AbstractIntroduction The identification and validation of biomarkers of chemotherapy sensitivity is critical in order to individual-ise therapy in breast cancer. We evaluated pathological com-plete response (pCR) to GAT, and its correlation with tumour biomarkers before and after neoadjuvant chemotherapy.Materials and methods Stage III (and stage II with T≥5 cm) breast cancer patients were included. Treatment consisted of adriamycin (40 mg/m2) day 1, and paclitaxel (150 mg/m2) followed by gemcitabine (2000 mg/m2) day 2, every 14 days for six cycles. Tissue from pre-treatment biopsy and surgery was evaluated for biologic markers by immunohis-tochemistry. Two XPD single nucleotide polymorphisms (SNP) were also analysed.Results Forty-six patients entered the trial. Median age was 49.5 years (range 31–72); 25 patients (54%) were pre-menopausal; 12 (26%) were ER-PgR-negative; pCR was observed in 17% (95% CI: 6.4–28.4) of patients. Sig-nificant differences in marker expression (mean±SD) in correlation to pathological response were only found in Ki-67. After treatment, tumours showed lower Ki-67-, surviv-ing- and pERK-positive cells. No correlation between XPD polymorphisms and pCR was found. The overall response rate was 89% (95% CI: 80.1–98.1). Fifteen patients (33%) underwent breast-conserving surgery. The most frequent

grade 3 or 4 toxicities were neutropenia (with one febrile neutropenia) and asthenia. Conclusion These results show an effective regimen with acceptable tolerability. Our data suggest that not only clas-sical markers (ER, Ki-67), but also survivin and pERK could be involved in the response to GAT, which may con-tribute to therapy individualisation in future study designs.

Keywords Breast cancer · Neoadjuvant chemotherapy · Adriamycin · Gemcitabine · Paclitaxel · Biomarker

Introduction

Neoadjuvant chemotherapy in patients with locally ad-vanced breast cancer results in satisfactory survival rates, while increasing the number of women undergoing conser-vative surgery [1–3]. Patients achieving pathological com-plete response (pCR) may have a longer disease-free and overall survival [4]. However, pCR rates generally do not exceed 30% [5–7]; only biological therapy in combination with chemotherapy, like trastuzumab, shows higher pCR rates [8]. Currently, the standard neoadjuvant treatment is a combination of anthracyclines and taxanes, however a definitive schedule has not been established yet; combina-tions or sequential administration with new drugs are being tested. New trials are ongoing to identify and validate bio-markers of chemotherapy sensitivity, and to better define the individual risk of relapse. Negative hormonal receptor status and greater proliferative activity are the parameters more consistently associated with high chemotherapy sen-sitivity. Nevertheless, the strength of this association is not sufficient to discriminate patients at different risk levels to provide individualised therapy.

The affiliations are listed at the end of the article

P. Sánchez-Rovira (Y)Medical Oncology DepartmentJaén Hospital ComplexAvda. del Ejército Español, 10ES-23007 Jaén, Spaine-mail: pedro.sanchez.rovira.sspa@juntadeandalucia.es

Clin Transl Oncol (2012) 14:430-436DOI 10.1007/s12094-012-0820-4

R E S E A R C H A R T I C L E S

Classical markers like ER and ki-67, but also survivin and pERK, could be involved in the pathological response to gemcitabine, adriamycin and paclitaxel (GAT) in locally advanced breast cancer patients: results from the GEICAM/2002-01 phase II study

Pedro Sánchez-Rovira · Antonio Antón · Agustí Barnadas · Amalia Velasco · María Lomas · María Rodríguez-Pinilla · José Luis Ramírez · César Ramírez · María José Ríos · Eva Castellá · Carmen García-Andrade · Belén San Antonio · Eva Carrasco · José Luis Palacios

Received: 4 August 2011 / Accepted: 17 September 2011

Clin Transl Oncol (2012) 14:430-436 431

Many DNA-damaging chemotherapy agents induce the formation of DNA lesions (adducts). Single nucleotide polymorphisms (SNPs) can impair the DNA repair mecha-nisms involved in removing DNA adducts. Understand-ing the correlation between DNA repair genotypes and response/survival in cancer patients may help to elucidate how certain polymorphic variants can influence chemo-therapy outcome. The neoadjuvant chemotherapy model is the ideal clinical setting to validate these concepts and to design tailored therapies.

We designed a phase II trial to evaluate the activity (pCR) of a combination with gemcitabine, adriamycin and paclitaxel (GAT), and the relationship between tumour bio-logical profile, patient genotypes and response to treatment.

The primary objective of this trial was to determine pCR rate after 6 cycles of GAT. Secondary objectives were to assess the correlation between several tumour markers and pCR; to evaluate changes in tumour marker profile be-fore and after treatment; to determine overall response rate (ORR); to characterise drug-related toxicities; and to assess the breast-conserving surgery rates and disease-free survival.

Materials and methods

Patient selection

Patients (18–75 years) with previously untreated, histologi-cally confirmed (by surgical or core biopsy) stage III breast cancer (or stage II with tumours ≥5 cm) were included. Pa-tients were required to have Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 or 1, confirmed normal cardiac function, and adequate bone marrow re-serve, hepatic and renal functions. Patients were excluded if they had inflammatory or metastatic breast cancer or any other severe or uncontrolled systemic disease. Adequate contraception and a negative pregnancy test were required for women of child-bearing potential. Patients with a previ-ous history of auricle-ventricular arrhythmia, congestive heart failure, second- or third-degree active heart block, or myocardial infarction within 6 months prior to the study entry were also excluded. This trial was approved by the local ethical review boards and the Spanish Ministry of Health, and was conducted in compliance with Good Clini-cal Practices and the tenets of the Declaration of Helsinki. All patients provided written informed consent before en-tering the study.

Treatment plan

Neoadjuvant chemotherapy consisted of adriamycin 40 mg/m2, administered on Day 1 as a 15-min intravenous (i.v.) infusion. Paclitaxel 150 mg/m2 was administered on Day 2 as a 3-h i.v. infusion followed by gemcitabine 2000 mg/m2 as a 60-min i.v. infusion, every two weeks for 6 cycles. Patients without progression underwent mastectomy or breast-conserving surgery plus axillary dissection accord-ing to the surgeon’s criteria.

Treatment with granulocyte colony stimulating factor (G-CSF) was allowed in the event of febrile neutropenia, grade 4 neutropenia lasting ≥5 days or infection during neutropenia. Secondary prophylaxis with G-CSF was al-lowed in the event of cycle delay due to toxicity. Within a cycle, therapy was delayed for a maximum of two weeks until neutrophils and platelets achieved established thresh-olds, and non-haematological toxicities (except for alope-cia and nausea/vomiting) returned to grade 1.

For subsequent cycles, a 25% dose reduction of the three drugs was made for prolonged neutropenia (neu-trophils <0.5×109/l lasting >5 days), febrile neutropenia, thrombocytopenia with haemorrhage, cycle delay lasting >1 week due to haematological toxicity and grade 3 non-haematological toxicity (except alopecia and nausea/vom-iting). GAT was discontinued for grade 3 cardiac toxicity and for any grade 4 non-haematological toxicity (except al-opecia and nausea/vomiting). Paclitaxel dose was reduced by 25% for grade 2 peripheral neuropathy or myalgia and by 50% for grade 3 myalgia. Paclitaxel dose was omitted in case of grade 3 peripheral neuropathy, maintaining the other 2 drugs with 25% dose reduction (paclitaxel could be resumed when recovery was achieved). Adriamycin and paclitaxel doses were reduced by 25% in case of grade 2 cardiac toxicity. Up to two dose reductions were allowed; once reduced, dose could not be escalated.

Patient assessments

Before study entry, all patients had a breast and axillary disease assessment by physical exam and mammogram or ultrasound, ECOG PS evaluation, surgical or core biopsy, complete blood cell count, serum chemistry, electrocardio-gram and left ventricular ejection fraction measurement. Additional tests were performed to exclude metastasis.

pCR was defined as no invasive cells identifiable in breast sections at surgery. Response was measured by physical exam and breast imaging before surgery and eval-uated according to the World Health Organization (WHO) criteria [9].

Patients were evaluated every cycle for adverse events and haematological and biochemical parameters. Toxicity was graded following the National Cancer Institute Com-mon Toxicity Criteria (NCI-CTC) version 2.0 [10], and the worst toxicity grade per patient per cycle was reported.

Tissue microarray construction

Haematoxylin and eosin-stained sections from core biop-sies (pretreatment) and surgical specimens (posttreatment) were marked on individual paraffin blocks. Two tissue cores (1 mm in diameter) were obtained from each speci-men. Additionally, 5 non-neoplastic breast tissue samples were included as controls, following the Kononen meth-odology [11]. A haematoxylin and eosin-stained section was reviewed to confirm the presence of morphologically representative areas of the original lesions.

432 Clin Transl Oncol (2012) 14:430-436

Immunohistochemistry

Immunohistochemical staining for oestrogen receptors (ER), progesterone receptors (PR), Ki-67, p53, cytokeratin 5/6 (CK5/6), pERK and survivin was performed by the EnVision method (DakoCytomation, Glostrup, Denmark) and evaluated simultaneously by two pathologists. Mono-clonal antibodies included mouse anti-human ER (SP1, MD; 1:50), PR (1A6, Novocastra; 1:10), CK 5/6 (D5/16 B4, DakoCytomation; 1:25), EGFR (EGFR.113, Novocas-tra; 1:10), Ki-67 (MIB-1, DakoCytomation; 1:50) and p53 (do-7, Novocastra; 1:25). In addition, polyclonal antisera against pERK (Cell Signaling, MA; 1:150) and survivin (Santa Cruz, 1:1000) were used. HER2 expression was studied using HercepTest (DakoCytomation).

The percentage of stained nuclei, independently of the intensity, was scored for ER, PgR, Ki-67, p53, survivin and pERK in each of the tumour cores, and the mean value was calculated. Likewise, the percentage of cells with cytoplas-mic stain was scored for CK5/6. For EGFR, only mem-branous staining was evaluated. A case was considered positive when the staining was found in more than 1% of cells for ER, PR, EGFR and CK5/6. Cases were considered to have low expression of Ki-67, survivin and pERK when staining was seen in ≤15% of cells. Cases were determined to have low p53 expression when staining was observed in ≤50% of cells. HER2 was evaluated according to the DAKO system of four categories (0 to 3+) proposed for the evaluation of HercepTest. Only 3+ staining was considered a positive result.

Tumours were considered to have a basal-like pheno-type if they were ER/HER2-negative and CK5/6 and/or EGFR-positive [12].

Sample collection and genotyping

Venous blood was collected at baseline for genomic DNA isolation. We analysed XPD Asp312Asn polymorphism in exon 10 (XPD10) and XPD Lys751Gln polymorphism in exon 23 (XPD23). Polymorphisms were assessed us-ing a 5´ nuclease allelic discrimination assay. Primers and probes for LysXPD751Gln and AspXPD312Asn are de-scribed elsewhere [13]. Fluorescence was measured before and after therapy. Data were analysed using an Allelic Dis-crimination Program (Applied Biosystems). For each poly-morphism, DNA samples were genotyped at least twice to confirm the results.

Statistical analyses

Enrolled patients who received at least one dose of any drug were evaluable for safety and efficacy. Simon’s method was used to calculate the sample size testing the null hypothesis (H0) that pCR rate equals 10% against the alternative hy-pothesis (H1) that pCR rate is ≥10%. Assuming an alpha error of 0.05 and a test power of 80% when the true pCR rate was 20%, 43 evaluable patients were required to be re-

cruited. All hypotheses were tested at an alpha level of 0.05 (two sided) using either the chi-squared or Fisher’s exact test for categorical variables; or the t-test for paired samples to compare differences between biomarker expression lev-els before and after treatment. The SPSS package version 15.0 was used for the statistical analysis of the study.

Results

Patient characteristics

Between March 2003 and July 2004, 46 patients from ten participating centres were included (Table 1). The median age was 49.5 years (range: 31–72 years); 54% of patients were pre-menopausal, and 46% were post-menopausal. Median tumour size was 6 cm (range 2–14), with 31 (67%) and 9 (20%) patients being T3 and T4, respectively. A to-tal of 67% presented stage IIIA, 22% stage IIIB and 11% stage IIB with tumours ≥5 cm. Most patients (93.5%) had ECOG PS of 0. Seventy-four percent of patients were hor-monal receptor positive.

All included patients were considered evaluable for safety and efficacy.

Table 1 Baseline characteristics of patients (N=46)

No. of patients (%)

Age (years) Median (range) 49.5 (31-72)Menopausal status Pre-menopausal 25 (54.3) Post-menopausal 21 (45.7)Tumour size Median, cm (range) 6.0 (2-14) T2 6 (13) T3 31 (67) T4a 9 (19.6)Histology Infiltrating ductal carcinoma 35 (76.1) Infiltrating lobular carcinoma 7 (15.2) Other 4 (8.7)Histological grade GX 20 (43.5) G1 2 (4.3) G2 13 (28.3) G3 11 (23.9)Hormonal receptor status ER and/or PR positive 34 (73.9) ER and PR negative 12 (26.1)Disease stage IIB 5 (10.9) IIIA 31 (67.4) IIIB 10 (21.7)ECOG PS 0 43 (93.5) 1 3 (6.5)

aFour patients were T4 not specified, four T4b, one T4d

Clin Transl Oncol (2012) 14:430-436 433

Efficacy

Eight patients (17%: [95% CI: 6.4–28.4]) achieved pCR in the breast; three of them were also free of invasive tumour in the axilla. None of the pCRs were observed in patients with lobular carcinoma.

The ORR was 89% (95% CI: 80.1–98.1), with 9 (19.6%) CRs, 32 (69.5%) PRs and 5 patients (10.9%) with stable disease. No patient progressed during treatment.

Surgery was performed in all patients. Breast conserva-tion was possible in 15 patients (33%), all of them with T2–T3 tumours.

With a median follow-up of 68.3 months (range: 7.4–81.5), 20 patients have had recurrence or death; the median DFS has not been reached yet.

Dose administration

A total of 273 cycles were administered, with a median of 6 cycles per patient (range: 5–6). Forty-three patients (93%) received the 6 planned cycles. Ninety cycles (33%) were delayed and six GAT doses were reduced (2%) in 6 patients (13%). The main reason for dose modifications was neutropenia (causing 58 cycle delays and 3 dose reductions). Asthenia, mucositis, anaemia, thrombocytopenia and infection were some of the other reasons causing dose adjustments or delays. The median dose intensity administered was 18 mg/m2/week for adri-amycin (88% of planned), 65 mg/m2/week for paclitaxel (87% of planned) and 896 mg/m2/week for gemcitabine (90% of planned).

Toxicity

Toxicities are shown in Table 2. The most frequent grade 3 or 4 haematological toxicity was neutropenia, which was seen in 32 patients (70%), in 19% of cycles; grade 4 neu-tropenia was reported in 13 patients (28%), in 6% of cy-cles. One patient had febrile neutropenia. Three additional patients had neutropenia complicated with infection (one grade 3 pneumonia and two grade 2 mycotic infections). Thirty-three patients (72%) received G-CSF. Anaemia was common but generally mild; five patients (11%) needed transfusion of packed red blood cells and 14 patients received erythropoietin. The majority of patients experi-enced asthenia, which was grade 3 in four (9%) patients. Gastrointestinal toxicity was generally mild; 4 patients (9%) had grade 3 mucositis. Eleven patients experienced non-neutropenic infections, including one grade 4 bilateral pneumonia and one catheter-related infection. One patient was hospitalised because of grade 4 pneumonitis. No pa-tient discontinued the trial due to toxicity.

Tumour markers

Immunohistochemical markers and pathological responsepCR rate was not higher in tumours negative for hormonal receptors, HER2-positive, Ck5/6-positive or p53-positive (Table 3). pCR occurred in 25% of cases with high prolif-eration index (Ki-67 >15%), but not in tumours with low proliferation index; this difference was not statistically significant.

Table 2 Haematological and non-haematological CTC-NCI toxicities version 2.0 (N=46)

No. of patients (%)Toxicity

Grade 1 Grade 2 Grade 3 Grade 4

Haematological Anaemia 20 (43.5) 23 (50.0) 3 (6.5) 0 Leucopenia 11 (23.9) 18 (39.1) 7 (15.2) 2 (4.3) Neutropeniaa 3 (6.5) 8 (17.4) 19 (41.3) 13 (28.3) Thrombocytopenia 26 (56.5) 4 (8.7) 2 (4.3) 0Non-haematological Alopecia 0 45 (97.8) – – Arthralgia 6 (13.0) 2 (4.3) 0 0 Asthenia 13 (28.3) 17 (37.0) 4 (8.7) 0 Cutaneous 5 (10.9) 4 (8.7) 1 (2.2) 0 Diarrhoea 10 (21.7) 5 (10.9) 1 (2.2) 0 Dispnoea 1 (2.2) 2 (4.3) 1 (2.2) 0 Phlebitis 1 (2.2) 2 (4.3) 0 0 Infection without neutropenia 6 (13.0) 3 (6.5) 1 (2.2) 1 (2.2)b

Myalgia 9 (19.6) 7 (15.2) 0 0 Peripheral neurotoxicity 13 (28.3) 5 (10.9) 0 0 Nausea/vomiting 8 (17.4) 17 (37.0) 2 (4.3) 0 Mucositis 14 (30.4) 13 (28.3) 4 (8.7) 0 Transaminases elevation 26 (56.5) 6 (13.0) 0 0

aOne patient had febrile neutropeniabBilateral pneumonia

434 Clin Transl Oncol (2012) 14:430-436

When marker expression (mean±SD) was compared between tumours with and without pCR, significant dif-ferences were found in the expression of Ki-67 (no pCR: 28.11±19.7; pCR: 50±23.7; p=0.023) and not in others; data on file.

Four out of 39 (10.3%) cases showed basal-like phe-notype. A numerically higher percentage of pCR was ob-served in tumours with a basal-like phenotype (2 out of 4, 50%) than in HER2-positive (3 out of 14, 21.4%) and ER-positive/HER2-negative tumours (2 out 11, 18.2%).

Immunohistochemical markers after treatmentKi-67, survivin and pERK expression showed a signifi-cantly lower percentage of positive cells after chemother-apy (Table 4). ER expression showed a numerically higher number of positive cells after treatment (though this differ-ence was non-significant). No changes were observed for PgR, HER2, p53, CK5/6 and EGFR (data not shown).

Patient genotypes and pathological response

Data were grouped as follows: Lys751Lys vs. Lys751Gln or Gln751Gln for XPD23 and Asp312Asp vs. Asp312Asn

or Asn312Asn for XPD10 (Table 5). Results showed a nu-merically lower number of pCR in the subgroup of patients with polymorphisms in XPD10 and XPD23.

Discussion

The primary endpoint of our study was the pCR rate with the addition of gemcitabine to a combination of adriamycin and paclitaxel in a dose-dense schedule. This combination was initially tested in metastatic breast cancer patients showing a high activity with an ORR of 80.4% and 36.5% CRs [14].

Our results, with an ORR of 89.1% and a pCR rate of 17%, show a significant activity, considering that 89.1% of patients were stage III with a median tumour size of 6 cm. The activity of this triplet combination, in terms of both ORR and pCR rate, is in the upper range in comparison to other combinations containing taxanes and/or anthra-cyclines and anti-metabolites [15, 16]. The GAT regimen allows for a shorter treatment duration before surgery when compared with sequential schedules and could potentially produce an increase in pCR rate, as shown in other studies [17], where a dose-dense schedule was compared with a three weekly schedule. In contrast, other studies [18] did not confirm this hypothesis, although the lower number of cycles could be the cause for that.

Our study confirms that GAT can be safely adminis-tered without any patient discontinuing due to toxicity and grade 3–4 neutropenia in only 19% of cycles. The low rate of conservative surgery observed (33%) may be explained by the large median tumour size at diagnosis.

The molecular signatures that predict response to a given chemotherapy are not well characterised. Elucidating molecular markers of efficacy for traditional agents, thus, may provide valuable information for a better selection of candidate patients.

One of the pretreatment markers studied here showed an association with pCR: Ki-67; in fact mean values of this

Table 3 Relationship between biomarker expression and pathological response

No. of patients (%)Marker (n)

no pCR pCR p

ER (38) Positive 22(84.6) 4 (15.4) Negative 9 (75.0) 3 (25.0) 0.656*PR (38) Positive 16 (84.2) 3 (15.8) Negative 15(78.9) 4 (21.1) 1.0*Ki-67 (34) Low 10 (100.0) 0 High 18 (75.0) 6 (25.0) 0.148*HER2 (39) Positive 11 (78.6) 3 (21.4) Negative 21 (84.0) 4 (16.0) 0.686*P53 (39) Low 26 (86.7) 4 (13.3) High 6 (66.7) 3 (33.3) 0.319*EGFR (33) Positive 1 (100.0) 0 Negative 27 (84.4) 5 (15.6) 1.0*CK5/6 (38) Positive 6 (75.0) 2 (25.0) Negative 25 (83.3) 5 (16.7) 0.624*pERK (35) Low 20 (83.3) 4 (16.7) High 8 (72.7) 3 (27.3) 0.652*Survivin (34) Low 19 (86.4) 3 (14.6) High 8 (66.7) 4 (33.3) 0.211*

ER=estrogen receptor, PR=progesterone receptor, *Fisher exact double-sided p-valueThe significance level is 0.05

Table 4 Changes in biomarker expression after chemotherapy

Marker (n) Mean SD p

ER (n=23) Baseline 32.0 29.7 After treatment 43.0 35.9 0.094*Ki-67 (n=24) Baseline 27.8 19.9 After treatment 16.7 23.6 0.016*Survivin (n=23) Baseline 14.4 12.2 After treatment 4.0 9.0 <0.001*pERK (n=24) Baseline 17.7 24.3 After treatment 1.1 2.4 0.003*

*t-test p-valueThe significance level is 0.05

Clin Transl Oncol (2012) 14:430-436 435

marker were higher in patients achieving a pCR (p=0.023). Correlation with other markers could have been hindered by the fact that pCR was a relatively uncommon event and that our trial was not powered to assess that. Our results are in agreement with previous studies that suggest that breast carcinomas with very high baseline proliferation may have better response to chemotherapy [19]. In our study, the mean Ki-67 index was significantly lower after treatment, suggesting that chemotherapy exerted a true anti-prolifer-ative effect on tumours in vivo. Reduced proliferation with chemotherapy may be partly due to increased apoptosis in actively proliferating cells such that the residual popula-tion would be enriched in Ki-67-negative cells [19, 20]. Survivin is a member of the apoptosis inhibitors (IAP) gene family, and is implicated in both apoptosis inhibition and mitosis regulation [21]. High survivin expression is frequently associated with poor prognosis in many cancers, although this association is not conclusive in breast cancer. It has been suggested that high survivin predicts for poor response to endocrine therapy, but for good response to chemotherapy [22]. Although here baseline levels of sur-vivin did not predict pCR, survivin expression was inten-sively reduced after treatment, suggesting the reduction of proliferating cells after chemotherapy.

Some studies have reported that ER-negative tumours are more chemosensitive to anthracycline-based neoadju-vant therapy than ER-positive tumours [23, 24]; however, we did not observe this association. Studies investigating the modulation of steroid receptor status by pre-operative chemotherapy reported no significant changes in ER and PgR [25] after primary chemotherapy, whereas others reported that 14% of ER-positive tumours shifted to ER-negative after treatment [26]. We observed a numerical increase in the number of ER-positive cells after chemo-therapy (p=0.09), suggesting an amplified chemoresis-tance.

The role of pERK in mediating response to chemo-therapy has not been studied in vivo. In vitro studies in hu-man non-small-cell lung cancer suggested that gemcitabine induced apoptotic cell death via phosphorylated activation of pERK; genetical or pharmacological inhibition of pERK

activation markedly blocked gemcitabine-induced cell death [27]. Here, this marker showed a noticeable reduc-tion after treatment, suggesting that apoptosis occurred preferentially in cells expressing pERK, which were elimi-nated by the treatment.

As for histological types, we did not find any pCR among infiltrating lobular carcinomas (ILC). These data agreed with recent reports indicating a very low sensitivity of ILC to different types of chemotherapy, which is related to the particular biological profile (ER-positive, low prolif-eration index) of ILCs [28, 29].

A recent report has suggested that breast cancer mo-lecular subtypes, as defined by cDNA microarrays, re-spond differently to pre-operative chemotherapy [30], with 45% of breast carcinomas with a basal-like phenotype or HER2-positive tumours achieving pCR when treated with paclitaxel followed by 5-fluorouracil, doxorubicin and cyclophosphamide. In contrast, only 6% of pCR were ob-served among luminal (ER-positive) tumours. In our study, we observed a numerically higher percentage of pCR in tumours with a basal-like phenotype (2 out of 4, 50%) in comparison with HER2-positive (3 out of 14, 21.4%) and ER-positive/HER2-negative tumours (2 out 11, 18.2%). These results suggest that basal-like breast carcinomas could represent tumours with high chemosensitivity in both the neoadjuvant and adjuvant settings.

Despite the influence of XPD polymorphisms in pre-dicting outcome in other tumour types, no relationship was observed between XPD SNPs and pathological response in our study.

This phase II study shows that the GAT combination is an active regimen in the neoadjuvant setting and there-fore could be an option in the neoadjuvant treatment of breast cancer patients with stage III disease, with moderate toxicity. Our data suggest that not only classical markers such as ER and Ki-67 but also survivin, pERK and tumour phenotype could be involved in the response to the GAT regimen.

Conflict of interest Belen San Antonio is a full-time Lilly employee and Eva Carrasco was a full-time Lilly employee by the time the trial was run. The other authors declare there have been no involvements that might raise the question of bias in the work reported or in the conclusions, implications or opinions stated.

Acknowledgements The authors want to acknowledge María del Carmen Cámara for her contribution to the conduct of study and María José Escudero for her contribution to the statistical analyses of the study. They also thank Dra. C. Crespo from Hospital Ramón y Cajal, Dr. C. Jara from Fundación Alcorcón, Dr. M. López-Vega from Hospital Universitario de Valdecilla, Dra. M.J. Godes from Hospital General de Valencia and Dra. L. Calvo from Hospital Universitario de A Coruña, Spain, for their valuable support in the inclusion of patients. The authors also thank Lilly S.A. and Bristol-Myers Squibb S.L. for their support of this study and Infociencia for assisting in the editing of the manuscript. This clinical trial was supported by Lilly S.A. and Bristol-Myers Squibb S.L. Both Companies provided fund-ing as well as the drugs for this study.

Table 5 Relationship between XPD23 and XPD10 polymorphisms and pathological response

Polymorphisms No. of patients (%)

pCR in breast No pCR in breast p

XPD23 Lys751Lys 1 (6.2) 15 (93.8) No Lys751Lys 5 (25.0) 15 (75.0) 0.1274*XPD10 Asp312Asp 1 (5.3) 18 (94.7) No Asp312Asp 5 (29.4) 12 (70.6) 0.0604*

*Fisher exact double-sided p-valueThe significance level is 0.05

436 Clin Transl Oncol (2012) 14:430-436

The affiliations

A. AntónMedical Oncology DepartmentMiguel Servet University HospitalZaragoza, Spain

A. BarnadasMedical Oncology DepartmentSant Creu i San Pau HospitalBarcelona, Spain

A. VelascoMedical Oncology DepartmentPrincesa University HospitalMadrid, Spain

M. LomasMedical Oncology DepartmentInfanta Cristina University HospitalBadajoz, Spain

M. Rodríguez-PinillaMolecular Pathology Program DepartmentCNIOMadrid, Spain

J.L. RamírezCatalá d’Oncología Institute Molecular Biology Cancer LaboratoryGermans Trias i Pujol HospitalBadalona, Barcelona, Spain

C. RamírezPathology DepartmentJaén Hospital ComplexJaén, Spain

M.J. RíosPathology DepartmentMiguel Servet University HospitalZaragoza, Spain

E. CastelláPathology DepartmentGermans Trias i Pujol HospitalBadalona, Barcelona, Spain

C. García-AndradePathology DepartmentSanta Cristina HospitalMadrid, Spain

B. San Antonio · E. CarrascoMedical DepartmentEli Lilly and CompanyMadrid, Spain

J.L. PalaciosPathology DepartmentVirgen del Rocío University HospitalSevilla, Spain

References

1. Fisher B, Brown A, Mamounas E et al (1997) 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 15: 2483–2493

2. Powles TJ, Hickish TF, Makris A et al (1995) Randomized trial of chemoendocrine therapy started before or after surgery for treatment of primary breast cancer. J Clin Oncol 13:547–552

3. Van der Hage JA, van de Velde CJ, Julien JP et al (2001) Preoperative chemotherapy in primary operable breast cancer: results from the European Organization for Research and Treatment of Can-cer trial 10902. J Clin Oncol 19:4224–4237

4. Wolmark N, Wang J, Mamounas E et al (2001) Preoperative chemotherapy in patients with oper-able breast cancer: nine-year results from National Surgical Adjuvant Breast and Bowel Project B-18. J Natl Cancer Inst Monogr 96–102

5. Bear HD, Anderson S, Brown A et al (2003) The effect on tumour response of adding sequential preoperative docetaxel to preoperative doxorubicin and cyclophosphamide: preliminary results from National Surgical Adjuvant Breast and Bowel Project Protocol B-27. J Clin Oncol 21:4165–4174

6. Green MC, Buzdar AU, Smith T et al (2005) Weekly paclitaxel improves pathologic complete remission in operable breast cancer when com-pared with paclitaxel once every 3 weeks. J Clin Oncol 23:5983-5992

7. Smith IC, Heys SD, Hutcheon AW et al (2002) Neoadjuvant chemotherapy in breast cancer: sig-nificantly enhanced response with docetaxel. J Clin Oncol 20:1456–1466

8. Coudert BP, Arnould L, Moreau et al (2006) Pre-operative systemic (neo-adjuvant) therapy with trastuzumab and docetaxel for Her2-over-expressing stage II or III breast cancer: results a multicenter Phase II trial. Ann Oncol 17:409–414

9. World Health Organization (1979) The WHO Handbook for Reporting Results of Cancer Treat-ment. Offset publication N° 48, Geneva, Switzer-land

10. Tumour National Cancer Institute (1998) Cancer Therapy Evaluation Program Common Toxicity

Criteria, version 2.0. Available at http://ctep.info.nih.gov/reporting/index.html

11. Kononen J, Bubendorf L, Kallioniemi A et al (1998) Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 4:844–847

12. Nielsen TO, Hsu FD, Jensen K et al (2004) Im-munohistochemical and clinical characterization of the basal-like subtype of invasive breast carci-noma. Clin Cancer Res 10:5367–5374

13. De las Peñas R, Sanchez-Ronco M, Alberola V et al (2006) Polymorphisms in DNA repair genes modulate survival in cisplatin/gemcitabine-treated non-small-cell lung cancer patients. Ann Oncol 17:668–675

14. Sánchez-Rovira P, Jaén A, González E et al (2001) Biweekly gemcitabine, doxorubicin and paclitaxel as first-line treatment in metastatic breast cancer. Final results from a phase II Trial. Oncology (Williston Park) 15:44–47

15. Lebowitz PF, Eng-Wong J, Swain SM et al (2004) A phase II trial of neoadjuvant docetaxel and capecitabine for locally advanced breast cancer. Clin Cancer Res 10:6764–6769

16. Conte PF, Donati S, Gennari A et al (2005) Prima-ry chemotherapy with gemcitabine, epirubicin and taxol (GET) in operable breast cancer: a phase II study. Br J Cancer 93:406–411

17. Moebus V, Jackisch C, Lueck HJ et al (2010) Intense dose-dense sequential chemotherapy with epirubicin, paclitaxel, and cyclophosphamide compared with conventionally scheduled chemo-therapy in high-risk primary breast cancer: mature results of an AGO phase III study. J Clin Oncol 28:2874–2880

18. Von Minckwitz G, Raab G, Caputo A et al (2005) Doxorubicin with cyclophosphamide followed by docetaxel every 21 days compared with doxoru-bicin and docetaxel every 14 days as preoperative treatment in operable breast cancer: the GEPAR-DUO study of the German Breast Group. J Clin Oncol 23:2676–2685

19. Urruticoechea A, Smith IE, Dowsett M (2005) Proliferation marker Ki-67 in early breast cancer. J Clin Oncol 23:7212–7220

20. Dowsett M, Ebbs SR, Dixon JM et al (2005) Biomarker changes during neoadjuvant anastro-zole, tamoxifen or the combination: influence of hormonal status and Her-2 in breast cancer–a

study from the IMPACT trialists. J Clin Oncol 23:2477–2492

21. Altieri DC (2003) Validating survivin as a cancer therapeutic target. Nat Rev Cancer 3:46–54

22. Span PN, Tjan-Heijnen VC, Manders P et al (2006) High survivin predicts a poor response to endocrine therapy, but a good response to chemo-therapy in advanced breast cancer. Breast Cancer Res Treat 98:223–230

23. Kuerer HM, Newman LA, Smith TL et al (1999) Clinical course of breast cancer patients with complete pathologic primary tumour and axillary lymph node response to doxorubicin-based neo-adjuvant chemotherapy. J Clin Oncol 17:460–469

24. Stearns V, Gallagher A, Kleer CG et al (2004) A pilot study to establish a clinical model to perform phase II studies of breast cancer chemopreventive agents in women at high risk with biomarkers as surrogate endpoints for activity. Clin Cancer Res 10:8332–8340

25. Schneider J, Lucas R, Sanchez J et al (2000) Modulation of molecular marker expression by in-duction chemotherapy in locally advanced breast cancer: correlation with the response to therapy and the expression of MDR1 and LRP. Anticancer Res 20:4373–4377

26. Taucher S, Rudas M, Gnant M et al (2003) Se-quential steroid hormone receptor measurements in primary breast cancer with and without in-tervening primary chemotherapy. Endocr Relat Cancer 10:91–98

27. Chang GC, Hsu SL, Tsai JR et al (2004) Extracel-lular signal-regulated kinase activation and Bcl-2 downregulation mediate apoptosis after gemcit-abine treatment partly via a p53-independent path-way. Eur J Pharmacol 502:169–183

28. Mathieu MC, Rouzier R, Llombart-Cussac A et al (2004) The poor responsiveness of infiltrating lobular breast carcinomas to neoadjuvant che-motherapy can be explained by their biological profile. Eur J Cancer 403:342–351

29. Cristofanilli M, Gonzalez-Angulo A, Sneige N et al (2005) Invasive lobular carcinoma classic type: response to primary chemotherapy and survival outcomes. J Clin Oncol 23:41–48

30. Rouzier R, Perou CM, Symmans WF et al (2005) Breast cancer molecular subtypes respond differ-ently to preoperative chemotherapy. Clin Cancer Res 11:5678–5685