Bhumsuk Keam Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea , Seock-Ah Im Department of Internal Medicine and Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea , Hee-Jun Kim Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea , Do-Youn Oh Department of Internal Medicine and Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea , Jee Hyun Kim Department of Internal Medicine and Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea , Se-Hoon Lee Department of Internal Medicine and Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea , Eui Kyu Chie Department of Radiation Oncology, Seoul National University College of Medicine, Seoul, Korea , Wonshik Han Department of Surgery, Seoul National University College of Medicine, Seoul, Korea , Dong-Wan Kim Department of Internal Medicine and Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea , Woo Kyung Moon Department of Radiology, Seoul National University College of Medicine, Seoul, Korea, , Tae-You Kim Department of Internal Medicine and Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea , In Ae Park Department of Pathology, Seoul National University College of Medicine, Seoul, Korea , Dong-Young Noh Department of Surgery, Seoul National University College of Medicine, Seoul, Korea , Dae Seog Heo Department of Internal Medicine and Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea , Sung Whan Ha Department of Radiation Oncology, Seoul National University College of Medicine, Seoul, Korea , Yung-Jue Bang Department of Internal Medicine and Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea

We also performed multivariate analysis (Table 6). Cox proportional hazard regression analysis for OS included statistically significant variables (initial clinical stage, ER, bcl-2, and triple negative). In multivariate analysis, initial clinical stage was the only significant independent prognostic factor to impact on OS (hazard ratio 3.597, p = 0.044).

Clinical significance of triple negative breast cancer

Forty seven patients (32.4%) of the 145 were triple negative breast cancer. Clinicopathologic variables according to triple negative are summarized in Table 7. Triple negative breast cancer patients showed statistically higher nuclear grade, and lower bcl-2 positive rate than non-triple negative breast cancer patients. A trend for high levels of Ki-67 was also observed in triple negative, although it did not reach statistical significance (p = 0.053). The pCR rate and clinical RR in triple negative were significantly higher (p = 0.005, p = 0.012, respectively). However, RFS and OS were significantly short in triple negative breast cancer patients (p < 0.001, p = 0.021, respectively). RFS and OS survival curves for triple negative and non-triple negative are shown in Figure 2. Because c-erbB2 positivity by immunohistochemistry was unclear, we conducted a second analysis considering 2+ as c-erbB2 positive. Using this definition of triple negative, the results were similar.

 

Discussion

The clinical course of breast cancer patients treated with neoadjuvant chemotherapy remains difficult to predict, because histologically homogeneous breast cancers vary in response to therapy and have divergent outcomes [23]. As a result, many researchers have tried to identify prognostic factors in order to give optimal individualized therapy in locally advanced breast cancer, as well as in early breast cancer. Currently, pCR is the most powerful prognostic factor for prolonged survival in neoadjuvant chemotherapy [3,5,6,24]. However, a significant proportion of patients with pCR have recurrent diseases [9]. Moreover, the prognostic factors for patients receiving neoadjuvant chemotherapy differ from those for patients who receive adjuvant or palliative chemotherapy, because pathologic parameters including tumor size and nodal status are changed by neoadjuvant chemotherapy [3]. Hence we tried to determine the additional predictive and prognostic markers for early relapse other than pCR in neoadjuvant setting.

In the present study, we found that a triple negative phenotype was a predictive marker for response in neoadjuvant docetaxel and doxorubicin chemotherapy. In addition, initial clinical stage, hormone receptor, histologic grade, bcl-2 and Ki-67 were all associated with RFS. In other published studies using non-anthracycline based chemoendocrine agents [11], it was reported that positive ER, absence of c-erbB2 and decrease in Ki-67 were associated with a good clinical response. Overexpression of p53 was also reported to be associated with a lower response rate to anthracycline based neoadjuvant chemotherapy [13,14,25] and to be an independent factor for poor survival [14,25]. In our results, overexpression of p53 failed to show clinical significance in neoadjuvant setting. However, p53 mutation which was associated with response to neoadjuvant chemotherapy [13] was not in agreement with p53 overexpression measured by quantitative immunohistochemistry. Additional mutational study of p53 is needed to clarify correlation between p53 and clinical outcomes. The predictive or prognostic value of bcl-2, apoptosis regulatory protein, remains controversial in neoadjuvant setting. In one study, higher bcl-2 expression was predictive for pCR [26], while other studies did not find any correlation between bcl-2 expression and clinical response [25,27]. Traditional prognostic makers such as nodal stage [28] and c-erbB2 [10,11] showed no prognostic value in our result. Relatively short follow up period of 18.6 months might partially explain this. As yet, these biologic markers are inconclusive, owing to heterogeneous chemotherapeutic regimens and the small sample size of extant studies. More studies should be carried out, to identify more precisely the prognostic markers in the neoadjuvant setting.

In our results, pCR which is considered to be the most powerful prognostic factor did not show significant prognostic value. Possible explanations for the weakened prognostic power of pCR are the relatively lower rate of pCR (7.6%), the short course of neoadjuvant chemotherapy, and the short duration of follow up (18.6 months). We conducted only three cycles of neoadjuvant chemotherapy, while other neoadjuvant regimens have been based on four to six cycles, and have shown higher pCR rates (8–26%) than our own study [24,29,30].

Optimal treatment after neoadjuvant chemotherapy remains still uncertain [31]. Unlike early breast cancer, it is not yet clear whether adjuvant therapy should be conducted according to initial clinical stage or post operative pathologic stage. In our homogeneous patient population, initial clinical stage was an independent prognostic factor for survival, while pathologic stage failed to reflect ultimate survival. This result was obtained by using accurate staging work up modalities, including breast MRI and chest computed tomography. In contrast, Carey et al [32] analyzed 135 patients with median follow up of 5 years and reported that pathologic stage after neoadjuvant chemotherapy was useful for predicting survival. Chollet et al [33] also reported prognostic value of residual tumor size and nodal status after neoadjuvant chemotherapy with median follow up of 9.3 year. However, despite short duration of follow up, our results showed statistical superiority of initial clinical stage in predicting survival. This result might give us useful information when determining post operative adjuvant therapy.

Triple negative breast cancer has been reported as being associated with a poor clinical outcome in early breast cancer [34,35]. In locally advanced breast cancer, there are limited data about response to chemotherapy and survival. In the present study, we found that triple negative breast cancer responded to neoadjuvant chemotherapy initially but then relapsed rapidly. Generally, tumor responsiveness to chemotherapy is believed to be associated with a longer survival. However, in triple negative phenotype, tumor responsiveness did not affect prolonged survival. In contrast, non-triple negative breast cancer did not show a marked response but progressed rather slowly. This paradoxical feature is consistent with other studies conducted in basal-like breast cancer, which was identified using gene expression profiling [36,37]. In the present study, we did not conduct gene expression profiling and hierachial cluster analysis. However, it is known that 80–90% of triple negative breast cancers by immunohistochemistry are basal-like subtypes by gene expression profiling [35] and have a similar clinical behavior, in addition [38,39]. It is notable that we were able to obtain useful predictive and prognostic information by simple immunohistochemistry without high cost.

It is not yet certain whether the poor prognosis of triple negative breast cancer is due to its aggressive feature or because of lack of targeted therapy, including adjuvant hormonal therapy and c-erbB2 targeted agents. We hypothesized that triple negative breast cancer itself seems to reflect more aggressive tumor biology and growth rate potential with high expression of Ki-67. Our data suggest that patients with triple negative breast cancer should be candidates for clinical trials to determine additional agents including antiangiogenic agents.

Conclusion

Several molecular markers play a role as predictive and prognostic factors in stage II and III breast cancer patients receiving neoadjuvant chemotherapy. We also confirmed the usefulness of initial clinical stage, as a predictor of survival. We found that triple negative phenotype was associated with shorter survival, even though it was associated with a higher response rate to neoadjuvant chemotherapy. These results might assist in identifying and understanding the importance of clinically useful markers in the neoadjuvant setting, and help to optimize treatments.