Abstract withdrawn

Breast cancer consists of multiple diff erent molecular subtypes and diff erent biological processes, and consequently diff erent molecular markers are associated with prognosis and chemotherapy sensitivity in the distinct disease subsets [1]. A large number of biological processes including cell cycle regulation, DNA replication, mitotic spindle checkpoint, and p53 function are strongly prognostic in ER+ cancers but not among ER– cancers [2,3]. Interestingly, the number of biological pathways, and therefore genes, that are associated with prognosis or treatment sensitivity are substantially larger and more consistent in ER+ cancers than among ER– tumors [1,4]. This implies that it is easier to discover prognostic and predictive markers for ER+ than for ER– cancers. In ER– cancers, the single most consistent, but still modestly accurate, good prognostic predictor is the presence of immune cell infi ltration [5]. Immune cell signatures are also associated with more favorable prognosis in highly proliferative ER+ cancers but not in ER+ cancers with low proliferation [6]. It is also increasingly clear that the same molecular marker can be associated with several diff erent outcome endpoints in various and often opposing manners. For example, high Ki67 expression is predictive of worse prognosis in the absence of any systemic therapy in ER+ cancers, but at the same time it is also predictive of higher sensitivity to chemotherapy. Similar opposing bidirectional associations with treatment response and prognosis exist for many other markers including histologic grade, Tau protein expression and almost all prognostic gene signatures [7]. It is important to be aware of these complex multi-directional interactions between molecular markers and various clinical endpoints that may also vary from breast cancer subtype to subtype. Ignoring these potential marker–disease subset–outcome interactions can lead to contradictory and confusing results across studies (due to diff erences in patient composition and heterogeneity of therapy between studies) and may also lead to the discovery of biomarkers that are clinically less useful (because of unrecognized subtype-restricted performance) [8,9]. References 1. Iwamoto T, Bianchini G, Booser D, et al.: Gene pathways associated with prognosis and chemotherapy sensitivity in molecular subtypes of breast cancer. J Natl Cancer Inst 2011, 103:264-272. 2. Bianchini G, Iwamoto T, Coutant C, et al.: Prognostic and therapeutic implications of distinct kinase expression patterns in diff erent subtypes of breast cancer. Cancer Res 2010, 70:8852-8862. 3. Coutant C, Rouzier R, Qi Y, et al.: Distinct p53 gene signatures are needed to predict prognosis and response to chemotherapy in ER-positive and ER-negative breast cancers. Clin Cancer Res 2011, 17:2591-2601. 4. Tordai A, Wang J, Andre F, et al.: Evaluation of biological pathways involved in chemotherapy response in breast cancer. Breast Cancer Res 2008, 10:1-9. 5. Karn T, Pusztai L, Ruckhaeberle E, et al.: Melanoma antigen family A identifi ed by the bimodality index defi nes a subset of triple negative breast cancers as candidates for immune response augmentation. Eur J Cancer 2011. doi:10.1016/j.ejca.2011.06.025. 6. Bianchini G, Qi Y, Hugo AR, et al.: Molecular anatomy of breast cancer stroma and its prognostic value in ER-positive and -negative cancers. J Clin Oncol 2010, 28:4316-4323. 7. Andre F, Hatzis C, Anderson K, et al.: Microtubule-associated protein-tau is a bifunctional predictor of endocrine sensitivity and chemotherapy resistance in estrogen receptor-positive breast cancer. Clin Cancer Res 2007, 13:2061-2067. 8. Pusztai L, Broglio K, Andre F, et al.: Eff ect of molecular disease subsets on disease-free survival in randomized adjuvant chemotherapy trials for estrogen-receptor positive breast cancer. J Clin Oncol 2008, 26:4679-4683. 9. Sotiriou C, Pusztai L: Gene expression signatures in breast cancer. N Engl J Med 2009, 360:790-800.

Triple-negative breast cancers (TNBCs) are among the most clinically challenging because of their poor prognosis and paucity of treatment options. In part through our genomic profi ling studies, breast cancer is now appreciated as being composed of multiple diseases. One of these diseases, the basal-like breast cancer (BLBC) subtype, is now known to represent a unique disease entity with a distinct etiology and biology. Over the years, BLBC has become more commonly known as TNBC because the majority of these tumors lack expression of ER, PR and HER2; however, not all TNBC are BLBC, and not all BLBC are TNBC. Recently, we discovered that a signifi cant subset of TNBC is comprised of a new subtype, the claudin-low, which is important because it is biologically distinct from BLBC and has a number of features reminiscent of mammary stem cells [1]. In addition, luminal A, luminal B, and HER2-enriched tumors are also identifi ed within TNBCs in various small proportions, which highlights the complexity of the clinically based classifi cation. We have explored the treatment sensitivity of the various intrinsic subtypes to neoadjuvant anthracycline/taxane-based chemotherapy using a large publicly available dataset [2]. Across all patients, and within TNBC, basal-like tumors were found associated with a higher likelihood of achieving a complete pathological response (pCR) than the rest of the subtypes, including the claudin-low. In multivariate logistic regression models for pCR prediction, we observed that the intrinsic molecular subtypes virtually always make the fi nal model, even if clinical variables and other genomic predictors are included. In addition, our analyses show that those tumors that achieve a pCR showed a better survival outcome than those that did not, regardless of their molecular subtype; this eff ect is much larger within the basal-like subtype, which is concordant with previous fi ndings. This intriguing association between residual disease after therapy and poor outcome in basal-like and claudin-low tumors points to intratumor cell heterogeneity as a possible explanation, where resistant and aggressive cell clones might already exist in the pretreated tumor. Our preliminary analyses using a combination of fl uorescent activated cell sorting and global gene expression on numerous preclinical models of basallike breast cancers including cell lines and primary tumor xenografts suggest the existence of at least two cell populations in many BLBC models. These diff erent cell populations are currently being tested for tumor-initiating cell activities, and additional studies focusing on these populations changing with treatment are also being performed. Poly(ADP-ribose) polymerase-1 (PARP-1) is a nuclear DNA-binding enzyme activated by DNA strand breaks and has a key role in the signalling of DNA single-strand breaks as part of the repair process.
In anti-cancer therapy, many agents cause DNA damage as their mechanism of cytotoxicity, and repair of damage is a cause of tumour resistance. Additionally in tumours where double-strand break repair is defective (for example in BRCA-related cancers) PARP inhibitors have potential single-agent activity. Thus, PARP-1 was identifi ed as a potential therapeutic target for cancer treatment and PARP inhibitors have entered the clinic both in combination with cytotoxic chemotherapy, as single agents in DNA repair defi cient tumours, and more recently in combination with radiotherapy. The fi rst PARP inhibitor to be given to cancer patients in 2003 was AG014699 (rucaparib), a tricyclic indole, which is a potent intravenous inhibitor of PARP. This phase I study had a pharmacodynamic endpoint of PARP inhibition in PBMCs, demonstrating for the fi rst time proof of mechanism of the class. Subsequently AZD2281 (olaparib) entered clinical trials as a single agent (2005), and demonstrated the proof of concept of synthetic lethality in BRCA defective tumours in two small phase II studies. Over the last 5 years seven further inhibitors have entered cancer clinical trials either as a single agent (MK4827) or in combination with various cytotoxic regiments (ABT888, veliparib; BSI-201, iniparib; CEP-9722; INO-1001; E7016, BMN 673) in late preclinical development. Initial exciting data suggesting that iniparib improved outcome in patients with triple-negative breast cancer in combination with chemotherapy have not been confi rmed in phase III studies, although there are clearly patients who benefi t from this agent. In terms of mechanism of action, iniparib diff ers from all the other compounds in the class that are competitive inhibitors at the NAD + binding site of PARP. Iniparib is postulated to have a diff erent mechanism of action and may not be a bona fi de PARP inhibitor. It has been a period of rapid clinical development of a new class of agents with exciting evidence of improved response rates in some tumour areas. This class of agents also presents some interesting challenges in clinical trial design, and mechanistic understanding. This presentation will overview the current clinical status of PARP inhibitors and will discuss these challenges and potential biomarker strategies. Evading immune destruction should be considered an emerging hallmark of cancer [1]. Highly immunogenic cancer cells can be eliminated in immunocompetent hosts as a result of the immunoediting process. Weakly immunogenic variants can grow and generate solid tumours. Regulatory T cells (Tregs) were found to be involved in the maintenance of the immune tolerance both preventing autoimmune disease and curtailing antitumour immune response. Modulation of immune response in cancer patients is the result of a balanced activity of Tregs and T-eff ector cells [2]. In cancer patients, an increased number of Tregs was found in blood and tumour tissue: it was demonstrated that Tregs suppress T-cell response and natural killer cell proliferation and function, thus interfering both with acquired and innate immunity. Upregulation of Tregs in the tumour bed can be associated with worse prognosis [3]. Drugs blocking function of Tregs increase activity of T eff ectors and, as a side eff ect, induce an autoimmune disease. Issues of biology and prognosis of breast cancer in the presence of a deregulation of the immune system need to be studied. The identifi cation of immunological and genetic features aff ecting immune response in patients with minimal tumour burden is the optimal background for development of clinical studies in the adjuvant setting. Research on tumour-associated antigens (TAAs) has identifi ed a large collection of peptide epitopes that have been and are being used for vaccination of cancer patients [4]. Several potential advantages of using peptidebased vaccines include: easy and relatively inexpensive production of synthetic peptides; the easy administration of peptides in a clinical setting; the possibility of treating only those patients whose tumours overexpress the target antigens; and the availability of in vitro or ex vivo assays that can assess patients' immune response to vaccine epitopes. The aim of future studies will be to assess the immunoreactivity of several antigens in a large series of breast cancer samples classifi ed according to molecular subtypes. Identifi cation of potential targets in subpopulations of patients with breast cancer may allow identifi cation of patients who are potential candidates for adjuvant therapeutic vaccines. It is our current thinking that patients with minimal residual disease after preoperative chemotherapy are the ideal setting to test the effi cacy of a vaccination strategy. To date, vaccines for breast cancer have been mainly used in end-stage disease. TAA antigens off er a novel opportunity for fostering vaccine development and therapy.

O5
Epithelial-mesenchymal transition as a mechanism for the progression of breast carcinoma JP Thiery 1,2 , WJ Sim 1 , K Chua 2 , R Huang 2 , S Mori 2 , T Tan  Epithelial-mesenchymal transition (EMT) is a major process controlling multiple events during development. EMT has been conserved throughout evolution to control morphogenetic events, such as the formation of the three primary germ layers during gastrulation. Most interestingly, signal transduction pathways have been remarkably conserved in many diff erent species. EMT pathways are also tightly connected to determination and diff erentiation programmes, and are reactivated in adult tissues following injury or exposure to toxic agents. EMT is likely to operate during the early stages of carcinoma invasion that lead to blood or lymph vessel intravasation. Mesenchymal-like carcinoma cells undergo a mesenchymal-toepithelial transition in distant sites from the primary tumour and eventually become micrometastatic. We have characterised bone marrow micrometastases from breast cancer patients and found that the detection of micrometastatic carcinoma cells was associated with poorer distant metastasis-free survival, local relapse-free survival, and overall survival. Despite high rates of adjuvant systemic treatment and breast irradiation in this series, disseminated carcinoma cells remain a prognostic factor, in favour of the resistance to treatment of locally or distant disseminated cancer cells in bone marrow-positive patients. In addition, we detected micrometastatic carcinoma cells in patients with T1 tumours, suggesting that dissemination occurs much earlier during tumour progression than is generally accepted. Thus, bone marrow micrometastases should become a very useful prognostic indicator for relapse, and an excellent surrogate marker for patient's response to treatment. The mesenchymal-like state of carcinoma confers stemness, protection from cell death, escape from immune response and, most importantly, resistance to conventional and targeted therapies. Current strategies based on the EMT concept are aimed at designing new therapeutic approaches that interfere with the plasticity of carcinoma cells. Our laboratory has devised a high-content, highthroughput screen for EMT. Several combinations of drugs have been shown to selectively inhibit EMT. This strategy may be used to interfere with tumour progression, particularly in breast carcinomas that have acquired resistance to conventional therapies. Breast cancer risk is increased in women who have attributes of the insulin resistance syndrome, such as obesity (postmenopausal), central obesity (premenopausal and postmenopausal), high endogenous insulin levels, clinical diabetes and physical inactivity. There is a large body of evidence that obesity is associated with a 25 to 50% relative increase in risk of breast cancer recurrence or death, with adverse eff ects that appear to be independent of hormone receptor status. Obesity, particularly when it is central, is strongly associated with insulin resistance in healthy individuals and breast cancer patients. Several studies have shown that higher insulin and/or C-peptide levels (a marker of insulin secretion), both of which are linked to insulin resistance, are associated with an increased risk of recurrence and death in women with early stage breast cancer, even in the absence of diabetes. Risk is increased twofold to threefold in those with insulin levels in the highest (versus lowest) quartile. Data from our group suggest that these prognostic associations of insulin are most marked in the fi rst 5 years post diagnosis. A role of insulin in breast cancer outcomes is biologically plausible given overexpression of insulin receptors (IR), most frequently the fetal form of the receptor (IR-A), by the majority of human breast cancers. IR-A often hybridizes with insulin-like growth factor 1 receptors to stimulate mitogenic signaling pathways; hybrid receptor activation has been associated with poor clinical outcomes. The current observational and preclinical evidence linking insulin to breast cancer is suffi ciently compelling that neoadjuvant and adjuvant intervention studies have been initiated to evaluate clinical anti-cancer eff ects of metformin, an agent that lowers insulin levels and has other potential non-insulin-mediated anti-cancer eff ects (mainly through activation of adenosine monophosphateactivated protein kinase). Early results from window of opportunity neoadjuvant studies suggest short-term, single-agent metformin (2 to 3 weeks) lowers insulin levels, reduces proliferation and increases apoptosis. NCIC MA32, an ongoing randomized, multicenter, placebocontrolled, adjuvant trial involving 3,582 women with early stage breast cancer, will provide more defi nitive evidence regarding potential anti-cancer eff ects. Additional studies of metformin in the metastatic setting are underway and/or planned. Because other factors (such as infl ammation, adipocytokines (leptin, adiponectin), higher estrogen levels) may also mediate prognostic eff ects of obesity and/or insulin resistance in breast cancer, additional research targeting these mediators as well as obesity per se is also needed. The phosphatidylinositol-3 kinase (PI3K) pathway is overall the most frequently mutated pathway in cancer, with mutation and/or amplifi cation of the genes encoding the PI3K catalytic subunits p110α (PIK3CA) and p110β (PIK3CB), the PI3K regulatory subunit p85α (PIK3R1), receptor tyrosine kinases (RTKs) such as HER2 (ERBB2) and FGFR1, the PI3K activator K-Ras, the PI3K eff ectors AKT1, AKT2, and PDK1, and loss of the lipid phosphatases PTEN and INPP4B. PI3K is activated by growth factor RTKs and G-protein-coupled receptors (GPCRs). PI3K activates Akt, which, in turn, phosphorylates and inactivates Tuberin (TSC2), a GTPase-activating protein of the Ras homologue Rheb. Inactivation of Tuberin allows GTP-bound Rheb to accumulate and activate the mTOR/ Raptor (TORC1) complex, which regulates protein synthesis and cell growth. mTOR also couples with Rictor to form the TORC2 complex, which phosphorylates and activates AKT. Class IA PI3K isoforms are heterodimeric lipid kinases that contain a p110 catalytic subunit and a p85 regulatory subunit. The three genes PIK3CA, PIK3CB, and PIK3CD encode the homologous p110α, p110β, and p110δ isozymes, respectively. Expression of p110δ is largely restricted to immune and hematopoietic cells, whereas p110α and p110β are ubiquitously expressed. p110α is essential for signaling and growth of tumors driven by PIK3CA mutations, RTKs, and/or mutant Ras, whereas p110β lies downstream of GPCRs and has been shown to mediate tumorigenesis in PTEN-defi cient cells. PIK3CA mutations are the most commonly known genetic alterations of this pathway in cancer, where ≥80% occur within the helical (E542K and E545K) and kinase (H1047R) domains of p110α. Such mutations confer increased catalytic activity through diff erent mechanisms, but both induce characteristics of cellular transformation including growth factor-independent and anchorage-independent growth, and resistance to anoikis. Several drugs targeting multiple levels of the PI3K network (that is, PI3K, AKT, mTOR) have been developed. A number of ATP-mimetics that bind competitively and reversibly to the ATP-binding pocket of p110 are in early clinical development. These include the pan-PI3K inhibitors BKM120, XL-147, PX-866, PKI-587, and GDC-0941, the p110αspecifi c inhibitors BYL719, GDC-0032, and INK-1117, the p110δ-specifi c inhibitor CAL-101, and the dual PI3K/mTOR inhibitors BEZ235, BGT226, PF-4691502, GDC-0980, and XL-765. The pan-PI3K and p110α-specifi c inhibitors are equally potent against oncogenic p110α mutants. The rationale for the development of isozyme-specifi c antagonists is to allow higher doses of anti-p110α and anti-p110β drugs to be delivered without incurring side eff ects caused by pan-PI3K inhibitors. Interim results from a phase I trial with the p110δ-specifi c inhibitor CAL-101 in patients with hematologic malignancies showed that treatment reduced P-AKT levels >90% in peripheral blood lymphocytes and induced objective clinical responses. Recently completed phase I trials with BKM120, BEZ235, and XL-147 showed that treatment partially inhibited PI3K as measured by levels of P-S6 and P-AKT in patients' skin or tumors, and 2-deoxy-2-[ 18 F]fl uoro-D-glucose uptake measured by PET. Main toxicities were rash, hyperglycemia, diarrhea, fatigue and, mood alterations. Few clinical responses were observed in patients with and without detectable PI3K pathway mutations, although screening for genetic lesions in this pathway was not comprehensive.

O8
Both allosteric and ATP-competitive pan-inhibitors of the three isoforms of AKT are also being developed. AZD5363, GDC-0068, GSK2141795, and GSK690693 are ATP-competitive compounds that have shown antitumor activity in preclinical models and recently entered phase I trials. Allosteric inhibitors such as MK-2206 bind to the AKT PH domain and/or hinge region to promote an inactive conformation of the AKT protein that is unable to bind to the plasma membrane. MK-2206 inhibits AKT signaling in vivo, and suppresses growth of breast cancer xenografts harboring PIK3CA mutations or ERBB2 amplifi cation. Phase I data showed that treatment with MK-2206 decreases levels of P-AKT, P-PRAS40, and P-GSK3β in tumor cells, peripheral blood mononuclear cells, and hair follicles. The mTOR kinase is a component of PI3K-driven oncogenesis that functions within two signaling complexes: TORC1 and TORC2 (described above). The macrolide rapamycin and its analogs form complexes with FK506-binding protein (FKBP12). This complex then binds to mTOR and inhibits the kinase activity of TORC1 but not TORC2. Laboratory-based studies of BRCA1 and BRCA2 demonstrated that loss of function of either gene resulted in signifi cantly increased susceptibility to certain forms of chemotherapy, including interstrand DNA cross-linking agents such as the platinum drugs and mitomycin C. More recently, loss of BRCA1 or BRCA2 function has also been shown to increase sensitivity to PARP inhibition, a fi nding made possible as a result of increased understanding of the DNA repair implications of BRCA1 or BRCA2 loss. To a large extent, these laboratory-based observations have now been verifi ed in clinical trials enrolling patients with hereditary breast cancer. The implications of the discovery of BRCA1 and BRCA2 for treatment options in sporadic breast cancer are more complex. Based on a series of striking phenotypic similarities between the majority of sporadic triple-negative breast cancers and most cancers that arise in BRCA1 heterozygotes, the hypothesis arose that perhaps many of these sporadic cancers might also share a similar lesion in DNA repair (BRCAness) with the BRCA1-related tumors. This notion has now been put to the test in ongoing clinical trials that treat sporadic triple-negative breast cancer patients with platinum agents, PARP inhibitors, or combinations. The current evidence for and against this hypothesis will be discussed. The newly discovered diff erential expression in numerous tissues, key cellular processes and multiple diseases for several families of long and short noncodingRNAs (ncRNAs, RNAs that do not codify for proteins but for RNAs with regulatory functions), including the already famous class of microRNAs (miRNAs), strongly suggest that the scientifi c and medical communities have signifi cantly underestimated the spectrum of ncRNAs whose altered expression has signifi cant consequences in diseases. miRNA and other short or long ncRNA alterations are involved in the initiation, progression and metastases of human breast cancer.
The main molecular alterations are represented by variations in gene expression, usually mild and with consequences for a vast number of target protein coding genes. The causes of the widespread diff erential expression of ncRNAs in malignant compared with normal cells can be explained by the location of these genes in cancer-associated genomic regions, by epigenetic mechanisms and by alterations in the processing machinery. miRNA and other short or long ncRNA expression profi ling of human breast tumors has identifi ed signatures associated with diagnosis, staging, progression, prognosis and response to treatment. In addition, profi ling has been exploited to identify ncRNAs that may represent downstream targets of activated oncogenic pathways or that are targeting protein coding genes involved in cancer. Recent studies proved that miRNAs and noncoding ultraconserved genes are main candidates for the elusive class of cancer predisposing genes and that other types of ncRNAs participate in the genetic puzzle giving rise to the malignant phenotype. Last, but not least, the shown expression correlations of these new ncRNAs with cancer metastatic potential and overall survival rates suggest that at least some member of these novel classes of molecules could potentially fi nd use as biomarkers or novel therapeutics in cancers and other diseases. Pertuzumab, a monoclonal antibody that inhibits HER2-HER3 heterodimers, appears to be eff ective when combined with trastuzumab ± chemotherapy. T-DM1, an antibody-drug conjugate, has also displayed remarkable activity in the setting of refractory disease and has limited toxicity. It is presently under investigation in multiple randomized trials. Neratinib is an oral irreversible tyrosine kinase inhibitor that targets HER1, HER2, and HER4. As a single agent, it appears to be more active than lapatinib, but is associated with more signifi cant toxicity. It, too, is under evaluation in phase III trials in the adjuvant and metastatic settings. A variety of other agents are under active study including the mTOR inhibitors, the PI3kinase inhibitors, angiogenesis inhibitors, and IGFR antagonists. It is likely that a number of new agents will be available for the treatment of HER2 + breast cancer in the next several years, and outcomes for this group of patients will continue to improve. during SOLTI-0701 that was eff ective and manageable. Doses can be escalated to 2,500 mg/m 2 and 800 mg/day or reduced to manage toxicity. Dose re-escalation after reduction is only allowed for sorafenib/ placebo. Guidelines detail prophylactic and symptomatic therapy for HFSR/HFS. Radiographic assessment is every 6 weeks for 36 weeks, then every 9 weeks. The primary endpoint is PFS. Secondary endpoints include overall survival, time to progression, overall response rate, and duration of response. Enrollment began in November 2010 and targets ~519 patients. Conclusion RESILIENCE will provide defi nitive PFS data for sorafenib + capecitabine as fi rst-line or second-line therapy in HER2-negative advanced BC and will better characterize the benefi t-to-risk profi le of this regimen. Luminal breast cancer constitutes almost all ER-positive tumours and as such constitutes around 75 to 80% of the disease. The luminal group is highly heterogeneous in terms of genetic aberrations such as mutations, amplifi cations/deletions and translocations, and also phenotypic characteristics such as proliferation and the expression of oestrogen-dependent proteins such as PgR, TFF1 and GREB1. While assessment of some of these molecular characteristics at presentation can act as a guide to outcome, there remains substantial uncertainty in prognostic and predictive algorithms. Our approach has been to study the biological relationships by applying specifi c suppressants of the synthesis of oestrogen -that is, aromatase inhibitors (AIs) -in the presurgical setting. The changes in proliferation (Ki67) that occur are related to treatment benefi t and the residual Ki67 to residual risk of recurrence. In addition, the molecular changes can be characterized as intermediate endpoints of response. The POETIC trial of 2 weeks' AI or not in the window of time between diagnosis and surgery has now recruited over 2,000 patients (August 2011). Biopsies taken before and after the AI are providing a uniquely powerful set of data to understand the mechanisms of response and resistance to oestrogen deprivation. Pilot work has indicated that although luminal B tumours have higher initial Ki67 levels, their antiproliferative response to an AI is proportionally similar to luminal A tumours, indicating a similar initial responsiveness but higher residual risk of recurrence. The development and implementation of biomarkers for the diagnosis and classifi cation of breast cancers and stratifi cation of breast cancer patients into clinically meaningful groups are essential for the realisation of individualised medicine. The accurate, robust and reproducible assignment of patients into subgroups of therapeutic relevance is of utmost importance. Breast cancer patient treatment decision-making currently relies on the analyses of a few immunohistochemical markers (for example, oestrogen receptor, progesterone receptor and HER2), fl uorescence and/ or chromogenic in situ hybridisation, protein analysis of lysates, and quantitative real-time PCR. It has become clear, however, that these markers are not suffi cient for the potential of individualised therapy to be fully realised. The advent of high-throughput technologies and their use in fundamental and translational research endeavours have led to the development of diagnostic markers, potential prognostic and predictive factors, and therapeutic targets, which ultimately will need to be incorporated in clinical practice. Some of the major challenges in this process are to determine the accuracy of the research hypothesis, the exclusion of potential biases, and to defi ne whether the reagents and methodologies are fi t for purpose. This requires not only a thorough assessment of the accuracy, robustness and reproducibility of the markers and the methods for their analysis, but also an adequate contextualisation of the validity of a given biomarker. For instance, immunohistochemistry has become one of the major tools for the identifi cation of expression of potential markers in cancer tissues; albeit at fi rst glance trivial to perform, immunohistochemical analysis can be aff ected by numerous parameters that can aff ect its accuracy. Likewise, several gene expression profi ling approaches for the identifi cation of molecular subtypes of breast cancer have been shown not to assign the same patients into the same molecular subgroups consistently. Interpretation of in situ hybridisation is also fraught with diffi culties. Discrepancies in the assessment of biomarkers have often been attributed to intra-tumour heterogeneity, without exclusion of sources of technical variation. The challenges for the interpretation of biomarker studies and validation of biomarkers in human tissues will be discussed.

Introduction
We have shown that nestin expression is higher in breast carcinoma with a basal phenotype [1] and collagen triple helix repeat containing 1 (CTHRC1) and periostin may predict bone metastasis of breast cancer [2]. Our aim therefore was to examine the simultaneous role of nestin and CTHRC1 in breast cancer progression. Methods Archival formalin-fi xed paraffi n-embedded 173 invasive breast cancer samples classifi ed into WHO histotypes and luminal, triplenegative and Her2 subtypes were immunohistochemically stained with CTHRC1, periostin, nestin and vimentin antibodies. Staining was evaluated with histoscore and neoangiogenesis was assessed as the number of nestin-positive new vessels. The degree of infl ammation was evaluated on HE-stained slides. Data were statistically processed by nonparametric Mann-Whitney U test, Spearman correlation coeffi cient and Pearson chi-square. Results Both CTHRC1 stromal (P = 0.013) and nestin epithelial expression (P = 0.001) were higher in the triple-negative subtype. We found strong association between nestin expression in cancer cells and CTHRC1 stromal expression in advanced stage patients (r = 0.614; P = 0.007). Nestin expression was also associated with vimentin expression in breast cancer cells (r = 0.491; P <0.001). Both nestin and vimentin showed positive association with degree of infl ammation, in particular in triple-negative patients (r φ = 0.422; r = 0.521; both P <0.001).
We observed higher nestin positivity in patients with lymph node metastases and high periostin stromal expression (P = 0.031). Conclusion For the fi rst time, we report an association between CTHRC1 and nestin expression in patients with advanced breast cancer. Further investigation is needed to better clarify their role in infl ammation and breast cancer progression. ≥200 were associated with lower PFS of 8.2 months (95% CI = 0.6 to 10.8) compared with those with <200, PFS 16.9 months (95% CI = 8.78 to NA), P = 0.003. See Figure 1. No diff erence was observed in OS. Fourteen patients (74%) that had stable disease/partial response decreased or maintained their CEC value. Baseline CTCs ≥5 was associated with a median PFS of 15.2 months (95% CI = 7.6 to 16.9). Twenty-two patients (92%) that had stable disease/partial response decreased or maintained their CTC value. The CTC level was not correlated with the CEC level, P = 0.74. Conclusion Our study suggests signifi cant correlations between high levels of baseline CECs and poor prognosis. Addiction of B to fi rst-line CT was related to a high reduction of CEC and CTC count. Introduction Status of the surgical margins following wide local excision for breast cancer remains one of the strongest predictors of local recurrence. In our practice, a margin of 1 mm and more is considered adequate. In this study, we aim to determine whether clinicopathological factors other than surgical margins contribute to the risk of local recurrence. Methods A retrospective review was performed of 548 consecutive patients who underwent wide local excision for invasive carcinoma or ductal carcinoma in situ (DCIS) from 1 January 2004 to 31 December 2008. Surgery was not routinely off ered to patients with margins of 1 mm or more. All patients with wide local excision received postoperative whole breast irradiation, inclusive of a boost to the tumour bed.

Introduction
Results Local recurrence developed in 20% of those with involved margins, as compared with 8.7% of those with close margins, and 5.4% of those with margins of 1 mm and more. Although local recurrence was more likely with an involved or close surgical margin, this reached only borderline signifi cance (P = 0.05). Oestrogen receptor (ER) status was found to be an independent predictor of local recurrence, with ER-negative tumours being three times more likely to recur (P < 0.01, OR = 0.30, 95% CI = 0.13 to 0.66). There was no correlation with a triplenegative phenotype or other clinicopathological factors. Conclusion A margin of 1 mm or more appears to be adequate following wide local excision. However, ER status emerged as a stronger predictor for local recurrence and alone remained signifi cant on multivariate analysis.   Median (95% CI) DFS (years) 8.1 (6.5 to 8.7) 6.4 (5.4 to 7.2) 9.2 (7.6 to 11.9) 5.6 (5.2 to 6.1)

Results
The median ki67 value for the ER + PRgroup was 17%. A ki67 cutoff at 14% discerns two groups of diff erent prognosis inside the luminal group (extracting HER2 + and PR -); and comparison of ki67 cutoff between 14 and 11% found overlapped CI (median: 6.31 (5.99 to 6.62) vs. 6.49 (6.21 to 6.78)). Conclusion Initial data support the use of a classical CMF regimen in TNBC patients. However, the biomarker of CMF responsiveness is needed for clinical practice. We confi rmed the bcl2 positivity as a predictor of CMF sensitivity in TNBC. Thus, validation of this marker in a larger study is needed. Higher chromosome 17 copy number was associated with better outcome, suggesting the importance of its assessment.

Introduction
In breast cancer (BC), several cytogenetic markers are routinely investigated (Her2, progesterone and estrogen receptors) that distribute patients into diff erent diagnostic groups, determine prognosis and predict eff ectiveness of therapy. Additional genetic markers can improve prediction or prognosis in BC patients. C-MYC and CCND1 gene aberrations frequently found in BC were chosen for this study.

Methods
The status of C-MYC, CCND1 genes, chromosomes 8 and 11 were determined on 74 patients from a group of 103 BC patients treated with herceptin in a palliative regime, using FISH assay on formalin-fi xed paraffi n-embedded tissue samples. One hundred nuclei per sample were analyzed in each sample. Clinical data were correlated with our cytogenetic results. Results Amplifi cations of CCND1 and C-MYC were occurring together in most cases (P <0.0001), especially in Her-2/neu-negative cases (P = 0.01, resp. 0.03). Progesterone receptor positivity was associated with CCND1, resp. C-MYC increased copy number (P = 0.039, resp. 0.038). Amplifi cation (ratio gene/chromosome ≥2.2) of CCND1, resp. C-MYC was determined in 8.1% (6/74), resp. 12.2% (9/74) cases. Polysomy Results After proper normalization, 123 out of 671 miRNAs showed a good correlation of their expression data between FFPE and FF tissue, and suffi cient analytical robustness (they are expressed in at least one-third of FFPE samples). In addition, we analyzed the expression of various markers with diagnostic value in breast cancer. As regards high-performance proteomics, the protocols developed generated over 6,000 MS/MS spectra, enabling the identifi cation of hundreds of proteins in each sample. Conclusion We have selected the most appropriate assays to study miRNA expression in breast cancer FFPE archived samples. The protocols developed allow proteome analysis of FFPE samples using the latest mass spectrometry equipment. The technologies implemented during the development of this project allow one to compare the expression data at both miRNA and protein levels to study breast cancer from an authentic system-biology perspective.
Introduction Acquired resistance to endocrine therapies remains a major clinical obstacle in hormone-sensitive breast tumors. The complexity of the underlying biological mechanisms remains poorly understood and the purpose of this study was to identify low abundant proteins and central pathways associated with tamoxifen resistance. Methods The global protein expression of the parental tamoxifensensitive MCF7S0.5 cell line and the tamoxifen-resistant TamR1 cell line were compared using SILAC labeling and quantitative mass spectrometry. Data were processed using MaxQuant, the global protein-protein interaction network was predicted using STRING and enriched pathways were identifi ed with KEGG analysis. Selected proteins diff erentially expressed were validated using western blotting and immunocytochemistry. Results Proteomic analysis identifi ed 5,370 proteins based on at least one unique peptide of which 4,448 proteins could be quantifi ed based on at least two peptides. Forty-one proteins were found to be diff erentially expressed more than threefold and eight proteins were validated by western blotting and immunocytochemistry as potential biomarkers associated with tamoxifen resistance. In total 539 proteins were diff erentially expressed 1.5-fold or more and could be subgrouped into kinases, transcription factors, receptor activity proteins, cell adhesion proteins, cell cycle proteins and stress responder proteins. We identifi ed several regulated proteins to be important in subnetworks that among others are involved in focal adhesion, DNA replication, apoptosis, and insulin and HER2 signaling pathway. Conclusion Novel low abundant proteins not previously associated with tamoxifen resistance have been identifi ed and validated using biochemical methods. At present the proteins are being validated on a panel of primary breast cancer biopsies from patients treated with tamoxifen monotherapy and with known clinical outcome. Our data also revealed several pathways associated to tamoxifen resistance. The importance of these pathways needs to be studied further.
Introduction Metastases are the major cause of cancer-related deaths, but the mechanisms of the metastatic process remain poorly understood. In recent years, the involvement of microRNAs (miRNAs) in cancer has become apparent, and the objective of this study was to identify miRNAs associated with breast cancer progression. Methods Global miRNA expression profi ling was performed on 47 tumor samples from 14 patients with paired samples from primary breast tumors and corresponding lymph node and distant metastases using LNA-enhanced miRNA microarrays. The identifi ed miRNA expression alterations were validated by real-time PCR, and tissue distribution of the miRNAs was visualized by in situ hybridization.

Results
The patients in which the miRNA profi le of the primary tumor and corresponding distant metastasis clustered in a unsupervised cluster analysis showed signifi cantly shorter intervals between the diagnosis of the primary tumor and distant metastasis (median 1.6 years) compared with those that did not cluster (median 11.3 years) (P <0.003). Fifteen miRNAs were identifi ed that were signifi cantly diff erentially expressed between primary tumors and corresponding distant metastases, including miR-9, miR-219-5p and four of the fi ve members of the miR-200 family involved in epithelial-mesenchymal transition. Tumor expression of miR-9 and miR-200b was confi rmed using in situ hybridization, which also verifi ed higher expression of these miRNAs in the distant metastases versus corresponding primary tumors. Conclusion Our results demonstrate alterations in miRNA expression at diff erent stages of disease progression in breast cancer, and suggest a direct involvement of the miR-200 family and miR-9 in the metastasis process.