An in silico approach helped to identify the best experimental design, population, and outcome for future randomized clinical trials.

OBJECTIVES: The main objective of our work was to compare different randomized clinical trial (RCT) experimental designs in terms of power, accuracy of the estimation of treatment effect, and number of patients receiving active treatment using in silico simulations. STUDY DESIGN AND SETTING: A virtual population of patients was simulated and randomized in potential clinical trials. Treatment effect was modeled using a dose-effect relation for quantitative or qualitative outcomes. Different experimental designs were considered, and performances between designs were compared. One thousand clinical trials were simulated for each design based on an example of modeled disease. RESULTS: According to simulation results, the number of patients needed to reach 80% power was 50 for crossover, 60 for parallel or randomized withdrawal, 65 for drop the loser (DL), and 70 for early escape or play the winner (PW). For a given sample size, each design had its own advantage: low duration (parallel, early escape), high statistical power and precision (crossover), and higher number of patients receiving the active treatment (PW and DL). CONCLUSION: Our approach can help to identify the best experimental design, population, and outcome for future RCTs. This may be particularly useful for drug development in rare diseases, theragnostic approaches, or personalized medicine.

A methodological framework for drug development in rare diseases.

INTRODUCTION: Developing orphan drugs is challenging because of their severity and the requisite for effective drugs. The small number of patients does not allow conducting adequately powered randomized controlled trials (RCTs). There is a need to develop high quality, ethically investigated, and appropriately authorized medicines, without subjecting patients to unnecessary trials. AIMS AND OBJECTIVES: The main aim is to develop generalizable framework for choosing the best-performing drug/endpoint/design combinations in orphan drug development using an in silico modeling and trial simulation approach. The two main objectives were (i) to provide a global strategy for each disease to identify the most relevant drugs to be evaluated in specific patients during phase III RCTs, (ii) and select the best design for each drug to be used in future RCTs. METHODOLOGICAL APPROACH: In silico phase III RCT simulation will be used to find the optimal trial design and was carried out in two steps: (i) statistical analysis of available clinical databases and (ii) integrative modeling that combines mathematical models for diseases with pharmacokinetic-pharmacodynamics models for the selected drug candidates. CONCLUSION: There is a need to speed up the process of orphan drug development, develop new methods for translational research and personalized medicine, and contribute to European Medicines Agency guidelines. The approach presented here offers many perspectives in clinical trial conception.

Experimental designs for small randomised clinical trials: an algorithm for choice.

BACKGROUND: Small clinical trials are necessary when there are difficulties in recruiting enough patients for conventional frequentist statistical analyses to provide an appropriate answer. These trials are often necessary for the study of rare diseases as well as specific study populations e.g. children. It has been estimated that there are between 6,000 and 8,000 rare diseases that cover a broad range of diseases and patients. In the European Union these diseases affect up to 30 million people, with about 50% of those affected being children. Therapies for treating these rare diseases need their efficacy and safety evaluated but due to the small number of potential trial participants, a standard randomised controlled trial is often not feasible. There are a number of alternative trial designs to the usual parallel group design, each of which offers specific advantages, but they also have specific limitations. Thus the choice of the most appropriate design is not simple. METHODS: PubMed was searched to identify publications about the characteristics of different trial designs that can be used in randomised, comparative small clinical trials. In addition, the contents tables from 11 journals were hand-searched. An algorithm was developed using decision nodes based on the characteristics of the identified trial designs. RESULTS: We identified 75 publications that reported the characteristics of 12 randomised, comparative trial designs that can be used in for the evaluation of therapies in orphan diseases. The main characteristics and the advantages and limitations of these designs were summarised and used to develop an algorithm that may be used to help select an appropriate design for a given clinical situation. We used examples from publications of given disease-treatment-outcome situations, in which the investigators had used a particular trial design, to illustrate the use of the algorithm for the identification of possible alternative designs. CONCLUSIONS: The algorithm that we propose could be a useful tool for the choice of an appropriate trial design in the development of orphan drugs for a given disease-treatment-outcome situation.

Patients with metastatic breast cancer leading to CD4+ T cell lymphopaenia have poor outcome.

BACKGROUND: Low lymphocyte count is a prognostic factor in cancer patients including metastatic breast cancer patients (MBC) but the relative role of each lymphocyte subtype is unclear in MBC. METHODS: The impact of lymphocyte subsets was analysed in two prospective MBC patients' cohorts. Cohort A patients (n=103) were included before the first line of chemotherapy and cohort B patients (n=101) were included after at least one line of chemotherapy. Extensive phenotypic analyses were performed on fresh whole blood. Plasma cytokines levels were measured using commercially available Luminex-based multiplex kits. Prognostic value of lymphocyte subsets and circulating cytokines was analysed. RESULTS: In both cohorts, severe lymphopaenia (<0.7 Giga/L) correlated with poor overall survival (OS) (median OS: 6.6 months versus 21.7 months in cohort A and 4.5 versus 9 months in cohort B). CD8(+), CD19(+) and CD56(+) T cell counts had no significant prognostic value for OS. After stratification (≤0.2, [0.20-0.45], >0.45 Giga/L), CD4 lymphopaenia appeared to be correlated with poor OS in both cohorts. Furthermore, severe CD4(+) lymphopaenia (≤0.2 Giga/L) was strongly correlated with poor OS in both cohorts (1.2 months versus 24.9 months in cohort A and 5.7 versus 13.1 months in cohort B). In multivariate analysis, after stratification CD4(+) lymphopaenia appeared to be an independent prognostic factor for OS in both cohorts. CD4(+) lymphopaenia correlated with low plasmatic levels of CCL22 that might directly contribute to CD4(+) lymphopaenia. CONCLUSIONS: CD4(+) lymphopaenia was associated with reduced OS in MBC patients regardless of the chemotherapy line. Decreased levels of plasmatic CCL22 may contribute to CD4(+) lymphopaenia.

Can treatment with Cocculine improve the control of chemotherapy-induced emesis in early breast cancer patients? A randomized, multi-centered, double-blind, placebo-controlled Phase III trial.

BACKGROUND: Chemotherapy induced nausea and vomiting (CINV) remains a major problem that seriously impairs the quality of life (QoL) in cancer patients receiving chemotherapy regimens. Complementary medicines, including homeopathy, are used by many patients with cancer, usually alongside with conventional treatment. A randomized, placebo-controlled Phase III study was conducted to evaluate the efficacy of a complex homeopathic medicine, Cocculine, in the control of CINV in non-metastatic breast cancer patients treated by standard chemotherapy regimens. METHODS: Chemotherapy-naïve patients with non-metastatic breast cancer scheduled to receive 6 cycles of chemotherapy including at least three initial cycles of FAC 50, FEC 100 or TAC were randomized to receive standard anti-emetic treatment plus either a complex homeopathic remedy (Cocculine, registered in France for treatment of nausea and travel sickness) or the matching placebo (NCT00409071 clinicaltrials.gov). The primary endpoint was nausea score measured after the 1st chemotherapy course using the FLIE questionnaire (Functional Living Index for Emesis) with 5-day recall. Secondary endpoints were: vomiting measured by the FLIE score, nausea and vomiting measured by patient self-evaluation (EVA) and investigator recording (NCI-CTC AE V3.0) and treatment compliance. RESULTS: From September 2005 to January 2008, 431 patients were randomized: 214 to Cocculine (C) and 217 to placebo (P). Patient characteristics were well-balanced between the 2 arms. Overall, compliance to study treatments was excellent and similar between the 2 arms. A total of 205 patients (50.9%; 103 patients in the placebo and 102 in the homeopathy arms) had nausea FLIE scores > 6 indicative of no impact of nausea on quality of life during the 1st chemotherapy course. There was no difference between the 2 arms when primary endpoint analysis was performed by chemotherapy stratum; or in the subgroup of patients with susceptibility to nausea and vomiting before inclusion. In addition, nausea, vomiting and global emesis FLIE scores were not statistically different at any time between the two study arms. The frequencies of severe (Grade ≥ 2) nausea and vomiting were low in our study (nausea: P: 17.6% vs C: 15.7%, p=0.62; vomiting: P: 10.8% vs C: 12.0%, p=0.72 during the first course). CONCLUSION: This double-blinded, placebo-controlled, randomised Phase III study showed that adding a complex homeopathic medicine (Cocculine) to standard anti-emetic prophylaxis does not improve the control of CINV in early breast cancer patients.

Impaired IFN-α production by plasmacytoid dendritic cells favors regulatory T-cell expansion that may contribute to breast cancer progression.

Infiltration and dysfunction of immune cells have been documented in many types of cancers. We previously reported that plasmacytoid dendritic cells (pDC) within primary breast tumors correlate with an unfavorable prognosis for patients. The role of pDC in cancer remains unclear but they have been shown to mediate immune tolerance in other pathophysiologic contexts. We postulated that pDC may interfere with antitumor immune response and favor tolerance in breast cancer. The present study was designed to decipher the mechanistic basis for the deleterious impact of pDC on the clinical outcome. Using fresh human breast tumor biopsies (N = 60 patients), we observed through multiparametric flow cytometry increased tumor-associated (TA) pDC (TApDC) rates in aggressive breast tumors, i.e., those with high mitotic index and the so-called triple-negative breast tumors (TNBT). Furthermore, TApDC expressed a partially activated phenotype and produced very low amounts of IFN-α following toll-like receptor activation in vitro compared with patients' blood pDC. Within breast tumors, TApDC colocalized and strongly correlated with TA regulatory T cells (TATreg), especially in TNBT. Of most importance, the selective suppression of IFN-α production endowed TApDC with the unique capacity to sustain FoxP3(+) Treg expansion, a capacity that was reverted by the addition of exogenous IFN-α. These findings indicate that IFN-α-deficient TApDC accumulating in aggressive tumors are involved in the expansion of TATreg in vivo, contributing to tumor immune tolerance and poor clinical outcome. Thus, targeting pDC to restore their IFN-α production may represent an attractive therapeutic strategy to overcome immune tolerance in breast cancer.

Permanent 125I-seed prostate brachytherapy: early prostate specific antigen value as a predictor of PSA bounce occurrence.

PURPOSE: To evaluate predictive factors for PSA bounce after 125I permanent seed prostate brachytherapy and identify criteria that distinguish between benign bounces and biochemical relapses. MATERIALS AND METHODS: Men treated with exclusive permanent 125I seed brachytherapy from November 1999, with at least a 36 months follow-up were included. Bounce was defined as an increase ≥ 0.2 ng/ml above the nadir, followed by a spontaneous return to the nadir. Biochemical failure (BF) was defined using the criteria of the Phoenix conference: nadir +2 ng/ml. RESULTS: 198 men were included. After a median follow-up of 63.9 months, 21 patients experienced a BF, and 35.9% had at least one bounce which occurred after a median period of 17 months after implantation (4-50). Bounce amplitude was 0.6 ng/ml (0.2-5.1), and duration was 13.6 months (4.0-44.9). In 12.5%, bounce magnitude exceeded the threshold defining BF. Age at the time of treatment and high PSA level assessed at 6 weeks were significantly correlated with bounce but not with BF. Bounce patients had a higher BF free survival than the others (100% versus 92%, p = 0,007). In case of PSA increase, PSA doubling time and velocity were not significantly different between bounce and BF patients. Bounces occurred significantly earlier than relapses and than nadir + 0.2 ng/ml in BF patients (17 vs 27.8 months, p < 0.0001). CONCLUSION: High PSA value assessed 6 weeks after brachytherapy and young age were significantly associated to a higher risk of bounces but not to BF. Long delays between brachytherapy and PSA increase are more indicative of BF.

Bevacizumab plus microtubule targeting agents in heavily pre-treated ovarian cancer patients: a retrospective study.

OBJECTIVES. As vascular endothelial growth factor (VEGF) is expressed in ovarian cancer, we assessed the efficacy and safety of bevacizumab (a monoclonal antibody targeting VEGF) plus microtubule targeting agents for heavily pre-treated ovarian carcinoma patients. METHODS. We retrospectively reviewed 43 patients with recurrent epithelial ovarian carcinoma. Combined treatment included bevacizumab with paclitaxel in 32 (74%), docetaxel in 10 (23%), and vinorelbine in one (2.3%) patients, respectively. RESULTS. The median number of combined treatment was six cycles (range 1-29). On RECIST criteria, the objective response rate (ORR) was 40% (16% CR and 24% PR). Clinical benefit (complete response [CR] plus partial response [PR] and stable disease [SD] lasting ≥ 3 months) was 74% (CI95%: 46.7-77%). Median duration of treatment and overall survival were 3.9 months (range 0.2-14.4 months) and 20.1 months (CI95%: 13.8-20.1) respectively. No toxic death was reported. Grade 3-4 toxicity occurred in 30% of patients. Gastrointestinal perforations and fistula occurred in 3 (7%) and 6 (14%) patients, respectively. CONCLUSION. Although being active in terms of ORR, bevacizumab plus microtubule targeting agents - mainly taxanes - leads to a high rate of gastro-intestinal perforations and fistula in heavily pre-treated ovarian carcinoma patients.

Final results of ERASME-4: a randomized trial of first-line docetaxel plus either capecitabine or epirubicin for metastatic breast cancer.

OBJECTIVE: To assess the efficacy of capecitabine plus docetaxel (XT) versus epirubicin plus docetaxel (ET) as first-line therapy for metastatic breast cancer (MBC). PATIENTS AND METHODS: Patients with no prior chemotherapy for MBC were randomized to 3-weekly cycles of either XT (capecitabine 1,000 mg/m(2) twice daily, days 1-14; docetaxel 75 mg/m(2), day 1) or ET (epirubicin 75 mg/m(2), day 1; docetaxel 75 mg/m(2), day 1). The primary endpoint was non-progression rate 6 months after randomization. The planned sample size was 106 patients based on a randomized, phase II selection design. RESULTS: Between April 2004 and January 2007, 68 patients were randomized, giving 82% power to select the best regimen according to a 6-month non-progression rate. Slow accrual led to premature study termination. Baseline characteristics were generally well balanced between arms. The 6-month non-progression rates were 75.8% with XT versus 65.7% with ET (p = 0.36). After 42 months' median follow-up, median progression-free survival was 12.4 versus 6.8 months, respectively (p = 0.040). The safety profiles were consistent with previous experience. CONCLUSION: Further larger studies are warranted to validate these results. Despite more grade 3 hand-foot syndrome, first-line XT may be a valid alternative to ET, potentially improving efficacy.

Quantitative and functional alterations of plasmacytoid dendritic cells contribute to immune tolerance in ovarian cancer.

In ovarian cancer, the immune system fails to eradicate established tumors partly due to the induction of immune tolerance within tumor microenvironment. In this study, we investigated the contribution of plasmacytoid dendritic cells (pDC) in the establishment of immune tolerance in a cohort of 44 ovarian cancer patients. In the tumor and malignant ascites, CD4(+)CD123(+)BDCA2(+) pDC were the most abundant dendritic cell subset; however, they were profoundly depleted in peripheral blood. The presence of pDC in primary ovarian cancer, but not ascites, was an independent prognostic factor associated with early relapse. Following chemotherapy, we observed a partial restoration of blood pDC levels in patients in complete remission. These findings show preferential recruitment of pDC into tumors where they express a partially mature phenotype that may reflect an in situ activation. Importantly, compared with pDC found in ascites or blood, tumor-associated pDC (TApDC) produced less IFN-α, TNF-α, IL-6, macrophage inflammatory protein-1ß, and RANTES in response to toll-like receptor stimulation, and alterations in pDC functions were mainly mediated through tumor-derived TNF-α and TGF-ß. Unlike ascites-derived pDC, TApDC induced IL-10 production from allogeneic naive CD4(+) T lymphocytes, suggesting the existence of a paracrine immunosuppressive loop. Taken together, our findings indicate that both local and systemic dysfunction of pDC play a critical role in the progression of ovarian cancer via induction of immune tolerance.

Validation of prognostic scores for survival in cancer patients beyond first-line therapy.

BACKGROUND: We aimed to validate prognostic scores for survival in patients undergoing chemotherapy for advanced or metastatic cancer after first-line treatment. METHODS: We previously described two models with good prognostic value based on a combination of Performance Status (PS) and either lactate dehydrogenase (LDH) level or lymphocyte count. These factors were evaluated for their ability to predict overall survival (OS) in a prospective cohort of 299 patients. Clinical and blood parameters were prospectively recorded. Candidate prognostic factors for OS with 0.05 significance level in univariate analysis were included in a multivariate Cox model. RESULTS: Median age was 59 years (range: 26-85). Primary tumor sites were breast (45%), lung (15%), ovaries (11%) and others (29%). The number of metastatic sites was 1 (29%), 2 (48%), >2 (23%). Median follow-up and median OS were 12 and 6 months, respectively. Multiple regression analysis confirmed that PS >1, lymphocyte count ≤700/µL and LDH >600 UI/L were independent predictors of short OS, as well as interleukin 6 (IL-6) level, serum albumin concentration and platelet count. CONCLUSIONS: Prognostic scores using PS plus LDH level or PS plus lymphocyte count were validated for predicting survival in metastatic cancer patients in relapse beyond first-line treatment. A score combining PS, LDH, lymphocyte and platelet count, serum albumin and IL-6 level was superior in determining patients' prognosis.

Factors of interrupting chemotherapy in patients with Advanced Non-Small-Cell Lung Cancer.

BACKGROUND: Little is known about prognosis of metastatic patients after receiving a first-line treatment and failure. Our group already showed in pre-treated patients enrolled in phase I clinical trials that a performance status (PS) > 2 and an LDH > 600 UI/L were independent prognostic factors. In this prospective study, which included 45 patients, we identified clinical and biological variables as outcome predictors in metastatic Non-Small Cell lung cancer after first line chemotherapy were identified. FINDINGS: Forty-five patients that were previously treated for metastatic disease from 12/2000 to 11/2005 in the comprehensive cancer centre (Centre Léon Bérard). Clinical assessment and blood parameters were recorded and considered. Patient prognostic factors for overall survival (OS) with a 0.05-significance level in univariate analysis were entered in a multivariate Cox model for further analysis.Patients' median age was 58.5 years (range: 37 - 76). Sixty two percent of the patients were PS = 0 or 1. After inclusion, nine patients received second-line (22.5%), and two received third-line chemotherapy (5%). Univariate analysis showed that the factors associated with reduced OS were: PS > 2, weight loss >10%, more than one line of chemotherapy treatment and abnormal blood parameters (hemoglobin (Hb), platelet and neutrophils counts). Multiple regression analysis confirmed that PS > 2 and abnormal hemoglobin were independent predictors for low overall survival. According to the presence of none (33%), 1 (37%) and 2 (30%) prognostic factors, median OS were 12, 5 and 2 months respectively. CONCLUSION: From this prospective study, both PS and anemia were found as independent determinants of survival, we found that both PS and anemia were independent determinants of survival. The combination of poor PS and anemia is an effective strategy to predict survival in the case of patients with metastatic NSCLC receiving further treatment after the first line.

Body weight change in women receiving adjuvant chemotherapy for breast cancer: a French prospective study.

BACKGROUND & AIMS: Adjuvant chemotherapy has frequently been associated with weight gain after breast cancer diagnosis. We aimed to prospectively evaluate body weight variations in French patients with early breast cancer. METHODS: This prospective observational study included 272 breast cancer patients who were candidates for adjuvant chemotherapy. Weight and body mass index were measured at baseline visit, then at 9 and 15 months from baseline (6 and 12-month post-chemotherapy). At baseline visit, information on the benefits of weight gain prevention and healthy diet was given by a dietician. Univariate logistic regression was performed to test the association between weight gain and potential predictive factors. RESULTS: Thirty percent of patients gained weight during the year before diagnosis, 26% were overweight and 15% were obese. At one year, the mean weight change was +1.5kg (SD=4.1) and +2.3% (SD=6.0); 60% of the cohort had gained weight, with a median increase of 3.9kg (SD=3.0) and 5.9% (SD=4.4). Reported weight gain during the year before diagnosis appears to be the only factor associated with the absence of post-chemotherapy weight gain (OR=0.54, 95% CI [0.31-0.95], p=0.034). CONCLUSION: Body weight increased in the post-chemotherapy period in French breast cancer survivors, even when given dietary recommendations. Appropriate weight management interventions with nutritional follow-up and physical activity programs are needed.

Revisiting the level of evidence in randomized controlled clinical trials: A simulation approach.

BACKGROUND: The Evidence Based Medicine (EBM) paradigm requires that results from Randomized Controlled Trials (RCTs) must be assessed for validity before being assimilated. However, evaluating available evidence is often still based on intuitive processes rather than on rigorous scientific analysis. OBJECTIVE: To establish a hierarchy among the different factors influencing the level of evidence of RCT results, using a Monte Carlo simulation. METHODS: The complete RCT model involved three submodels: i) the input-output submodel for the prediction of events (using the sigmoid dose-response relationship as the basic model), ii) the execution submodel for deviations from a randomized, controlled two-arm parallel trial related to either patient-specific or investigator-specific elements or both: placebo or nocebo effect, errors of measurement, effect of concomitant therapy, regression to the mean phenomenon, blinding process, loss to follow-up and randomization process, iii) the covariate distribution submodel. RESULTS: The most important factors influencing discrepancies in the true-to-observed odds ratio were the blinding process, the measurement errors (affecting either the therapeutic or the adverse effects), the placebo effect, the effect of concomitant therapies and to a less extent the randomization process. Whereas the randomization process remained the only relevant factor in double-blinded trials, the hierarchy of other factors was modified according to the type of blinding. CONCLUSION: In RCTs, the hierarchy of confounding factors differs according to the type of blinding and the current short list of components of the strength of evidence (poorly concealed randomization and lack of blinding) appears to be incomplete.

Outcome of recurrent and metastatic small cell carcinoma of the bladder.

BACKGROUND: Bladder small cell carcinoma is an uncommon tumour. Through a retrospective study we will present the evolution of recurrent and metastatic disease and outcome of patients treated at Léon-Bérard Cancer Centre. METHODS: Only 15 patients having recurrent or metastatic bladder small cell carcinoma were treated at Léon-Bérard Cancer Centre between 1996 and 2007. The patients were divided in two groups: a mixed small cell carcinoma group (9 patients) and a pure small cell carcinoma group (6 patients). All the records and informations related to treatment and outcome of the 15 patients were retrospectively analyzed. Various characteristics of small cell carcinoma were investigated. RESULTS: The median age of the 15 patients having recurrent or metastatic bladder small cell carcinoma and treated at Léon-Bérard Cancer Centre was 63 years and the disease was at stage IV for all cases. Nine patients were treated by chemotherapy. Four patients were treated by local radiotherapy (3 with radiotherapy without previous surgery and 1 with surgery followed by radiotherapy) and chemotherapy. One patient was treated by whole brain radiotherapy. And one patient died before treatment. After 52.4 months median follow up, 12 patients died. Median overall survival was 7.6 months. Survival probability at 1 year was 33%. Median overall survival was 9.9 months in the mixed small cell carcinoma group, and was only 4.6 months in the pure small cell carcinoma group. Survival probability at 1 year in the mixed small cell carcinoma group was 44% as compared to 17% in the pure small cell carcinoma group (Log-rank test: p = 0.228). CONCLUSION: Recurrent and metastatic bladder small cell carcinoma is associated with very poor prognosis. The pure bladder small cell carcinoma appears to have poorer outcome than the mixed bladder small cell carcinoma. Chemotherapy using platinum drugs is a mainstay treatment.

A candidate molecular signature associated with tamoxifen failure in primary breast cancer.

INTRODUCTION: Few markers are available that can predict response to tamoxifen treatment in estrogen receptor (ER)-positive breast cancers. Identification of such markers would be clinically useful. We attempted to identify molecular markers associated with tamoxifen failure in breast cancer. METHODS: Eighteen initially ER-positive patients treated with tamoxifen requiring salvage surgery (tamoxifen failure [TF] patients) were compared with 17 patients who were disease free 5 years after surgery plus tamoxifen adjuvant therapy (control patients). cDNA microarray, real-time quantitative PCR, and immunohistochemistry on tissue microarrays were used to generate and confirm a gene signature associated with tamoxifen failure. An independent series of 33 breast tumor samples from patients who relapsed (n = 14) or did not relapse (n = 19) under tamoxifen treatment from a different geographic location was subsequently used to explore the gene expression signature identified. RESULTS: Using a screening set of 18 tumor samples (from eight control patients and 10 TF patients), a 47-gene signature discriminating between TF and control samples was identified using cDNA arrays. In addition to ESR1/ERalpha, the top-ranked genes selected by statistical cross-analyses were MET, FOS, SNCG, IGFBP4, and BCL2, which were subsequently validated in a larger set of tumor samples (from 17 control patients and 18 TF patients). Confirmation at the protein level by tissue microarray immunohistochemistry was observed for ER-alpha, gamma-synuclein, and insulin-like growth factor binding protein 4 proteins in the 35 original samples. In an independent series of breast tumor samples (19 nonrelapsing and 14 relapsing), reduced expression of ESR1/ERalpha, IGFBP4, SNCG, BCL2, and FOS was observed in the relapsing group and was associated with a shorter overall survival. Low mRNA expression levels of ESR1/ERalpha, BCL2, and FOS were also associated with a shorter relapse-free survival (RFS). Using a Cox multivariate regression analysis, we identified BCL2 and FOS as independent prognostic markers associated with RFS. Finally, the BCL2/FOS signature was demonstrated to have more accurate prognostic value for RFS than ESR1/ERalpha alone (likelihood ratio test). CONCLUSIONS: We identified molecular markers including a BCL2/FOS signature associated with tamoxifen failure; these markers may have clinical potential in the management of ER-positive breast cancer.