Meta-analysis confirms achieving pathological complete response after neoadjuvant chemotherapy predicts favourable prognosis for breast cancer patients

EUROPEAN JOURNAL OF CANCER

4 7 ( 2 0 1 1 ) 2 0 8 4 –2 0 9 0

available at www.sciencedirect.com

journal homepage: www.ejconline.com

Review

Meta-analysis confirms achieving pathological complete response after neoadjuvant chemotherapy predicts favourable prognosis for breast cancer patients Xiangnan Kong a, Meena S. Moran b, Ning Zhang a, Bruce Haffty c, Qifeng Yang

a,c,*

a

Department of Breast Surgery, Qilu Hospital, Shandong University, School of Medicine, Wenhua West Road No. 107, Ji’nan, Shandong 250012, PR China b Department of Therapeutic Radiology, Yale University School of Medicine, New Haven, CT, USA c Department of Radiation Oncology, UMDNJ-Robert Wood Johnson School of Medicine and the Cancer Institute of New Jersey, New Brunswick, NJ, USA

A R T I C L E I N F O

A B S T R A C T

Article history:

Neoadjuvant chemotherapy (NAC) has become a widely accepted method of sequencing

Received 4 January 2011

systemic therapy for breast cancer treatment. While ‘response to chemotherapy’ in the

Received in revised form 7 June 2011

neoadjuvant setting has been utilised to predict prognosis, the published data are inconsis-

Accepted 7 June 2011

tent. The present meta-analysis was conducted to determine whether the pathologic

Available online 5 July 2011

response to NAC predicts for outcomes. Papers were selected from the PubMed database based on defined inclusion and exclusion criteria. Parameters such as number/percentage

Keywords:

of patients having pCR and outcome statistics (i.e. overall survival (OS), disease-free

Breast cancer

survival (DFS), relapse-free survival (RFS)) were collected. The analysis included 16 studies

Neoadjuvant chemotherapy

with 3776 patients. The summary odds ratio (OR) estimating the association of OS with pCR

Pathological complete response

was 3.44 (95% confidence interval [95%CI]: 2.45–4.84), with similar findings for DFS

Prognosis

(OR = 3.41, 95%CI: 2.54–4.58) and RFS (OR = 2.45, 95%CI: 1.59–3.80). No obvious statistical

Meta-analysis

heterogeneity was detected. Funnel plots and Egger’s tests did not reveal publication bias.

Prognostic indicator

This meta-analysis confirms that pathologic response is a prognostic indicator for RFS, DFS and OS and suggests that patients achieving pCR after NAC have favourable outcomes. Ó 2011 Elsevier Ltd. All rights reserved.

1.

Introduction

Invasive breast cancers are one of the most common types of malignancies in women internationally and are generally considered chemo-sensitive tumours. While the historic standard for treatment has been surgery followed by systemic therapy,1 more recently, chemotherapy is being delivered in the neoadjuvant setting prior to definitive surgery and is now considered a widely accepted option of sequencing.

Neoadjuvant chemotherapy (NAC) was first described for patients with locally advanced breast cancer (LABC) in the late 1970’s.2 In recent years, the use of NAC has been extended to earlier-stage breast cancers,2,3 with the intention to downstage the tumour and potentially enabling breast conserving surgery (BCS) for patients who may have otherwise required mastectomy. These objectives have been substantiated in many studies.3–7 It is postulated that NAC may diminish the potential of micrometastases in breast cancer patients,

* Corresponding author: Address: Department of Breast Surgery, Qilu Hospital, Shandong University, School of Medicine, Wenhua West Road No. 107, Ji’nan, Shandong 250012, PR China. Tel./fax: +86 531 82169268. E-mail address: [email protected] (Q. Yang). 0959-8049/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejca.2011.06.014

EUROPEAN JOURNAL OF CANCER

which is supported by animal model data that suggest the enhanced survival is related to the improved control of micrometastases.8,9 In the clinical setting, NAC appears to be equivalent to, if not better than, systemic therapy delivered adjuvantly after surgery. Furthermore, the use of NAC improves the rates of BCS for patients wanting to preserve their breast(s).3–7 However, standard prognostic indicators such as number of involved lymph nodes or size of tumour at surgical resection are no longer applicable in the neoadjuvant setting, when systemic therapy often downstages and potentially eradicates the disease. We now are learning that the response of tumour to NAC may be prognostic for outcomes; for example, patients who achieve a pathologic complete response (pCR) after NAC may have better overall survival compared with matched patients having only a partial pathological response (pPR).10 Unfortunately, there are limited high quality prospective data and the current publications have not been consistent in their conclusions. Whereas some publications suggest that achieving a pCR improves survival rates; others report no significant difference in survival between complete and partial responders.11 Identification of prognostic variables for patients receiving NAC is critical to breast cancer multi-disciplinary management, to determine if additional therapy is warranted. With strong data to support the correlation of outcomes with clinical-pathologic features in the neoadjuvant setting, we may be able to avoid additional post-surgical interventions in a patient who was otherwise high risk at presentation. Thus, pinpointing specific prognostic factors will be very clinically meaningful for physicians treating breast cancer. The aim of current meta-analysis was to analyse the existing data to determine if pathologic response predicts for outcomes in breast cancer patients who have received NAC.

2.

Materials and methods

2.1.

Literature search

Eligible articles were searched in the PubMed database from January 1985 to November 2010. Publications with the following search words in the title, abstract or keywords were included: breast cancer; neoadjuvant chemotherapy; preoperative chemotherapy; pathological complete remission; pathological complete response; pathologic response; prognosis; outcome; survival. Only studies written in the English language were included.

2.2.

4 7 ( 2 0 1 1 ) 2 0 8 4 –2 0 9 0

2085

ies in which additional post-surgical adjuvant treatment was delivered (i.e. radiotherapy) were excluded so that adjuvant treatment factors did not confound results. Lastly, any investigations which included data from other published studies were excluded.

2.3.

Data extraction

The following information was recorded on each eligible trial: authors’ names, year of publication, duration of follow-up, total number of patients, number and percentages of patients achieving pCR and pPR, systemic therapy regimen delivered, primary TNM stage, survival data (percentages and number of events) and results of toxicity parameters reported. If the authors found a correlation between response and outcome, this finding was also recorded.

2.4.

Statistical analysis

Summary statistics of patients achieving pCR after NAC with ORs and 95% CI were calculated using the fixed effect model/ Mantel–Haenszel method when there was minimal heterogeneity in the variables amongst studies and the DerSimonian– Laird method (random effect model) when there was significant heterogeneity. Each publication was weighted according to the size of the sample. The v2 and I2 test methods were utilised for the between-study heterogeneity of the ORs. Statistically significant differences were defined as <0.05 for v2 P and greater than 50% for the I2 test. Forest plots were generated using standard techniques to summarise the included studies with horizontal lines representing 95% CI, the area of each square representing the weighting and the positions of each square demonstrating the OR point estimate. The overall summary estimate under fixed effects with its CI was shown. The vertical line was at the null value (OR = 1.0). Funnel plots and Egger’s tests were created using standard techniques for detecting publication bias. The funnel plots displayed the OR associated with pCR in each study. Each OR was reported on a log scale against its standard error. The vertical line indicated the pooled estimate of the overall OR, with the sloping lines representing the expected 95% CI for a given SE. All statistical analyses were performed using Review Manager (version 5.0.25 for Windows; Cochrane Collaboration, Oxford) and Stata (version 11; Stata Corporation, College Station, TX, USA).

3.

Results

3.1.

Eligible studies

Inclusion and exclusion criteria

All of the following criteria had to be met for a publication to be included in this analysis: (a) NAC had to be delivered before loco regional therapy of breast cancer; (b) detailed statistics had to be reported for both complete and partial responders (i.e. exact patient numbers and percentage of pCR and pPR, detailed outcomes such as percentages of overall survival [OS], disease-free survival [DFS] or relapse-free survival [RFS], in addition to p values); (c) the definition of what constituted a pCR had to be explicit. Publications that did not have all three criteria were excluded from analysis in order to avoid significant heterogeneity between studies. Furthermore, stud-

Results of the literature search identified 317 papers based on the search words. Of these, 36 were potentially eligible studies based on title and abstract review, which mentioned identifying correlation of pCR and prognosis. Further critical review of each of these 36 publications excluded 11 studies for inclusion/summarisation of other studies within their manuscript, and 9 studies for concluding pCR may be associated with better prognosis without providing adequate, detailed statistics. Thus, based on the criteria described above, 16 publications were eligible for inclusion in this meta-analysis.5,10–24

2086

3.2.

EUROPEAN JOURNAL OF CANCER

Study characteristics

Ultimately, a total of 3776 cases from the 16 publications which had pathological results and survival data were included in this meta-analysis. Characteristics of these studies are detailed in Table 1. Based on these studies, the percentage of patients achieving pCR after NAC from 1998 to 2010 ranged from 10% to 25%, excluding some with only dozens of patients. All systemic therapy delivered typically consisted of an anthracycline and/or taxane with other chemotherapeutic agents, in accordance with the standard of care for NAC.

3.3.

Summary percentage of achieving pCR after NAC

From the 16 studies and 3776 patients included in this analysis, 647 were reported to achieve pCR. The aggregate percentage of pCR was, therefore, 17.1%.

3.4.

pCR and survival results

4 7 ( 2 0 1 1 ) 2 0 8 4 –2 0 9 0

studies, 3 had small sample sizes; when excluding these 3, the OR ranged from 1.12 to 7.56. The overall summary estimate OR was 3.44 (95% CI: 2.45–4.84). Heterogeneity testing revealed I2 = 43% and v2 = 17.43, p = 0.07. Twelve studies (n = 2683 patients) provided DFS data.5,11–15,17–20,22,24 The OR and 95% CI for each study and the summary OR are shown in Fig. 2. The ORs from each of the 12 studies ranged from 1.31 to 17.29. Excluding the 1 study with small sample size, the remaining 11 studies had ORs ranging between 1.31 and 6.97. The overall summary estimate OR was 3.41 (95% CI: 2.54–4.58), with no obvious evidence of heterogeneity (I2 = 34%, v2 = 16.63, p = 0.12). Four articles (n = 1341 cases) provided RFS data.5,10,21,23 The OR and 95% CI for each study and the summary OR are shown in Fig. 3. The ORs of the 4 studies were ranged from 0.81 to 3.98. The overall summary estimate OR was 2.45 (95% CI: 1.59–3.80), with no significant evidence of heterogeneity (I2 = 35%, v2 = 4.64, p = 0.20).

3.5.

Only 11 of the 16 studies provided OS data (n = 3182 patients).5,10–13,15–17,19,21,24 The OR and 95% CI for each study and the summary OR are shown in Fig. 1. The individual ORs of the 11 articles ranged from 1.12 to 19.58. Of these 11

cCR and survival analysis

Three studies provided data of clinical CR and survival. Huang 200919 and Ezzat 200415 provided OS and DFS data, and Chang 201023 gave RFS result. Due to the lack of sufficient data, these statistical analyses could not be completed.

Table 1 – Characteristics of the eligible studies. Study

Duration of follow-up

No. of Neoadjuvant patients chemotherapy treatment

TNM stage

Percentage of pCR

survival data OS DFS RFS

Fisher et al.5 Honkoop et al.12 Kuerer et al.13

Overall 60 months Median 32 months Median 58 months

743ª 42 372

Anthracycline + others Anthracycline + others Anthracycline + others

Dieras et al.14 Ezzat et al.15 Ring et al.16 Abrial et al.17 Andre et al.10

Median Median Median Median Median

200b 126c 439d 710e 534

Toi et al.18

Median 40 months

202f

Huang et al.19 Eralp et al.11 Frasci et al.20

Median 63.4 months Median 43 months Median 41 months

119g 110 74

Sikov et al.21 Jung et al.22

Median 28 months Median 61.5 months

55h 66

Chang et al.23 Median 22.8 months Al-Tweigeri et al.24 Overall 60 months

74i 59

Anthracycline Taxane + others Anthracycline + others Anthracycline + taxane Anthracycline + taxane + others Anthracycline + taxane + others Anthracycline + others Anthracycline + taxane Anthracycline + taxane + others Taxane + others Anthracycline + taxane + others Taxane + others Anthracycline + taxane + others

a

31 months 37.5 months 53 months 91.2 months 31.2 months

Not reported IIIA, IIIB IIA, IIB, IIIA, IIIB, IV Not reported IIA, IIB, IIIA, IIIB I, II, III II, III I, II, III

12% 54.8% 12%

X X X

14.6% 23.6% 12% 13% 19%

X X X X

Not reported

25%

IIB, IIIA, IIIB, IIIC II, III Not reported

18.5% 11.7% 62%

X X

Not reported II, III

45% 15.2%

X

Not reported Not reported

26.8% 24%

In Fisher et al. (1998)5 study, only 683 patients had pathological results that can ascertain whether achieving pCR. In Dieras et al. (2004)14 study, only 192 patients had pathological results that can ascertain whether achieving pCR. c In Ezzat et al. (2004)15 study, only 123 patients had pathological results that can ascertain whether achieving pCR. d In Ring et al. (2004)16 study, 4 patients were lost during follow-up. e In Abrial et al. (2005)17 study, only 656 patients had pathological results that can ascertain whether achieving pCR f In Toi et al. (2008)18 study, 191 patients had pathological results and survival statistics. g In Huang et al. (2009)19 study, only 115 patients had pathological results that can ascertain whether achieving pCR. h In Sikov et al. (2009)21 study, only 53 patients had pathological results that can ascertain whether achieving pCR. i In Chang et al. (2010)23 study, only 71 patients had pathological results that can ascertain whether achieving pCR. b

X X X

X

X X X X X X X X X X X

X

X

EUROPEAN JOURNAL OF CANCER

4 7 ( 2 0 1 1 ) 2 0 8 4 –2 0 9 0

2087

Fig. 1 – Forest plots of odds ratio (OR) for overall survival (OS) from eligible studies of breast cancer associated with achieving pCR. Horizontal lines represent 95% CI. The area of each square represents the weighting and the positions of each square demonstrate the OR point estimate.

Fig. 2 – Forest plots of odds ratio (OR) for disease-free survival (DFS) from eligible studies of breast cancer associated with achieving pCR. Horizontal lines represent 95% CI. The area of each square represents the weighting and the positions of each square demonstrate the OR point estimate.

3.6.

Toxicity analysis

All 16 eligible studies investigated the toxicity of NAC. After receiving NAC, there were no treatment-related deaths, and almost no cardiac toxicity has been reported. The three most common haematologic toxicities were neutropenia, anaemia and thrombocytopenia. And the most common non-haematologic toxicity was nausea and vomiting. Alopecia occurred in the vast majority of patients.

3.7.

Publication bias

Funnel plots were generated to detect for potential publication bias, and are shown in Figs. 4 and 5. In order to diminish

the subjective bias, Egger’s tests were also performed to detect publication bias. Results of Egger’s tests showed that there was no statistical publication bias amongst the 3 studies used for analysing pCR and survival data (Egger’s test p-values = 0.811, 0.297 and 0.758, respectively).

4.

Discussion

NAC is being increasingly utilised for breast cancer to downstage tumours, increase resectability, and potentially make a mastectomy-requiring patient eligible for breast conserving surgery. Despite these benefits, an inherent simultaneous impediment with this approach is the inability to utilise

2088

EUROPEAN JOURNAL OF CANCER

4 7 ( 2 0 1 1 ) 2 0 8 4 –2 0 9 0

Fig. 3 – Forest plots of odds ratio (OR) for relapse-free survival (RFS) from eligible studies of breast cancer associated with achieving pCR. Horizontal lines represent 95% CI. The area of each square represents the weighting and the positions of each square demonstrate the OR point estimate. traditional pathologic features such as number of lymph nodes involved, margin status, primary tumour size etc. for prognostic purposes. Therefore, a different set of prognostic markers need to be defined and validated for patients receiving neoadjuvant chemotherapy, since accurate prediction of prognosis is critical in guiding additional adjuvant treatment needs. A recent publication from the National Cancer Institute Conference Statement on neoadjuvant chemotherapy suggests that initial T/N stage at presentation and response to neoadjuvant chemotherapy should be utilised to guide further loco regional and possibly systemic therapy.25 But as noted by the authors, data from high quality prospective studies are limited. While currently, many surgical and radiation management decisions for breast cancer patients treated with NAC are based in part on the extent of the disease at diagnosis, additional prognostic indicators from the post-chemotherapy period need to be identified and validated before consensus recommendations can be made, and treatment can be tailored to individual patients. Currently, there remains discordance in the published literature regarding whether achieving a pCR is prognostic for outcomes after NAC. Many of the existing studies are small with limited follow-up. Thus, the primary goal of

this meta-analysis was to investigate whether achieving pCR correlates to outcomes in breast cancer patients receiving NAC, and to further elucidate which specific outcome parameters the pathologic response may predict for. Based on the combined data of 16 eligible published studies, our findings suggest that a pCR is prognostic for OS (OR = 3.44, 95%CI: 2.45–4.84), DFS (OR = 3.41, 95%CI: 2.54–4.58) and RFS (OR = 2.45, 95% CI: 1.59– 3.80). Of the studies included in our analysis, we did not find significant statistical publication bias. However, it is important to note that this finding may be influenced by studies that were excluded from this analysis and were not accounted for. Additionally, while we did not find any obvious statistical heterogeneity amongst the studies included in this analysis, there were some notable clinical differences among the eligible studies (i.e. types of regimens used, TNM stages, molecular subtypes of the patients included, and duration of follow-up) which warrant further discussion. The NAC agents varied amongst the included studies but consistently, the vast majority of patients received a regimen that consisted of an anthracycline and/or taxane with other agents, and followed the NCCN Clinical Practice Guidelines in Oncology-Breast Cancer (V.3.2010). There are multiple

Fig. 4 – Funnel plot detect publication bias in the study of relationship between pCR and overall survival (OS). Each OR was reported on a log scale against its standard error. The vertical line indicated the pooled estimate of the overall OR, with the sloping lines representing the expected 95% CI for a given SE.

Fig. 5 – Funnel plot detect publication bias in the study of relationship between pCR and disease-free survival (DFS). Each OR was reported on a log scale against its standard error. The vertical line indicated the pooled estimate of the overall OR, with the sloping lines representing the expected 95% CI for a given SE.

EUROPEAN JOURNAL OF CANCER

studies that suggest that differences in the chemotherapy agents utilised in the neoadjuvant setting affect pCR rates; for example, the addition of a taxane to anthracycline-based neoadjuvant regimen has resulted in higher pCR rates and a survival benefit.26 Furthermore, studies assessing the frequency, dosing and sequencing in the NAC setting suggest that these factors may also play a role in the pCR rates.27,28 Though we were not able to assess the effects of the varying regimens on outcomes, future analysis of these parameters would be a worthwhile future endeavour. Another consideration of this dataset is the various stages of diseases that were included, which may potentially lead to small, qualitative heterogeneity. Most of the trials had varying stages of disease at presentation and ideally, investigating the complete response rates by stage would allow for us to draw more powerful conclusions; unfortunately, we were limited in the (lack of) detailed reported data of pathologic responses as a function of stage. Furthermore, there was variability in follow-up amongst studies, with many publications reporting less than 5-year (often estimated) outcomes. Given the natural history of invasive breast cancers, longer follow-up is required. There are now several published series evaluating the relationship between molecular subtype and pCR after NAC. Compared to patients with hormone receptor positive disease, patients with hormone receptor negative disease, human epidermal growth factor receptor-2 positive (HER2+) tumours and triple negative tumours who receive neoadjuvant chemotherapy appear to have much higher rates of pCR.29 These data cohesively suggest that underlying molecular subtypes likely play a role in pathologic response rates. Therefore, molecular subtyping is yet another parameter that was not consistently reported across all studies and was not accounted for in this meta-analysis. In summary, while comparison of outcomes between trials and individual series is difficult because of the heterogeneity of the study design and patient populations, our meta-analysis, conducted with adherence to stringent inclusion and exclusion criteria, suggests that achieving a complete pathologic response following NAC is indeed a predictor of overall survival, disease-free survival and relapse-free survival.

5.

Conclusions

The main purpose of our study was to pool existing data to determine whether pathologic response after neoadjuvant chemotherapy predicts for outcomes. We conclude that the pathologic response is prognostic for relapse-free, disease-free and overall survival, but were unable to draw any conclusions regarding the prognostic value of clinical response due to the lack of sufficient data. As we proceed through this new era of individualised treatment, identification of clinically relevant prognostic variables in the neoadjuvant setting will be critical to delivering tailored therapy, improving upon outcomes and decreasing toxicity. Additional prospective, controlled studies of NAC assessing long-term outcomes are warranted.

Conflict of interest statement None declared.

4 7 ( 2 0 1 1 ) 2 0 8 4 –2 0 9 0

2089

Acknowledgements This Project is supported by Programme for New Century Excellent Talents in University, Key Project of Chinese Ministry of Education (No. 108080), National Natural Science Foundation of China (Nos. 30772133, 81072150) and Independent Innovation Foundation of Shandong University (IIFSDU, No. 2009JQ007) to Prof. Q. Yang.

R E F E R E N C E S

1. He W. Observation of therapeutic effectiveness of neoadjuvant chemotherapy in breast cancer. Chinese Community Doctors 2010;24(12):79. 2. De Lena M, Zucali R, Viganotti G, Valagussa P, Bonadonna G. Combined chemotherapy-radiotherapy approach in locally advanced (T3b–T4) breast cancer. Cancer Chemother Pharmacol 1978;1(1):53–9. 3. Beasley GM, Olson Jr JA. What’s new in neoadjuvant therapy for breast cancer? Adv Surg 2010;44:199–228. 4. Fisher B, Brown A, Mamounas E, et al. Effect of preoperative chemotherapy on local-regional disease in women with operable breast cancer: findings from National Surgical Adjuvant Breast and Bowel Project B-18. J Clin Oncol 1997;15(7):2483–93. 5. Fisher B, Bryant J, Wolmark N, et al. Effect of preoperative chemotherapy on the outcome of women with operable breast cancer. J Clin Oncol 1998;16(8):2672–85. 6. Gianni L, Baselga J, Eiermann W, et al. Feasibility and tolerability of sequential doxorubicin/paclitaxel followed by cyclophosphamide, methotrexate, and fluorouracil and its effects on tumor response as preoperative therapy. Clin Cancer Res 2005;11(24 Pt 1):8715–21. 7. Wolmark N, Wang J, Mamounas E, Bryant J, Fisher B. Preoperative chemotherapy in patients with operable breast cancer: nine-year results from National Surgical Adjuvant Breast and Bowel Project B-18. J Natl Cancer Inst Monogr 2001;30:96–102. 8. Fisher B, Saffer E, Rudock C, Coyle J, Gunduz N. Effect of local or systemic treatment prior to primary tumor removal on the production and response to a serum growth-stimulating factor in mice. Cancer Res 1989;49(8):2002–4. 9. Bear HD, Anderson S, Smith RE, et al. Sequential preoperative or postoperative docetaxel added to preoperative doxorubicin plus cyclophosphamide for operable breast cancer: National Surgical Adjuvant Breast and Bowel Project Protocol B-27. J Clin Oncol 2006;24(13):2019–27. 10. Andre F, Mazouni C, Liedtke C, et al. HER2 expression and efficacy of preoperative paclitaxel/FAC chemotherapy in breast cancer. Breast Cancer Res Treat 2008;108(2):183–90. 11. Eralp Y, Smith TL, Altundag K, et al. Clinical features associated with a favorable outcome following neoadjuvant chemotherapy in women with localized breast cancer aged 35 years or younger. J Cancer Res Clin Oncol 2009;135(1):141–8. 12. Honkoop AH, van Diest PJ, de Jong JS, et al. Prognostic role of clinical, pathological and biological characteristics in patients with locally advanced breast cancer. Br J Cancer 1998;77(4):621–6. 13. Kuerer HM, Newman LA, Smith TL, et al. Clinical course of breast cancer patients with complete pathologic primary tumor and axillary lymph node response to doxorubicinbased neoadjuvant chemotherapy. J Clin Oncol 1999;17(2):460–9.

2090

EUROPEAN JOURNAL OF CANCER

14. Dieras V, Fumoleau P, Romieu G, et al. Randomized parallel study of doxorubicin plus paclitaxel and doxorubicin plus cyclophosphamide as neoadjuvant treatment of patients with breast cancer. J Clin Oncol 2004;22(24):4958–65. 15. Ezzat AA, Ibrahim EM, Ajarim DS, et al. Phase II study of neoadjuvant paclitaxel and cisplatin for operable and locally advanced breast cancer: analysis of 126 patients. Br J Cancer 2004;90(5):968–74. 16. Ring AE, Smith IE, Ashley S, Fulford LG, Lakhani SR. Oestrogen receptor status, pathological complete response and prognosis in patients receiving neoadjuvant chemotherapy for early breast cancer. Br J Cancer 2004;91(12):2012–7. 17. Abrial SC, Penault-Llorca F, Delva R, et al. High prognostic significance of residual disease after neoadjuvant chemotherapy: a retrospective study in 710 patients with operable breast cancer. Breast Cancer Res Treat 2005;94(3):255–63. 18. Toi M, Nakamura S, Kuroi K, et al. Phase II study of preoperative sequential FEC and docetaxel predicts of pathological response and disease free survival. Breast Cancer Res Treat 2008;110(3):531–9. 19. Huang O, Chen C, Wu J, et al. Retrospective analysis of 119 Chinese noninflammatory locally advanced breast cancer cases treated with intravenous combination of vinorelbine and epirubicin as a neoadjuvant chemotherapy: a median follow-up of 63.4 months. BMC Cancer 2009;9:375. 20. Frasci G, Comella P, Rinaldo M, et al. Preoperative weekly cisplatin-epirubicin-paclitaxel with G-CSF support in triplenegative large operable breast cancer. Ann Oncol 2009;20(7):1185–92. 21. Sikov WM, Dizon DS, Strenger R, et al. Frequent pathologic complete responses in aggressive stages II to III breast cancers with every-4-week carboplatin and weekly paclitaxel

4 7 ( 2 0 1 1 ) 2 0 8 4 –2 0 9 0

22.

23.

24.

25.

26.

27.

28.

29.

with or without trastuzumab: a Brown University Oncology Group Study. J Clin Oncol 2009;27(28):4693–700. Jung SY, Kim SK, Nam BH, et al. Prognostic Impact of [18F] FDG-PET in operable breast cancer treated with neoadjuvant chemotherapy. Ann Surg Oncol 2010;17(1):247–53. Chang HR, Glaspy J, Allison MA, et al. Differential response of triple-negative breast cancer to a docetaxel and carboplatinbased neoadjuvant treatment. Cancer 2010;116(18):4227–37. Al-Tweigeri TA, Ajarim DS, Alsayed AA, et al. Prospective phase II study of neoadjuvant doxorubicin followed by cisplatin/docetaxel in locally advanced breast cancer. Med Oncol 2010;27(3):571–7. Buchholz TA, Lehman CD, Harris JR, et al. Statement of the science concerning locoregional treatments after preoperative chemotherapy for breast cancer: a National Cancer Institute conference. J Clin Oncol 2008;26(5):791–7. Heys SD, Sarkar T, Hutcheon AW. Primary docetaxel chemotherapy in patients with breast cancer: impact on response and survival. Breast Cancer Res Treat 2005;90(2):169–85. von Minckwitz G, Raab G, Caputo A, et al. Doxorubicin with cyclophosphamide followed by docetaxel every 21 days compared with doxorubicin and docetaxel every 14 days as preoperative treatment in operable breast cancer: the GEPARDUO study of the German Breast Group. J Clin Oncol 2005;23(12):2676–85. Untch M, Mobus V, Kuhn W, et al. Intensive dose-dense compared with conventionally scheduled preoperative chemotherapy for high-risk primary breast cancer. J Clin Oncol 2009;27(18):2938–45. Liedtke C, Mazouni C, Hess KR, et al. Response to neoadjuvant therapy and long-term survival in patients with triple-negative breast cancer. J Clin Oncol 2008;26(8):1275–81.