This Article Abstract Full Text (PDF) Erratum (v25,p3790) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Loi, S. Articles by Sotiriou, C. Search for Related Content PubMed PubMed Citation Articles by Loi, S. Articles by Sotiriou, C. Related Articles Related Correspondence Social Bookmarking                            What's this?

Definition of Clinically Distinct Molecular Subtypes in Estrogen Receptor–Positive Breast Carcinomas Through Genomic Grade

Sherene Loi, Benjamin Haibe-Kains, Christine Desmedt, Françoise Lallemand, Andrew M. Tutt, Cheryl Gillet, Paul Ellis, Adrian Harris, Jonas Bergh, John A. Foekens, Jan G.M. Klijn, Denis Larsimont, Marc Buyse, Gianluca Bontempi, Mauro Delorenzi, Martine J. Piccart, Christos Sotiriou

From the Jules Bordet Institute; Machine Learning Group, Université Libre de Bruxelles; International Drug and Development Institute, Brussels, Belgium; Peter MacCallum Cancer Center, Melbourne, Australia; Guys Hospital, London; John Radcliffe Hospital, Oxford, United Kingdom; Karolinska Institute, Stockholm, Sweden; Erasmus Medical Center, Daniel den Hoed Cancer Center, Rotterdam, the Netherlands; National Center of Competence in Research Molecular Oncology, Swiss Institute of Cancer Research and Swiss Institute of Bioinformatics, Epalinges, Switzerland

Address reprint requests to Christos Sotiriou, MD, PhD, Translational Research Unit, Jules Bordet Institute, Université Libre de Bruxelles, 121 Blvd de Waterloo, 1000 Brussels, Belgium; e-mail: christos.sotiriou{at}bordet.be

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Purpose A number of microarray studies have reported distinct molecular profiles of breast cancers (BC), such as basal-like, ErbB2-like, and two to three luminal-like subtypes. These were associated with different clinical outcomes. However, although the basal and the ErbB2 subtypes are repeatedly recognized, identification of estrogen receptor (ER) –positive subtypes has been inconsistent. Therefore, refinement of their molecular definition is needed.

Materials and Methods We have previously reported a gene expression grade index (GGI), which defines histologic grade based on gene expression profiles. Using this algorithm, we assigned ER-positive BC to either high–or low–genomic grade subgroups and compared these with previously reported ER-positive molecular classifications. As further validation, we classified 666 ER-positive samples into subtypes and assessed their clinical outcome.

Results Two ER-positive molecular subgroups (high and low genomic grade) could be defined using the GGI. Despite tracking a single biologic pathway, these were highly comparable to the previously described luminal A and B classification and significantly correlated to the risk groups produced using the 21-gene recurrence score. The two subtypes were associated with statistically distinct clinical outcome in both systemically untreated and tamoxifen-treated populations.

Conclusion The use of genomic grade can identify two clinically distinct ER-positive molecular subtypes in a simple and highly reproducible manner across multiple data sets. This study emphasizes the important role of proliferation-related genes in predicting prognosis in ER-positive BC.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Since the advent of microarray analysis, a number of studies have reported that breast cancers (BCs) can be classified into molecular subtypes on the basis of distinct gene expression profiles. These subgroups were associated with different disease outcomes, suggesting a biologic basis behind the clinical heterogeneity of BC.^1-4 Up to five subclasses of BCs (basal-like, ErbB2-like, and luminal A, B, and C) have been proposed based on an intrinsic gene set, with the basal-like subtype, which was mainly estrogen receptor (ER) negative, experiencing a worse outcome compared with the luminal subtypes, which were mainly ER positive.^1,4.

Although the basal-like and ErbB2 subtypes are repeatedly recognized in independent data sets, discernment of the ER-positive subtypes has been difficult. Sorlie et al² used a nearest centroid predictor based on an intrinsic gene set to identify the luminal subtypes in independent data sets. However, many samples were considered unclassified because they had a low correlation with the subtype centroids,² and definition of the intrinsic gene set has been inconsistent.^1,2,4,5 Recently, a set of 822 estrogen-induced genes was found to be able to predict long-term outcome in ER-positive tumors better than the previously described luminal A/B classification.⁶ However, this extensive gene list does not yet facilitate itself as a clinical test. The recurrence score (RS) is a commercially available test (Oncotype Dx; Genomic Health, Inc, Redwood City, CA) that uses a complex mathematical algorithm to stratify ER-positive patients.⁷ The lack of a simple diagnostic test has limited the implementation of a molecular classification into the clinical setting where such stratification could result in the identification of therapeutics that have differential effects in the various subgroups.

Overexpression of cell cycle genes is known to be associated with poor clinical outcome. Perou et al⁴ originally reported a proliferation cluster of genes that correlated with cellular proliferation rates and was noted to have considerable variation between the subgroups.⁸ Taking this concept further, we have previously defined a gene expression grade index (GGI).⁹ The GGI is an algorithm derived to enumerate the degree of similarity between a tumor sample and histologic grade. In our previous study, genomic grading using the GGI was a stronger predictor of clinical outcome than histologic grade in systemically untreated BC patients.⁹ The aim of this study was to determine whether genomic grade could aid in the task of refining ER-positive molecular subtypes and, in particular, how a classification based on high or low genomic grade compares with previously described molecular classifications.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Further descriptions of the methods are available in the Appendix (online only).

Tumor Samples
Our own data set consisted of 335 early-stage BC samples; 86 of these have been previously used in another study (methods described in Sotiriou et al,⁹ with the raw data available at the Gene Expression Omnibus repository database [http://www.ncbi.nlm.nih.gov/geo/], accession code GSE2990). Our previously unpublished data set consisted of 249 samples from patients, all of whom had received adjuvant tamoxifen only (henceforth referred to as the tamoxifen-treated data set). The raw data for the tamoxifen-treated data set are also available at the Gene Expression Omnibus database.

For our unpublished data set, microarray analysis was performed with Affymetrix U133A Genechips (Affymetrix, Santa Clara, CA). This data set contained samples from the John Radcliffe Hospital, Oxford, United Kingdom; Guys Hospital, London, United Kingdom; and Uppsala University Hospital, Uppsala, Sweden. Each hospital's institutional ethics board approved the use of the tissue material, and written informed consent was obtained.

We also used five other publicly available data sets in our analysis, described in the following recent publications: Sorlie et al,^1,2 van de Vijver et al,¹⁰ Wang et al,¹¹ and Sotiriou et al.³ The survival analyses involved (1) only ER-positive tumors from patients who had received no systemic adjuvant treatment to visualize the natural disease history of the subgroups defined by GGI (total of 417 ER-positive patients, henceforth referred to as the untreated data set; van de Vijver et al,¹⁰ n = 122; Wang et al,¹¹ n = 209; and 86 samples from our previously published data set) and (2) the tamoxifen-treated data set (n = 249) to gauge the effect of adjuvant tamoxifen monotherapy. All clinical data are listed in Appendix Tables A1 and A2 (online only).

Data Analysis
ER and progesterone receptor expression levels. Eligible samples were selected according to a positive ER status by ligand-binding assay. The cutoff value for classification as positive or negative for ER and progesterone receptor (PgR) was 10 fmol/mg protein. A positive ER level was confirmed by microarray expression levels using probe sets 205225_at and 208305_at, representing the ESR1 and PgR genes, respectively. For division into hormone receptor–rich and –poor groups for the survival analyses, the median expression level of these probe sets was used as the cutoff.

Histologic grade. Histologic grade was based on the Elston-Ellis grading system.¹² A single pathologist graded the Oxford and the Swedish populations. Central pathology review grading was available for the data set from van de Vijver et al.¹⁰ The Wang et al¹¹ data set was graded by local pathologists.

Index based on the expression of proliferation-related genes to quantify genomic grade: GGI. We have previously described a GGI.⁹ Briefly, this index is a scoring system based on the average expression levels of genes associated with histologic grade. A high GGI corresponds to high grade, and a low GGI corresponds to a low grade. Although it is possible to use the GGI as a continuous variable, we dichotomized the GGI into high–and low–genomic grade subgroups for the generation of Kaplan-Meier curves and hazard ratio estimation between risk groups. The cutoff used was chosen to maximize the separation between histologic grades 1 and 3, and no survival information was used to optimize the high- and low-risk groups.

The probe sets of the Affymetrix U133A GeneChip were mapped to other microarray platforms by matching the official gene symbol.⁹ Of the 128 probe sets (representing 97 genes) that comprise the GGI, we could map 48 and 63 sets in the Sorlie et al¹ 2001 and Sorlie et al² 2003 data sets, respectively; 49 in the Sotiriou et al³ data set; and 122 in the van de Vijver et al¹⁰ data set.

Comparison With Other ER-Positive Molecular Classifications
Hierarchical clustering with the intrinsic gene set. We investigated the relationship between the GGI and the luminal A and B and normal-like subtypes described by Perou et al⁴ using an intrinsic gene set and the data sets of Sorlie et al^1,2 and van de Vijver et al as reported in Chang et al.¹³

The RS model. Using the microarray data and the 21-gene RS algorithm developed by Paik et al,⁷ the three risk groups were produced in the tamoxifen-treated data set. The low- and intermediate-risk categories were combined into one low-risk group because their clinical outcomes were not significantly different.

Statistical Analysis
Wilcoxon signed rank tests were performed to compare the GGI values of the BC subtypes. Correlation between the RS and the GGI was performed using a Pearson correlation. Survival curves were visualized using Kaplan-Meier plots and compared using log-rank tests using the end point of time to distant metastasis (TDM), which is often used as a surrogate marker for BC-specific survival. The univariate and multivariate hazard ratios were estimated using Cox regression analysis. All survival analyses were stratified by data set. The P values in the multivariate analysis were based on Wald tests. In analyses involving one or more clinical variables, a patient was excluded if the value of one variable was missing. All statistical tests were two sided. Time-dependent receiver operating characteristic curves were computed using the method described by Heagerty et al.¹⁴ Statistical analysis was performed using SPSS Statistical Software Package, version 13.0 (SPSS Inc, Chicago, IL) and the R software package (version 2006, http://www.R-project.org).

The relationship between survival and some continuous variables was assessed using a variant of a method introduced to compute the estimated survival for individual rate of distant metastases plots.^15-17 We have plotted the estimated proportion of distant metastasis with respect to the GGI, ESR1, and PgR using a Cox model fitted with only the variable under study.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Applying the GGI to the Luminal Subtypes
To investigate the expression of the genes in the GGI in relation to the previously reported molecular subtypes, expression data were extracted and clustered with the samples ordered according to subtypes as presented in the original articles¹ (Fig 1A). In general, the ER-negative subtypes, the basal-like and ErbB2-like subtypes, had high GGI values or were classified as having high genomic grade (GGI 0). In contrast, almost all of the ER-positive luminal A subtypes, which had the best clinical outcome, had low GGI values. These GGI values represented low genomic grade similar to the normal-like subtype. However, compared with the luminal A subtype, the luminal B and C tumors had significantly higher GGI values (P < .01; Appendix Table A3, online only). From these results, we can observe that the luminal subtypes were associated with a diverse range of GGI values, whereas the majority of ER-negative tumors were predominantly of high genomic grade.

View larger version (20K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 1. (A and B) Applying genomic grade to the previously reported molecular subtypes. Box plots of the gene expression grade index (GGI) score (median and range) are placed below each subtype. Samples previously undefined by the intrinsic gene set are shown in gray. (C) GGI values of the five subtypes and unclassified samples from the data set of van de Vijver et al.10

Generation of Two ER-Positive Subtypes Defined by Genomic Grade and Assessment of Their Clinical Relevance
We went on to further validate this concept in a data set of 666 ER-positive BC samples. We assigned each sample to either a high–or low–genomic grade subtype depending on their GGI value. To determine whether our classification produced subgroups that were of clinical importance, Kaplan-Meier survival analyses were performed comparing the ER-positive tumors according to their GGI score (high v low genomic grade) and expression levels of ER and PgR for both untreated and tamoxifen only–treated samples (Fig 2).

View larger version (17K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 2. Kaplan-Meier survival curves for time to distant metastasis (TDM) for gene expression grade index (GGI; high v low), quantitative ESR1 expression levels (rich v poor), and PgR expression levels (rich v poor). (A, B, and C) The results for the untreated data set (n = 417). (D, E, and F) The results for the tamoxifen-treated data set (n = 249). PgR, progesterone receptor.

As further demonstration of the prognostic value of genomic grade in ER-positive tumors, we generated figures displaying the estimated expected rate of distant recurrence as a continuous function of the GGI and compared this with continuous levels of ESR1 and PgR for each data set used in this analysis (Fig 3; Appendix Fig A2, online only). As shown, the 10-year estimated expected rate of developing distant metastases increased steeply as the GGI score increased, indicating high discriminatory value in ER-positive tumors.

View larger version (21K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 3. Plots of the rate of the development of distant metastasis in estrogen receptor (ER) –positive patients. (A) Results for the systemically untreated data set (Wang et al11 data set). (B) Results for the tamoxifen-treated data set. The x-axis indicates the individual score per patient, the solid lines indicate median scores, and the dashed lines indicate the 25% and 75% CIs. GGI, gene expression grade index; PgR, progesterone receptor.

View larger version (11K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 4. Time-dependent receiver operating characteristic curves for the gene expression grade index (GGI) and the 21-gene recurrence score (RS) for metastases within 10 years for the tamoxifen-treated data set. Area under the curve is 0.7 for GGI and 0.69 for RS. P value shows curves are not significantly different.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

As shown in Figure 1, despite using a different approach to produce molecular subgroups, we have shown that classifications overlap significantly with the previously reported molecular classification that was produced using the intrinsic gene set and hierarchical cluster analysis.^1-3 Interestingly, the GGI also is significantly correlated with and performs similarly to the RS algorithm (using the microarray data) in the tamoxifen-treated data set. However, the use of the GGI has a number of potential advantages. First, its value is simply calculated by averaging the expression level of (up to) 97 genes. As demonstrated, a result can be achieved by using only a subset of genes with little loss in performance, independent of the platform used.⁹ Second, the biologic function of the set of genes used to define the GGI is understood because these genes were selected through their correlation with histologic grade. Because there seems to be significant concordance in outcome predictions among several different gene sets used for prognostication,²⁴ the identification of a common biologic phenomenon is of utmost importance. Therefore, the use of genomic grade to stratify ER-positive patients may facilitate both the introduction of a molecular subtype classification into clinical practice and focused scientific research.

Currently, there are several molecular signatures that claim to be able to predict prognosis in ER-positive BC patients treated with tamoxifen.^6,7,11,25-27 Proliferation-related genes seem to be the essential component of these existing gene signatures.⁹ Dai et al²⁸ have previously reported that increased cell proliferation was associated with poor outcome in a subset of BC patients. In the RS,⁷ the proliferation set of genes has the highest weight or coefficient in the formula, indicating their crucial importance determining prognosis.⁷ Similarly, another molecular classification using only estrogen-regulated genes to divide ER-positive BC showed a significant association of the poor outcome subgroup with histologic grade and high expression of proliferation genes.⁶ In this study, we attempt to reconcile the subgroups generated by genomic grade and those previously reported by finding a common biologic thread that may help explain the heterogeneity behind ER-positive BC. In the future, it will be important to further compare the different categorizations to better understand their relationships and to develop an optimal, widely accepted, and easily implemented stratification within tumors classified as endocrine responsive. A prospectively acquired cohort will be also important to eliminate the selection bias that is inevitable with archived frozen tumor samples, as seen with the lack of significance for nodal status and the relatively poor 10-year TDM in the tamoxifen-treated data set compared with what would be expected from a population-based series.²⁹

In conclusion, the use of genomic grade as defined by the GGI can distinguish two subtypes within ER-positive BCs in a reliable and reproducible manner across multiple data sets. In the future, stratification by subtype in prospective clinical trials may elucidate important diverse effects of the various endocrines, chemotherapies, and biologic agents. At present, it is unclear whether the high–genomic grade subgroup will benefit from chemotherapy or alternative antiestrogen agents. In addition, further focused biologic investigation into the upstream oncogenic pathways that drive the cell cycle machinery could be beneficial in developing new agents to treat the high-grade subgroup.³⁰

	AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
The authors indicated no potential conflicts of interest.

	AUTHOR CONTRIBUTIONS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Conception and design: Sherene Loi, Benjamin Haibe-Kains, Christine Desmedt, Mauro Delorenzi, Martine J. Piccart, Christos Sotiriou

Financial support: Martine J. Piccart

Administrative support: Sherene Loi, Benjamin Haibe-Kains, Christine Desmedt, Cheryl Gillett, Paul Ellis, Jonas Bergh, Martine J. Piccart, Christos Sotiriou

Provision of study materials or patients: Sherene Loi, Christine Desmedt, Françoise Lallemand, Andrew M. Tutt, Cheryl Gillett, Paul Ellis, Adrian Harris, Jonas Bergh, Denis Larsimont, Martine J. Piccart, Christos Sotiriou

Collection and assembly of data: Sherene Loi, Benjamin Haibe-Kains, Christine Desmedt, Françoise Lallemand, Andrew M. Tutt, Cheryl Gillett, Paul Ellis, Adrian Harris, Jonas Bergh, John A. Foekens, Jan G.M. Klijn, Denis Larsimont, Mauro Delorenzi, Christos Sotiriou

Data analysis and interpretation: Sherene Loi, Benjamin Haibe-Kains, Christine Desmedt, Adrian Harris, John A. Foekens, Jan G.M. Klijn, Denis Larsimont, Marc Buyse, Gianluca Bontempi, Mauro Delorenzi, Martine J. Piccart, Christos Sotiriou

Manuscript writing: Sherene Loi, Benjamin Haibe-Kains, Christine Desmedt, John A. Foekens, Jan G.M. Klijn, Gianluca Bontempi, Mauro Delorenzi, Martine J. Piccart, Christos Sotiriou

Final approval of manuscript: Sherene Loi, Benjamin Haibe-Kains, Christine Desmedt, Andrew M. Tutt, Adrian Harris, John A. Foekens, Jan G.M. Klijn, Marc Buyse, Gianluca Bontempi, Mauro Delorenzi, Martine J. Piccart, Christos Sotiriou

	Appendix

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Supplementary Methods
Tumor samples. Samples from Oxford and London were shipped to the Jules Bordet Institute in Brussels, Belgium, where RNA was extracted and samples were hybridized. For the samples from Uppsala, RNA was extracted at the Karolinska Institute, Stockholm, Sweden, and samples were hybridized at the Genome Institute of Singapore in Singapore. The quality of the RNA obtained from each tumor sample was assessed via the RNA profile generated by the Agilent bioanalyzer (Agilent, Santa Clara, CA). RNA extraction, amplification, hybridization, and image scanning were performed according to standard Affymetrix protocols (Affymetrix, Santa Clara, CA). Gene expression values from the CEL files were normalized by use of the standard quantile normalization method in Robust Multiarray Analysis (Bolstad BM, Irizarry RA, Astrand M, et al. Bioinformatics 19:185-193, 2003). Each population was normalized separately. Each hospital's institutional ethics board approved the use of the tissue material, and written informed consent was obtained.

Derivation of the cutoff for the gene expression grade index. The cutoff defining the gene expression grade index (GGI) low- and high-risk groups is computed from the information provided by histologic grade. This cutoff is the equidistant point between the means of the GGI of histologic grade 1 and grade 3 patient samples. We can classify all of the patients as low or high risk according to their GGI value using a cross-validation procedure and obtain almost identical classification (Sotiriou C, Wirapati P, Loi S, et al. J Natl Cancer Inst 98:262-272, 2006; data not shown). The cutoff was selected to separate, as much as possible, histologic grades 1 and 3 without using any survival information. Therefore, this cutoff was not clinically optimized but was adequate to demonstrate the prognostic abilities of the GGI in estrogen receptor (ER) –positive breast cancer samples. Therefore, a better cutoff could be found to classify patients in low- and high-risk groups; however, this clinically optimized cutoff is difficult to estimate and should be studied in a large clinical trial.

Rate of distant metastases plots. The x-axis represents the GGI or the expression levels of the ESR1 and PgR genes, and each point shows the individual patient levels. All continuous values are scaled to be able to pool all the values in a meta-analysis. The GGI is scaled by dividing each GGI value by the distance between the means of GGI from histologic grade 1 and grade 3 patients. The expression levels are scaled to improve comparison between different data sets. The expression levels of ESR1 and PgR are scaled using the following formula and centered by the median of each population separately.

where j = 0,..., n, n being the number of patients.

Comparison with other ER–positive molecular classifications: Hierarchical clustering with the intrinsic gene set. We used the Cluster program to perform average linkage hierarchical cluster analysis (Eisen MB, Spellman PT, Brown PO, et al. Proc Natl Acad Sci U S A 95:14863-14868, 1998) after median centering of each gene in the GGI using an uncentered Pearson correlation as similarity measurement. The cluster results were viewed using TreeView (Java Tree View, Boston, MA). Expression data were downloaded and extracted from data sets of Sorlie et al (Sorlie T, Perou CM, Tibshirani R, et al. Proc Natl Acad Sci U S A 98:10869-10874, 2001; Sorlie T, Tibshirani R, Parker J, et al. Proc Natl Acad Sci U S A 100:8418-8423, 2003). The samples were ordered according to subtype as presented in the figures of the original publications (Sorlie T, Perou CM, Tibshirani R, et al. Proc Natl Acad Sci U S A 98:10869-10874, 2001; Sorlie T, Tibshirani R, Parker J, et al. Proc Natl Acad Sci U S A 100:8418-8423, 2003; Sotiriou C, Neo SY, McShane LM, et al. Proc Natl Acad Sci U S A 100:10393-10398, 2003) to investigate the relationship between the expression of the genes in the GGI and the previously reported molecular breast cancer subtypes. For the data set of van de Vijver et al (van de Vijver MJ, He YD, van't Veer LJ, et al. N Engl J Med 347:1999-2009, 2002), the five-class intrinsic gene signature was assigned by matching the expression value of the intrinsic genes to the nearest expression centroid of the five classes, as previously published (Chang HY, Nuyten DS, Sneddon JB, et al. Proc Natl Acad Sci U S A 102:3738-3743, 2005).

Comparison with other ER-positive molecular classifications: The recurrence score model. To classify the tamoxifen-treated data set using the recurrence score (RS) model predictor, we used the 21 genes and RS algorithm as described in Paik et al (Paik S, Shak S, Tang G, et al. N Engl J Med 351:2817-2826, 2004). Briefly, the microarray expression data of the required 16 genes were normalized to the five reference genes. We then used these values and the algorithm to generate an RS score for each sample. The same cutoffs were used (scores of 0 to 18, 19 to 31, or > 31) to assign each patient to a low-, intermediate-, or high-risk category.

View larger version (16K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig A1. Kaplan-Meier curves of subtypes from the van de Vijver et al data set (van de Vijver MJ, He YD, van't Veer LJ, et al. N Engl J Med 347:1999-2009, 2002).

View larger version (21K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig A2. Plots of the rate of the development of distant metastasis as a continuous function of gene expression grade index (GGI), ESR1 expression levels, and PgR expression levels. (A) Results from the systemically untreated data set (our data set). (B) Results from the van de Vijver et al (van de Vijver MJ, He YD, van't Veer LJ, et al. N Engl J Med 347:1999-2009, 2002) data set. The x-axis indicates the individual score per patient. The solid lines indicate the median scores for each variable, and the dashed lines indicate the 25% and 75% CIs. Only estrogen receptor–positive patients were used in this analysis.

	NOTES

 
Supported by grants from the Jean-Claude Heuson Breast Cancer Foundation (S.L., C.S.), the Belgian National Foundation for Research (FNRS; B.H.-K., C.D., C.S.), the E. Lauder Breast Cancer Foundation (C.S.), and the MEDIC Foundation (C.S.).

S.L., B.H.-K., and C.D. contributed equally to this article.

Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
1. Sorlie T, Perou CM, Tibshirani R, et al: Gene expression patterns of breast carcinomas distinguish tumour subclasses with clinical implications. Proc Natl Acad Sci U S A 98:10869-10874, 2001[Abstract/Free Full Text]

2. Sorlie T, Tibshirani R, Parker J, et al: Repeated observation of breast tumour subtypes in independent gene expression data sets. Proc Natl Acad Sci U S A 100:8418-8423, 2003[Abstract/Free Full Text]

3. Sotiriou C, Neo SY, McShane LM, et al: Breast cancer classification and prognosis based on gene expression profiles from a population-based study. Proc Natl Acad Sci U S A 100:10393-10398, 2003[Abstract/Free Full Text]

4. Perou CM, Sorlie T, Elsen MB, et al: Molecular portraits of human breast tumours. Nature 406:747-752, 2000[CrossRef][Medline]

5. Hu Z, Fan C, Oh DS, et al: The molecular portraits of breast tumors are conserved across microarray platforms. BMC Genomics 7:96, 2006[CrossRef][Medline]

6. Oh DS, Troester MA, Usary J, et al: Estrogen-regulated genes predict survival in hormone receptor-positive breast cancers. J Clin Oncol 24:1656-1664, 2006[Abstract/Free Full Text]

7. Paik S, Shak S, Tang G, et al: A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med 351:2817-2826, 2004[Abstract/Free Full Text]

8. Sorlie T: Molecular portraits of breast cancer: Tumour subtypes as distinct disease entities. Eur J Cancer 40:2667-2675, 2004[CrossRef][Medline]

9. Sotiriou C, Wirapati P, Loi S, et al: Gene expression profiling in breast cancer: Understanding the molecular basis of intermediate histologic grade to improve prognosis. J Natl Cancer Inst 98:262-272, 2006[Abstract/Free Full Text]

10. van de Vijver MJ, He YD, van't Veer LJ, et al: A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999-2009, 2002[Abstract/Free Full Text]

11. Wang Y, Klijn J, Zhang Y, et al: Gene-expression profiles to predict distant metastasis of lymph-node-negative primary breast cancer. Lancet 365:671-679, 2005[Medline]

12. Elston CW, Ellis IO: Pathological prognostic factors in breast cancer: I. The value of histological grade in breast cancer—Experience from a large study with long-term follow-up. Histopathology 19:403-410, 1991[Medline]

13. Chang HY, Nuyten DS, Sneddon JB, et al: Robustness, scalability, and integration of a wound-response gene expression signature in predicting breast cancer survival. Proc Natl Acad Sci U S A 102:3738-3743, 2005[Abstract/Free Full Text]

14. Heagerty PJ, Lumley T, Pepe MS: Time-dependent ROC curves for censored survival data and a diagnostic marker. Biometrics 56:337-344, 2000[CrossRef][Medline]

15. Therneau TM, Grambasch PM: Modeling Survival Data: Extending the Cox Model. New York, NY, Springer, 2000

16. Valenta Z, Weissfeld L: Estimation of the survival function of Gray's piecewise-constant time-varying coefficients model. Stat Med 21:717-727, 2002[CrossRef][Medline]

17. Gray RJ: Flexible methods for analyzing survival data using splines, with applications to breast cancer prognosis. J Am Stat Assoc 87:942-951, 1992[CrossRef]

18. Goss PE, Ingle JN, Martino S, et al: A randomized trial of letrozole in postmenopausal women after five years of tamoxifen therapy for early-stage breast cancer. N Engl J Med 349:1793-1802, 2003[Abstract/Free Full Text]

19. Boccardo F, Rubagotti A, Puntoni M, et al: Switching to anastrozole versus continued tamoxifen treatment of early breast cancer: Preliminary results of the Italian Tamoxifen Anastrozole Trial. J Clin Oncol 23:5138-5147, 2005[Abstract/Free Full Text]

20. Thurlimann B, Keshaviah A, Coates AS, et al: A comparison of letrozole and tamoxifen in postmenopausal women with early breast cancer. N Engl J Med 353:2747-2757, 2005[Abstract/Free Full Text]

21. Baum M, Buzdar A, Cuzick J, et al: Anastrozole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal women with early-stage breast cancer: Results of the ATAC (Arimidex, Tamoxifen Alone or in Combination) trial efficacy and safety update analyses. Cancer 98:1802-1810, 2003[CrossRef][Medline]

22. Coombes RC, Hall E, Gibson LJ, et al: A randomized trial of exemestane after two to three years of tamoxifen therapy in postmenopausal women with primary breast cancer. N Engl J Med 350:1081-1092, 2004[Abstract/Free Full Text]

23. Jakesz R, Jonat W, Gnant M, et al: Switching of postmenopausal women with endocrine-responsive early breast cancer to anastrozole after 2 years’ adjuvant tamoxifen: Combined results of ABCSG trial 8 and ARNO 95 trial. Lancet 366:455-462, 2005[CrossRef][Medline]

24. Fan C, Oh DS, Wessels L, et al: Concordance among gene-expression-based predictors for breast cancer. N Engl J Med 355:560-569, 2006[Abstract/Free Full Text]

25. van 't Veer LJ, Dai H, van de Vijver MJ, et al: Gene expression profiling predicts clinical outcome of breast cancer. Nature 415:530-536, 2002[CrossRef][Medline]

26. Ma XJ, Wang Z, Ryan PD, et al: A two-gene expression ratio predicts clinical outcome in breast cancer patients treated with tamoxifen. Cancer Cell 5:607-616, 2004[CrossRef][Medline]

27. Loi S, Piccart MJ, Desmedt C, et al: Prediction of early distant relapses on tamoxifen in early-stage breast cancer (BC): A potential tool for adjuvant aromatase inhibitor (AI) tailoring. J Clin Oncol 23:6s, 2005 (suppl, abstr)

28. Dai H, van't Veer L, Lamb J, et al: A cell proliferation signature is a marker of extremely poor outcome in a subpopulation of breast cancer patients. Cancer Res 65:4059-4066, 2005[Abstract/Free Full Text]

29. Early Breast Cancer Trialists’ Collaborative Group: Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: An overview of the randomised trials. Lancet 365:1687-1717, 2005[CrossRef][Medline]

30. Wilcken NR, Prall OW, Musgrove EA, et al: Inducible overexpression of cyclin D1 in breast cancer cells reverses the growth-inhibitory effects of antiestrogens. Clin Cancer Res 3:849-854, 1997[Abstract]

Submitted May 4, 2006; accepted January 2, 2007.

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

Related Correspondence

• Expression Profiling in Breast Carcinoma: New Insights on Old Prognostic Factors?
  Marc Debled, Gaëtan MacGrogan, Isabelle de Mascarel, Véronique Brouste, Hervé Bonnefoi, and Louis Mauriac
  JCO 2007 25: 4316-4317 [Full Text]

This article has been cited by other articles:

				O. Gluz, C. Liedtke, N. Gottschalk, L. Pusztai, U. Nitz, and N. Harbeck Triple-negative breast cancer--current status and future directions Ann. Onc., November 9, 2009; (2009) mdp492v1. [Abstract] [Full Text] [PDF]

				G. Frasci, G. D'Aiuto, P. Comella, M. D'Aiuto, M. Di Bonito, P. Ruffolo, G. Iodice, A. Petrillo, S. Lastoria, P. Oliviero, et al. Preoperative weekly cisplatin, epirubicin, and paclitaxel (PET) improves prognosis in locally advanced breast cancer patients: an update of the Southern Italy Cooperative Oncology Group (SICOG) randomised trial 9908 Ann. Onc., October 8, 2009; (2009) mdp356v1. [Abstract] [Full Text] [PDF]

				S. H. Dairkee, A. Sayeed, G. Luciani, S. Champion, Z. Meng, L. R. Jakkula, H. S. Feiler, J. W. Gray, and D. H. Moore Immutable Functional Attributes of Histologic Grade Revealed by Context-Independent Gene Expression in Primary Breast Cancer Cells Cancer Res., October 1, 2009; 69(19): 7826 - 7834. [Abstract] [Full Text] [PDF]

				A. Torkamani and N. J. Schork Prestige centrality-based functional outlier detection in gene expression analysis Bioinformatics, September 1, 2009; 25(17): 2222 - 2228. [Abstract] [Full Text] [PDF]

				H. J. Burstein, I. Souter, H. A. D'Alessandro, and D. C. Sgroi Case 25-2009 -- A 36-Year-Old Woman with Hormone-Receptor-Positive Breast Cancer N. Engl. J. Med., August 13, 2009; 361(7): 699 - 707. [Full Text] [PDF]

				F. Correa Geyer and J. S. Reis-Filho Microarray-based Gene Expression Profiling as a Clinical Tool for Breast Cancer Management: Are We There Yet? International Journal of Surgical Pathology, August 1, 2009; 17(4): 285 - 302. [Abstract] [PDF]

				C. Liedtke, C. Hatzis, W. F. Symmans, C. Desmedt, B. Haibe-Kains, V. Valero, H. Kuerer, G. N. Hortobagyi, M. Piccart-Gebhart, C. Sotiriou, et al. Genomic Grade Index Is Associated With Response to Chemotherapy in Patients With Breast Cancer J. Clin. Oncol., July 1, 2009; 27(19): 3185 - 3191. [Abstract] [Full Text] [PDF]

				C. H. Barrios, C. Sampaio, J. Vinholes, and R. Caponero What is the role of chemotherapy in estrogen receptor-positive, advanced breast cancer? Ann. Onc., July 1, 2009; 20(7): 1157 - 1162. [Abstract] [Full Text] [PDF]

				A. M. Sieuwerts, J. Kraan, J. Bolt-De Vries, P. Van Der Spoel, F. Elstrodt, M. Smid, M. Timmermans, B. Mostert, M. Schutte, J. W. M. Martens, et al. Response: Re: Anti-Epithelial Cell Adhesion Molecule Antibodies and the Detection of Circulating Normal-Like Breast Tumor Cells J Natl Cancer Inst, June 16, 2009; 101(12): 896 - 897. [Full Text] [PDF]

				S. P. Pitroda, N. N. Khodarev, M. A. Beckett, D. W. Kufe, and R. R. Weichselbaum From the Cover: MUC1-induced alterations in a lipid metabolic gene network predict response of human breast cancers to tamoxifen treatment PNAS, April 7, 2009; 106(14): 5837 - 5841. [Abstract] [Full Text] [PDF]

				M. Schmidt, J. G. Hengstler, C. von Torne, H. Koelbl, and M. C. Gehrmann Coordinates in the Universe of Node-Negative Breast Cancer Revisited Cancer Res., April 1, 2009; 69(7): 2695 - 2698. [Abstract] [Full Text] [PDF]

				J. S. Parker, M. Mullins, M. C.U. Cheang, S. Leung, D. Voduc, T. Vickery, S. Davies, C. Fauron, X. He, Z. Hu, et al. Supervised Risk Predictor of Breast Cancer Based on Intrinsic Subtypes J. Clin. Oncol., March 10, 2009; 27(8): 1160 - 1167. [Abstract] [Full Text] [PDF]

				C. Sotiriou and L. Pusztai Gene-Expression Signatures in Breast Cancer N. Engl. J. Med., February 19, 2009; 360(8): 790 - 800. [Full Text] [PDF]

				S Badve and H Nakshatri Oestrogen-receptor-positive breast cancer: towards bridging histopathological and molecular classifications J. Clin. Pathol., January 1, 2009; 62(1): 6 - 12. [Abstract] [Full Text] [PDF]

				P. Bhat-Nakshatri, G. Wang, H. Appaiah, N. Luktuke, J. S. Carroll, T. R. Geistlinger, M. Brown, S. Badve, Y. Liu, and H. Nakshatri AKT Alters Genome-Wide Estrogen Receptor {alpha} Binding and Impacts Estrogen Signaling in Breast Cancer Mol. Cell. Biol., December 15, 2008; 28(24): 7487 - 7503. [Abstract] [Full Text] [PDF]

				M. Dowsett and A. K. Dunbier Emerging Biomarkers and New Understanding of Traditional Markers in Personalized Therapy for Breast Cancer Clin. Cancer Res., December 15, 2008; 14(24): 8019 - 8026. [Abstract] [Full Text] [PDF]

				D. T. Ross, C.-y. Kim, G. Tang, O. L. Bohn, R. A. Beck, B. Z. Ring, R. S. Seitz, S. Paik, J. P. Costantino, and N. Wolmark Chemosensitivity and Stratification by a Five Monoclonal Antibody Immunohistochemistry Test in the NSABP B14 and B20 Trials Clin. Cancer Res., October 15, 2008; 14(20): 6602 - 6609. [Abstract] [Full Text] [PDF]

				L. R. Webster, S.-F. Lee, C. Ringland, A. L. Morey, A. M. Hanby, G. Morgan, K. Byth, P. A. Mote, P. J. Provan, I. O. Ellis, et al. Poor-Prognosis Estrogen Receptor-Positive Breast Cancer Identified by Histopathologic Subclassification Clin. Cancer Res., October 15, 2008; 14(20): 6625 - 6633. [Abstract] [Full Text] [PDF]

				B. Haibe-Kains, C. Desmedt, C. Sotiriou, and G. Bontempi A comparative study of survival models for breast cancer prognostication based on microarray data: does a single gene beat them all? Bioinformatics, October 1, 2008; 24(19): 2200 - 2208. [Abstract] [Full Text] [PDF]

				M. P. Goetz, V. J. Suman, F. J. Couch, M. M. Ames, J. M. Rae, M. G. Erlander, X.-J. Ma, D. C. Sgroi, C. A. Reynolds, W. L. Lingle, et al. Cytochrome P450 2D6 and Homeobox 13/Interleukin-17B Receptor: Combining Inherited and Tumor Gene Markers for Prediction of Tamoxifen Resistance Clin. Cancer Res., September 15, 2008; 14(18): 5864 - 5868. [Abstract] [Full Text] [PDF]

				C. Desmedt, B. Haibe-Kains, P. Wirapati, M. Buyse, D. Larsimont, G. Bontempi, M. Delorenzi, M. Piccart, and C. Sotiriou Biological Processes Associated with Breast Cancer Clinical Outcome Depend on the Molecular Subtypes Clin. Cancer Res., August 15, 2008; 14(16): 5158 - 5165. [Abstract] [Full Text] [PDF]

				C. R. Acharya and A. Potti Documenting Biospecimen Conditions in Reports of Studies--Reply JAMA, August 13, 2008; 300(6): 651 - 651. [Full Text] [PDF]

				M. Schmidt, D. Bohm, C. von Torne, E. Steiner, A. Puhl, H. Pilch, H.-A. Lehr, J. G. Hengstler, H. Kolbl, and M. Gehrmann The Humoral Immune System Has a Key Prognostic Impact in Node-Negative Breast Cancer Cancer Res., July 1, 2008; 68(13): 5405 - 5413. [Abstract] [Full Text] [PDF]

				X.-J. Ma, R. Salunga, S. Dahiya, W. Wang, E. Carney, V. Durbecq, A. Harris, P. Goss, C. Sotiriou, M. Erlander, et al. A Five-Gene Molecular Grade Index and HOXB13:IL17BR Are Complementary Prognostic Factors in Early Stage Breast Cancer Clin. Cancer Res., May 1, 2008; 14(9): 2601 - 2608. [Abstract] [Full Text] [PDF]

				J. S. Ross, C. Hatzis, W. F. Symmans, L. Pusztai, and G. N. Hortobagyi Commercialized Multigene Predictors of Clinical Outcome for Breast Cancer Oncologist, May 1, 2008; 13(5): 477 - 493. [Abstract] [Full Text] [PDF]

				C. R. Acharya, D. S. Hsu, C. K. Anders, A. Anguiano, K. H. Salter, K. S. Walters, R. C. Redman, S. A. Tuchman, C. A. Moylan, S. Mukherjee, et al. Gene Expression Signatures, Clinicopathological Features, and Individualized Therapy in Breast Cancer JAMA, April 2, 2008; 299(13): 1574 - 1587. [Abstract] [Full Text] [PDF]

				S. H. Dairkee, J. Seok, S. Champion, A. Sayeed, M. Mindrinos, W. Xiao, R. W. Davis, and W. H. Goodson Bisphenol A Induces a Profile of Tumor Aggressiveness in High-Risk Cells from Breast Cancer Patients Cancer Res., April 1, 2008; 68(7): 2076 - 2080. [Abstract] [Full Text] [PDF]

				M. Chanrion, V. Negre, H. Fontaine, N. Salvetat, F. Bibeau, G. M. Grogan, L. Mauriac, D. Katsaros, F. Molina, C. Theillet, et al. A Gene Expression Signature that Can Predict the Recurrence of Tamoxifen-Treated Primary Breast Cancer Clin. Cancer Res., March 15, 2008; 14(6): 1744 - 1752. [Abstract] [Full Text] [PDF]

				I. G. Yulug and B. Gur-Dedeoglu Functional genomics in translational cancer research: focus on breast cancer Brief Funct Genomic Proteomic, March 7, 2008; (2008) eln009v1. [Abstract] [Full Text] [PDF]

				C. Camps, F. M. Buffa, S. Colella, J. Moore, C. Sotiriou, H. Sheldon, A. L. Harris, J. M. Gleadle, and J. Ragoussis hsa-miR-210 Is Induced by Hypoxia and Is an Independent Prognostic Factor in Breast Cancer Clin. Cancer Res., March 1, 2008; 14(5): 1340 - 1348. [Abstract] [Full Text] [PDF]

				N. U. Lin and E. P. Winer Advances in Adjuvant Endocrine Therapy for Postmenopausal Women J. Clin. Oncol., February 10, 2008; 26(5): 798 - 805. [Abstract] [Full Text] [PDF]

				P. Finetti, N. Cervera, E. Charafe-Jauffret, C. Chabannon, C. Charpin, M. Chaffanet, J. Jacquemier, P. Viens, D. Birnbaum, and F. Bertucci Sixteen-Kinase Gene Expression Identifies Luminal Breast Cancers with Poor Prognosis Cancer Res., February 1, 2008; 68(3): 767 - 776. [Abstract] [Full Text] [PDF]

				J. Adelaide, P. Finetti, I. Bekhouche, L. Repellini, J. Geneix, F. Sircoulomb, E. Charafe-Jauffret, N. Cervera, J. Desplans, D. Parzy, et al. Integrated Profiling of Basal and Luminal Breast Cancers Cancer Res., December 15, 2007; 67(24): 11565 - 11575. [Abstract] [Full Text] [PDF]

				W. Schroth, L. Antoniadou, P. Fritz, M. Schwab, T. Muerdter, U. M. Zanger, W. Simon, M. Eichelbaum, and H. Brauch Breast Cancer Treatment Outcome With Adjuvant Tamoxifen Relative to Patient CYP2D6 and CYP2C19 Genotypes J. Clin. Oncol., November 20, 2007; 25(33): 5187 - 5193. [Abstract] [Full Text] [PDF]

				M. Debled, G. MacGrogan, I. de Mascarel, V. Brouste, H. Bonnefoi, and L. Mauriac Expression Profiling in Breast Carcinoma: New Insights on Old Prognostic Factors? J. Clin. Oncol., September 20, 2007; 25(27): 4316 - 4317. [Full Text] [PDF]

				S. Loi, B. Haibe-Kains, C. Desmedt, and C. Sotiriou In Reply J. Clin. Oncol., September 20, 2007; 25(27): 4317 - 4318. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Erratum (v25,p3790) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Loi, S. Articles by Sotiriou, C. Search for Related Content PubMed PubMed Citation Articles by Loi, S. Articles by Sotiriou, C. Related Articles Related Correspondence Social Bookmarking                            What's this?