According to ASCO guideline and the latest St Gallen consensus, triple-negative breast cancer (TNBC) occurs only if there is no expression of estrogen receptor (ER-negative), progesterone receptor (PR-negative), and there is neither expression nor amplification of human epidermal growth factor receptor 2 (HER2-negative) in a tumor [¹, ²]. Apart from these clinicopatological markers and classical classification of breast cancer subtypes, there are molecular gene tests, which allow dividing TNBC into two main subtypes—more common basal-like and claudin-low [³]. Still, molecular classification has still no influence over clinical management. TNBC accounts for about 9–21 % of all breast cancers including patients with stage I–IV breast cancer [⁴–⁸]. In the past, this rate was higher because it included cases with ER/PR less than 10 %.

TNBC patients have poorer outcomes compared with other cancer subtypes [⁷–¹⁴]. They are at higher risk of early recurrence, mainly in the lungs, brain, and soft tissue [⁴, ⁶, ¹³–¹⁸]. The highest risk of relapse is between the first and third year after primary treatment. In cases of recurrence, the survival is shorter than in non-TNBC patients [⁹, ¹³, ¹⁴, ¹⁹, ²⁰]. However, TNBC is more sensitive to chemotherapy. The rate of pathological complete remission (pCR) after neoadjuvant chemotherapy is higher than in other breast cancer subtypes [¹¹, ¹³, ¹⁹]. On the other hand, methods of the treatment in this group of patients are still limited in clinical practice because of the lack of molecular targets. The adjuvant treatment is usually recommended in TNBC and should include anthracyclines, taxanes, and an alkylating agent [², ²¹].

The aim of this study was to analyze recurrence pattern in order to determine the prognostic factors of recurrence and overall survival in a group of consecutive 228 TNBC patients treated at Cancer Center and Institute of Oncology in Warsaw, Poland, between the years 2005 and 2006.

The old term for TNBC included tumors with low expression of hormonal receptors (ER/PR < 10 %). There are a lot of publications with data referring to this old classification making the rate of TNBC tumors to be higher in the past than nowadays. In our study, this rate was 9 % and is comparable to other studies (9–21 %) [⁴–⁸]. The rate of TNBC patients was in some studies even higher, but they assessed other groups of TNBC patients (e.g. African American patients, only neoadjuvant setting or patients with stage I–III breast cancer) [¹², ¹³, ²²].

The frequency of nodal involvement at diagnosis in TNBC patients differs in studies with conflicting results. Lin et al. demonstrated recently that TNBC tumors were less likely to be lymph node-positive, and a similar outcome was reported in other studies (positive lymph nodes: 38 and 41 %, respectively) [¹⁴, ²²]. Contrary, in our study, 55 % of TNBC patients had lymph node involvement, which is consistent with the frequency of 54.4 % reported in other study [⁹]. Nodal status issue remains unresolved.

Neoadjuvant chemotherapy consisting of anthracyclines and taxanes has been widely accepted as standard therapy of locally advanced breast cancer [²¹]. Patients with TNBC have increased pCR rates compared with non-TNBC. This rate was 29 % in patients who received neoadjuvant anthracycline-based chemotherapy and 38 % after anthracycline and taxane combined treatment [¹¹, ¹⁹]. In our study, two-thirds of patients received anthracycline-taxane neoadjuvant chemotherapy and only 15 % of them achieved pCR—recurrence occurred in almost half of patients in this small group. However, such result is doubtful because there were only 7 patients who achieved pCR (all of them received chemotherapy containing taxanes); the result was lower than normal. In addition, it was reported that if pCR was achieved, patients with TNBC and non-TNBC would have similar survival [¹³, ¹⁹].

TNBC has a characteristic pattern of recurrence. Dent et al. [⁹] reported that in their study the risk of recurrence rose sharply from the date of diagnosis, peaked at 1–3 year interval and then dropped quickly. Similarly, in another study, the risk of relapse was strongly time-dependent and dramatically higher for TNBC patients during the first 3 years after diagnosis [¹³]. The same phenomenon was observed in our study. Additionally, we determined the sequence of metastases development in different sites. Local relapse and metastases to the brain and visceral organs occurred in the first 3 years, but the risk of bone metastases declined from diagnosis and then rose slightly again after 5 years of observation (Fig. 1). The brain and lungs were the most common sites of recurrence which is in partial agreement with the results from other studies. Lin et al. [¹⁶] analyzed the sites of distant recurrence in 116 metastatic TNBC patients and reported that the majority of metastases were in lungs and liver. The brain was the third most common site of recurrence. However, in recently published study, TNBC tumors were associated with a greater risk of brain and lung metastases [¹⁴]. In contrast, in other studies, the rates of local relapse or bone metastases were the highest [⁴, ²³]. The frequency of liver metastases in our study was low; this result was consistent with the observation from a previous study [²⁴]. According to initial staging in our study, the recurrences were 1 % (1 patient), 36 % (25 patients), and 63 % (44 patients), respectively, for patients with stages I, II, and III breast cancer. These data were similar to results described by Alarcon-Rozas et al. [²⁵] (7.5, 32.5, and 60 %, respectively).

After a 6 year follow-up, DFS and OS in our study were 68 and 62 %, respectively. These results accord with a previous study which did not include TNBC patients with stage IV; 5-year DFS and OS were 68.2 and 74.5 %, respectively [²⁶]. However, in the above-mentioned Kaplan et al. [⁷] study, 5-year relapse-free survival (RFS) and OS in TNBC patients were 84 and 81 %, respectively. This result may be dependent on the difference in the staging of breast cancer between studies—in our study, majority of patients had more advanced disease (more patients with stage III and less with stage I), whereas nearly 80 % of patients in Kaplan study presented with stage I or II.

In a recent retrospective study, TNBC metastatic patients were divided into two subgroups by RFS [²⁷]. The analysis showed that patients with RFS ≥ 3 years had better outcomes—higher disease control rate (DCR), longer progression-free survival (PFS) to first-line palliative chemotherapy, and longer OS than those with RFS < 3 years (DCR 55 vs. 77 %, p = 0.022; median PFS 3.6 vs. 7.7 months, p = 0.0001; median OS 17.4 vs. 42.0 months, p = 0.0003). In our study, 155 patients (71 %) had RFS ≥ 3 years. Only 7 patients experienced recurrence after 3 years from primary diagnosis and 3 of them died. On the other hand, almost all patients with RFS < 3 years passed away (59 of 63 patients).

Furthermore, patients with brain metastases have poor outcomes. The median survival from brain metastases has been reported to be between 2.9 and 4.9 months, compared to 6.3 months in our study [⁶, ¹⁵–¹⁸, ²⁸]. Brain metastases were the first site of recurrence in 19 patients (8.3 %) (in some of them metastases occurred concurrently in other sites). None of the patients developed brain metastases at the diagnosis of breast cancer. On the other hand, in a recent study, the incidence of brain metastases as the first site of recurrence in TNBC patients initially at stage I–III was 4.7 % [²⁹]. The incidence of brain metastases as the first site of relapse in the in the above-mentioned study led by Park was much more common in patients with shorter RFS than in those with longer RFS (16 vs. 3 %, p = 0.047) [²⁷]. Similar results were found in our study—in the same group of patients (metastatic TNBC, initially presented with stages I–III), brain metastases were more likely to develop in patients with RFS < 3 years then with RFS ≥ 3 years (24 and 3 %, respectively).

The most relevant factor responsible for survival in this study was tumor size. The hazard ratio (HR) of recurrence in patients with a tumor >5 cm was 16 times higher than among patients with a tumor <2 cm. A hazard for death was also elevated in patients with large tumors (HR = 8.21 in tumors >5 cm). These results were in agreement with a previous report where the tumor size was the most important prognostic factor in TNBC patients [⁵]. Even small, node-negative (T1N0) TNBC tumors appear to have a higher recurrence rate, which was documented in some studies as well [³⁰, ³¹]. According to these observations, TNBC patients should be given more aggressive treatment, even if they are in a low-risk category. Apart from tumor size, the patients’ prognosis in our and other studies depended also on nodal status [⁸, ³²]. In a recent study, survival did not differ among patients with N1, N2, and N3 [³³]. In contrast, some analysis showed that TNM staging was not sufficient for predicting therapeutic outcome for TNBC patients due to the different biology of this breast cancer subtype [³⁴].

It is known that TNBC affects younger women. However, one question that should be addressed is whether younger age is associated with poor prognosis in this group of patients. The results of studies are ambiguous. Kassam et al. [³⁵] reported that TNBC patients <50 years had worse outcome. In other study, age was not related to prognosis [⁸]. On the other hand, Ovcaricek et al. [²⁶] showed that age >65 years was an independent prognostic factor for DFS and that the risk of recurrence was 1.79-fold higher in older patients than in younger patients. Interestingly, in the univariate analysis of our study, older patients had shorter survival (age > 65 vs. < 65 years, p = 0.036), but this result lost its independent prognostic value in the multivariate analysis. Only 55 % of patients older than 65 years received adjuvant/neoadjuvant chemotherapy (among stage I–III TNBC), whereas this rate in younger patients was 91 %. This can explain our founding at least partially.

To the best of our knowledge, this is the largest study evaluating TNBC in Poland. Recurrence rates for different sites were presented by cumulative hazard rates and annual hazard rates. However, longer time of observation is necessary to complete and verify these results, especially in terms of recurrence in different sites.

Patients and oncologists always have a dilemma on how to cope with this aggressive disease. For patients, the diagnosis of breast cancer is fearsome; so when they know additionally that they suffer from TNBC—a subtype with poor outcomes—this situation is often more stressful. We recommend that this group of patients should be offered to participate in clinical trials with novel agents.

TNBC patients have a unique pattern of relapse, which occurs mostly in the first 3 years following diagnosis. The most common sites of recurrence were brain and lungs. Tumor size was independent prognostic factor for prognosis, and this feature should be mainly considered in the management of TNBC patients. This group of patients should receive aggressive adjuvant therapy to prevent early recurrence or death. Further prospective clinical trials are needed to identify the most efficient therapy in order to improve survival outcomes.

The authors declare that they have no conflict of interest.