, p = 0.132). Moreover, no differences were observed when comparing the OS of MCB with that of each of the other histologies (MCB vs. HG-IDC, p = 0.246; MCB vs. TN-IDC, p = 0.255, and MCB vs. TN-ILC, p = 0.387). Conclusions Neither OS nor DFS differ between patients with MCB and those with other histologies with unfavorable immunohistochemical factors.


Background
Breast cancer (BC) is the most common malignant neoplasia in women in Mexico in terms of incidence and mortality [1,2]. Approximately 85% of BC cases involve invasive ductal carcinoma (IDC) or invasive lobular carcinoma (ILC). The remaining 15% of cases include other types of BC with specific differentiation, including metaplastic carcinoma of the breast (MCB) [3].
This study aimed to determine whether an MCB diagnosis confers poor prognosis with respect to overall survival (OS) and disease-free survival (DFS) compared with other common histologies that are considered unfavorable (high-grade invasive ductal carcinoma [HG-IDC], triple-negative invasive ductal carcinoma (TN-IDC), and triple-negative invasive lobular carcinoma [TN-ILC]).

Methods
Patients were selected from the database of the Mammary Tumors Department of the National Institute of Cancerology (INCan) in Mexico City between January 2005 and January 2013. We obtained the clinical file records of female patients diagnosed with MCB, IDC, or ILC using conventional histology by one of the authors. We excluded all patients with mixed histologies, those with a diagnosis of carcinoma in situ, those whose clinical file was incomplete, and those who had another histology or synchronous tumors in another organ.
From the IDC patient group, we created two subgroups, one defined as HG-IDC, to which patients with Scarff-Bloom-Richardson grading system 8 and 9 were assigned, and another group who had an IHC report of TN (estrogen receptors [ER] and progesterone receptors [PR] with nuclear staining of <1%, human epidermal receptor [HER2/neu] with IHC of ≤1, or fluorescence in situ hybridization at a ratio of HER2:centromere enumerator probe [CEP] 17 of <1.8). We paired subjects by age and clinical stage (CS), selecting two subjects for each type of MCB case. From the ILC patient group, we included classic and pleomorphic variants and selected patients with an IHC report of TN as previously described.
We obtained and analyzed the following variables: age; tumor size; lymph node status; distant metastasis; CS; the presence of hormonal receptors (HRs) and HER2/neu; use of induction therapy; pathological response; surgery performed; adjuvant therapy; and recurrence, progression and survival of the patients through the last follow-up appointment or date of death. For this final variable, we located patients or spoke with a family member by telephone to determine the patient's status.
We conducted descriptive statistics for the demographic variables and report central tendency measures. We conducted univariate analysis of means using Student's t test for continuous variables and the chi-square or Fisher exact test for ordinal variables.
We calculated the follow-up of each patient. DFS was defined as the period between the date at which the patient's initial multimodal treatment (surgical, radiotherapeutic, and/or systemic) terminated and the date of recurrence or the patient's last appointment. Recurrence was defined as the presence of disease >6 months after final treatment. Progression was defined as the presence of disease 6 months before treatment was finalized. OS was defined as the period between the date treatment ended and the date of the patient's last appointment or death.
DFS and OS were analyzed by the Kaplan-Meier survival curve method, and the curves were compared using the log-rank test. We performed multivariate analysis to ascertain which variables exerted an influence on OS and DFS using the Cox proportional hazards model. Statistical significance was defined as a p value of <0.05. We used the statistical software program SPSS ® 2012 for Windows (SPSS Inc., Austin, TX, USA) for data analysis. The Director of Research of the INCan gave ethical approval for the collection and use of the data for this study.
This study was reviewed and approved by the Institutional Review Board of the National Cancer Institute of Mexico (approval number REV/20/14), and has been performed in compliance with the Helsinki Declaration. Since this study collected only archived data and materials, without using personal data from patients, we were not required by our federal legislation to obtain an informed consent from patients, as well as for phone calls, when preservation of confidentiality is assured. During the telephone call, the patient or the family member that answered was informed that a retrospective analysis was being performed and that the actual status was necessary for completing the information, also was established that all information will be kept confidential in accordance to the legislation and will be used only for this evaluation.

Results
We identified 32 (0.6%) patients with a diagnosis of MCB among 5,440 patients diagnosed with BC. Of these 32 patients, we excluded 8 due to incomplete clinical records or having mixed histologies, resulting in a total of 24 patients for the final analysis. Of a total of 487 patients with an ILC diagnosis, we obtained 37 (7.6%) patients with TN-ILC. The patients with MCB were paired by age and CS at a ratio of 1:2 with patients with HG-IDC and TN-IDC, resulting in 48 patients for each group. All of the patients' general clinicopathological characteristics are described in Table 1. We analyzed 157 patients in total, with an average follow-up of 40 months (range, 2.3-97.5 months).
We conducted a multivariate analysis with using a Cox proportional hazards model. In the comparative analysis, a significant difference (p <0.05) was observed, and none of the variables independently influenced OS or DFS.

Discussion
Presently, BC is the most common neoplasm diagnosed in Mexico [2]; thus, infrequent histologies are observed more frequently than in other countries. One such histological type is MCB, and little is known about the clinicopathological characteristics and prognosis of MCB. Table 4 depicts the most representative series comparing MCB with other histologies.
In this study, MCB represented 0.6% of all BC diagnosed, a percentage similar to those reported in other studies [5,7,10,11,22]. In our series, the mean age at diagnosis of MCB was >40 years, similarly to other studies [5,7,8,[10][11][12]22,24,30]. All of the MCB and other BC patients exhibited tumors >5 cm. Furthermore, lymph node metastasis was observed in the majority of cases, and no difference in the rate of lymph node metastasis was observed with respect to other histologies. This finding can be explained by the nature of the study and the fact that Mexican female patients tend to present at an advanced stage. Previous studies have established that MCB tends to be diagnosed with larger tumors [7,22]. However, Bae et al. found no difference in tumor size between MCB and TN-IDC (p = 0.144) [33]. Lymph node metastasis has been shown to be less frequent in MCB [31,32], but not all studies have confirmed this observation [22,33]. In our study, 8.3% of the patient with MCB exhibited distant metastasis. However, our results with respect to MCB are consistent with the literature [22,31,34].
MCB was less likely to exhibit HR expression, with expression detected in 0-17% of cases [15,16,22]. Mourad et al. suggested that the absence of HR in MCB is due to the absence of an extensive glandular component [35]. Additionally, HER2/neu overexpression has been described within a range of 4-16% [31,33]. In our study, most patients with MBC were HR-negative or HER2/neu-negative, unlike patients with HG-IDC. This description is consistent with that of other studies [12]. Park et al. described that 84% of MCB cases were TN, as were 20.1% of IDC cases (p <0.001) [34]. Due to the rarity of MCB, few studies have evaluated the response to induction therapy [12,28]. In a series of 39 patients in which 9 patients were given induction CT, 8 exhibited progression, and all died due to the disease [28]. In our study, only half of the MCB patients received CT induction. Of patients who had a response, 25% had a pCR; however, 50% experienced progression during induction. In our institution, most patients who experience progression during induction CT have received radiotherapy with cisplatin as a radiosensitizer. Importantly, most MCB patients who exhibited progression had been treated with FAC-paclitaxel and concomitant cisplatin and radiotherapy, indicating that this histology responds poorly to this systemic treatment. This response may be associated with the intrinsic characteristics of MCB. According to Weigelt, MCB can be molecularly classified as claudin-low and basal-like [36], implying a behavior and systemic therapy response similar to those of MBC and mesenchymal neoplasias [37].
The group in which more BCSs were conducted was the TN-IDC group, followed by the HG-IDC group; notably, neither MCB nor TN-ILC patients were submitted to BCS, likely because these patients were diagnosed with larger tumors and exhibited poorer responses to induction CT. Song et al. also reported more mastectomies in MCB patients relative to TN-IDC patients (92.7 vs. 80.92%; p = 0.054) [32]; however, Bae et al. did not observe this difference (78.7 vs. 66.1%; p = 0.09) [33]. Of the patients who received adjuvant chemotherapy, nearly 50% received CT. This percentage can be explained by the fact that the patients who received induction CT were not included ( Table 2). The percentage of patients who received RT was similar, with a slightly higher trend in MCB and TN-ILC compared with HG-IDC and TN-IDC (Table 2), possibly because BCS was less frequent  among patients with MCB and TN-ILC. There is controversy in the literature concerning whether patients with MCB tend to receive more RT than other histologies [31][32][33]. The mean DFS of MCB was 48.52 months, with a 5-year DFS of 52.1%. No significant differences were observed in DFS between MCB and each of the other histologies analyzed (Figure 2). The mean OS of patients with MCB was 59.77 months, with a 5-year OS of 72.2%. When comparing the OS of the groups studied, a difference was observed between the four groups (p = 0.027). However, when comparing the OS of MCB with each group, this difference was not detected (Figure 1). In patients with MCB, mean OS decreased to 40 months (72.2 and 73%, respectively); thus, patients with MCB stop dying, whereas patients with TN-ILC continue to do so (80-month OS of 72.2% vs. 44.3%, respectively) ( Figure 1) [39]. However, both studies compared cancers based only on the hormone receptors status, without knowing the HER2/ neu status or type of CT applied.   To our knowledge, this is the first study on MCB reported in the literature in Latin America and particularly in a Mexican population; furthermore, this is the first study to compare MCB with TN-ILC. Our study has several limitations, such as its retrospective design, the number of MCB cases, the focus only on IHC characteristics, and the relatively short mean follow-up. Therefore, future studies should consider the molecular characteristics of each of the tumors.

Conclusions
MCB is an infrequent entity and thus is rarely studied. MCB tends to exhibit less BCS, likely due to the CS at which it is diagnosed, its TN receptors and because it exhibits disease progression. However, our study demonstrated that the OS and DFS in patients with MCB do not differ from those of patients with the most common BC histologies with poor IHC profiles. Future studies should determine whether the molecular characteristics contribute to the prognosis of this type of BC.