Aims The prognosis of metaplastic breast cancer (MBC) is reportedly worse than that of triple-negative invasive ductal carcinoma (TN-IDC), but the determinants of poor prognosis are not yet known.
Methods Patients from two Korean cancer centres were included in this study (67 MBC and 520 TN-IDC). Characteristics of the two disease groups, including clinical parameters, histological features, chemoresponsiveness, disease recurrence and survival estimates, were evaluated.
Results MBC presented with larger tumours, more frequent distant metastasis and higher histological grade compared with TN-IDC (p<0.001). All but nine patients with MBC had triple-negative disease. Disease-free survival and overall survival (OS) of MBC were worse than TN-IDC (p<0.001). Multivariable analysis of disease-free survival revealed MBC type as an independent prognostic factor (HR 2.53; 95% CI 1.32 to 4.84) along with lymph node metastasis and implementation of breast conserving surgery. For OS, MBC type remained a significant prognostic factor (HR 2.56; 95% CI 1.18 to 5.54). Chemoresponsiveness of MBC and TN-IDC were similar in both neoadjuvant (p=1.000) and advanced disease settings (p=0.508). For a given MBC type, risk factors for disease recurrence included the presence of a squamous component (HR 4.0; 95% CI 1.46 to 10.99) and lymph node metastasis (HR 4.76; 95% CI 1.67 to 13.60); the risk factor for OS was initial distant metastasis (HR 10.77; 95% CI 2.59 to 44.76).
Conclusions MBC had worse survival outcomes compared with TN-IDC. Poor prognosis for MBC was likely caused by frequent recurrence with high initial stage and the unique biology of MBC itself.
- invasive ductal carcinoma
- metaplastic carcinoma
- triple-negative breast cancer
- breast cancer
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- invasive ductal carcinoma
- metaplastic carcinoma
- triple-negative breast cancer
- breast cancer
Invasive breast cancer has been classified with regard to morphologic characteristics, biologic phenotypes and, more recently, gene expression profiles.1 Following the development of agents with specific molecular targets, classification according to particular molecular expressions, including the oestrogen receptor (ER), progesterone receptor (PgR) and human epidermal growth factor receptor 2 (HER2), has become of great importance. In particular, triple-negative breast cancer (TNBC), lacking ER, PgR and HER2 overexpression, is associated with an unfavourable prognosis without any benefits from molecular targeting.2–4
Metaplastic breast cancer (MBC), a rare subtype accounting for <1% of invasive breast cancers, is characterised by high-grade carcinoma with heterogeneous metaplastic components and is usually ‘triple negative’.5 6 Recently, Lien et al7 demonstrated differential upregulation of epithelial-mesenchymal transition-related genes in MBC compared with invasive ductal carcinoma, while Hennessy et al8 showed that MBC was distinct from basal-like cancers. We previously reported that MBC presents with large tumours, aggressive pathological features and frequent initial distant metastasis resulting in worse prognosis compared with typical triple-negative invasive ductal carcinoma (TN-IDC).9
In the current study, we aimed to determine what might explain the difference in outcomes between MBC and other TNBC with regard to clinical parameters, histological features and response to chemotherapy in both neoadjuvant and metastatic settings. To obtain a relatively large number of MBC cases for our analysis, patients were pooled from the expanded breast cancer registries of two institutions in Korea.
Patients with pathological specimens classified as MBC were selected from two Korean cancer centres, the National Cancer Center, Korea and Severance Hospital of Yonsei University, between January 2001 and December 2009. In total, 67 patients with MBC were enrolled, including seven (10.4%) patients with de novo stage IV disease (figure 1). To compare clinical parameters, 520 TN-IDC were also selected from the database of National Cancer Center during the same period. Therapeutic strategies were made according to the tumour stage and receptor expression.
We evaluated conventional clinicopathological factors according to previously described methods.9 Immunohistochemistry (IHC) was performed using paraffin-embedded tissues and antibodies against ER (SP1; Ventana), PgR (1E2; Ventana), HER2 (polyclonal; DAKO), p53 (Bp53-11; Ventana) and Ki-67 (MIB-1; DAKO). A cut-off value of ≥10% of stained nuclei by IHC analysis or a value of ≥2/7, according to the modified Allred scoring system, was used to define ER and PgR positivity. HER2 expression was scored as 0–3, according to DAKO Hercep test recommendations. If the result of IHC was 2+, HER2 positivity was determined based on gene amplification by fluorescence in situ hybridisation using a cut-off ratio of 2.0. Results for p53 and Ki-67 expression were reported as percentages of IHC-stained cells. For the statistical analysis, p53>25% and Ki67≥20% were considered to be high expression levels. Tumour grade was determined using the modified Scarff–Bloom–Richardson classification.10 Pathological review of the MBC cases was performed by two experts (RJY and KYM), and tumour stage was assessed according to the 7th edition of the American Joint Committee on Cancer staging system.11
Response to first-line chemotherapy in metastatic settings
Tumour response assessment in neoadjuvant and metastatic settings was performed using RECIST 1.0.12 Pathological complete response (CR) was defined as the complete disappearance of invasive cancer in the breast and lymph nodes. The objective response rate (ORR) was defined as the percentage of patients who showed a complete response (CR) or partial response. The clinical benefit rate was the proportion of patients with CR, partial response and stable disease for >6 months.
Comparison of parameters between the two groups was performed using χ2 or t-tests. Survival analysis was conducted with the Kaplan–Meier method and the log-rank test. Overall survival (OS) was calculated as the duration from the first day of initial treatment to death or the last follow-up date. Disease-free survival (DFS) was defined as the duration from the day of definitive surgery to the first date on which disease recurrence was documented. p Values <0.05 were regarded as significant. For multivariable analysis, Cox proportional hazard model was applied and all significant parameters confirmed by univariate analysis and other confounding factors were included.
The characteristics of patients are described in table 1. MBC presented with larger tumours (p<0.001) and more frequent distant metastases at diagnosis (p=0.003) compared with TN-IDC, but lymph node metastasis did not differ between the two groups (p=0.592). MBC showed significantly higher histological grade than TN-IDC (p<0.001). Ki-67 and p53 (over)expression were similar between MBC and TN-IDC (p=0.057 and p=0.458).
All but nine MBC were triple-negative, with the exception of three ER- and/or PgR-positive, five HER2-positive, and one hormone receptor and HER2-positive tumours. Although the application of surgical treatment (p=0.052) and adjuvant chemotherapy (p=0.365) did not differ between MBC and TN-IDC, radiotherapy (p=0.027) and breast conserving surgery (BCS) (p<0.001) were less frequently performed in MBC than TN-IDC.
The histological features of all MBC cases were reviewed by two pathologists, and the results are shown in table 2. Two cases previously misdiagnosed as MBC were excluded after review. Thirty-five cases (52.2%) had squamous differentiation, which was the most common subtype. Eight cases (11.9%) were of mixed subtype, featuring various metaplastic components.
Survival outcome and prognostic factors in patients with MBC versus TN-IDC
During the median follow-up of 52.2 months (range 0.4–107.8 months), 104 patients experienced disease recurrence (figure 1). Loco-regional recurrence without distant metastasis was observed in four (6%) MBC and six (1.2%) TN-IDC. Among the patients with distant metastasis, 20 (29.9%) MBC and 64 (12.3%) TN-IDC showed visceral organ involvement. Nine (13.4%) MBC and 20 (3.8%) TN-IDC had multiple metastases in more than three sites. Finally, 89 patients (22 (32.8%) MBC and 67 (12.9%) TN-IDC) died, of whom 73 died of cancer progression. Two deaths were due to associated infection without evidence of cancer progression, and the cause of death in 14 patients was not specified.
MBC showed worse DFS (5-year rates, 45.6% vs 81.6%; p<0.001) and OS (5-year rates, 53.7% vs 84.6%; p<0.001) (figure 2). The survival estimates remained unchanged when 13 patients with non-triple-negative MBC were excluded. When we performed stage-matched analyses, worse prognosis of MBC than TN-IDC of all stages was noted, but significantly worse in stage II diseases (p<0.001 for both DFS and OS) (supplementary table 1).
To investigate the reason why MBC showed poor outcome, univariate and multivariable analyses were performed with many clinicopathological factors as shown in table 3. MBC type itself compared with TN-IDC (HR 2.53, 95% CI 1.32 to 4.84) was revealed as an independent prognostic factor for disease recurrence with lymph node metastasis (HR 3.97, 95% CI 2.51 to 6.28) and implementation of BCS (HR 0.45, 95% CI 0.28 to 0.72). For OS, MBC type itself remained significant (HR 2.56, 95% CI 1.18 to 5.54) after adjustment with confounding factors. Other prognostic factors associated with OS were lymph node positivity (HR 4.03, 95% CI 2.31 to 7.01), distant metastasis (HR 3.71, 95% CI 1.48 to 9.28), BCS (HR 0.52, 95% CI 0.30 to 0.90) and adjuvant chemotherapy (HR 0.51, 95% CI 0.29 to 0.92).
The proportion of patients who underwent BCS significantly correlated with initial stage and included 90.2% of those with stage I, 83.1% of those with stage II, 51.0% of those with stage III and 30.8% of those with stage IV (p<0.001). Adjuvant radiation therapy was more frequently delivered to patients with BCS compared those with mastectomy (90.6% vs 32.2%; p<0.001). Although HER2 positivity was associated with worse DFS and OS in six MBC patients who did not receive adjuvant anti-HER2 therapy, it was not significant in multivariable analysis.
Comparison of response to chemotherapy in patients with MBC or TN-IDC
Among patients with stage I–III disease, 10/ 60 (16.7%) with MBC and 104/508 (20.5%) TN-IDC received anthracycline- and taxane-based neoadjuvant chemotherapy. The rate of pathological CR in both the primary tumour and axillary lymph nodes was similar in the two groups (10.0% vs 12.5%; p=1.000).
Of 30 MBC and 93 TN-IDC who developed recurrence or de novo stage IV disease, 25 and 85 patients, respectively, received palliative chemotherapy, with four non-overlapping regimens: anthracycline-based (17%), taxane-based (35%), capecitabine-containing (30%) and others (17%). The distribution of regimens between the MBC and TN-IDC groups was not different (p=0.280). The mean number of chemotherapy cycles was 5.56 (range, 2–15) for MBC and 5.03 (range, 1–25) for TN-IDC (p=0.591). ORR (38.9% vs 47.5%; p=0.508) and clinical benefit rate (50% vs 52.5%; p=0.841) were similar between the two groups. Subgroup analysis according to chemotherapy regimen did not show significant differences in ORR between the groups.
Determinants of poor outcomes in 23 patients with recurrent MBC
Among 60 patients with stage I–III MBC, 23 experienced disease recurrences. On univariate analysis of DFS (table 4), tumour size, lymph node involvement, HER2 status, squamous features of MBC histology, type of surgery and adjuvant radiation were significant predictive factors for disease recurrence in patients with MBC. Over half of the mastectomy group (51.7%) received radiation, which was associated with better DFS, although this did not reach statistical significance (6.9 vs 21.5 months; p=0.110). The histological feature of squamous component in MBC correlated with worse DFS compared with MBC without such components (HR 4.0; 95% CI 1.46 to 10.99). However, both univariate and multivariable analyses determined that it was not a significant factor for OS (p=0.104). Only distant metastasis at diagnosis was a significant prognostic factor for OS in multivariable analysis (HR 10.77; 95% CI 1.2.59 to 44.76).
The present study demonstrated the dismal prognosis of MBC relative to TN-IDC, as we had previously reported in a smaller MBC population.9 Although most of the studies of patients with MBC considered only a few cases due to its rarity of incidence,13–15 Tseng and Martinez recently reported a 63.3% 5-year OS rate of MBC collected from the Surveillance, Epidemiology, and End Results database,16 which is similar to our current finding of a 53.7% 5-year OS rate. When compared with TN-IDC, both DFS and OS were worse in MBC in our study. Okada et al reported a worse prognosis for MBC compared with invasive ductal or lobular carcinoma.17
MBC was classified into three or five histological subtypes based on the presence of heterogeneous elements by Wargotz et al18–22 and Oberman.23 WHO has categorised MBC into purely epithelial, mixed epithelial and mesenchymal types,24 but the relationship between pathological subtype and prognosis remains largely unclear. Yamaguchi et al25 reported significantly higher metastatic risks in high-grade spindle cell and pure squamous cell groups than in matrix-producing carcinomas and low-grade spindle cell groups, suggesting prognostic implications of histological subtypes. Although the presented study has limitations, which include the retrospective nature of our evaluation and a lack of uniformity of the treatment modalities, current data indicated that the presence of squamous elements was a risk factor for disease recurrence and poor prognosis, as also reported by Okada et al.17
Liedtke et al26 reported that TNBC had poor clinical outcomes despite its favourable response to chemotherapy. In a series of 255 patients with stage I–III TNBC, OS was significantly worse than that of 863 patients with other subtypes (3-year rates, 74% vs 89%) despite a higher pathological CR rate to neoadjuvant chemotherapy (22% vs 11%). In our data, pathological CR rates to neoadjuvant chemotherapy in patients with stage I–III MBC or TN-IDC were similar and around 10%. Moreover, we did not find any significant differences in responsiveness to preferred chemotherapy between the MBC and TN-IDC groups in a palliative setting. This is in contrast to the report by Rayson et al,27 which observed only one PR among seven patients with metastatic MBC. We must note that the current study included more patients with MBC who received relatively newer chemotherapeutic regimens containing taxane and capecitabine compared with the earlier study,27 which had considered various chemotherapeutic agents in a much smaller sample size.
In terms of possible explanations for the worse outcomes of MBC, the comparable response with chemotherapy in neoadjuvant and metastatic settings suggests that variation in chemoresponsiveness is not a sufficient explanation. Although this should be confirmed prospectively in cohorts that have undergone similar regimens, such studies will be difficult due to the rarity of MBC. The current study suggests an important role for adjuvant radiation in patients with mastectomy to prevent disease recurrence, although this was not supported by the multivariable analysis. This result is supported by the recent report by Tseng and Martinez,16 which observed that adjuvant radiation therapy given to 580 (38.6%) of 1501 patients with MBC from the Surveillance, Epidemiology, and End Results database provided an OS benefit in patients with lumpectomy and certain high risk patients with mastectomy. Current practices in the management of patients with MBC do not consider the unique biology of MBC, and the therapeutic role of adjuvant radiation is yet to be validated.
We must describe a few limitations of our study. Treatment for recurrent MBC was heterogeneous in this series owing to its retrospective data collection from two hospitals. We adopted the cut-off value of 10% by IHC to define ER/PgR positivity, although recent recommendation of threshold for endocrine therapy has been changed from 10% to 1%.28 Based on this new concept of ‘triple-negativity’, further study may be warranted to achieve a consensus on clinical course of patients with MBC and therapeutic strategies for them.
In conclusion, we demonstrated worse survival outcomes of patients with MBC compared with those with TN-IDC in an expanded study population. Poor prognosis for MBC is likely to be caused by frequent recurrence with the unique biology of MBC itself as well as high initial stage.
The present study reported worse clinical outcomes of metaplastic breast cancer (MBC) compared with triple-negative invasive ductal carcinoma of breast.
Poor prognosis for MBC was likely caused by frequent recurrences with high initial stage and the unique biology of MBC itself, rather than relative resistance to chemotherapeutic agents.
Among MBC patients, the histological feature of squamous component correlated with poor disease-free survival along with lymph node metastasis.
Statistical advices were given by the statistician BH Nam at National Cancer Cencer, Korea.
Funding No direct funding was received for the study.
Competing interests None.
Patient consent This is a retrospective study and we did not perform any invasive procedures or additional studies. The results of pathological report aimed at therapeutic procedures and clinical information were only used, although again it was reviewed retrospectively. It was exempted from consent by IRB.
Ethics approval Ethics approval was provided by National Cancer Center, Korea.
Provenance and peer review Not commissioned; externally peer reviewed.