Genomic Grade Index (GGI) is a 97-gene signature that improves histologic grade (HG) classification in invasive breast carcinoma. In this prospective study we sought to evaluate the feasibility of performing GGI in routine clinical practice and its impact on treatment recommendations.
Patients with pT1pT2 or operable pT3, N0-3 invasive breast carcinoma were recruited from 8 centers in Belgium. Fresh surgical samples were sent at room temperature in the MapQuant Dx™ PathKit for centralized genomic analysis. Genomic profiles were determined using Affymetrix U133 Plus 2.0 and GGI calculated using the MapQuant Dx® protocol, which defines tumors as low or high Genomic Grade (GG-1 and GG-3 respectively).
180 pts were recruited and 155 were eligible. The MapQuant test was performed in 142 cases and GGI was obtained in 78% of cases (n=111). Reasons for failures were 15 samples with <30% of invasive tumor cells (11%), 15 with insufficient RNA quality (10%), and 1 failed hybridization (<1%). For tumors with an available representative sample (≥ 30% inv. tumor cells) (n=127), the success rate was 87.5%. GGI reclassified 69% of the 54 HG2 tumors as GG-1 (54%) or GG-3 (46%). Changes in treatment recommendations occurred mainly in the subset of HG2 tumors reclassified into GG-3, with increased use of chemotherapy in this subset.
The use of GGI is feasible in routine clinical practice and impacts treatment decisions in early-stage breast cancer.
Citation: Metzger-Filho O, Catteau A, Michiels S, Buyse M, Ignatiadis M, Saini KS, et al. (2013) Genomic Grade Index (GGI): Feasibility in Routine Practice and Impact on Treatment Decisions in Early Breast Cancer. PLoS ONE8(8): e66848. https://doi.org/10.1371/journal.pone.0066848
Editor: Sharon A. Glynn, National University of Ireland Galway, Ireland
Received: July 10, 2012; Accepted: May 10, 2013; Published: August 19, 2013
Copyright: © 2013 Metzger-Filho et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The authors have no external support or funding to report.
Competing interests: The authors have the following interests. Ipsogen SA provided financial support for the conduct of this study and is the employer of Aurélie Catteau, Virginie Fasolo, Sihem Naji and Hélčne Peyro-Saint-Paul. Drs. Sotiriou and Piccart are co-inventors of patent and patent applications on the CGI owned by Universite Libre de Bruxelles and licensed to Ipsogen SA. Virginie Fasolo is co-inventor of patent applications jointly owned by Universite Libre de Bruxelles and Ipsoegn SA and licensed to Ipsogen SA. The name of the Patent is "METHODS FOR THE DIAGNOSIS, THE DETERMINATION OF THE GRADE OF A SOLID TUMOR AND THE PROGNOSIS OF A SUBJECT SUFFERING FROM CANCER". There are no further patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors.
Integrating the results of genomic tests with classic clinico-pathologic variables to make treatment decisions represents a major advance in early breast cancer . Several gene expression signatures have been developed with the aim of refining the prognosis of patients diagnosed with this disease . Retrospective analyses have demonstrated the ability of genomic signatures to classify patients into risk categories, thereby adding prognostic value to classic clinico-pathologic variables [2–4]. Moreover, such signatures provide predictive information, identifying the subgroups of patients most likely to benefit from therapeutic interventions such as chemotherapy [5,6]. The value of these signatures for use in clinical practice, however, is still under investigation in three prospective phase III clinical trials [7–9].
A major concern is the feasibility of incorporating genomic signatures into routine clinical procedures. As for any practice-changing technology, a careful evaluation of the logistics associated with its use is needed in order to ensure future applicability. This is particularly true for molecular assays requiring fresh or frozen tumor samples, which add complexity to routine surgical and pathology procedures.
GGI is a 97-gene microarray signature developed to refine histologic grade (HG) assessment, mainly capturing the expression of proliferation genes. In a seminal study, GGI was shown to improve the prognostic value of classic Elston-Ellis histologic grading in patients diagnosed with early breast cancer [3,10]. GGI is particularly useful for segregating HG2 tumors into low (GG-1) or high (GG-3) genomic grade with clinical behavior similar to HG1 and HG3 respectively, and has been recommended as a complement to histologic grade by the 2009 St Gallen consensus . Further analyses have demonstrated the ability of GGI to categorize estrogen receptor (ER)-positive tumors into two subtypes with distinct clinical outcomes based on the differences in their proliferation rates [12,13].
Evidence for GGI predictive value was provided in a retrospective analysis of patients treated with neoadjuvant chemotherapy . High GGI was associated with increased sensitivity to neoadjuvant anthracycline and taxane-based chemotherapy in both ER-positive and ER-negative breast cancer subsets. This finding was confirmed in a recent pooled analysis of almost 1000 patients .
The present study prospectively evaluated both the feasibility of implementing GGI in eight community-based hospitals and the impact of GGI results on treatment decisions.
Patients and Methods
All patients provided written informed consent prior to enrolment. The protocol was approved by the Jules Bordet Institute’s central ethics committee, and local ethics committees reviewed the protocol at each participating institution. The study was registered with the European Community’s Drug Regulatory Authorities Clinical Trial System (EudraCT 2009-015521-36). The protocol for this trial and supporting CONSORT checklist are available as supporting information; see Checklist S1 and Protocol S1. Eligible patients included women with histologically confirmed pT1-pT2 or pT3 operable primary invasive breast cancer with no more than three positive nodes (defined as lymph node metastases >2mm) and absence of distant metastasis. Patients with multifocal tumors were eligible, provided that a sample from each tumor was taken for the GGI test. Patients with in situ disease were eligible if invasive carcinoma was present. Patients with previous invasive cancer within the past five years were ineligible, except for patients treated adequately for carcinoma in situ of the cervix or non-melanoma skin cancer.
The primary objective of this study was to evaluate the feasibility of implementing GGI in community hospitals in Belgium for breast cancer patients with node negative and 1-3 node positive early breast cancer. As defined in the protocol, GGI would be considered a feasible genomic test if results were obtained in > 70% of evaluated patients. The secondary objective was to evaluate the impact of GGI on adjuvant treatment decisions for patients with early breast cancer. This was done by comparing physicians’ treatment recommendations before having knowledge of the GGI test results to recommendations with a hypothetical GG-1 and GG-3 result and to the treatment ultimately administered after discussion with the patient.
Participating centers and study material
Participating hospitals were required to have multidisciplinary breast cancer care structures and a tumor board in place and at least one dedicated physician (surgeon, pathologist, or medical oncologist) as a local coordinator. Eight Belgian hospitals participated in the study, and the logistics of the study was coordinated by the Jules Bordet Institute. At the initiation of the study, an instructional meeting was organized with all participating hospitals to discuss the logistics and their integration into routine clinical practice. An instructional video in which the full procedure – from sampling to shipping was demonstrated by an experienced pathologist was also provided to each pathologist involved in the study. Ready-to-use sample collection MapQuant Dx™ PathKits (Ipsogen) containing all necessary material for tumor sampling were provided to collect and ship tumors.
Tumor sampling and shipment
Fresh tumor specimens were sampled prior to adding any fixative and within one hour of surgery. Investigators were instructed to select a fragment from the periphery of the tumor where the tumor is the most proliferative, and to avoid its center, which is often necrotic and fibrotic. The minimum size required was 3mm diameter. Alternatively, frozen samples could be provided. In this case, 10x20μm thick sections per tumor sample were required. Samples (fresh or frozen) were immersed immediately into the RNAlater® preserving solution (Ambion Inc., Austin, TX, USA) and kept at 4°C (up to 3 weeks) or at room temperature (up to one week) until shipped.
One Hematoxylin and Eosin (H&E) slide representative of the tumor sample was also required in order to check the content of invasive tumor cells. The H&E slide was prepared from a section adjacent to the sample taken for the test. The MapQuant Dx PathKits containing single tumor samples, a representative H&E slide, and a sample report form (SRF) (for pathology laboratory information, study number, sampling procedure, storage condition, sampling date, shipping date, and kit-use checking controls) were sent at room temperature to a central laboratory (DNA Vision SA, Charleroi, Belgium), where the microarray analyses were performed.
Microarray analysis and GGI determination
Central pathology review was performed on all tumor samples received at the central laboratory. The H&E slides received had to contain at least 30% of invasive tumor cells for further analysis. RNA was extracted and purified with the RNeasy MiniKit® according to the manufacturer’s instructions (Qiagen). RNA was quantified and qualitatively assessed on triplicate samples on a NanoDrop® ND-1000 UV–Vis Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) and Agilent 2100 Bioanalyzer (Agilent Technology, Santa Clara, CA). Samples with ≥ 100 ng and a RIN ≥ 7 were qualified for microarray analyses performed on Affymetrix U133 Plus 2.0 arrays according to MapQuant Dx conditions (GEO accession number GSE43365). GGI was computed using the Ipsogen MapQuant Dx protocol (Table S1), which defines tumors as low GGI (GG-1) or high GGI (GG-3), based on GGI values. To ensure the robustness and accuracy of genomic grading as per the MapQuant protocol, when the GGI value of a sample could not be ascribed with a probability ≥ 75% to either GG-1 or GG-3, it was defined as equivocal (Eq).
At each participating center, a Case Report Form 1 (CRF section 1) was filled-in after surgery and pathological assessment. The records described the following: patient characteristics; definitive tumor histology; histologic grade (assessed according to Elston-Ellis recommendations); ER, progesterone receptor (PR) and HER2 status; tumor availability for the genomic test; and systemic adjuvant treatment recommendations by the treating physicians. Treatment recommendations were based on 1) patient and tumor characteristics, and 2) patient and tumor characteristics plus hypothetical GGI results (GG-1 or GG-3), each center using its usual routine decision algorithm. The actual GGI results were provided to participating centers only after receipt of the completed CRF section 1. Once this information was received, the physicians discussed treatment recommendations with their patients. The final treatment prescribed was then reported on CRF section 2, which was sent back to the coordinating center.
To test the success rate of reporting GGI, a sample size of 137 was calculated to have a power of 85% at a one-sided significance level of 5% using an empirical estimate of variance (null hypothesis H0: p = 0.70 vs. alternative hypothesis HA: p > 0.70) when the true success rate is equal to the prespecified value of 0.80 under the alternative (using an empirical estimate of variance). The original critical value from which the null hypothesis could be rejected was 106 successes out of 137. Descriptive analyses were performed to classify patients according to clinico-pathologic characteristics, HG and GGI, initial treatment recommendations, treatment recommendations after GGI results, and final treatment prescribed to patients. Concordance between HG and GGI was defined as the proportion of patients classified as HG1/GG-1 and HG3/GG-3. Statistical analysis was performed using SAS version 9.2
Between February and December 2010, 180 patients were enrolled across eight participating hospitals in Belgium. Following initial breast cancer surgery, a total of 25 patients were excluded because of the presence of more than three positive nodes (n = 22) or non-invasive tumor (n = 3).
The clinico-pathologic characteristics of the 155 eligible patients are summarized in table 1 and outlined here: the median age was 38 years (range 31-90); most patients were considered to be in either perfect health (43%) or to have only minor problems (33%); the median tumor diameter was 1.6 cm (range 0.6 cm to 5.5 cm); most patients had T1 tumors (68%) and node-negative disease (79%); and histological grade classified 38 tumors as HG1 (24%), 77 as HG2 (50%) and 40 as HG3 (26%).
|Age at diagnosis|
|Median [min-max]||38 [31-90]|
|Perfect health||66 (43%)|
|Minor problems||51 (33%)|
|Average for age||21 (13%)|
|Major problems||17 (11%)|
|Tumor size (cm)|
|Median [Min-Max]||1.6 [0.6-5.5]|
|1-3 positive||33 (21%)|
|ER pos / PR pos||115 (74%)|
|ER pos / PR neg||20 (13%)|
|ER neg / PR neg||20 (13%)|
A flow diagram of recruited patients and selected tumors is depicted in Figure 1. Out of 155 eligible patients, a total of 13 patients could not be retained for further genomic analysis (8 tumors estimated too small for sampling; no sample available for 3 patients; and 2 tumor samples sent without H&E slide). The median tumor size of the 8 tumors that were not sampled was 0.8 cm (range 0.6 cm to 1 cm). The 142 evaluated patients for whom a tumor sample and representative H&E slide were provided were thus included for the primary endpoint analysis. However, 15 tumors were excluded after central pathology assessment because of insufficient invasive tumor cell content (<30%), and an additional 15 tumor samples failed the RNA qualification. Out of the 112 remaining samples, hybridization was successful for 111 (99%).
Abbreviations: inv, invasive; N, nodal status.
With 111 successes out of 142 (78.2%, with a lower bound of the exact one-sided 95% confidence interval provided by 71.7%) the null hypothesis that the true probability is equal to 70% is rejected with a one-sided p-value of 0.017. In the subset of patients with ≥30% invasive tumor cells in the representative slide, GGI was obtained with a success rate of 87.5% (111/127).
Comparison between histologic grade and genomic grade classification
Cross-tabulations of histologic and genomic grade for the 111 patients with a GGI result are shown in table 2. A total of 21 (19%) tumors were classified as HG1, 54 (49%) as HG2 and 36 (32%) as HG3. GGI classified 39 (35%) tumors as GG-1, 48 (43%) as GG-3, and 24 (22%) as equivocal. Within the HG2 subgroup representing 49% of the cohort, GGI reclassified 69% of the cases (n=37) into GG-1 (n =20, 54%) and GG-3 (n = 17, 46%), respectively. The overall concordance between HG1 (n=15)/GG-1 (n=19) and HG3 (n=30)/GG-3 (n=31) was 90%. GG-Equivocal results were not included in the overall concordance estimate as it is non-informative for treatment decision.
Impact of genomic grade on treatment decisions
We first analyzed the potential impact of GGI on treatment decisions in the subgroup of patients with ER-positive/HER2-negative breast cancer where tumor grade is one of the key factors driving chemotherapy decision. Out of 155 eligible patients (described in Figure 1), 126 were classified as ER-positive and HER2-negative. Among these, 38 (30%) tumors were classified as HG1, 67 (53%) as HG2 and 21 (17%) as HG3. The physicians’ recommendations for adjuvant chemotherapy based on clinico-pathologic features were compared to their recommendations for chemotherapy based on clinico-pathologic features plus hypothetical GGI results (either GG-1 or GG-3). Physicians recommended chemotherapy to 33% of the patients based on clinico-pathologic features. Adding a hypothetical GG-1 to the clinico-pathologic features decreased the chemotherapy indications to 25%, while a hypothetical GG-3 result increased them to 54% as detailed in table 3. A similar analysis of the HG2 subgroup (n = 68) revealed that physicians recommended chemotherapy to 31% of patients based on clinico-pathologic features, to 24% based on clinico-pathologic features plus hypothetical GG-1 and to 59% when based on clinico-pathologic features plus hypothetical GG-3.
|Treatment recommendation||Clinico-pathologic characteristics (n, %)||Hypothetical GG-1 (n, %)||Hypothetical GG-3 (n, %)|
|CHT||42 (33%)||32 (25%)||68 (54%)|
|HT||84 (67%)||94 (75%)||58 (46%)|
As a second step, we evaluated the impact of GGI on final treatment decisions for patients diagnosed with HG2/ER-positive/HER2-negative breast cancers reclassified either as GG-1 or GG-3. Out of 111 patients for which the GGI test was available, 32 patients were diagnosed with HG2/ER-positive/HER2-negative breast cancer and reclassified as GG-1 (n=19) or GG-3 (n=13) as shown in Figure 2. One of the 3 HG2/GG-1 patients initially proposed for chemo-hormonotherapy (CHT) was finally treated with hormonotherapy only, while one patient initially recommended for hormonotherapy finally received CHT. By contrast, while CHT was initially recommended for 7 patients out of the 13 HG2/GG-3 (54%), 11 received CHT (85%). No changes in treatment recommendation occurred in the subset of patients with ER-positive breast cancer with HG1 or HG3 and reclassified into GGI equivocal.
The results of this prospective study demonstrate that measuring GGI is feasible in routine clinical practice. Collecting fresh tumor samples, shipping them to a central laboratory for gene expression profiling, and obtaining GGI results was achieved successfully in 78% of evaluated cases. In the subset of patients with adequate tumor sampling (representative H&E slide provided with ≥30% invasive tumor cells) GGI was feasible in 87.5% of cases. Overall concordance for HG1/GG-1 and HG3/GG-3 was 90%, and GGI reclassified 69% HG2 tumors, thereby providing additional information in cases where histologic grade is uninformative. This HG2 reclassification rate was identical to the one observed in a recent study in which MapQuant GGI was performed on a retrospective series of 163 early breast cancer samples .
Previous studies evaluating the feasibility of integrating genomic microarray technologies requiring fresh or frozen tumor tissue into the early treatment of breast cancer have reported similar success rates [17,18]. The 70-gene prognostic signature (MammaPrintTM) was prospectively evaluated in a study involving 16 community-based hospitals in the Netherlands, and the results were successfully provided for 73% of eligible patients (427 of 585) . In a subsequent pilot study, MammaPrintTM achieved a success rate of 72% (46 out of 64) , and reinforced the conduct of the completely enrolled 6600-patient phase III study MINDACT (Microarray in Node-negative and 1-3 node-positive disease may Avoid ChemoTherapy), evaluating whether there is benefit to combining MammaPrintTM with clinic-pathologic features for treatment decision-making in early breast cancer .
When compared to previous reports [17,18], it is clear that our study has been conducted in an era in which genomics are being increasingly implemented in clinical practice. This is especially true in the United States, where the 21-gene recurrence score (Oncotype DX™) is endorsed by the American Society of Clinical Oncology as a useful tool to predict prognosis and risk of recurrence in patients with node-negative, ER-positive breast cancer, and reimbursed . The fact that genomic tests can improve our understanding of breast cancer biology and that such genomic tests are currently not reimbursed in Europe probably influenced investigators to rapidly enroll patients in our study. The 180 early-stage breast cancer patients were enrolled within only 10 months in 8 community-based hospitals. The oncology-tertiary center, Jules Bordet Institute, coordinated the study without enrolling patients. This indicates that genomic tests can easily be implemented in routine practice.
In regard to treatment plans, participating physicians used the GGI results in combination with their usual treatment algorithm, i.e., without pre-specified treatment guidance. Tumor boards were asked to establish their treatment strategy based on their usual clinico-pathologic features, and then using the hypothetical and final GGI results. The final treatment decision was ultimately made after discussion with each patient. The exercise based on hypothetical GGI results performed before informing investigators of the final results demonstrated that physicians are willing to increase chemotherapy indications for those classified as GG-3 and to decrease chemotherapy indications for those classified GG-1
Upon receiving the final GG results, physicians changed their treatment recommendations mostly for patients with discordant GGI and HG results, therefore mainly for the HG2 cases, which represents the optimal target population for the Genomic Grade Index. However, the small numbers of HG2 cases in our study, which primary objective was to assess feasibility of the GGI in the routine of a breast cancer care centre, limits our ability to draw definitive conclusions. The difference between the hypothetical scenarios and the final treatment decisions is likely to reflect patient preferences. Similar results were observed in a prospective study evaluating the impact of Oncotype DX™ on the treatment selections of medical oncologists and patients . Out of a total of 89 early stage breast cancer cases, medical oncologists changed their pre- and post-test treatment recommendations from chemotherapy plus hormonotherapy to hormonotherapy in 22.5% of cases, but when patient preferences were taken into consideration, the same changes occurred in only 10.1% of the cases.
In conclusion, the present prospective study demonstrates that determining GGI is achievable in community hospitals, which reinforces the feasibility of incorporating genomic technologies into routine clinical practice. In addition, integrating GGI into clinico-pathologic risk assessment may guide physicians and patients towards better individualization of treatment decisions for early stage breast cancer. However, the clinical usefulness of genomic tools is still uncertain and is under evaluation in prospective randomized clinical trials.
Study Protocol - MapQuant Dx Genomic Grade: feasibility in routine practice and impact of tumor grade quantification on treatment decision-making in early breast cancer patients.
Conceived and designed the experiments: OM MB MP CS. Performed the experiments: OM MB SM VF. Analyzed the data: OM SM MB MI MP CS. Contributed reagents/materials/analysis tools: OM AC SM MB MI KMS EA VF SN JLC PD MC PC VJ JPK DL VR DF NC PV BV HPSP MP CS. Wrote the manuscript: OM SM CS.
- 1. Sotiriou C, Pusztai L (2009) Gene-expression signatures in breast cancer. N Engl J Med 360: 790-800. doi:10.1056/NEJMra0801289. PubMed: 19228622.
- 2. van de Vijver MJ, He YD, van’t Veer LJ, Dai H, Hart AA et al. (2002) A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347: 1999-2009. doi:10.1056/NEJMoa021967. PubMed: 12490681.
- 3. Sotiriou C, Wirapati P, Loi S, Harris A, Fox S et al. (2006) Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis. J Natl Cancer Inst 98: 262-272. doi:10.1093/jnci/djj052. PubMed: 16478745.
- 4. Paik S, Shak S, Tang G, Kim C, Baker J et al. (2004) A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med 351: 2817-2826. doi:10.1056/NEJMoa041588. PubMed: 15591335.
- 5. Albain KS, Barlow WE, Shak S, Hortobagyi GN, Livingston RB et al. (2010) Prognostic and predictive value of the 21-gene recurrence score assay in postmenopausal women with node-positive, oestrogen-receptor-positive breast cancer on chemotherapy: a retrospective analysis of a randomised trial. Lancet Oncol 11: 55-65. doi:10.1016/S1470-2045(09)70314-6. PubMed: 20005174.
- 6. Paik S, Tang G, Shak S, Kim C, Baker J et al. (2006) Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor-positive breast cancer. J Clin Oncol 24: 3726-3734. doi:10.1200/JCO.2005.04.7985. PubMed: 16720680.
- 7. Cardoso F, Van’t Veer L, Rutgers E, Loi S, Mook S et al. (2008) Clinical application of the 70-gene profile: the MINDACT trial. J Clin Oncol 26: 729-735. doi:10.1200/JCO.2007.14.3222. PubMed: 18258980.
- 8. Sparano JA, Paik S (2008) Development of the 21-gene assay and its application in clinical practice and clinical trials. J Clin Oncol 26: 721-728. doi:10.1200/JCO.2007.15.1068. PubMed: 18258979.
- 9. Hayes DF (2012) Targeting Adjuvant Chemotherapy: A Good Idea That Needs to Be Proven! J Clin Oncol, 30: 1264–7. PubMed: 22355050.
- 10. Filho OM, Ignatiadis M, Sotiriou C (2010) Genomic Grade Index: An important tool for assessing breast cancer tumor grade and prognosis. Crit Rev Oncol/Hematol 77: 20-29.
- 11. Goldhirsch A, Ingle JN, Gelber RD, Coates AS, Thürlimann B et al. (2009) Thresholds for therapies: highlights of the St Gallen International Expert Consensus on the primary therapy of early breast cancer 2009. Ann Oncol 20: 1319-1329. doi:10.1093/annonc/mdp322. PubMed: 19535820.
- 12. Loi S, Haibe-Kains B, Desmedt C, Lallemand F, Tutt AM et al. (2007) Definition of clinically distinct molecular subtypes in estrogen receptor-positive breast carcinomas through genomic grade. J Clin Oncol 25: 1239-1246. doi:10.1200/JCO.2006.07.1522. PubMed: 17401012.
- 13. Metzger-Filho O, Michiels S, Bertucci F, Catteau A, Salgado R et al. (2012) Genomic grade adds prognostic value in invasive lobular carcinoma. Ann Oncol, 24: 377–84. PubMed: 23028037.
- 14. Liedtke C, Hatzis C, Symmans WF, Desmedt C, Haibe-Kains B et al. (2009) Genomic grade index is associated with response to chemotherapy in patients with breast cancer. J Clin Oncol 27: 3185-3191. doi:10.1200/JCO.2008.18.5934. PubMed: 19364972.
- 15. Ignatiadis M, Singhal SK, Desmedt C, Haibe-Kains B, Criscitiello C et al. (2012) Gene modules and response to neoadjuvant chemotherapy in breast cancer subtypes: a pooled analysis. J Clin Oncol 30: 1996-2004. doi:10.1200/JCO.2011.39.5624. PubMed: 22508827.
- 16. Reyal F, Bollet MA, Caly M, Gentien D, Carpentier S et al. (2012) Respective prognostic value of genomic grade and histological proliferation markers in early stage (pN0) breast carcinoma. PLOS ONE 7: e35184. doi:10.1371/journal.pone.0035184. PubMed: 22529987.
- 17. Mook S, Bonnefoi H, Pruneri G, Larsimont D, Jaskiewicz J et al. (2009) Daily clinical practice of fresh tumour tissue freezing and gene expression profiling; logistics pilot study preceding the MINDACT trial. Eur J Cancer 45: 1201-1208. doi:10.1016/j.ejca.2009.01.004. PubMed: 19232484.
- 18. Bueno-de-Mesquita JM, van Harten WH, Retel VP, van’t Veer LJ, van Dam FS et al. (2007) Use of 70-gene signature to predict prognosis of patients with node-negative breast cancer: a prospective community-based feasibility study (RASTER). Lancet Oncol 8: 1079-1087. doi:10.1016/S1470-2045(07)70346-7. PubMed: 18042430.
- 19. Harris L, Fritsche H, Mennel R, Norton L, Ravdin P et al. (2007) American Society of Clinical Oncology 2007 update of recommendations for the use of tumor markers in breast cancer. J Clin Oncol 25: 5287-5312.
- 20. Lo SS, Mumby PB, Norton J, Rychlik K, Smerage J et al. (2010) Prospective multicenter study of the impact of the 21-gene recurrence score assay on medical oncologist and patient adjuvant breast cancer treatment selection. J Clin Oncol 28: 1671-1676. doi:10.1200/JCO.2008.20.2119. PubMed: 20065191.