www.nature.com/bjc CONSENSUS STATEMENT Recommendations from the European Commission Initiative on Breast Cancer for multigene testing to guide the use of adjuvant chemotherapy in patients with early breast cancer, hormone receptor positive, HER-2 negative Paolo Giorgi Rossi1, Annette Lebeau2, Carlos Canelo-Aybar3,4, Zuleika Saz-Parkinson 5,6, Cecily Quinn7, Miranda Langendam8, Helen Mcgarrigle9, Sue Warman10, David Rigau3, Pablo Alonso-Coello3, Mireille Broeders11,12, Axel Graewingholt13, Margarita Posso3,14,15, Stephen Duffy 16, Holger J. Schünemann17 and the ECIBC Contributor Group BACKGROUND: Predicting the risk of recurrence and response to chemotherapy in women with early breast cancer is crucial to optimise adjuvant treatment. Despite the common practice of using multigene tests to predict recurrence, existing recommendations are inconsistent. Our aim was to formulate healthcare recommendations for the question “Should multigene tests be used in women who have early invasive breast cancer, hormone receptor-positive, HER2-negative, to guide the use of adjuvant chemotherapy?” METHODS: The European Commission Initiative on Breast Cancer (ECIBC) Guidelines Development Group (GDG), a multidisciplinary guideline panel including experts and three patients, developed recommendations informed by systematic reviews of the evidence. Grading of Recommendations Assessment, Development and Evaluation (GRADE) Evidence to Decision frameworks were used. Four multigene tests were evaluated: the 21-gene recurrence score (21-RS), the 70-gene signature (70-GS), the PAM50 risk of recurrence score (PAM50-RORS), and the 12-gene molecular score (12-MS). RESULTS: Five studies (2 marker-based design RCTs, two treatment interaction design RCTs and 1 pooled individual data analysis from observational studies) were included; no eligible studies on PAM50-RORS or 12-MS were identified and the GDG did not formulate recommendations for these tests. CONCLUSIONS: The ECIBC GDG suggests the use of the 21-RS for lymph node-negative women (conditional recommendation, very low certainty of evidence), recognising that benefits are probably larger in women at high risk of recurrence based on clinical characteristics. The ECIBC GDG suggests the use of the 70-GS for women at high clinical risk (conditional recommendation, low certainty of evidence), and recommends not using 70-GS in women at low clinical risk (strong recommendation, low certainty of evidence). British Journal of Cancer (2021) 124:1503–1512; https://doi.org/10.1038/s41416-020-01247-z BACKGROUND (PR)-positive), human epidermal growth factor receptor 2 Breast cancer is the most frequently diagnosed cancer among (HER2)-negative breast cancer represents about 70% of breast women.1 In the European Union, including UK, 404,920 women cancer diagnosed in western countries.3 At the time of diagnosis, were diagnosed with breast cancer and 98,755 died because around 60% of this type of cancer has not spread to lymph of this disease in 2018.2 Hormone receptor (HoR)-positive nodes,3 and approximately 15% of these women will develop a (i.e. oestrogen receptor (ER)- and/or progesterone receptor recurrence within 10 years if treated with adjuvant endocrine 1 Azienda Unità Sanitaria Locale—IRCCS di Reggio Emilia, Reggio Emilia, Italy; 2Department of Pathology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; 3 Iberoamerican Cochrane Center, Biomedical Research Institute (IIB Sant Pau-CIBERESP), Barcelona, Spain; 4Department of Paediatrics, Obstetrics and Gynaecology, Preventive Medicine, and Public Health, PhD Programme in Methodology of Biomedical Research and Public Health, Universitat Autònoma de Barcelona, Bellaterra, Spain; 5European Commission, Joint Research Centre (JRC), Ispra, Italy; 6Instituto de Salud Carlos III, Health Technology Assessment Agency, Avenida Monforte de Lemos 5, Madrid, Spain; 7St. Vincent’s University Hospital, Dublin, Ireland; 8Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Amsterdam UMC, University of Amsterdam, Amsterdam Public Health Institute, Amsterdam, The Netherlands; 9Cardiff and Vale UHB - General Surgery, Cardiff, UK; 10Havyatt Lodge, Havyatt Road, Langford, North Somerset, UK; 11 Department for Health Evidence, Radboud University Medical Center, Nijmegen, the Netherlands; 12Dutch Expert Centre for Screening, Nijmegen, the Netherlands; 13Radiologie am Theater, Paderborn, NRW, Germany; 14Department of Epidemiology and Evaluation, IMIM (Hospital del Mar Medical Research Institute), Barcelona, Spain; 15Research Network on Health Services in Chronic Diseases (REDISSEC), Barcelona, Spain; 16Centre for Cancer Prevention, Queen Mary University of London, Charterhouse Square, London, UK and 17 Michael G. DeGroote Cochrane Canada and McGRADE Centres; Department of Health Research Methods, Evidence and Impact, McMaster University Health Sciences Centre, Hamilton, Ontario, Canada Correspondence: Zuleika Saz-Parkinson (zuleika.saz-parkinson@ec.europa.eu) Members of the ECIBC Contributor Group are listed above Acknowledgements. These authors contributed equally: Paolo Giorgi Rossi, Annette Lebeau Received: 1 October 2020 Revised: 10 December 2020 Accepted: 17 December 2020 Published online: 18 February 2021 © The Author(s) 2021 Published by Springer Nature on behalf of Cancer Research UK Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1504 Table 1. Structured clinical questions. Population Intervention Comparison Outcomes Patients with hormone receptor-positive, 70-gene signature to decide chemotherapy: • According to the considered clinical trial • Overall survival HER2- negative, lymph node-negative or women with low clinical risk do not receive designs the comparison would be either: • Disease-free up to 3 lymph nodes-positive invasive chemotherapy independently from 70-gene • All patients receive chemotherapy survival breast cancer. signature; women with high clinical risk will • All patients with low clinical risk do not • Adverse effects receive chemotherapy only if at high receive chemotherapy and those with high of drugs genomic risk clinical risk subgroup receive • Quality of life. 21-gene recurrence score to decide chemotherapy. chemotherapy, in two alternative scenarios: Direct comparisons between the different • Women with low or intermediate genomic risk multigene tests were not performed. will not receive chemotherapy, women with high genomic risk will; • Only women with high clinical risk will be tested for genomic risk and receive chemotherapy if genomic risk is high; women with low clinical risk will not receive chemotherapy. Initially PICOs covered four multigene tests, but two of them, the PAM50 risk of recurrence score (RORS) and 12-gene molecular score (12-MS), were excluded because of study selection criteria. These selection criteria allowed no direct comparison between the different multigene tests. therapy alone.4,5 The risk of recurrence could be reduced by the adjuvant chemotherapy” was structured following the Population, addition of chemotherapy.5 However, given the relatively low risk of Intervention, Comparison and Outcomes (PICO) format (Table 1). recurrence and the partial effectiveness of chemotherapy in these The outcomes were also prioritised by the GDG using a nine-point women, most would be over-treated if all received chemotherapy. scale (7 to 9 critical; 4 to 6 important; 1 to 3 of limited importance), 1234567890();,: The same rationale would apply to women with HoR-positive, as suggested by the Grading of Recommendations Assessment, HER2-negative invasive breast cancer with 1–3 positive lymph Development and Evaluation (GRADE) approach.11,18 The GDG nodes.6 Several prognostic factors, including clinical-pathological decided not to attempt any head-to-head comparisons between features such as age, tumour size, percentage of ER- and PR-positive the different tests. cells as well as Ki67-index,7–9 predict the risk of recurrence and can help identify women who would benefit the most from Systematic review chemotherapy. Although these factors have been shown to Data sources and searches. MEDLINE (May 2018), EMBASE (May discriminate different prognostic groups, they showed no or 2018) and CENTRAL (May 2018) databases were searched, using minimal predictive value on the response to chemotherapy.10 pre-defined algorithms, for both systematic reviews and individual In the last 15 years, different tests have been developed to studies (Supplementary Table 2); this original search was stratify patients with early breast cancer into different risk of continuously run up to October 2018. Lists of references of the recurrence groups by analysing the activity of various genes. included studies were reviewed and members of the GDG were Although these multigene tests use diverse techniques (RT-PCR, requested to provide additional studies. microarray, and others) and diverse target gene combinations, they all focus on genes involved in cell proliferation. The tests Study selection. Randomised controlled trials (RCT) and cohort provide recurrence risk profiles categorised in different ways. studies (including pooled analyses of studies), either from Some tests have been explicitly proposed to provide additional prospective or retrospective analysis, of stored specimen samples information to clinical-pathological features, as the 12-MS and the were included as long as they applied any of the four tests as PAM50-RORS. Based on the results of the MINDACT trial,6 the predictive markers for guiding the use of adjuvant chemotherapy application of the 70-GS also takes into account clinical prognostic (Supplementary Fig. 1) characteristics, while the 21-RS has been proposed to substitute A predictive marker identifies the differential benefit of a clinical risk-based treatment decisions. treatment based on the marker status. Thus, we included the The European Commission Initiative on Breast Cancer (ECIBC) following assessment approaches: (a) Marker-based strategy design: aims to provide evidence-based recommendations for screening patients are assigned to a treatment arm depending on whether and diagnosis of breast cancer.11 The Guidelines Development they received treatment (i.e. endocrine therapy or endocrine plus Group (GDG) of the ECIBC prioritised a clinical question on the use chemotherapy) according to the test results or according to usual of multigene tests to guide the use of adjuvant chemotherapy in clinical practice. The predictive value is assessed by comparing the HoR-positive, HER2-negative and lymph node-negative or up to 3 outcomes from the testing-based arm versus the non-testing arm; lymph nodes-positive invasive breast cancer. Four multigene tests, (b) Treatment interaction design: patients are divided into groups used to stratify women with breast cancer into different groups based on the marker status (i.e. high and low marker status). Then according to recurrence risk,12–17 are included in the clinical they are allocated to receive endocrine therapy or endocrine question (Supplementary Table 1): 21-RS (Oncotype DX, Genomic treatment plus chemotherapy. The predictive value is assessed by Health Inc), 12-MS (EndoPredict, Myriad Genetics Inc), PAM50- observing the relative efficacy of treatment differences between RORS (Prosigna test, NanoString Technologies Inc.), and 70-GS marker status and treatment assignments. (MammaPrint; Agendia Inc). Direct comparison between different Studies that only reported prognosis data based on marker tests is beyond the scope of these recommendations. status (without considering differential treatment effect), indivi- dual observational studies, abstracts or conference communica- tions not published as full text articles, and articles published in a METHODS language other than English were excluded. Structured question and outcome prioritisation With respect to economic evidence, cost-utility, cost-benefit, and The clinical question “Should multigene tests be used in patients cost-consequences, analyses were included if conducted within who have HoR-positive, HER-2 negative, lymph node-negative or up clinical trials, as well as observational and modelling studies, to 3 lymph nodes-positive invasive breast cancer to guide the use of published in English during the last decade (Supplementary Table 3). Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1505 After a calibration process, each reviewer (CCA and KP) assessed low clinical and high genomic risk, as well as those with low titles and abstracts for eligibility. Subsequently, two reviewers (CCA clinical and low genomic risk, do not benefit from adjuvant and KP), independently, reviewed the full text of all the pre-selected chemotherapy. For sensitivity analyses, in women of low references. Discrepancies were solved either by consensus or with clinical and high genomic risk, two different assumptions the help of a third reviewer (DR) (Supplementary Fig. 2 and were used to estimate the benefits: (a) The advantage of Supplementary Table 4). receiving adjuvant chemotherapy is equivalent to that observed in the MINDACT trial6 at five years, and the effect Data extraction and risk of bias assessment. Two reviewers (CCA is maintained at 10 years; (b) the advantage from adjuvant and KP) independently assessed risk of bias and extracted the chemotherapy is equivalent to that observed by Paik and following information from each study: first author, year of colleagues26 for all women at high genomic risk, indepen- publication, country, study design, inclusion and exclusion criteria, dent of their clinical risk. Distributions of the clinical risk number of patients, age, participants’ characteristics and priori- within the multigene risk strata are those reported in the tised outcomes. TAILORx trial.23 The risk of bias of the included RCTs was assessed using the Cochrane Risk of Bias tool for randomised trials.19 Cohort studies For the 70-GS, we focused only on comparing strategy 2 with were assessed with the “Risk Of Bias In Non-randomised Studies - scenario 2 in which women at high clinical risk would be tested of Interventions-I” (ROBINS-I) tool.20 For economic evaluations, one and/or treated, because the evidence from the MINDACT trial reviewer (MP) screened the search results and used the NICE indicates a very small benefit, if any, from adjuvant chemotherapy methodology checklist to assess applicability and methodological in women with low clinical risk, independent of their genomic limitations.21 Studies with poor applicability and/or high risk of risk.6 bias were excluded (Supplementary Fig. 2). Evidence to decision and recommendation formulation Data analysis. Descriptive statistics were used to summarise the The process the ECIBC GDG used to formulate recommendations characteristics of the included patients across studies. The effect has been described in a dedicated article published elsewhere.11 measures for prioritised outcomes and their corresponding 95% In brief, a subgroup of GDG members including experts on the confidence intervals (CIs) were reported as presented in individual topic and an informed patient (the so-called PICO responsible studies. unit), took primary responsibility for the review and completion of the first draft of the SoF tables and the EtD frameworks, Certainty of the evidence. The certainty of evidence per outcome conducted initially by the systematic review team. The frameworks and overall certainty was rated using the GRADE approach. For were used in the meetings to help the complete GDG formulate each recommendation, the GDG received a Summary of Findings the recommendations. Subsequently they were reviewed by a (SoF) table and a first draft of an evidence to decision framework technical team from the Joint Research Centre, the PICO (EtD).22 responsible unit and the systematic review team. Finally, the recommendations and frameworks were approved by the GDG. Comparison scenarios and modelling A simple deterministic decision tree model without discounting was built by PGR with input from the rest of the GDG to estimate the RESULTS downstream consequences of testing patients with the multigene Included studies tests versus different scenarios of usual care (Supplementary We included five studies (Supplementary Fig. 1): two RCTs,6,23 two Fig. 3). For the 21-RS, the general model assumptions were: the secondary analyses of stored tissue blocks collected from former population of eligible women was divided into the three risk parent clinical trials26,27 and one pooled analysis of observational groups, as reported in the TAILORx trial at recruitment until 2008 studies12 from four previously reported validation studies, (14% low risk of recurrence, 68% intermediate, 18% high).23 Rate including unpublished data (Supplementary Table 5).28–30 of events, observed in the RCTs, were applied to the simulated usual care arms. Clinical risk of recurrence was classified as low 21 gene recurrence score and high according to the modified AdjuvantOnline!Score.24,25 Treatment interaction design studies. Paik and colleagues26 Results are based on a fixed observation time of 10 years. provided estimates for distant recurrence free survival in patients Two strategies for implementing the multigene test to guide with lymph node-negative breast cancer stratified into three levels the use of adjuvant chemotherapy were considered as interven- of the 21-RS risk groups. Adding chemotherapy to endocrine tions (Supplementary Fig. 3a): therapy, compared to endocrine therapy alone, may have a different effect on recurrence across groups, i.e. a larger effect in 1. All women would undergo multigene testing and adjuvant women with higher 21-RS, but the evidence is very uncertain: chemotherapy would be given accordingly (only to those hazard ratio (HR) of 1.31 (95% CI 0.46–3.78), 0.61 (95% CI classified in the high genomic risk group, i.e. with a 0.24–1.59) and 0.26 (95% CI 0.13–0.53) in low, intermediate and score ≥26). high risk groups, respectively (Supplementary Table 6). 2. Only women with high clinical risk would undergo multi- Albain and colleagues27 included stored tumour specimens for gene testing, and only those with high genomic risk would genomic testing of postmenopausal women with HoR-positive, receive chemotherapy. Women with low clinical risk would node-positive breast cancer. They performed an analysis adjusted not receive it. by the number of positive nodes that suggests no benefit for Two scenarios were considered as usual care comparators (“C” chemotherapy on disease free survival (DFS) in the low genomic of the PICO framework) (Supplementary Fig. 3b): risk group (HR = 1.02; 95% CI 0.54–1.93) and a potential advantage in the high genomic risk (HR = 0.59, 95% CI 0.35–1.01) (Supple- 1. All women would be referred to adjuvant chemotherapy mentary Table 6). The authors refer similar results for overall (assuming 18.4% would not comply, i.e. the proportion of survival (OS). women not receiving chemotherapy among those assigned to the treatment arm in TAILORx).23 Marker-based strategy. Sparano and colleagues23 provided 2. Women would receive adjuvant chemotherapy only if the results for several disease-free survival (DFS)-related outcomes clinical risk is high. The model assumes that women with among women with an intermediate genomic risk group (11 to 26 Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1506 risk score) allocated to either endocrine therapy alone or Anticipated outcomes for different comparisons between the 21-gene recurrence score testing strategies (interventions) and comparator scenarios (no testing) per 1000 women with chemotherapy plus endocrine therapy. The as-treated results difference Absolute suggest little to no difference in the risk of recurrence with chemotherapy plus endocrine therapy for invasive DFS (HR 1.14; −211 −1 0 0 0 95% CI 0.99–1.31). For distant metastases, local recurrence and OS, similar results were observed (Supplementary Table 6). 70-GS Treatment interaction design studies. Knauer and colleagues12 Comparator scenario 2b described results from a pooled database analysis with a median follow-up time of 7.1 years. Patients with a low and high genomic 314 79 27 39 24 risk who received chemotherapy may have a lower risk of recurrence than those with endocrine therapy alone, but the evidence is very uncertain (HR 0.26; 95% CI 0.03–2.02 and HR 0.35; 95% CI 0.17–0.71, respectively). The results for mortality were Intervention consistent with the observed pattern of the risk of recurrence but strategy 2a the evidence was also very uncertain (HR 0.58; 95% CI 0.07–4.98; HR 0.21; 95%CI 0.07–0.59, respectively) (Supplementary Table 7). 103 80 27 39 24 Marker-based strategy. Cardoso and colleagues6 reported DFS and OS among patients in the clinical/genomic discordant-risk groups which were allocated to receive either chemotherapy in difference Absolute addition to endocrine therapy or endocrine therapy alone. Women with high clinical risk and low genomic risk may have −134 −1 0 0 0 an increase of DFS (HR 0.64; 95% CI 0.43–0.95), of distant metastases free survival (HR 0.65; 95% CI 0.38–1.10) and of OS (HR 0.63; 95% CI 0.29–1.37) (Supplementary Table 7). The group of low clinical risk and high genomic risk showed imprecise effects and Comparator uncertain evidence for DFS (HR 0.74; 95% CI 0.40–1.39), distant scenario 2b metastases free survival (HR 0.90; 95% CI 0.40–2.01) and OS (HR 0.72; 95% CI 0.23–2.24).6 314 79 27 39 24 Modelling for predicting impact of testing on patient’s outcomes Depending on the different treatment scenarios (all women are referred to chemotherapy or only women with high clinical risk are Intervention treated with chemotherapy) and genetic testing strategies (genetic strategy 1a testing carried out in all women or testing only those with high hormone receptor-positive, HER2-negative, lymph node-negative invasive breast cancer. clinical risk), the number of women who avoid chemotherapy by 180 80 28 39 24 using the 21-RS would change from more than 600 to about 200. Survival outcomes did not change substantially (Table 2) for the 21- RS based on the benefits from adding chemotherapy in the MINDACT trial.6 However, on the assumption that all women with difference Absolute high genomic score would obtain the same benefits from adding chemotherapy as observed by Paik and colleagues,26 independently −636 2 1 1 1 from their clinical risk, the intervention could potentially prevent 37 distant metastases compared to a scenario in which only women with high clinical risk would be treated with chemotherapy. On the other hand, we considered for the 70-GS a two-step Comparator strategy according to the results of the MINDACT trial, testing only scenario 1b women with high clinical risk.6 Consequently, the only scenario considered was one in which only high-risk women would receive 820 78 26 38 23 chemotherapy (Table 3). The use of the 70-GS would result in an avoidance of chemotherapy in about 230 women out of 1000 associated with small increase of recurrences. According to Supplementary Fig. 3b. Intervention According to Supplementary Fig. 3a. Results from the systematic review of economic evidence strategy 1a From the primary literature search and from the two identified systematic reviews,31,32 12 cost-effectiveness evaluations were identified (Supplementary Fig. 2 and Supplementary Table 4).33–44 180 80 27 39 24 The GDG agreed that these economic evaluations used models that were not directly applicable to the clinical question of Distant recurrence interest. Therefore, cost-effectiveness was evaluated considering Local and distant Invasive disease the benefits and harms estimated using the GDG’s ad-hoc model Chemotherapy described above (Supplementary Fig. 3) and the costs reported by Number of recurrence recurrence the studies included in the literature review. Eight studies reported Outcome Table 2. Deaths costs for the use of the 21-RS in women with negative lymph nodes,33–41 whereas three reported costs for women with up to 3 positive lymph nodes.42–44 The reported costs of the 21-RS were b a Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1507 EUR 3180 per patient in five out of the 11 studies included. These was included. A negative impact on equity was considered a costs did not show significant differences between countries or potential concern (Table 4). over time. For the 70-GS, two studies reported costs of EUR 3153 For women with HoR-positive, HER2-negative, node-negative per patient for the use of this assay in women with negative invasive breast cancer, the ECIBC GDG suggests the use of the 21- lymph nodes.40,41 RS to guide the use of chemotherapy (conditional recommenda- tion, very low certainty of the evidence, Table 5). The recommen- Certainty of evidence dation is conditional because the certainty of evidence was very The overall certainty of the evidence was rated as low to very low. low and the downstream consequences of avoiding chemother- The main concerns across studies were risk of bias, indirectness of apy were not quantified, thus making the balance of benefits and trial populations and imprecision (Supplementary Tables 6 and 7). harms difficult to determine, together with the large resource The evidence was also downgraded for indirectness due to the (costs) requirements (Table 4, see also https://healthcare-quality. assumptions used for the model that implied the use of evidence jrc.ec.europa.eu/sites/default/files/Guidelines/EtDs/Updated/ from one population in another population and from different ECIBC_GLs_EtD_21_gene_recurrence_score.pdf). duration of follow-up across studies. The GDG did not consider women with node-positive invasive breast cancer in this recommendation, because they were not Evidence to decision frameworks included in the TAILORx trial,23 the main source for model 21-RS. The GDG judged the anticipated desirable effects (i.e. the parameters. The GDG also stated that sub-populations with high avoided chemotherapy treatments) of using the test to guide clinical risk (defined according to AdjuvantOnline!)24,25 may chemotherapy to be large and the undesirable effects (i.e. increase experience larger net desirable consequences and provide a more in recurrence) trivial, with very low certainty of the evidence. The favourable cost-effectiveness profile (Fig. 1). On the other hand, costs were considered large, though no cost-effectiveness study women with low clinical risk may experience smaller or no net desirable consequences. Indirect evidence from the MINDACT trial using the 70-GS supports that conclusion. In fact, in this trial there Table 3. Anticipated outcomes for the comparison between the are very small, if any, benefits from chemotherapy in low clinical 70-gene signature assay testing strategy (intervention) and risk women, independently of the genomic risk.6 comparator scenario (no testing) per 1000 women with hormone New relevant results have been published on the 21-RS since receptor-positive, HER2-negative, lymph node-negative or up to the systematic review used for this recommendation was 3 lymph nodes-positive invasive breast cancer. conducted.45,46 Recent data from the TAILORx trial stratified by age and clinical risk have been published.45 The authors suggest Outcome Intervention Comparator Absolute that women below 50 with an intermediate genomic risk score strategy 2a scenario 2b difference could have a benefit from adding chemotherapy to endocrine Treated women 270 501.4 −232 therapy if their clinical risk is high but the absence of a dose Distant disease 53 48 4.5 response and very imprecise effect estimates suggest that chance could play a major role. Furthermore, the reported analysis Disease free 100 93 7 suggests that data for evaluating the potentially most efficient Deaths 30 26 3.5 two-step testing strategy is missing.47,48 Mariotto and colleagues49 a According to supplementary Fig. 3a. showed that application of the 21-RS risk to decide whether or not b According to supplementary Fig. 3b. to provide chemotherapy would produce savings in the actual US real clinical practice. Another analysis using the same data Table 4. Judgements by the Guideline Development Group (GDG) in Evidence to decision framework for the question: Should multigene tests be used in patients who have hormone receptor-positive, HER-2 negative, lymph node-negative or up to 3 lymph nodes-positive invasive breast cancer to guide the use of adjuvant chemotherapy? 21-gene recurrence score (limited to women 70-gene signature assay with negative lymph nodes) Low clinical risk High clinical risk Problem Yes Yes Yes Desirable effects Large Trivial Large Undesirable effects Trivial Trivial Small Certainty of evidence Very low Low Low Values Probably no important uncertainty or Probably no important Probably no important variability uncertainty or variability uncertainty or variability Balance of effects Probably favours the intervention Favours the comparison Probably favours the intervention Resources required Large costs Large costs Large savings Certainty of evidence of required Very low Very low Very low resources Cost effectiveness No included studies No included studies No included studies Equity Probably reduced Probably reduced Probably reduced Acceptability Varies Varies Varies Feasibility Varies Varies Varies Final recommendation Conditional in favour of the intervention Strong against intervention Conditional in favour of the intervention Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1508 Table 5. European Commission Initiative on Breast Cancer guidelines development group recommendations on the use of multigene tests to guide the use of adjuvant chemotherapy in patients with early breast cancer, hormone receptor-positive, HER-2 negative. 21-gene recurrence score 70-gene signature Women with low clinical riska Women with high clinical riska Recommendation For women with hormone receptor- For women with hormone receptor- For women with hormone receptor- positive, HER2-negative, lymph node- positive, HER2-negative, lymph node- positive, HER2-negative, lymph node- negative invasive breast cancer, the negative or up to 3 lymph nodes- negative or up to 3 lymph nodes- ECIBC’s Guidelines Development Group positive invasive breast cancer at low positive invasive breast cancer at high (GDG) suggests using the 21-gene clinical risk, the ECIBC’s Guidelines clinical risk, the ECIBC’s Guidelines recurrence score to guide the use of Development Group (GDG) recommends Development Group (GDG) suggests chemotherapy. not using the 70-gene signature test to using the 70-gene signature test to guide the use of chemotherapy. guide the use of chemotherapy. Strength Conditional recommendation for the Strong recommendation against the Conditional recommendation for the intervention intervention intervention Very low certainty of the evidence Low certainty of the evidence Low certainty of the evidence Sub-group The GDG did not consider women with The proportion of women with 2 or 3 The proportion of women with 2 or 3 considerations node-positive invasive breast cancer to node-positive breast cancer was small, node-positive breast cancer was small, be included in this recommendation. so the results may be less clear in this so the results may be less clear in this Women with high clinical riska and low subgroup. subgroup. genomic risk (larger tumour diameter and higher grade) may experience larger net desirable consequences and provide a better cost-benefit profile. Women with low clinical riska and high genomic risk may experience smaller or no net desirable consequences. Indirect evidence from other gene based testing (e.g. 70-gene signature) supports that conclusion. a For definitions of low and high clinical risk, see Supplementary Table 8. indicates that savings would be much larger if testing would be high clinical risk and suggests the use of the 70-GS only for performed according to a two-step strategy.50 The GDG, for the women at high clinical risk. moment, judged that the new evidence is consistent with the recommendation. Strengths and weaknesses of the study The strength of the recommendations includes the ECIBC’s 70-GS. In light of the results from the MINDACT trial,6 the GDG adherence to the requirements for trustworthy development of decided to split the recommendation according to clinical risk of guidelines.51–53 Previously, we described some limitations of our the population under study (at low and high clinical risk, Table 5 guidelines.11,54 The weakness of the deterministic decision tree and Fig. 1). In the low clinical risk group the GDG recommends model used is that it is, to some extent, a simplistic approach and against using the 70-GS testing to guide the use of chemotherapy some assumptions are questionable (i.e. negligible effects in low (strong recommendation, very low certainty of the evidence) as clinical risk, same effects in studies with different duration of there are no apparent benefits and there are very large costs (EUR follow-up). Furthermore, we did not actually quantify the side 3153 per patient) (https://healthcare-quality.jrc.ec.europa.eu/sites/ effects of chemotherapy, considering that avoiding any unneces- default/files/Guidelines/EtDs/Updated/ECIBC_GLs_EtD_70_gene_ sary chemotherapy was a desirable effect. testing_low_risk.pdf). For women with HoR-positive, HER2-negative, node-negative or Relation to other guidelines up to 3 lymph nodes-positive invasive breast cancer at high NICE recently published guidance on multigene testing.55 The clinical risk, the ECIBC GDG suggests using the 70-GS test to guide panel decided to evaluate the evidence for the four commercially the use of chemotherapy. The judgments favoured the interven- available multigene tests included in the ECBIC question, and for tion in the high clinical risk population due to the moderate the IHC4 + C test.56,57 The NICE guidelines did not follow the desirable effect, a balance that probably favours the use of 70-GS GRADE methodology and also had a different goal, i.e. deciding testing, and the large savings (Table 4). The recommendation is which tests should be funded by the UK National Health Service. conditional mainly because of the low certainty of the evidence The NICE 21-RS recommendation is similar to the ECIBC about the effects. The GDG also stated that the proportion of recommendation, a conditional recommendation limited to those women with 2 or 3 positive lymph nodes was small, patients in which the risk of distant recurrence is intermediate, therefore making the results less clear in this subgroup (https:// using a validated tool such as PREDICT58 or the Nottingham healthcare-quality.jrc.ec.europa.eu/sites/default/files/Guidelines/ Prognostic Index (Table 6). This approach is based on the EtDs/Updated/ECIBC_GLs_EtD_70_gene_testing_high_risk.pdf). assumption that in some patients the clinical and pathological prognostic parameters are consistent (low or high risk), and that multigene testing does not provide additional information. In DISCUSSION particular, the NICE panel recommended against the use of the 70- Statement of principal findings GS, based on cost-effectiveness considerations (Table 6). Unfortu- The ECIBC GDG suggests the use of 21-RS in lymph node-negative nately, the NICE guideline does not allow comparison of our women, recognising that benefits are probably larger in women at estimates of desirable and undesirable health effects. Some Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1509 21-GENE RECURRENCE SCORE 70-GENE SIGNATURE Testing all strategy Two steps strategy: Clinical risk Two steps strategy: Clinical risk assessment followed by multigene assessment followed by testing only if clinical risk is high multigene testing only if clinical risk is high Women HoR+, HER2–, T1 or T2, lymph node- Women HoR+, HER2–, T1 or T2, lymph node- Women HoR+, HER2–, T1 or negative negative T2, node-negative or up to 3 nodes-positive GDG suggests to perform genomic risk assessment Clinical risk assessment Clinical risk assessment Low genomic Intermediate High risk genomic risk genomic risk Low clinical High clinical risk risk Low clinical High clinical risk risk Endocrine Endocrine Endocrine therapy therapy therapy plus Endocrine GDG suggests to Endocrine chemotherapy GDG suggests therapy perform genomic treatment to perform risk assessment genomic risk assessment Low genomic Intermediate High genomic risk genomi crisk risk Low genomic High genomic risk risk Endocrine Endocrine Endocrine therapy therapy plus therapy chemotherapy Endocrine Endocrine therapy therapy plus chemotherapy Fig. 1 Flow charts reporting the possible uses of 21-gene recurrence score and for the 70-gene signature test to guide the use of adjuvant chemotherapy in patients with early invasive breast cancer, hormone receptor-positive, HER-2 negative. Two strategies are proposed for 21-gene recurrence score, the first in which all women are tested for genomic risk assessment and treated accordingly, the second in which only women with high clinical risk are tested for genomic assessment, while those at low clinical risk are referred to endocrine therapy alone without genomic risk assessment. According to sub-group considerations reported by the GDG, the latter strategy is probably more cost effective and women might experience larger net desirable consequences. For the 70-gene signature only a two-step strategy is proposed where only women at high clinical risk are tested for genomic risk; testing women at low clinical risk is not recommended. Table 6. Synopsis of American Society of Clinical Oncology (ASCO)59–61 and the UK National Institute for Health and Care Excellence (NICE)55 recommendations on 21-gene recurrence score and 70-gene signature assay in hormone receptor-positive, HER2-negative, lymph node-negative or up to 3 nodes-positive invasive breast cancer. ASCO58,60 NICE54 21-gene recurrence score Low clinical risk Strong recommendation High clinical risk Strong recommendation Recommended if: clinical risk is “intermediate” according to the PREDICT tool53 or the Nottingham Prognostic Index, i.e. the additional benefit of chemotherapy is between 3 and 5% increase in survival, the decision on the therapy will depend on the test result, the test is provided at reduced price. Negative lymph nodes Strong recommendation Recommendation applies to both lymph node-negative patients 1 to 3 positive lymph nodes Not recommended and lymph node-positive patients, restricted to micro-metastases ASCO60 NICE55 70-gene signature assay Low clinical risk Strong against Recommendation against because not cost effective High clinical risk Strong in favour Negative lymph nodes Strong in favour 1 to 3 positive lymph nodes Moderate in favour Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1510 methodological differences might explain the divergent results: the presented evidence evaluation, this would enable evidence- since real practice varies across Europe, we preferred to use based and transparent recommendations based on prognostic theoretical scenarios as comparators and interventions not cohort studies, randomised or not, that predict the recurrence risk accounting for non-compliance, while the NICE model used real of different subgroups. We are currently working on a healthcare practice as comparator (the prevalence of chemotherapy used in question on the significance of Ki67 using this strategy. In a next this group of women in the UK), and as intervention (a change in step, such an approach might also make it possible to evaluate the probability of receiving chemotherapy given the test result). multigene tests in the assessment for which we found no usable Furthermore, the model used by NICE assumed that the tests are evidence in the predictive search strategy used here, but for which only prognostically relevant but are not predictive of the response data on the prognostic value are available.13–17,64 to treatment, i.e. they calculated a constant HR of 0.77 for chemotherapy in addition to endocrine therapy compared to endocrine therapy alone in all risk groups. In contrast, the ECIBC ECIBC CONTRIBUTOR GROUP GDG judged the benefits to be trivial or small in the low clinical Mariangela Autelitano18, Bettina Borisch19, Xavier Castells20, Edoardo Colzani21, Jan risk group. Finally, for the 21-RS a commercial-in-confidence Daneš22, Patricia Fitzpatrick23, Livia Giordano24, Solveig Hofvind25, Lydia Ioannidou- discounted test cost was used to model cost-effectiveness, while Mouzaka26, Susan Knox27, Lennarth Nystrom28, Elena Parmelli29, Elsa Perez30, Alberto Torresin31, Ruben Van Engen32, Cary Van Landsveld-Verhoeven32, Ken Young33 for the 70-GS the regular market price was used. It is worth noting that despite NICE stating that the major benefit of the genetic 18 Cancer Registry of Milan, Milan, Italy; 19Institute of Global Health. University of testing strategies would be a reduction of chemotherapy, the cost Geneva, Geneva, Switzerland; 20IMIM, Barcelona, Spain; 21European Centre for Disease models predict health benefits only if chemotherapy is increased. Control and prevention (ECDC), Solna, Sweden; 22Charles University in Prague, The American Society of Clinical Oncology (ASCO) also provided Prague, Czech Republic; 23National Screening Service, Dublin, Ireland; 24CPO- recommendations on the use of multigene tests.59–61 Despite a Piedmont - AOU Citta` della Salute e della Scienza, Torino, Italy; 25Cancer Registry different methodological approach, the direction of the recom- of Norway, Oslo, Norway; 26University of Athens Medical School, Athens, Greece; 27 mendations is the same, but the strength is not (Table 6). There Europa Donna, Milan, Italy; 28Umea University, Umea, Sweden; 29European are differences in the grading of certainty of the evidence, Commission, Joint Research Centre, Ispra, Italy; 30University Hospital Dr. Josep Trueta, Girona, Spain; 31Ospedale Niguarda Ca’ Granda, Milan, Italy; 32Dutch Reference Centre considered as high by the ASCO panellists, while the GDG valued for Screening, Nijmegen, The Netherlands and 33National Coordinating Centre for the the evidence as very low for the 21-RS and low for the 70-GS. Physics of Mammography, Guildford, UK Unfortunately, we were unable to deduce the details of the ASCO evidence rating approach and also the criteria and judgments that were used to determine the strength of the recommendations. ACKNOWLEDGEMENTS Unlike us, ASCO and NICE made recommendations on the 12-MS To Kevin Pacheco (K.P.) for his support in the systematic review of effectiveness of the and the PAM50-RORS to guide treatment decisions,55,62 mainly multigene tests described in this publication. because they did not exclude studies based on prognostic results only. AUTHOR CONTRIBUTIONS Meaning of the study All authors contributed to the definition of the research protocol. C.C.A., D.R., P.A.C. The implications of our recommendations are context dependent. and M.P. were responsible for conducting the systematic review. P.G.R. and A.L. The criteria used for making decisions on the provision or not of prepared the first draft of the article. All authors contributed to the interpretation and adjuvant treatment differ between countries. Therefore, the cost- reporting of the results. All authors revised the manuscript and provided comments benefit profile of introducing one of the multigene tests might on subsequent versions of the article. All authors read and approved the final manuscript prior to submission and are accountable for all aspects of the work. also vary across countries. Decreasing costs for the tests would support a more widespread use. For these considerations, the GDG decided not to establish a threshold of recurrence risk to ADDITIONAL INFORMATION recommend genomic risk assessment, or a threshold for adding Ethics approval and consent to participate Not applicable. adjuvant chemotherapy, since these thresholds are context specific. Consent to publish Not applicable. In conclusion, the ECIBC GDG recommendations for or against the use of 21-RS and 70-RS are justified based on the judgments made. The transparency of our approach allows understanding Data availability All data sources used during this study are described in this published article and its additional information files. The datasets analysed are the rationale for making different recommendations for the two available from the corresponding author on reasonable request. tests and risk groups. Competing interests Members of the Guideline Development Group (GDG) do not Unanswered questions and future research receive financial compensation for their work but are reimbursed by the E.C. for Data protection issues may be of relevance for the 21-RS because travel-related expenses for the meetings organised by the JRC. Dr. Giorgi Rossi as processing of samples is centralised in one US lab and requires former-PI of an independent study on HPV-based cervical cancer screening, funded shipping samples abroad. Furthermore, transparent information by the Italian Ministry of Health, data owner, conducted negotiations with Roche on test results is not available and reproducibility has been diagnostics, Hologic-Genprobe, Becton-Dickinson to obtain reagents at reduced price questioned.63 or for free; the reagents obtained were not used in his institution. Dr. Lebeau reports The GDG recommends research on exploring in what sub- grants and reimbursement for travel-related expenses related to consultancy from groups the use of 21-RS would have larger anticipated benefits as Roche Pharma AG, reimbursement for travel-related expenses related to consultancy well as carrying out longer follow-up studies for 70-GS. The from Novartis Oncology, and grants from BioNTech Diagnostics GmbH outside the submitted work. Dr. Saz-Parkinson was employed by the European Commission, recommendations will be updated according to the ECIBC mon- coordinating the ECIBC Guidelines Development Group. Dr. Quinn is Chair of the itoring strategy in place (https://healthcare-quality.jrc.ec.europa. European Working Group for Breast Screening Pathology (EWGBSP). Various eu/discover-ecibc/methodologies/guidelines-updating). companies have provided some sponsorship to the EWGBSP for group meetings. Furthermore, the GDG is exploring the possibility of evaluating Dr. Gräwingholt is the responsible radiologist for screening unit Paderborn, Germany, biomarkers that may assist decision making regarding the consultant radiologist for screening programs in Switzerland, and consultant administration of adjuvant chemotherapy on the basis of their radiologist for Hellenic School of Senology. Dr. Canelo-Aybar, Dr. Rigau, Dr. Posso ability to identify women with a sufficiently low risk of relapse that Rivera, and Dr. Alonso-Coello reports that his institution received payments from the would allow them to be spared from chemotherapy. In contrast to European Commission to develop the systematic reviews informing the Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1511 recommendations. Authors not named here have disclosed no conflicts of interest. combined analysis of patients from the Austrian breast and colorectal cancer Disclosures can also be viewed at www.acponline.org/authors/icmje/ study group 8 and arimidex, tamoxifen alone or in combination randomized trials ConflictOfInterestForms.do?msNum=M18-3445. using the PAM50 risk of recurrence score. J. Clin. Oncol. 33, 916–922 (2015). 16. Buus, R., Sestak, I., Kronenwett, R., Denkert, C., Dubsky, P., Krappmann, K. et al. Funding information This work was supported by the European Commission. The Comparison of EndoPredict and EPclin with oncotype DX recurrence score for E.C. did not have any role in the study design, collection, analysis and interpretation prediction of risk of distant recurrence after endocrine therapy. J. Natl Cancer Inst. of the data. The researchers were independent of the funders and all authors, 108, djw149 (2016). external and internal, had full access to all of the data (including statistical reports 17. Sestak, I., Buus, R., Cuzick, J., Dubsky, P., Kronenwett, R., Denkert, C. et al. Com- and tables) in the study and take responsibility for the integrity of the data and the parison of the performance of 6 prognostic signatures for estrogen receptor- accuracy of the data analysis.Dissemination declarationResults will be disseminated positive breast cancer: a secondary analysis of a randomized clinical trial. JAMA through the ECIBC webpage (https://healthcare-quality.jrc.ec.europa.eu/european- Oncol. 4, 545–553 (2018). breast-cancer-guidelines) and specific women versions of these recommendations 18. Guyatt, G. H., Oxman, A. D., Kunz, R., Atkins, D., Brozek, J., Vist, G. et al. GRADE will also be made available to facilitate access to lay people. guidelines: 2. Framing the question and deciding on important outcomes. J. Clin. Epidemiol. 64, 395–400 (2011). Supplementary information The online version contains supplementary material 19. Higgins, J. P., Altman, D. G., Gøtzsche, P. C., Jüni, P., Moher, D., Oxman, A. D. et al. available at https://doi.org/10.1038/s41416-020-01247-z. The cochrane collaboration’s tool for assessing risk of bias inrandomised trials. BMJ 343, d5928 (2011). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims 20. Sterne, J. A., Hernán, M. A., Reeves, B. C., Savovic, J., Berkman, N. D., Viswanathan, in published maps and institutional affiliations. M. et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ 355, i4919 (2016). 21. NICE (National Institute for Health and Care Excellence). The guidelines manual: appendix G. NICE methodology checklist for economic evaluations. https://www.nice. org.uk/process/pmg6/resources/the-guidelines-manual-appendices-bi-2549703709/ REFERENCES chapter/appendix-g-methodology-checklist-economic-evaluations (2012). 1. Ferlay, J. E. M., Lam, F., Colombet, M., Mery, L., Piñeros, M., Znaor, A. et al. Global 22. Alonso-Coello, P., Oxman, A. D., Moberg, J., Brignardello-Petersen, R., Akl, E. A., Cancer Observatory: Cancer Today (International Agency for Research on Cancer, Davoli, M. et al. GRADE evidence to decision (EtD) frameworks: a systematic and Lyon, France, 2018). transparent approach to making well informed healthcare choices. 2: clinical 2. ECIS. European Cancer Information System From https://ecis.jrc.ec.europa.eu practice guidelines. BMJ 353, i2089 (2016). (2019). 23. Sparano, J. A., Gray, R. J., Makower, D. F., Pritchard, K. I., Albain, K. S., Hayes, D. F. 3. Howlader, N., Altekruse, S. F., Li, C. I., Chen, V. W., Clarke, C. A., Ries, L. A. G. et al. US et al. Adjuvant chemotherapy guided by a 21-gene expression assay in breast incidence of breast cancer subtypes defined by joint hormone receptor and cancer. N. Engl. J. Med. 379, 111–121 (2018). HER2 status. J. Natl Cancer Inst. 28, 106 (2014). 24. Ravdin, P. M., Siminoff, L. A., Davis, G. J., Mercer, M. B., Hewlett, J., Gerson, N. et al. 4. Early Breast Cancer Trialists’ Collaborative Group. Effects of chemotherapy and Computer program to assist in making decisions about adjuvant therapy for hormonal therapy for early breast cancer on recurrence and 15-year survival: an women with early breast cancer. J. Clin. Oncol. 19, 980–991 (2001). overview of the randomised trials. Lancet 365, 1687–1717 (2005). 25. Olivotto, I. A., Bajdik, C. D., Ravdin, P. M., Speers, C. H., Coldman, A. J., Norris, B. D. 5. Early Breast Cancer Trialists’ Collaborative Group. Relevance of breast cancer et al. Population-based validation of the prognostic model ADJUVANT! for early hormone receptors and other factors to the efficacy of adjuvant tamoxifen: breast cancer. J. Clin. Oncol. 23, 2716–2725 (2005). patient-level meta-analysis of randomised trials. Lancet 378, 771–784 (2011). 26. Paik, S., Tang, G., Shak, S., Kim, C., Baker, J., Kim, W. et al. Gene expression and 6. Cardoso, F., van’t Veer, L. J., Bogaerts, J., Slaets, L., Viale, G., Delaloge, S. et al. 70- benefit of chemotherapy in women with node-negative, estrogen receptor- gene signature as an aid to treatment decisions in early-stage breast cancer. N. positive breast cancer. J. Clin. Oncol. 24, 3726–3734 (2006). Engl. J. Med. 375, 717–729 (2016). 27. Albain, K. S., Barlow, W. E., Shak, S., Hortobagyi, G. N., Livingston, R. B., Yeh, I. T. 7. Goldhirsch, A., Wood, W. C., Coates, A. S., Gelber, R. D., Thurlimann, B., Senn, H. J. et al. Breast cancer intergroup of north america. prognostic and predictive value et al. Strategies for subtypes–dealing with the diversity of breast cancer: high- of the 21-gene recurrence score assay in a randomized trial of chemotherapy for lights of the St. Gallen International Expert Consensus on the Primary Therapy of postmenopausal, node-positive, estrogen receptor-positive breast cancer. Lancet Early Breast Cancer 2011. Ann. Oncol. 22, 1736–1747 (2011). Oncol. 11, 55–65 (2010). 8. Coates, A. S., Winer, E. P., Goldhirsch, A., Gelbert, R. D., Gnant, M., Piccart-Gebhart, 28. van de Vijver, M. J., He, Y. D., van’t Veer, L. J., Dai, H., Hart, A. A. M., Voskuil, D. W. M. et al. Tailoring therapies–improving the management of early breast cancer: St et al. A gene-expression signature as a predictor of survival in breast cancer. N. Gallen International Expert Consensus on the Primary Therapy of Early Breast Engl. J. Med. 347, 1999–2009 (2002). Cancer 2015. Ann. Oncol. 26, 1533–1546 (2015).. 29. Bueno-de-Mesquita, J. M., Linn, S. C., Keijzer, R., Wesseling, J., Nuyten, D. S. A., van 9. Petrelli, F., Viale, G., Cabiddu, M. & Barni, S. Prognostic value of different cut-off Krimpen, C. et al. Validation of 70-gene prognosis signature in node-negative levels of Ki-67 in breast cancer: a systematic review and meta-analysis of 64,196 breast cancer. Breast Cancer Res. Treat. 117, 483–495 (2009). patients. Breast Cancer Res. Treat. 153, 477–491 (2015). 30. Mook, S., Schmidt, M. K., Viale, G., Pruneri, G., Eekhout, I., Floore, A. et al. The 70- 10. Early Breast Cancer Trialists’ Collaborative, Group. Comparisons between different gene prognosis-signature predicts disease outcome in breast cancer patients polychemotherapy regimens for early breast cancer: meta-analyses of long-term with 1–3 positive lymph nodes in an independent validation study. Breast Cancer outcome among 100,000 women in 123 randomised trials. Lancet 379, 432–444 Res. Treat. 116, 295–302 (2009). (2012). 31. Blok, E. J., Bastiaanet, E., van den Hout, W. B., Liefers, G. J., Smit, V. T. H. B. M., 11. Schünemann, H. J., Lerda, D., Dimitrova, N., Alonso-Coello, P., Gräwingholt, A., Kroep, J. R. et al. Systematic review of the clinical and economic value of gene Quinn, C. et al. Methods for development of the European Commission Initiative expression profiles for invasive early breast cancer available in Europe. Cancer on Breast Cancer Guidelines: Recommendations in the Era of Guideline Trans- Treat. Rev. 62, 74–90 (2018). parency. Ann. Intern. Med. https://doi.org/10.7326/M18-3445 (2019). 32. Wang, S. Y., Dang, W., Richman, I., Mougalian, S. S., Evans, S. B. & Gross, C. P. Cost- 12. Knauer, M., Mook, S., Rutgers, E. J., Bender, R. A., Hauptmann, M., van de Vijver, M. effectiveness analyses of the 21-gene assay in breast cancer: systematic review J. et al. The predictive value of the 70-gene signature for adjuvant chemotherapy and critical appraisal. J. Clin. Oncol. 36, 1619–1627 (2018). in early breast cancer. Breast Cancer Res. Treat. 120, 655–661 (2010). 33. Davidson, J. A., Cromwell, I., Ellard, S. L., Lohrisch, C., Gelmon, K. A., Shenkier, T. 13. Dowsett, M., Sestak, I., Lopez-Knowles, E., Sidhu, K., Dunbier, A. K., Cowens, J. W. et al. A prospective clinical utility and pharmacoeconomic study of the impact of et al. Comparison of PAM50 risk of recurrence score with oncotype DX and IHC4 the 21-gene Recurrence Score(R) assay in oestrogen receptor positive node for predicting risk of distant recurrence after endocrine therapy. J Clin Oncol. 31, negative breast cancer. Eur. J. Cancer 49, 2469–2475 (2013). 2783–90. (2013). 34. Jahn, B., Rochau, U., Kurzthaler, C., Hubalek, M., Miksad, R., Sroczynski, G. et al. 14. Gnant, M., Sestak, I., Filipits, M., Dowsett, M., Balic, M., Lopez-Knowles, E. et al. Personalized treatment of women with early breast cancer: a risk-group specific Identifying clinically relevant prognostic subgroups of postmenopausal women cost-effectiveness analysis of adjuvant chemotherapy accounting for companion with node-positive hormone receptor-positive early-stage breast cancer treated prognostic tests OncotypeDX and Adjuvant!Online. BMC Cancer 17, 685 (2017). with endocrine therapy: a combined analysis of ABCSG-8 and ATAC using the 35. Katz, G., Romano, O., Foa, C., Vataire, A. L., Chantelard, J. V., Hervé, R. et al. Eco- PAM50 risk of recurrence score and intrinsic subtype. Ann. Oncol. 26, 1685–1691 nomic impact of gene expression profiling in patients with early-stage breast (2015). cancer in France. PLoS ONE 10, e0128880 (2018). 15. Sestak, I., Cuzick, J., Dowsett, M., Lopez-Knowles, E., Filipits, M., Dubsky, P. et al. 36. Lyman, G. H., Cosler, L. E., Kuderer, N. M. & Homberger, J. Impact of a 21-gene RT- Prediction of late distant recurrence after 5 years of endocrine treatment: a PCR assay on treatment decisions in early-stage breast cancer: An economic Recommendations from the European Commission Initiative on Breast Cancer. . . P Giorgi Rossi et al. 1512 analysis based on prognostic and predictive validation studies. Cancer 109, 54. Schünemann, H. J., Lerda, D., Quinn, C., Follmann, M., Alonso-Coello, P., Giorgi Rossi 1011–1018 (2007). P. et al. Breast cancer screening and diagnosis: a synopsis of the European breast 37. Ontario Health Technology Advisory Committee, Medical Advisory Secretariat. guidelines. Ann. Intern. Med. https://doi.org/10.7326/M19-2125, in press (2021). Gene expression profiling for guiding adjuvant chemotherapy decisions in 55. NICE National Institute for Health Care and Excellence (2018). Tumour profiling women with early breast cancer: an evidence- based and economic analysis. Ont. tests to guide adjuvant chemotherapy decisions in early breast cancer. Diag- Health Technol. Assess Ser 10, 1–57 (2010). nostic Guidance DG34. https://www.nice.org.uk/guidance/dg34/resources/ 38. Paulden, M., Franek, J., Pham, B., Bedard, P. L., Trudeau, M. & Krahn, M. Cost- tumour-profiling-tests-to-guide-adjuvant-chemotherapy-decisions-in-early- effectiveness of the 21-gene assay for guiding adjuvant chemotherapy decisions breast-cancer-pdf-1053750722245 (2020). in early breast cancer. Value Health 16, 729–739 (2013). 56. Cuzick, J., Dowsett, M., Pineda, S., Wale, C., Salter, J., Quinn, E. et al. Prognostic 39. Vataire, A. L., Laas, E., Aballea, S., Gligorov, J., Rouzier, R. & Chéreau, E. value of a combined estrogen receptor, progesterone receptor, Ki-67, and human Cost-effectiveness of a chemotherapy predictive test. Bull Cancer 99, 907–914 epidermal growth factor receptor 2 immunohistochemical score and comparison (2012). with the Genomic Health recurrence score in early breast cancer. J. Clin. Oncol. 40. Kondo, M., Hoshi, S. L., Ishiguro, H. & Toi, M. Economic evaluation of the 70-gene 29, 4273–4278 (2011). prognosis-signature (MammaPrint) in hormone receptor-positive, lymph node- 57. Barton, S., Zabaglo, L., A’Hern, R., Turner, N., Ferguson, T., O´Neill, S. et al. negative, human epidermal growth factor receptor type 2-negative early stage Assessment of the contribution of the IHC4 + C score to decision making in breast cancer in Japan. Breast Cancer Res. Treat. 133, 759–68. (2012). clinical practice in early breast cancer. Br. J. Cancer 106, 1760–1765 (2012). 41. Ward, S., Scope, A., Rafia, R., Pandor, A., Harnan, S., Evans, P. et al. Gene expression 58. Candido Dos Reis, F. J., Wishart, G. C., Dicks, E. M., Greenberg, D., Rashbass, J., profiling and expanded immunohistochemistry tests to guide the use of adjuvant Schmidt, M. K. et al. An updated PREDICT breast cancer prognostication and chemotherapy in breast cancer management: a systematic review and cost- treatment benefit prediction model with independent validation. Breast Cancer effectiveness analysis. Health Technol. Assess 17, 1–302 (2013). Res. 19, 58 (2017). 42. Blohmer, J. U., Rezai, M., Kummel, S., Kühn, T., Warm, M., Friedrich, K. et al. Using 59. Harris, L. N., Ismaila, N., McShane, L. M., Andre, F., Collyar, D. E., Gonzalez-Angulo, A. the 21-gene assay to guide adjuvant chemotherapy decision-making in early- M. et al. Use of biomarkers to guide decisions on adjuvant systemic therapy for stage breast cancer: a cost-effectiveness evaluation in the German setting. J. Med. women with early-stage invasive breast cancer: American Society of Clinical Econ. 16, 30–40 (2013). Oncology Clinical Practice Guideline. J. Clin. Oncol. 34, 1134–1150 (2016). 43. Nerich, V., Curtit, E., Bazan, F., Montcuquet, P., Villanueva, C., Chaigneau, L. et al. 60. Krop, I., Ismaila, N., Andre, F., Bast, R. C., Barlow, W., Collyar, D. E. et al. Use of Economic assessment of the routine use of Oncotype DX assay for early breast biomarkers to guide decisions on adjuvant systemic therapy for women with cancer in Franche-Comte region. Bull Cancer 101, 681–689 (2014). early-stage invasive breast cancer: American Society of Clinical Oncology Clinical 44. Vanderlaan, B. F., Broder, M. S., Chang, E. Y., Oratz, R. & Bentley, T. G. K. Cost- Practice Guideline Focused Update. J. Clin. Oncol. 35, 2838–2847 (2017). effectiveness of 21-gene assay in node-positive, early-stage breast cancer. Am. J. 61. Andre, F., Ismaila, N., Henry, N. L., Somerfield, M. R., Bast, R. C., Barlow, W. et al. Use Manag. Care 17, 455–464 (2011). of biomarkers to guide decisions on adjuvant systemic therapy for women with 45. Sparano, J. A., Gray, R. J., Makower, D. F., Albain, K. S., Saphner, T. J., Badve, S. S. early-stage invasive breast cancer: ASCO Clinical Practice Guideline Update- et al. Clinical outcomes in early breast cancer with a high 21-gene recurrence Integration of Results From TAILORx. J. Clin. Oncol. 37, 1956–1964 (2019). score of 26 to 100 assigned to adjuvant chemotherapy plus endocrine therapy: a 62. Harnan, S., Tappenden, P., Cooper, K., Stevens, J., Bessey, A., Rafia, R. et al. Tumour secondary analysis of the TAILORx randomized clinical trial. JAMA Oncol. https:// profiling tests to guide adjuvant chemotherapy decisions in early breast cancer: doi.org/10.1001/jamaoncol.2019.4794 (in press). a systematic review and economic analysis. Health Technol Assess 23, 1–328 46. Sparano, J. A., Gray, R. J., Ravdin, P. M., Makower, D. F., Pritchard, K. I., Albain, K. S. (2019). et al. Clinical and genomic risk to guide the use of adjuvant therapy for breast 63. Schildgen, V., Warm, M., Brockmann, M. & Schildgen, O. Oncotype DX breast cancer. N. Engl. J. Med. 380, 2395–2405 (2019). cancer recurrence score resists inter-assay reproducibility with RT(2)-profiler 47. Giorgi Rossi, P., Lebeau, A. & Schünemann, H. J. Clinical and genomic risk in multiplex RT-PCR. Sci. Rep. 9, 20266 (2019). adjuvant therapy for breast cancer. N. Engl. J. Med. 381, 1289–1290 (2019). 64. Sestak, I., Martín, M., Dubsky, P., Kronenwett, R., Rojo, F., Cuzick, J. et al. Prediction 48. Dowsett, M. & Turner, N. Estimating risk of recurrence for early breast cancer: of chemotherapy benefit by EndoPredict in patients with breast cancer who integrating clinical and genomic risk. J. Clin. Oncol. 37, 689–692 (2019). received adjuvant endocrine therapy plus chemotherapy or endocrine therapy 49. Mariotto, A., Jayasekerea, J., Petkov, V., Schechter, C. B., Enewold, L., Helzlsouer, K. alone. Breast Cancer Res Treat. 176, 377–386 (2019). J. et al. Expected monetary impact of oncotype DX score-concordant systemic breast cancer therapy based on the TAILORx trial. J. Natl Cancer Inst. 112, 154–160 (2020). Open Access This article is licensed under a Creative Commons 50. Giorgi Rossi, P. & Paci, E. Re: expected monetary impact of oncotype DX score- Attribution 4.0 International License, which permits use, sharing, concordant systemic breast cancer therapy based on the TAILORx trial. J. Natl adaptation, distribution and reproduction in any medium or format, as long as you give Cancer Inst. https://doi.org/10.1093/jnci/djz125, in press (2021). appropriate credit to the original author(s) and the source, provide a link to the Creative 51. Oxman, A. D., Fretheim, A. & Schünemann, H. J. SURE. Improving the use of Commons license, and indicate if changes were made. The images or other third party research evidence in guideline development: introduction. Health Res. Policy Syst. 4, material in this article are included in the article’s Creative Commons license, unless 12 (2006). indicated otherwise in a credit line to the material. If material is not included in the 52. Schünemann, H. J., Wiercioch, W., Etxeandia, I., Falavigna, M., Santesso, N., article’s Creative Commons license and your intended use is not permitted by statutory Mustafa, R. et al. Guidelines 2.0: systematic development of a comprehensive regulation or exceeds the permitted use, you will need to obtain permission directly checklist for a successful guideline enterprise. CMAJ 186, E123–E142 (2014). from the copyright holder. To view a copy of this license, visit http://creativecommons. 53. Woolf, S., Schünemann, H. J., Eccles, M. P., Grimshaw, J. M. & Shekelle, P. Devel- org/licenses/by/4.0/. oping clinical practice guidelines: types of evidence and outcomes; values and economics, synthesis, grading, and presentation and deriving recommendations. Implement Sci. 7, 61 (2012). © The Author(s) 2021