This information collection is a core HTA, i.e. an extensive analysis of one or more health technologies using all nine domains of the HTA Core Model. The core HTA is intended to be used as an information base for local (e.g. national or regional) HTAs.

Prognostic tests for breast cancer recurrence (uPA/PAI-1 [FEMTELLE], MammaPrint, Oncotype DX )

UPA/PAI-1 (FEMTELLE), MammaPrint, Oncotype DX compared to Standard of care in selecting treatment for Breast cancer recurrence in females

(See detailed scope below)

HTA Core Model Application for Diagnostic Technologies (1.1)
Core HTA
Tom Jefferson (age.na.s, Italy), Nicola Vicari (age.na.s, Italy), Heike Raatz (SNHTA, Switzerland)
Sarah Baggaley, NICE (Health problem and current use); Antonio Migliore, Agenas (Description and technical characteristics); Iris Pasternack, THL-FINOHTA (Safety); Mirjana Huic, AAZ (Clinical effectiveness), Isaura Vieira, INFARMED (Costs and economic evaluation); Dario Sacchini, A.Gemelli (Ethical analysis); Jennifer Butt, NICE (Organisational aspects); Marco Marchetti, A.Gemelli (Social and Legal aspects)
Agenzia nationale per i servizi sanitari regionali (age.na.s), Italy
A. Gemelli (Italy), AAZ (Croatia), Agenas (Italy), AHTAPol (Poland), AVALIA-t (Spain), INFARMED (Portugal), IPH-RS (Slovenia), NICE (United Kingdom), Regione Veneto (Italy), SNHTA (Switzerland), THL (Finland), UMIT (Austria)
13.6.2011 14.00.00
31.1.2013 18.05.00
Jefferson T, Vicari N, Raatz H [eds.]. Prognostic tests for breast cancer recurrence (uPA/PAI-1 [FEMTELLE], MammaPrint, Oncotype DX ) [Core HTA], Agenzia nationale per i servizi sanitari regionali (age.na.s), Italy ; 2013. [cited 7 February 2023]. Available from: http://corehta.info/ViewCover.aspx?id=113

Prognostic tests for breast cancer recurrence (uPA/PAI-1 [FEMTELLE], MammaPrint, Oncotype DX )

Health Problem and Current Use of the Technology >>

Collection summary


Given its high impact on the healthcare service, the management of breast cancer is a relevant issue for all European Union (EU) countries. The three tests assessed in this core health technology assessment (HTA)— uPA/PAI-1 (FEMTELLE®, American Diagnostica) based on immunostaining techniques and MammaPrint® (Agendia) and Oncotype DX® (Genomic Health), based on gene expression profiling—measure multiple markers within the tumour that may indicate how the tumour is likely to develop. The potential clinical utility of the tests lies in their ability to discriminate between patients who will benefit to a greater or lesser extent from a therapeutic intervention. The assessment of technologies of this type could be of interest to all the EU member states.

The uPA/PAI-1, MammaPrint and Oncotype DX tests are intended to predict the likelihood of breast cancer recurrence in women and to support the tailoring of treatment to the individual patient, which may reduce the number of women receiving chemotherapy and experiencing the associated side-effects. Specifically, the uPA/PAI-1 test is intended for women with newly diagnosed lymph node-negative breast cancer. MammaPrint is intended as a prognostic test for women with lymph node-negative and lymph node-positive breast cancer with a tumour size of 5 cm or less, or for women who are 61 years of age or less, with oestrogen receptor-positive or oestrogen receptor-negative, lymph node-negative breast cancer. The Oncotype DX assay is intended for use for women with newly diagnosed lymph node-negative or lymph node-positive, oestrogen receptor-positive, human epidermal growth factor receptor 2 (HER2)-negative invasive breast cancer.


Safety of the technology (SAF)

The three tests under evaluation share general safety concerns about aspects of environmental safety and surgical pathology practice, such as sample contamination, delays in transfer of samples to the testing laboratory, incorrect labelling and other features that may affect the reliability of the result and thus patient safety. There is insufficient evidence, from the 12 included studies, of possible anxiety or other psychosocial harms caused to the patients by the tests.

Effectiveness of the technology (EFF)

Studies assessing the prognostic/predictive accuracy of the tests were excluded. The 15 poor quality observational studies included did not provide any comparative direct evidence on the effects of treatment with and without the use of the index test. They assessed issues such as quality of life, anxiety and patient satisfaction. There is a clear need for manufacturers to communicate at an early stage with the European Medicines Agency (EMA), the US Food and Drug Administration (FDA) and health technology assessment (HTA) bodies to obtain so-called “early scientific advice” when designing randomised controlled trials (RCTs).

Costs, economic evaluation of the technology (ECO)

If dominance of the prognostic tests for breast cancer recurrence (PTBCRs) were to be shown in prospective independent comparative studies, introduction of these tests might avoid unnecessary use of expensive chemotherapy with its associated adverse effects for women who would derive little or no benefit from it. The tests might reduce the mortality rate in high-risk women who would have benefitted from chemotherapy. However, the key effectiveness data needed to draw conclusions about cost-effectiveness ratios is missing. The generation of appropriate effectiveness data should make the estimation of such ratios possible in the future.

Ethical aspects of the technology (ETH)

Direct evidence of improved outcomes is lacking for all three tests, but studies of components of clinical utility might provide indirect evidence. There is encouraging indirect evidence for Oncotype DX, and plausibility for potential use of MammaPrint and, possibly, the uPA/PAI-1 test. It seems plausible that more women will benefit (i.e. avoid unnecessary chemotherapy), but there is the potential for significant harms among a small number of low or intermediate risk women (who might have benefitted from chemotherapy), possibly resulting in breast cancer recurrence or death. Currently, there are insufficient data to confidently estimate these risks and benefits. In addition, it is difficult to determine what proportion of women with moderate to high risk, based on conventional risk assessments, will have a “low enough” score to affect their decision about chemotherapy. The use of the tests raises the question of the extent to which patients are prepared to participate in informed decision making about their care.

Organisational aspects of the technology (ORG)

The introduction of PTBCRs into the clinical pathways of women with breast cancer is unlikely to have significant organisational impacts. However attention should be paid to costs, and to communication between provider units and patients.

Social aspects of the technology (SOC)

If PTBCRs are introduced, emotional and psychological support is likely to be useful, especially if results of these prognostic tests are not in accord with those of standard clinicopathological prognostic factors analysis. Feelings of distress and anxiety are frequently reported by patients undergoing prognostic tests and are related to the consequences of the tests on treatment decisions. It remains unclear whether and how these feelings affect the social areas of the patient. Overall knowledge about some aspects of PTBCRs, such as harms, is low.

Legal aspects of the technology (LEG)

Legislation/regulation on three aspects should be developed at the European level. (1) the absence of an ad hoc European directive on medical devices—in vitro predictive tests, (2) the absence of the FEMTELLE uPA/PAI-1 test, MammaPrint and Oncotype DX from the Eudamed register (European registry) and (3) the need to ensure equitable access to these tests.

Closing Remarks

The Core Model is not intended to provide a cookbook solution to all problems but to suggest a way in which information can be assembled and structured, and to facilitate local adaptation. The information is assembled around the nine domains, each with several result cards in which questions and possible answers are reported.

The reasons for having a standardised but flexible content and layout are rooted in the way HTA is carried out in the EU HTA institutions and in the philosophy of the first EUnetHTA Joint Action (JA1) production experiment.

HTA is a complex multidisciplinary activity addressing a very complex reality – that of healthcare. Uniformly standardised evidence-based methods of conducting assessments for each domain do not exist (Corio M, Paone S, Ferroni E, Meier H, Jefferson TO, Cerbo M.  – Systematic review of the methodological instruments used in Health Technology Assessment. Rome, July 2011.). There are sometimes variations across and within Member States in how things are done and which aspects of the evaluation are privileged. This is especially so for the “softer”, more context-dependent domains such as the ethical and social domains.

Currently there are only three applications of the HTA Core Model (medical and surgical interventions, diagnostic technologies, and screening technologies). In this work the HTA Core Model for diagnostic technologies was used to assess the clinical effectiveness of the three prognostic tests. However there are fundamental differences between diagnostic and prognostic tests. Because of these differences, a number of the HTA Core assessment element questions for diagnostic technologies are not suited to prognostic technologies. Prognostic/predictive accuracy was not assessed in this HTA Core Model.

This test would represents a useful lesson for methodological development in EUnetHTA Joint Action 2.

Collection methodology


To produce a Core health technology assessment (HTA) assessing the effects of prognostic tests for breast cancer recurrence (PTBCRs) based on the EUnetHTA Core Model and working within the Collaborative Model 1 (COLMOD 1) organisational framework.


The work was based on the HTA Core Model on diagnostic technologies which was developed during the EUnetHTA Project 2006–2008.

The first step was the selection of the technology to be assessed using the Core Model; this phase was carried out through a three-step process that included surveys and questionnaires to WP4 partners by email. At the same time, the Collaborative Model to be used in this Core HTA was chosen by WP4 Partners.

Then there was the check of Partners’ availability to take the lead, as Primary Investigator, in one of the nine evaluation domains. At the same time, the nine Domain Teams were built up in accordance with Partners’ preferences and some general guidelines (i.e.: “each WP4/B AP should be involved in at least one domain, indicating its interest for at least one domain”).

Finally the specific work plan was shared, in accordance with the general WP4 3-year work plan and objectives. This specific work plan included the phases scheduled in the “HTA Core Model Handbook” (Production of Core HTAs and structured HTA information).

An editorial team was set up for discussion and major decisions on basic principles and solutions related to the content of the core HTA. The editorial team was chaired by Tom Jefferson (Agenas), vice-chaired by Heike Raatz (SNHTA), and composed of all the primary investigators of domains.

To allow collaboration between partners a draft protocol for Core Model use was agreed by the researchers involved. The research questions for each of the nine domains of the Core Model were formulated and the corresponding relevant assessment elements (AE) were selected.

Overlaps within domains were identified and assigned exclusively to one domain, by mutual agreement.

The research strategy was carried out by one of the domain team, collecting input from the others.

Evidence from published and manufacturer sources was identified, retrieved, assessed, and included according to pre-specified criteria and summarised to answer each AE. Each domain assessment was made by researchers from different institutions led by a primary investigator (COLMOD1); researchers from different WP4 Partners reviewed and commented on the Core HTA.

Introduction to collection

This section provides background information on the preparation and development of the Core HTA on prognostic tests for breast cancer recurrence (PTBCRs; uPA/PAI-1 [FEMTELLE®, American Diagnostica], MammaPrint® [Agendia] and Oncotype DX® [Genomic Health] ). The core HTA document was produced during the course of the first EUnetHTA Joint Action (JA1) 2010–2012.

The idea behind EUnetHTA’s Core Model is to provide a framework for structuring relevant HTA information while at the same time facilitating local use and adaptation of the information or guiding its production.

The Model is based on nine dimensions or “domains” of evaluation:

  1. Health problem and current use of the technology (CUR)
  2. Description and technical characteristics of technology (TEC)
  3. Safety (SAF)
  4. Effectiveness  (EFF)
  5. Costs and economic evaluation  (ECO)
  6. Ethical analysis (ETH)
  7. Organisational aspects (ORG)
  8. Social aspects (SOC)
  9. Legal aspects (LEG)

In the following nine documents the PTBCRs were assessed using the HTA Core Model Application for Diagnostic Technologies (1.1), developed during the EUnetHTA project 2006–2008.

The PTBCR Core HTA was prepared using an experimental Collaborative Model (COLMOD). The PTBCR Core HTA was prepared using COLMOD 1 in which groups of researchers from 22 different HTA Institutions (55 researchers in all) produced the domain texts. The experimental organisational model added a strong element of challenge but probably helped forge strong links across participants.

In the last months an intensive programme of interviews and consultations will elicit comments and feedback, both from those who contributed to the Core HTA and from those who read it for the first time. This validation plan includes an internal audit within the WorkPackage 4 teams during which each partner will validate the Core HTA they did not produce and the Core HTA production process (collaborative models, on-line tool, etc.).

At the same time, as scheduled in the 3-year work plan, the  Core HTA will be sent to the Stakeholder Advisory Group (SAG) for review and feedback before the final Public Consultation phase, during which the Core HTAs will be published in the On-line Tool and Service.

The results from the Validation Phase and SAG consultation will help us in amending the introduction of the 2 Core HTAs and in preparing a first methodological guidance for the EUnetHTA Joint Action 2.

This process, we hope, will provide invaluable information to improve the product and will help us to understand whether our efforts have been worthwhile.

The following contributed to the preparation of the document:


  • A. Gemelli, Italy
  • AAZ, Croatia
  • AETSA, Spain
  • Agenas,  Italy
  • AHTAPol, Poland
  • ASSR-RER, Italy
  • AVALIA-t, Spain
  • CAHIAQ, Spain
  • GÖG, Austria
  • HIQA, Ireland
  • IER (Slovenia)
  • INFARMED, Portugal
  • IQWIG, Germany
  • IPH-RS, Slovenia
  • Laziosanità, Italy
  • LBI-HTA, Austria
  • NICE, United Kingdom
  • Regione Veneto, Italy
  • SDU/CAST, Denmark
  • SNHTA, Switzerland
  • THL, Finland
  • UMIT, Austria
  • UTA, Estonia


  • Cari Almazan (CAHIAQ)
  • Ana Bação (Infarmed)
  • Sarah Baggaley (NICE)
  • Luciana Ballini (ASSR-RER)
  • Lidia Becla (AHTAPol)
  • Hanan Bell (NICE)
  • Jennifer Butt (NICE)
  • Angelica Carletto (A. Gemelli)
  • Marina Casini (A.Gemelli)
  • Emilio Chiarolla (Agenas)
  • Americo Cicchetti (A. Gemelli)
  • Mirella Corio (Agenas)
  • Chiara Filippi (Regione Veneto)
  • Teresa Gasparetto (Regione Veneto)
  • Massimo Gion (Regione Veneto)
  • Paolo Giorgi Rossi (Laziosanità)
  • Patricia Harrington (HIQA)
  • Mirjana Huic (AAZ)
  • Marjetka Jelenc (IPH-RS)
  • Gurleen Jhuti (NICE)
  • Eva-Maria Kernstock (GÖG)
  • Vesna Kovač (IER)
  • Stefan Lange (IQWIG)
  • Anne Lee (SDU/CAST)
  • Kristi Liiv (UTA)
  • Aurora Llanos-Mendez (AETSA)
  • Alessandro Lo Scalzo (Agenas)
  • Marco Marchetti (A.Gemelli)
  • Mirella Marlow (NICE)
  • Stefan Mathis-Endenhofer (LBI-HTA)
  • Emanuela Midolo (A.Gemelli)
  • Antonio Migliore (Agenas)
  • Roberta Minacori (A. Gemelli)
  • Judite Neves (Infarmed)
  • Monica O'Neill (HIQA)
  • Marco Oradei (A.Gemelli)
  • Monica Orzel (AHTAPol)
  • Anna Panasiuk (AHTAPol)
  • Iris Pasternack (THL)
  • Maria Rosaria Perrini (Agenas)
  • Heike Raatz (SNHTA)
  • Alexandra Ramssl-Sauer (GÖG)
  • Pietro Refolo, (A.Gemelli)
  • Matteo Ruggeri (A.Gemelli)
  • Katarzyna Sejbuk (AHTAPol)
  • Dario Sacchini (A.Gemelli)
  • Narine Sahakyan (UMIT)
  • Stefan Sauerland (IQWIG)
  • Petra Schnell-Inderst (UMIT)
  • Eva Turk (IPH-RS)
  • LeonorVarelaLema (AVALIA-t)
  • Isaura Vieira (INFARMED)
  • Luca Vignatelli (ASSR-RER)
  • Siw Waffenschmidt (IQWIG)
  • Claudia Wild (LBI-HTA)


TechnologyuPA/PAI-1 (FEMTELLE), MammaPrint, Oncotype DX

Urokinase plasminogen activator /plasminogen activator inhibitor 1 ELISA (uPA/PAI-1) is a registered enzyme-linked immunoassay (ELISA) kit (FEMTELLE) for the analysis of uPA/PAI-1 in fresh frozen tissue and is being provided by American Diagnostica Inc. It is CE marked in Europe but for research use only in the USA. Other commercial ELISA kits for separate in-house analysis of uPA and/or PAI-1 are available from different suppliers. These also use samples other than tissue and are also used for indications other than cancer {1}.

Technical details:

- Inspection of unfixed tissue

- Removal of a representative piece of tumour tissue (>50 mg)

- Freezing of the unfixed tissue (-20°C or colder)

- Storage of the frozen tissue (-20°C or colder) possible up to 3 weeks

Clinical Laboratory (Pathology, Hospital)

- Transport of frozen tumour tissue on dry ice

- Extraction of uPA and PAI-1


- Transfer of test results to physician

Costs for FEMTELLE including preparation, shipping and analysis of samples in a qualified laboratory amount to €400 (http://www.hkk.de/info/aktuelles/brustkrebs_tumorprognosetest). In house analysis with separate ELISA kits costs about €200.

Possible logistic issues to consider are {2}:

- Relatively large samples are needed. Given that the mean tumour size is <2 cm in many centres, this means that a substantial part of the tissue may be lacking for light microscopic investigation.

- Many centres no longer routinely freeze breast tissue and therefore lack the expensive equipment for this process.

Oncotype DX (Genomic Health) quantifies gene expression for 21 genes in breast cancer tissue by real-time reverse transcriptase-polymerase chain reaction (RT-PCR).

MammaPrint (Agendia) is a gene expression profiling platform based on microarray technology which uses a 70-gene expression profile {3}. The sample studied is fresh or frozen tissue. It has received 510(k) clearance from the FDA (premarket notification for medical devices), which also covers the use of Asuragen's RNARetain®, a room temperature, molecular fixative that supersedes freezing the tissue before shipment to the central US laboratory (www.agendia.com).

The test requires a fresh sample of tissue  composed of a minimum of 30% malignant cells and must be received by the company in their kit within 5 days of obtaining the material. The MammaPrint assay was developed on the basis of research initially conducted at the Netherlands Cancer Institute (Amsterdam) and collaborating institutions. Primary tumours from 117 patients with axillary lymph node-negative primary breast cancer were analysed on oligonucleotide microarrays. The data were subjected to supervised classification to establish a 70-gene RNA expression profile that correlated with a relatively short interval to distant metastases. [from NICE protocol and ASCO guideline]

Oncotype DX and MammaPrint have been evaluated and large-scale studies (TAILORx and MINDACT) are underway. The German Working Group for Gynecological Oncology1 (AGO) and the American Society of Clinical Oncology (ASCO) have recommended uPA/PAI-1 as risk-group-classification markers for routine clinical decision making in node-negative breast cancer, alongside established clinical and histomorphological factors.

Oncotype DX is recommended for node negative, oestrogen receptor-positive women and MammaPrint is applied in all early breast cancers. The tests are expensive: MammaPrint costs €2675 and Oncotype DX, US $3400.

RT-PCR and microarray analysis usually cost US $3500 or more. Oncotype and MammaPrint are not routinely covered by German statutory health insurance. MammaPrint is covered by Medicare and Medicaid in the USA (Pharmacogenomics Reporter: 23 December 2009; www.genomeweb.com.)

MeSH Terms:

There are no MeSH-Terms for Oncotype DX and MammaPrint.

Intended use of the technologyDefining an existing health condition in further detail to assist selection of appropriate or optimal treatment

Assessment of risk of breast cancer recurrence

Target condition
Breast cancer recurrence
Target condition description

Assessment of risk of breast cancer recurrence and likelihood of benefit from adjuvant treatment (particularly chemotherapy).

As testing for oestrogen receptor positivity is already considered to be part of the standard of care using these tests to decide on adjunctive treatment with Tamoxifen will not be considered part of the study question.

Target population

Target population sex: Female. Target population age: Any age except fetuses. Target population group: Patients who have the target condition.

Target population description

Women with invasive breast cancer in whom adjunctive treatment might be indicated

ComparisonStandard of care

Standard care without any of the three index tests (uPA/PAI-1, MammaPrint, Oncotype DX).

Depending on manpower and time resources the three index tests may also be compared with each other.

Health Problem and Current Use of the Technology >>