MR360 - Early Breast Cancer Trialists' Collaborative Group

The Early Breast Cancer Trialists' Collaborative Group (EBCTCG) was created in 1985 by researchers at the Clinical Trial Service Unit (CTSU), University of Oxford. The membership of EBCTCG consists of research groups which share their trial data for the purpose of the meta-analyses that assess the benefits and risks of treatments for early breast cancer. The EBCTCG Secretariat are members of staff at CTSU. They identify trials, contact collaborators to request data, process & analyse data and, with input from the collaborators, publish reports of the analyses.

Under previous versions of this Agreement the University of Oxford received identifiable MRIS Members and Posting, Cause of Death, Flagging Current Status and Cohort Event Notification Reports for the purpose of a research project referred to as Early Breast Cancer Trialists Collaborative Group (EBCTCG).

All data received under previous iterations of this Agreement relates to women who have been diagnosed with operable breast cancer (or breast cancer which might become operable through the use of neo-adjuvant therapy) and enrolled in one of seven randomised trials (listed below) comparing treatments for breast cancer, with recurrence or death as a principal outcome.

Going forward, the University of Oxford will receive pseudonymised Civil Registration (Deaths), Demographics, and Cancer Registration Data.

RANDOMISED TRIALS:

• Ovarian Irradiation Trial, Part I: Entry JUN-1948 to DEC-1950; 189 entered; (J Faculty Radio 10: 175-80 1959)

• Addenbrooke's Stage II: Entry OCT-1958 to MAY-1965 233 entered; (Lancet 2:1086-7 1971)

• Cardiff Trial: Entry SEP-1967 to JUN-1973; 200 entered; (Ann Surg Oncol 6(5):455-60 1999)

• Lister Trial: Entry SEP-1969 to SEP-1976; 534 entered; (Ann R Coll Surg Engl 63(4)239-43 1981)

• Regional Breast Study 1 & 2: Entry MAR-1970 to OCT-1975; 1012 entered;; 279 participants still alive in 1990 flagged in 2001 (Br J Surg 69(12) 693-6 1995)

• King's/Cambridge Radiotherapy (C.R.C. 1): Entry MAY-1970 to JUN-1975; 2800 entered;; a number of UK participants were flagged in 1999 (World J Surg. 1994 Jan-Feb;18(1):117-22)

• Tamoxifen trial: Entry NOV-1976 to JUL-1982; 1005 entered; (Br J Cancer 62(Suppl 12):16 1990)

• CMF/Tamoxifen UK: Entry MAY-1978 to MAR-1984; 118 entered; (Lancet 2 (8412):1148-9 1984)

• C.R.C. Adjuvant Breast Trial (C.R.C. 2): Entry SEP-1980 to APR-1986; 2230 entered;; a number of UK participants were flagged in 1999 (Br J Cancer 57(6):604-7 1988)

• Conservation Trial: Entry OCT-1982 to DEC-1987; 708 entered; (Clin Oncol R Coll Radiol 5(5):278-83 1993)

The original cohort consisted of 9029 individuals, of which an estimated 1385 are still alive. Appropriate legal permissions are in place for the University of Oxford to retain and process data on the remaining 1385 individuals that are still alive.

In the current iteration of this Agreement (v8), the University of Oxford will receive data on the remaining 1385 individuals, as well as the same data for a further 17,638 participants from eight more randomised trials, listed below:

1. Yorkshire Conservation Trial: Entry JUL-1986 to JUN-1990; 175 entered; Principal Investigator Prof David J Dodwell (Clin Oncol 17:618-622 2005)

2. ALMANAC Trial: Entry NOV-1999 to OCT-2003; 1031 entered; Principal Investigator Prof Robert E Mansel (J Clin Oncol 27:1177-83 2009)

3. East Anglia Sentinel Node Biopsy Trial: Entry NOV-1999 to FEB-2003 (298 entered; Principal Investigator Prof A D Purushotham (J Clin Oncol 23(19):4312-21 2005)

4. BASO II: MAR-1992 to OCT-2000; 1158 entered; Principal Investigator Prof Roger Blamey (Eur J Cancer 49(10):2294-302 2013)

5. C.R.C. Under 50s Trial (part of 'ZIPP'): Entry OCT-1987 to MAR-1999; 208 entered; Principal Investigator Prof M Baum (J Natl Cancer Inst 101(5):341-9 2009)

6. C.R.C. Over 50s Trial: Entry JAN-1987 to FEB-1997; 3888 entered; Principal Investigator Prof M Baum (J Clin Oncol 29(13):1657-63 2011)

7. A.N.Z. DCIS Trial: Entry MAY-1990 to OCT-1998; 1514 entered in UK; Principal Investigator in UK Prof Jack Cuzick (Lancet Oncol 12(1):21-9 2011)

8. ATAC Trial: Entry JUL-1996 to MAR-2000; 9366 entered; Principal Investigator Prof M Baum (Lancet Oncol 11(12):1135-41 2010)

Additional approval from CAG to receive data on the participants of the above trial has been granted.

Breast cancer is a common condition and if some widely practicable treatment could be shown to produce even a moderate reduction of 10 to 15% in mortality this could prevent the deaths from breast cancer of thousands of women each year. Hence, it is important to be able to distinguish between a treatment which produces a modest but real effect on mortality and one that has little or no effect. If such differences are to be assessed reliably, both moderate random errors and moderate biases must be avoided. Systematic overviews of all relevant randomised clinical trials can help in both respects, but to avoid introducing any bias it is important to get as complete follow up information as possible on all randomised patients in all such trials. This avoids undue emphasis on just the more promising (or just the less promising) results. Any possible long-term toxic effects of treatment can also be evaluated by looking at cause-specific mortality of patients who die from causes other than breast cancer and by comparing the incidence of second cancers at other sites in treated and untreated patients.

The most reliable way of ascertaining the vital status of lost patients (i.e., whether they are alive or deceased), of obtaining certified causes of death and ensuring long term mortality follow up is to use central registries. Any data thus obtained can be used to update the results of the relevant trial, so that the overview analyses can be performed more accurately. Data on the incidence of second cancers can also be obtained through the cancer registries and used to monitor long term toxicity.

This study began in 1985, when individual patient data from many randomised trials worldwide was provided to the EBCTCG secretariat for a systematic overview of treatment effects. This first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five-year mortality and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patients. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.

Further cycles of the collaboration were needed to help discover whether these differences persist, to help discover which types of patient are most likely to benefit from such treatments (and which variants of such treatments are most promising), to help assess the various other types of treatment for early breast cancer that have been studied in randomised trials, and to evaluate any long term toxicity of treatments. Therefore, it was intended that the overview data would be updated every five years with data collection taking place in the periods: January 1989 to April 1990, January 1994 to April 1995, January 1999 to April 2000, etc.

The EBCTCG has informed the treatment of women with breast cancer worldwide and will continue to do so into the future. For instance, the results informed the National Institutes of Health (NIH) consensus development conference on the treatment of early breast cancer in 2000. Subsequently, the 2005 report on chemotherapy and endocrine therapy showed the substantial effects on 15-year survival of the chemotherapy regimens and hormonal regimens (such as at least 5 years of tamoxifen in women with oestrogen-positive (ER+) disease) that were tested in the 1980s. The 2005 report on surgery and radiotherapy showed that treatments that substantially improve local control have little effect on breast cancer mortality during the first few years, but definite beneficial effects by 15 years. Updates of these analyses in the 2010s have added further weight to the evidence and further informed national and international guidelines on therapy, including those from the National Institute of Health and Care Excellence (NICE) in the UK. In the current decade, overviews of bisphosphonates, aromatase inhibitors, neoadjuvant chemotherapy and dose-intense chemotherapy, as well as targeted, partial and axillary radiotherapy will provide guidance to healthcare systems globally.

Identifiable patient data have been supplied by NHS Digital to CTSU for an original set of flagged trials. The aggregate number of participants in these trials was about 9,000 in 11 randomised controlled trials; there are currently fewer than 1,385 survivors from 7 of these trials. These trials were undertaken between 1948 and 1987 and since that time the participants have been followed up through their health records. During this period, the CTSU has been acting under the instruction of the trialists; the trials into which these participants were enrolled ceased follow-up and the trialists then gave EBCTCG authorisation to continue to receive follow-up of their outcomes to include in the EBCTCG meta-analyses.

Once received by CTSU, these identifiable data have been kept separate from the EBCTCG database by confining their use to the CTSU Clinical Trials Follow-up Service for EBCTCG (CTSU-ctfs ). The CTSU-ctfs is a separate service provided by CTSU with the purpose of processing personally identifiable information received from NHS-Digital. The CTSU-ctfs carries out the linkage activities between data from NHS-Digital and the original trial data to provide de-identified data to EBCTCG, holding any identifiable information in a separate, secure file store.

The data received from NHS Digital has been combined with an international cohort of women in meta-analyses of randomised controlled trials in the fields of surgery (18,000), radiotherapy versus surgery (51,000), targeted radiotherapy (26,000) and endocrine therapy (8000), this meta-analyses has been carried out by the EBCTCG at the University of Oxford. The wider EBCTCG project includes many more contemporary treatments and targeted therapies, and overall will collect data from more than 725,000 women for around 80 distinct comparisons. Most of the trialists in the EBCTCG collaboration will provide follow-up of individual patient mortality and cancer events directly to the EBCTCG every 5 years until they stop collecting this data. Where registry data is available all possible follow-up information will be sought.

The identifiable fields of live participants held are considered necessary for linkage with NHS data by CTSU. Once the correct link is made between this identifying information and the patient identities held by NHS Digital, these identifiers can be destroyed. No identifiable data is sent to EBCTCG.

The overarching purpose of this Agreement, and previous versions, is to provide outcomes by linkage with NHS Digital data for UK clinical trials so that the EBCTCG can perform individual participant data meta-analyses of the comparative effects of different treatments for women with breast cancer or death from breast cancer. The primary outcome measure of the EBCTCG analyses is time to death from breast cancer. The secondary outcome measures are time to recurrence of breast cancer, time to death from any cause, time to death from non-breast-cancer causes, time to second cancer and type of second cancer.

The sole data controller is the University of Oxford, and all data are processed in the Nuffield Department of Population Health (NDPH) at the University of Oxford. No other organisation processes the data.

The Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) Secretariat at NDPH, University of Oxford receives de-identified data from CTSU-ctfs and analyses them for inclusion in the EBCTCG meta-analyses. The EBCTCG comprises a further 213 organisations from around the world, which contribute data to the overview and are involved with the production of EBCTCG publications. None of these organisations make decisions in how NHS Digital data are processed, and do not have access to NHS Digital data.

The lawful basis for processing the data is GDPR article 6 (1) (e): ‘Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller’. The EBCTCG overview and meta-analyses report on the efficacy and safety of trial therapies and, through dissemination via treatment guidelines to physicians, inform future treatments for individuals with early breast cancer. Therefore processing the data is in the public interest.

The application also falls under Article 9 (2) (j) of EU GDPR, which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. Processing data from the participants of trials in the treatment of breast cancer is in the public interest and is necessary for scientific and research purposes in order to provide secondary analyses of the harms and benefits of breast cancer treatments. This makes full use of the available information for the possible benefit of future breast cancer patients, minimising research waste, while not identifying participants in the trials beyond what is essential for processing.

It is not considered that there is a risk of potential harm to the public from the dissemination of this information. Security measures in place at the University of Oxford ensure data privacy so that the likelihood of a breach of the participant’s privacy is extremely low. This has been confirmed by a data protection impact assessment at the University of Oxford.

Reports detailing the EBCTCG project’s findings will be published, subject to acceptance, in high impact peer-reviewed journals such as The Lancet or Journal of Clinical Oncology. The EBCTCG publications have been cited many thousands of times. The most recent reports published by the collaboration are listed at

https://www.ctsu.ox.ac.uk/research/the-early-breast-cancer-trialists-collaborative-group-ebctcg/ebctcg-publications

Reports of early findings and final results of the analyses will be presented in seminars at the University of Oxford and closed meetings between members of the collaborative group, including the 5-yearly EBCTCG main meeting and the annual EBCTCG Steering Committee meeting, which takes place in September each year. The three current patient members of the EBCTCG steering committee remain active in their roles with the EBCTCG and contribute substantially to steering committee and main meetings, with additional regular meetings to discuss the current analyses. They also disseminate the findings to their respective national and international groups. This level of PPI activity is expected to continue for the life of the EBCTCG project.

Findings from the project will be presented at the San Antonio Breast Cancer Symposium (SABCS) in December each year, and the American Society of Oncology (ASCO) meeting in June each year, and other national and international breast cancer and general cancer symposia.

The EBCTCG and Oxford University web pages will be updated with findings from the EBCTCG project.

All outputs to publications, presentations and web pages will be of aggregate data only with small numbers suppressed in line with HES Analysis Guidance.

Dissemination of project results/outputs is achieved primarily by the EBCTCG in the form of reports and presentations. On submission of a manuscript to a peer-reviewed journal, the funders, Cancer Research UK (CRUK) are informed so they can draft a press release if needed. The readership of EBCTCG reports consists mainly of academic researchers and clinicians, although material is also of interest to members of the public. Among the EBCTCG and its steering committee are leaders in the field of breast cancer who are instrumental in updating and disseminating treatment guidelines for early breast cancer. The EBCTCG Steering committee has three patient representatives who disseminate the findings throughout the international community of breast cancer patients.

All Intellectual Property arising from the analyses and the results of the analyses are owned by the University of Oxford.

In order to increase the impact and accessibility of the findings and in line with NDPH publications policy all EBCTCG results are published with Open Access.

Each research question is updated at approximately 5-year intervals. EBCTCG research questions address the benefits and risks of treating early breast cancer with chemotherapies, surgery, radiotherapy, endocrine therapies and therapies targeted towards molecular pathways and cancer dissemination. Research questions specific to the trials currently flagged by NHS Digital data are:

Axilliary radiotherapy vs not

Accelerated partial breast irradiation (APBI) to whole breast vs tumour bed only

Axillary biopsy with/without axillary radiotherapy

Axillary dissection vs not

Axillary radiotherapy vs axiillary dissection

Chemotherapy with / without tamoxifen

More vs less surgery to axilla

Nodal surgery vs radiotherapy

Ovarian irradiation vs not

Perioperative chemotherapy vs not

Perioperative chemotherapy vs tamoxifen

Radical mastectomy vs modified radical mastectomy

Tamoxifen vs not

Tamoxifen vs ovarian irradiation / oophorectomy

Target dates are not stated for the EBCTCG publications since they are subject to the availability of clinical trial data from many sources. Publication might be delayed while waiting for data in the interests of producing meaningful reports.

New EBCTCG publications are widely reported via press releases, websites, TV and radio.

EBCTCG Publications that use NHS Digital data, currently in progress and in planning, are:

1. A meta-analysis of trials of surgery with or without tamoxifen versus tamoxifen alone in women over the age of 70 will be submitted to the Journal of Clinical Oncology in 2021

2. A meta-analyses on endocrine therapy.

3. A report on ovarian suppression therapy in early breast cancer is in progress

4. A meta-analysis is being finalised of radiotherapy to the regional nodes (internal mammary chain, supraclavicular region and axilla) to obtain as precise an understanding as possible of the potential benefits and harms to evaluate the balance of disease control against late effects of radiotherapy.

5. Meta-analyses are being conducted of different radiotherapy treatments to the axilla, following clinical staging or sentinel lymph node biopsy (SLNB).

6. Data are being collected for a meta-analysis of whole versus partial breast radiotherapy.

Publications which are referenced in the Yielded Benefits below are:

1. Early Breast Cancer Trialists’ Collaborative Group (2014). Effect of radiotherapy after mastectomy and axillary surgery on 10-year recurrence and 20-year breast cancer mortality: meta-analysis of individual patient data for 8135 women in 22 randomised trials.

Lancet. 383:2127-2135. DOI: 10.1016/S0140-6736(14)60488-8.

2. Early Breast Cancer Trialists' Collaborative Group (2015). Aromatase inhibitors versus tamoxifen in early breast cancer: patient-level meta-analysis of the randomised trials. Lancet 386:1341-52. DOI: 10.1016/s0140-6736(15)61074-1

3. Early Breast Cancer Trialists' Collaborative Group (2015). Adjuvant bisphosphonate treatment in early breast cancer: meta-analyses of individual patient data from randomised trials. Lancet 386:1353-61. DOI: 10.1016/S0140-6736(15)60908-4

EBCTCG papers published between January 2008 and May 2017 were primary references in important updates to each of the major international guidelines for the management of early breast cancer. These include those produced by the National Institute for Health and Care Excellence (NICE), the UK’s Royal College of Radiologists (RCR) and the Scottish Intercollegiate Guidelines Network (SIGN); the European Society for Medical Oncology (ESO-ESMO), the Japanese Breast Cancer Society and the St Gallen international consensus on breast cancer; the American National Comprehensive Cancer Network (NCCN), Cancer Care Ontario and the American Society of Clinical Oncology (ASCO), and the American Society for Radiation Oncology (ASTRO). These updated guidelines brought in new recommendations (detailed in the Yielded Benefits) based on EBCTCG evidence, to improve the survival of breast cancer patients.

EBCTCG reports have been published in major medical journals and by September 2020 the 23 published reports (1985-2019) had been cited more than 27,000 times. Twelve of these reports, on ovarian ablation, chemotherapy, endocrine therapy (tamoxifen), surgery and radiotherapy, include data from NHS Digital (reference list).

The 1996 Lancet report on effects of ovarian ablation on recurrence and death showed that 15-year survival was significantly improved in premenopausal women receiving ovarian ablation in surgery or irradiation in trials that began before 1980. The planned update of this report for 2021 will include trials that began between 1987 and 2009.

The 2005 Lancet report on systemic therapies updated reports from 1998 and earlier to show the substantial effects on 15-year survival of the chemotherapy regimens (e.g., About 6 months of anthracycline-based chemotherapy) and endocrine regimens (eg.5 years of tamoxifen) that were being tested in the 1980s.

The 2012 Lancet chemotherapy report updated evidence from 2005, bringing together data from 100,000 women in 123 randomised trials of chemotherapy, showing that chemotherapy can reduce breast cancer mortality not only in ER-negative but also in ER-positive disease, and showing benefits of taxane-based over standard anthracycline chemotherapy.

The 2005 Lancet Report on surgery and radiotherapy updated reports from 1995 and 2000 to show that regimens that substantially reduce 5-year local recurrence rates have little effect on 5-year breast cancer mortality, but moderately reduce 15-year breast cancer mortality.

The 2011 Lancet report on 20,000 women in 20 randomised trials of about 5 years of tamoxifen versus no tamoxifen, showed a highly significant reduction of about a third in breast cancer mortality not only during years 0-4 and 5-9 after starting treatment but also during years 10-14. Tamoxifen is effective whether or not chemotherapy has been given and, importantly, even in weakly ER positive disease.

The long-term nature of the treatment benefits found in the above reports underlines the importance for EBCTCG of obtaining extended follow-up from older trials. For example, surgery and radiotherapy trial follow-up ending at five years would not have detected the 15-year reduction in mortality. Also, if 5-year tamoxifen trial follow-up had ended at 10 years the improvement in survival with 5 years of tamoxifen after 15 years would be missed. The update planned for 2021 of the ovarian suppression meta-analysis, which includes some of the oldest trials flagged by EBCTCG, is likely to provide important information about the long-term effects of this treatment.

The impact of the EBCTCG meta-analyses has been the dissemination of evidence of several moderate treatment effects that together have almost halved breast cancer mortality rates for women aged 35-69 since 1990. In addition, by demonstrating the need for big data in randomised trials the EBCTCG has fostered respect for the value of conducting much larger randomised controlled trials than was customary, in order to assess the effects of treatment properly.

Through conducting large-scale population-level studies, EBCTCG researchers have quantified the long-term benefits and risks of different breast cancer treatments, providing conclusive evidence where there was previously uncertainty. These discoveries have already changed clinical practice, ultimately saving patient lives and improving secondary outcomes. A series of meta-analyses of breast cancer-related clinical trials worldwide generated robust evidence into the effectiveness of treatments. These discoveries have been translated into clinical practice, ultimately benefitting the millions of women who are diagnosed with breast cancer worldwide each year. In particular, these studies identified a more effective treatment (aromatase inhibitors) than the standard endocrine therapy; identified new sub-groups of patients that would benefit from radiotherapy treatment; and demonstrated that bisphosphonates reduce the risk of bone metastasis and death from breast cancer.

The Early Breast Cancer Trialists' Collaborative Group (EBCTCG), based at the University of Oxford has since the 1980s been using individual patient data meta-analysis to assess the long-term benefits and side-effects of different treatment options for early breast cancer, collaborating with breast cancer trialists worldwide to collect long-term outcome data. This puts it in a unique position to assess early and late effects of treatment (either beneficial or harmful), which are not assessed as reliably within the original clinical trial reports. These large datasets allow definitive estimates of treatment effects overall and exploration of any differences in the effects of treatments in different subgroups of women, or between treatments within the same class, analyses which cannot be performed reliably using individual clinical trial publications. The EBCTCG notably provided the first conclusive evidence of the effectiveness of tamoxifen in the mid 1980s and has since produced evidence on other treatment modalities. Notable recent findings from the EBCTCG are:

a) A meta-analysis of 8,000 women in 22 trials of radiotherapy after mastectomy (published 2014) showed that radiotherapy reduced breast cancer recurrence and mortality not only in women whose breast cancer had spread to many lymph nodes but also in those with spread to only 1-3 axillary lymph nodes.

b) A meta-analysis (published 2015) using individual data on almost 32,000 women determined conclusively that aromatase inhibitors (which block oestrogen production) are a more effective treatment for breast cancer than tamoxifen in postmenopausal women. Specifically, aromatase inhibitors were found to reduce recurrence rates by 30% and 10-year mortality rates by 15% compared with tamoxifen.

c) A meta-analysis of 24 trials of the use of bisphosphonate therapy (19,000 women) (published 2015) demonstrated that bisphosphonates reduce the risk of bone metastasis and death from breast cancer in postmenopausal women. Furthermore, bisphosphonates also strengthen bones and effectively reduce damage from osteoporosis caused as a side effect by aromatase inhibitors, therefore adding to their clinical benefit.

In these studies it is important to weigh up benefits and risks which may not occur at the same time: in particular, radiotherapy and treatment with drugs such as anthracyclines can have potential late harms; whereas any benefit of long term treatment with endocrine therapy (such as aromatase inhibitors) is likely to persist may years after surgery. The data from original trial publications may only tell part of the story, and long-term flagging is essential to provide a fully balanced and nuanced exploration of benefits and risks, and their relative timings.

Changes in clinical practice to improve breast cancer treatment

Speaking of the 2014 update to the NICE guidelines on Early and locally advanced breast cancer: diagnosis and management, the Director for Guidelines for NICE wrote:

Widespread adoption of aromatase inhibitors:

Following the publication of a meta-analysis of aromatase inhibitors versus tamoxifen aromatase inhibitors have been adopted as standard-of-care: as a result, starting endocrine therapy with an aromatase inhibitor is now recommended for postmenopausal women with breast cancer whereas previously they were managed with either tamoxifen or an aromatase inhibitor. A study in 2019 found the majority of older women (55+) received aromatase inhibitors [G].

Extension of radiotherapy treatment:

Life-saving radiotherapy treatment is now recommended to a wider range of breast cancer patients, including those whose cancer spreads to 1-3 axillary lymph nodes.

Reducing deaths and adverse side-effects using bisphosphonates

The results of the meta-analysis prompted an immediate policy change, with bisphosphonates now being globally recommended for breast cancer patients, when they were not before. An increase in bisphosphonates use was confirmed by a survey in March 2020 by the charity Breast Cancer Now, finding that 94% of NHS Trusts who have a breast cancer service were routinely prescribing them and a further 3% in process of making them available.

The results presented in the Lancet were recognised by policy makers and used to lobby for improved patient access to bisphosphonates. For instance, during a 2017 debate in the House of Commons on Breast Cancer Drugs, Baroness Blackwood argued in favour of bisphosphonates being licensed as treatment for breast cancer patients, saying ‘research in The Lancet in 2015...found that bisphosphonates can be used to help women who are being treated for early breast cancer after the menopause by reducing the risk of the breast cancer spreading to the bone by 28%’. At least 35,700 postmenopausal women are diagnosed with primary breast cancer in the UK each year. Routine treatment with bisphosphonates prevents 1,180 women from dying from breast cancer annually: equivalent to one in ten breast cancer deaths with an overall net NHS saving of £5.09m per annual cohort of patients.

Ovarian Ablation:

The long-term benefits of ovarian ablation in women under 50 in the absence of other adjuvant treatments were identified, with confirmation that there is no significant benefit in older women.

Radiotherapy and Surgery:

Life-saving radiotherapy treatment is now recommended to a wider range of breast cancer patients, including those whose cancer spreads to 1-3 axillary lymph nodes. Less extensive mastectomy was not found to make a large difference in mortality, at least during the first 10 years following treatment.

Chemotherapy and Hormonal Therapy:

Chemotherapy was demonstrated to give a one-third mortality reduction compared with no chemotherapy, depending on absolute risks. Tamoxifen was found to improve survival and has become a standard therapy in premenopausal women with estrogen receptor-positive early breast cancer. The Chemotherapy / Hormonal therapy and Ovarian Ablation reports also provided evidence for the NICE guidelines on Early and Locally Advanced Breast Cancer update of 2009.

The Early Breast Cancer Trialists' Collaborative Group (EBCTCG) was created in 1985 by researchers at the Clinical Trial Service Unit (CTSU), University of Oxford. The membership of EBCTCG consists of research groups which share their trial data for the purpose of the meta-analyses that assess the benefits and risks of treatments for early breast cancer. The EBCTCG Secretariat are members of staff at CTSU. They identify trials, contact collaborators to request data, process & analyse data and, with input from the collaborators, publish reports of the analyses.

Under previous versions of this Agreement the University of Oxford received identifiable MRIS Members and Posting, Cause of Death, Flagging Current Status and Cohort Event Notification Reports for the purpose of a research project referred to as Early Breast Cancer Trialists Collaborative Group (EBCTCG).

All data received under previous iterations of this Agreement relates to women who have been diagnosed with operable breast cancer (or breast cancer which might become operable through the use of neo-adjuvant therapy) and enrolled in one of seven randomised trials (listed below) comparing treatments for breast cancer, with recurrence or death as a principal outcome.

Going forward, the University of Oxford will receive pseudonymised Civil Registration (Deaths), Demographics, and Cancer Registration Data.

RANDOMISED TRIALS:

• Ovarian Irradiation Trial, Part I: Entry JUN-1948 to DEC-1950; 189 entered; (J Faculty Radio 10: 175-80 1959)

• Addenbrooke's Stage II: Entry OCT-1958 to MAY-1965 233 entered; (Lancet 2:1086-7 1971)

• Cardiff Trial: Entry SEP-1967 to JUN-1973; 200 entered; (Ann Surg Oncol 6(5):455-60 1999)

• Lister Trial: Entry SEP-1969 to SEP-1976; 534 entered; (Ann R Coll Surg Engl 63(4)239-43 1981)

• Regional Breast Study 1 & 2: Entry MAR-1970 to OCT-1975; 1012 entered;; 279 participants still alive in 1990 flagged in 2001 (Br J Surg 69(12) 693-6 1995)

• King's/Cambridge Radiotherapy (C.R.C. 1): Entry MAY-1970 to JUN-1975; 2800 entered;; a number of UK participants were flagged in 1999 (World J Surg. 1994 Jan-Feb;18(1):117-22)

• Tamoxifen trial: Entry NOV-1976 to JUL-1982; 1005 entered; (Br J Cancer 62(Suppl 12):16 1990)

• CMF/Tamoxifen UK: Entry MAY-1978 to MAR-1984; 118 entered; (Lancet 2 (8412):1148-9 1984)

• C.R.C. Adjuvant Breast Trial (C.R.C. 2): Entry SEP-1980 to APR-1986; 2230 entered;; a number of UK participants were flagged in 1999 (Br J Cancer 57(6):604-7 1988)

• Conservation Trial: Entry OCT-1982 to DEC-1987; 708 entered; (Clin Oncol R Coll Radiol 5(5):278-83 1993)

The original cohort consisted of 9029 individuals, of which an estimated 1385 are still alive. Appropriate legal permissions are in place for the University of Oxford to retain and process data on the remaining 1385 individuals that are still alive.

In the current iteration of this Agreement (v8), the University of Oxford will receive data on the remaining 1385 individuals, as well as the same data for a further 17,638 participants from eight more randomised trials, listed below:

1. Yorkshire Conservation Trial: Entry JUL-1986 to JUN-1990; 175 entered; Principal Investigator Prof David J Dodwell (Clin Oncol 17:618-622 2005)

2. ALMANAC Trial: Entry NOV-1999 to OCT-2003; 1031 entered; Principal Investigator Prof Robert E Mansel (J Clin Oncol 27:1177-83 2009)

3. East Anglia Sentinel Node Biopsy Trial: Entry NOV-1999 to FEB-2003 (298 entered; Principal Investigator Prof A D Purushotham (J Clin Oncol 23(19):4312-21 2005)

4. BASO II: MAR-1992 to OCT-2000; 1158 entered; Principal Investigator Prof Roger Blamey (Eur J Cancer 49(10):2294-302 2013)

5. C.R.C. Under 50s Trial (part of 'ZIPP'): Entry OCT-1987 to MAR-1999; 208 entered; Principal Investigator Prof M Baum (J Natl Cancer Inst 101(5):341-9 2009)

6. C.R.C. Over 50s Trial: Entry JAN-1987 to FEB-1997; 3888 entered; Principal Investigator Prof M Baum (J Clin Oncol 29(13):1657-63 2011)

7. A.N.Z. DCIS Trial: Entry MAY-1990 to OCT-1998; 1514 entered in UK; Principal Investigator in UK Prof Jack Cuzick (Lancet Oncol 12(1):21-9 2011)

8. ATAC Trial: Entry JUL-1996 to MAR-2000; 9366 entered; Principal Investigator Prof M Baum (Lancet Oncol 11(12):1135-41 2010)

Additional approval from CAG to receive data on the participants of the above trial has been granted.

Breast cancer is a common condition and if some widely practicable treatment could be shown to produce even a moderate reduction of 10 to 15% in mortality this could prevent the deaths from breast cancer of thousands of women each year. Hence, it is important to be able to distinguish between a treatment which produces a modest but real effect on mortality and one that has little or no effect. If such differences are to be assessed reliably, both moderate random errors and moderate biases must be avoided. Systematic overviews of all relevant randomised clinical trials can help in both respects, but to avoid introducing any bias it is important to get as complete follow up information as possible on all randomised patients in all such trials. This avoids undue emphasis on just the more promising (or just the less promising) results. Any possible long-term toxic effects of treatment can also be evaluated by looking at cause-specific mortality of patients who die from causes other than breast cancer and by comparing the incidence of second cancers at other sites in treated and untreated patients.

The most reliable way of ascertaining the vital status of lost patients (i.e., whether they are alive or deceased), of obtaining certified causes of death and ensuring long term mortality follow up is to use central registries. Any data thus obtained can be used to update the results of the relevant trial, so that the overview analyses can be performed more accurately. Data on the incidence of second cancers can also be obtained through the cancer registries and used to monitor long term toxicity.

This study began in 1985, when individual patient data from many randomised trials worldwide was provided to the EBCTCG secretariat for a systematic overview of treatment effects. This first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five-year mortality and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patients. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.

Further cycles of the collaboration were needed to help discover whether these differences persist, to help discover which types of patient are most likely to benefit from such treatments (and which variants of such treatments are most promising), to help assess the various other types of treatment for early breast cancer that have been studied in randomised trials, and to evaluate any long term toxicity of treatments. Therefore, it was intended that the overview data would be updated every five years with data collection taking place in the periods: January 1989 to April 1990, January 1994 to April 1995, January 1999 to April 2000, etc.

The EBCTCG has informed the treatment of women with breast cancer worldwide and will continue to do so into the future. For instance, the results informed the National Institutes of Health (NIH) consensus development conference on the treatment of early breast cancer in 2000. Subsequently, the 2005 report on chemotherapy and endocrine therapy showed the substantial effects on 15-year survival of the chemotherapy regimens and hormonal regimens (such as at least 5 years of tamoxifen in women with oestrogen-positive (ER+) disease) that were tested in the 1980s. The 2005 report on surgery and radiotherapy showed that treatments that substantially improve local control have little effect on breast cancer mortality during the first few years, but definite beneficial effects by 15 years. Updates of these analyses in the 2010s have added further weight to the evidence and further informed national and international guidelines on therapy, including those from the National Institute of Health and Care Excellence (NICE) in the UK. In the current decade, overviews of bisphosphonates, aromatase inhibitors, neoadjuvant chemotherapy and dose-intense chemotherapy, as well as targeted, partial and axillary radiotherapy will provide guidance to healthcare systems globally.

Identifiable patient data have been supplied by NHS Digital to CTSU for an original set of flagged trials. The aggregate number of participants in these trials was about 9,000 in 11 randomised controlled trials; there are currently fewer than 1,385 survivors from 7 of these trials. These trials were undertaken between 1948 and 1987 and since that time the participants have been followed up through their health records. During this period, the CTSU has been acting under the instruction of the trialists; the trials into which these participants were enrolled ceased follow-up and the trialists then gave EBCTCG authorisation to continue to receive follow-up of their outcomes to include in the EBCTCG meta-analyses.

Once received by CTSU, these identifiable data have been kept separate from the EBCTCG database by confining their use to the CTSU Clinical Trials Follow-up Service for EBCTCG (CTSU-ctfs ). The CTSU-ctfs is a separate service provided by CTSU with the purpose of processing personally identifiable information received from NHS-Digital. The CTSU-ctfs carries out the linkage activities between data from NHS-Digital and the original trial data to provide de-identified data to EBCTCG, holding any identifiable information in a separate, secure file store.

The data received from NHS Digital has been combined with an international cohort of women in meta-analyses of randomised controlled trials in the fields of surgery (18,000), radiotherapy versus surgery (51,000), targeted radiotherapy (26,000) and endocrine therapy (8000), this meta-analyses has been carried out by the EBCTCG at the University of Oxford. The wider EBCTCG project includes many more contemporary treatments and targeted therapies, and overall will collect data from more than 725,000 women for around 80 distinct comparisons. Most of the trialists in the EBCTCG collaboration will provide follow-up of individual patient mortality and cancer events directly to the EBCTCG every 5 years until they stop collecting this data. Where registry data is available all possible follow-up information will be sought.

The identifiable fields of live participants held are considered necessary for linkage with NHS data by CTSU. Once the correct link is made between this identifying information and the patient identities held by NHS Digital, these identifiers can be destroyed. No identifiable data is sent to EBCTCG.

The overarching purpose of this Agreement, and previous versions, is to provide outcomes by linkage with NHS Digital data for UK clinical trials so that the EBCTCG can perform individual participant data meta-analyses of the comparative effects of different treatments for women with breast cancer or death from breast cancer. The primary outcome measure of the EBCTCG analyses is time to death from breast cancer. The secondary outcome measures are time to recurrence of breast cancer, time to death from any cause, time to death from non-breast-cancer causes, time to second cancer and type of second cancer.

The sole data controller is the University of Oxford, and all data are processed in the Nuffield Department of Population Health (NDPH) at the University of Oxford. No other organisation processes the data.

The Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) Secretariat at NDPH, University of Oxford receives de-identified data from CTSU-ctfs and analyses them for inclusion in the EBCTCG meta-analyses. The EBCTCG comprises a further 213 organisations from around the world, which contribute data to the overview and are involved with the production of EBCTCG publications. None of these organisations make decisions in how NHS Digital data are processed, and do not have access to NHS Digital data.

The lawful basis for processing the data is GDPR article 6 (1) (e): ‘Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller’. The EBCTCG overview and meta-analyses report on the efficacy and safety of trial therapies and, through dissemination via treatment guidelines to physicians, inform future treatments for individuals with early breast cancer. Therefore processing the data is in the public interest.

The application also falls under Article 9 (2) (j) of EU GDPR, which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. Processing data from the participants of trials in the treatment of breast cancer is in the public interest and is necessary for scientific and research purposes in order to provide secondary analyses of the harms and benefits of breast cancer treatments. This makes full use of the available information for the possible benefit of future breast cancer patients, minimising research waste, while not identifying participants in the trials beyond what is essential for processing.

It is not considered that there is a risk of potential harm to the public from the dissemination of this information. Security measures in place at the University of Oxford ensure data privacy so that the likelihood of a breach of the participant’s privacy is extremely low. This has been confirmed by a data protection impact assessment at the University of Oxford.

Reports detailing the EBCTCG project’s findings will be published, subject to acceptance, in high impact peer-reviewed journals such as The Lancet or Journal of Clinical Oncology. The EBCTCG publications have been cited many thousands of times. The most recent reports published by the collaboration are listed at

https://www.ctsu.ox.ac.uk/research/the-early-breast-cancer-trialists-collaborative-group-ebctcg/ebctcg-publications

Reports of early findings and final results of the analyses will be presented in seminars at the University of Oxford and closed meetings between members of the collaborative group, including the 5-yearly EBCTCG main meeting and the annual EBCTCG Steering Committee meeting, which takes place in September each year. The three current patient members of the EBCTCG steering committee remain active in their roles with the EBCTCG and contribute substantially to steering committee and main meetings, with additional regular meetings to discuss the current analyses. They also disseminate the findings to their respective national and international groups. This level of PPI activity is expected to continue for the life of the EBCTCG project.

Findings from the project will be presented at the San Antonio Breast Cancer Symposium (SABCS) in December each year, and the American Society of Oncology (ASCO) meeting in June each year, and other national and international breast cancer and general cancer symposia.

The EBCTCG and Oxford University web pages will be updated with findings from the EBCTCG project.

All outputs to publications, presentations and web pages will be of aggregate data only with small numbers suppressed in line with HES Analysis Guidance.

Dissemination of project results/outputs is achieved primarily by the EBCTCG in the form of reports and presentations. On submission of a manuscript to a peer-reviewed journal, the funders, Cancer Research UK (CRUK) are informed so they can draft a press release if needed. The readership of EBCTCG reports consists mainly of academic researchers and clinicians, although material is also of interest to members of the public. Among the EBCTCG and its steering committee are leaders in the field of breast cancer who are instrumental in updating and disseminating treatment guidelines for early breast cancer. The EBCTCG Steering committee has three patient representatives who disseminate the findings throughout the international community of breast cancer patients.

All Intellectual Property arising from the analyses and the results of the analyses are owned by the University of Oxford.

In order to increase the impact and accessibility of the findings and in line with NDPH publications policy all EBCTCG results are published with Open Access.

Each research question is updated at approximately 5-year intervals. EBCTCG research questions address the benefits and risks of treating early breast cancer with chemotherapies, surgery, radiotherapy, endocrine therapies and therapies targeted towards molecular pathways and cancer dissemination. Research questions specific to the trials currently flagged by NHS Digital data are:

Axilliary radiotherapy vs not

Accelerated partial breast irradiation (APBI) to whole breast vs tumour bed only

Axillary biopsy with/without axillary radiotherapy

Axillary dissection vs not

Axillary radiotherapy vs axiillary dissection

Chemotherapy with / without tamoxifen

More vs less surgery to axilla

Nodal surgery vs radiotherapy

Ovarian irradiation vs not

Perioperative chemotherapy vs not

Perioperative chemotherapy vs tamoxifen

Radical mastectomy vs modified radical mastectomy

Tamoxifen vs not

Tamoxifen vs ovarian irradiation / oophorectomy

Target dates are not stated for the EBCTCG publications since they are subject to the availability of clinical trial data from many sources. Publication might be delayed while waiting for data in the interests of producing meaningful reports.

New EBCTCG publications are widely reported via press releases, websites, TV and radio.

EBCTCG Publications that use NHS Digital data, currently in progress and in planning, are:

1. A meta-analysis of trials of surgery with or without tamoxifen versus tamoxifen alone in women over the age of 70 will be submitted to the Journal of Clinical Oncology in 2021

2. A meta-analyses on endocrine therapy.

3. A report on ovarian suppression therapy in early breast cancer is in progress

4. A meta-analysis is being finalised of radiotherapy to the regional nodes (internal mammary chain, supraclavicular region and axilla) to obtain as precise an understanding as possible of the potential benefits and harms to evaluate the balance of disease control against late effects of radiotherapy.

5. Meta-analyses are being conducted of different radiotherapy treatments to the axilla, following clinical staging or sentinel lymph node biopsy (SLNB).

6. Data are being collected for a meta-analysis of whole versus partial breast radiotherapy.

Publications which are referenced in the Yielded Benefits below are:

1. Early Breast Cancer Trialists’ Collaborative Group (2014). Effect of radiotherapy after mastectomy and axillary surgery on 10-year recurrence and 20-year breast cancer mortality: meta-analysis of individual patient data for 8135 women in 22 randomised trials.

Lancet. 383:2127-2135. DOI: 10.1016/S0140-6736(14)60488-8.

2. Early Breast Cancer Trialists' Collaborative Group (2015). Aromatase inhibitors versus tamoxifen in early breast cancer: patient-level meta-analysis of the randomised trials. Lancet 386:1341-52. DOI: 10.1016/s0140-6736(15)61074-1

3. Early Breast Cancer Trialists' Collaborative Group (2015). Adjuvant bisphosphonate treatment in early breast cancer: meta-analyses of individual patient data from randomised trials. Lancet 386:1353-61. DOI: 10.1016/S0140-6736(15)60908-4

EBCTCG papers published between January 2008 and May 2017 were primary references in important updates to each of the major international guidelines for the management of early breast cancer. These include those produced by the National Institute for Health and Care Excellence (NICE), the UK’s Royal College of Radiologists (RCR) and the Scottish Intercollegiate Guidelines Network (SIGN); the European Society for Medical Oncology (ESO-ESMO), the Japanese Breast Cancer Society and the St Gallen international consensus on breast cancer; the American National Comprehensive Cancer Network (NCCN), Cancer Care Ontario and the American Society of Clinical Oncology (ASCO), and the American Society for Radiation Oncology (ASTRO). These updated guidelines brought in new recommendations (detailed in the Yielded Benefits) based on EBCTCG evidence, to improve the survival of breast cancer patients.

EBCTCG reports have been published in major medical journals and by September 2020 the 23 published reports (1985-2019) had been cited more than 27,000 times. Twelve of these reports, on ovarian ablation, chemotherapy, endocrine therapy (tamoxifen), surgery and radiotherapy, include data from NHS Digital (reference list).

The 1996 Lancet report on effects of ovarian ablation on recurrence and death showed that 15-year survival was significantly improved in premenopausal women receiving ovarian ablation in surgery or irradiation in trials that began before 1980. The planned update of this report for 2021 will include trials that began between 1987 and 2009.

The 2005 Lancet report on systemic therapies updated reports from 1998 and earlier to show the substantial effects on 15-year survival of the chemotherapy regimens (e.g., About 6 months of anthracycline-based chemotherapy) and endocrine regimens (eg.5 years of tamoxifen) that were being tested in the 1980s.

The 2012 Lancet chemotherapy report updated evidence from 2005, bringing together data from 100,000 women in 123 randomised trials of chemotherapy, showing that chemotherapy can reduce breast cancer mortality not only in ER-negative but also in ER-positive disease, and showing benefits of taxane-based over standard anthracycline chemotherapy.

The 2005 Lancet Report on surgery and radiotherapy updated reports from 1995 and 2000 to show that regimens that substantially reduce 5-year local recurrence rates have little effect on 5-year breast cancer mortality, but moderately reduce 15-year breast cancer mortality.

The 2011 Lancet report on 20,000 women in 20 randomised trials of about 5 years of tamoxifen versus no tamoxifen, showed a highly significant reduction of about a third in breast cancer mortality not only during years 0-4 and 5-9 after starting treatment but also during years 10-14. Tamoxifen is effective whether or not chemotherapy has been given and, importantly, even in weakly ER positive disease.

The long-term nature of the treatment benefits found in the above reports underlines the importance for EBCTCG of obtaining extended follow-up from older trials. For example, surgery and radiotherapy trial follow-up ending at five years would not have detected the 15-year reduction in mortality. Also, if 5-year tamoxifen trial follow-up had ended at 10 years the improvement in survival with 5 years of tamoxifen after 15 years would be missed. The update planned for 2021 of the ovarian suppression meta-analysis, which includes some of the oldest trials flagged by EBCTCG, is likely to provide important information about the long-term effects of this treatment.

The impact of the EBCTCG meta-analyses has been the dissemination of evidence of several moderate treatment effects that together have almost halved breast cancer mortality rates for women aged 35-69 since 1990. In addition, by demonstrating the need for big data in randomised trials the EBCTCG has fostered respect for the value of conducting much larger randomised controlled trials than was customary, in order to assess the effects of treatment properly.

Through conducting large-scale population-level studies, EBCTCG researchers have quantified the long-term benefits and risks of different breast cancer treatments, providing conclusive evidence where there was previously uncertainty. These discoveries have already changed clinical practice, ultimately saving patient lives and improving secondary outcomes. A series of meta-analyses of breast cancer-related clinical trials worldwide generated robust evidence into the effectiveness of treatments. These discoveries have been translated into clinical practice, ultimately benefitting the millions of women who are diagnosed with breast cancer worldwide each year. In particular, these studies identified a more effective treatment (aromatase inhibitors) than the standard endocrine therapy; identified new sub-groups of patients that would benefit from radiotherapy treatment; and demonstrated that bisphosphonates reduce the risk of bone metastasis and death from breast cancer.

The Early Breast Cancer Trialists' Collaborative Group (EBCTCG), based at the University of Oxford has since the 1980s been using individual patient data meta-analysis to assess the long-term benefits and side-effects of different treatment options for early breast cancer, collaborating with breast cancer trialists worldwide to collect long-term outcome data. This puts it in a unique position to assess early and late effects of treatment (either beneficial or harmful), which are not assessed as reliably within the original clinical trial reports. These large datasets allow definitive estimates of treatment effects overall and exploration of any differences in the effects of treatments in different subgroups of women, or between treatments within the same class, analyses which cannot be performed reliably using individual clinical trial publications. The EBCTCG notably provided the first conclusive evidence of the effectiveness of tamoxifen in the mid 1980s and has since produced evidence on other treatment modalities. Notable recent findings from the EBCTCG are:

a) A meta-analysis of 8,000 women in 22 trials of radiotherapy after mastectomy (published 2014) showed that radiotherapy reduced breast cancer recurrence and mortality not only in women whose breast cancer had spread to many lymph nodes but also in those with spread to only 1-3 axillary lymph nodes.

b) A meta-analysis (published 2015) using individual data on almost 32,000 women determined conclusively that aromatase inhibitors (which block oestrogen production) are a more effective treatment for breast cancer than tamoxifen in postmenopausal women. Specifically, aromatase inhibitors were found to reduce recurrence rates by 30% and 10-year mortality rates by 15% compared with tamoxifen.

c) A meta-analysis of 24 trials of the use of bisphosphonate therapy (19,000 women) (published 2015) demonstrated that bisphosphonates reduce the risk of bone metastasis and death from breast cancer in postmenopausal women. Furthermore, bisphosphonates also strengthen bones and effectively reduce damage from osteoporosis caused as a side effect by aromatase inhibitors, therefore adding to their clinical benefit.

In these studies it is important to weigh up benefits and risks which may not occur at the same time: in particular, radiotherapy and treatment with drugs such as anthracyclines can have potential late harms; whereas any benefit of long term treatment with endocrine therapy (such as aromatase inhibitors) is likely to persist may years after surgery. The data from original trial publications may only tell part of the story, and long-term flagging is essential to provide a fully balanced and nuanced exploration of benefits and risks, and their relative timings.

Changes in clinical practice to improve breast cancer treatment

Speaking of the 2014 update to the NICE guidelines on Early and locally advanced breast cancer: diagnosis and management, the Director for Guidelines for NICE wrote:

Widespread adoption of aromatase inhibitors:

Following the publication of a meta-analysis of aromatase inhibitors versus tamoxifen aromatase inhibitors have been adopted as standard-of-care: as a result, starting endocrine therapy with an aromatase inhibitor is now recommended for postmenopausal women with breast cancer whereas previously they were managed with either tamoxifen or an aromatase inhibitor. A study in 2019 found the majority of older women (55+) received aromatase inhibitors [G].

Extension of radiotherapy treatment:

Life-saving radiotherapy treatment is now recommended to a wider range of breast cancer patients, including those whose cancer spreads to 1-3 axillary lymph nodes.

Reducing deaths and adverse side-effects using bisphosphonates

The results of the meta-analysis prompted an immediate policy change, with bisphosphonates now being globally recommended for breast cancer patients, when they were not before. An increase in bisphosphonates use was confirmed by a survey in March 2020 by the charity Breast Cancer Now, finding that 94% of NHS Trusts who have a breast cancer service were routinely prescribing them and a further 3% in process of making them available.

The results presented in the Lancet were recognised by policy makers and used to lobby for improved patient access to bisphosphonates. For instance, during a 2017 debate in the House of Commons on Breast Cancer Drugs, Baroness Blackwood argued in favour of bisphosphonates being licensed as treatment for breast cancer patients, saying ‘research in The Lancet in 2015...found that bisphosphonates can be used to help women who are being treated for early breast cancer after the menopause by reducing the risk of the breast cancer spreading to the bone by 28%’. At least 35,700 postmenopausal women are diagnosed with primary breast cancer in the UK each year. Routine treatment with bisphosphonates prevents 1,180 women from dying from breast cancer annually: equivalent to one in ten breast cancer deaths with an overall net NHS saving of £5.09m per annual cohort of patients.

Ovarian Ablation:

The long-term benefits of ovarian ablation in women under 50 in the absence of other adjuvant treatments were identified, with confirmation that there is no significant benefit in older women.

Radiotherapy and Surgery:

Life-saving radiotherapy treatment is now recommended to a wider range of breast cancer patients, including those whose cancer spreads to 1-3 axillary lymph nodes. Less extensive mastectomy was not found to make a large difference in mortality, at least during the first 10 years following treatment.

Chemotherapy and Hormonal Therapy:

Chemotherapy was demonstrated to give a one-third mortality reduction compared with no chemotherapy, depending on absolute risks. Tamoxifen was found to improve survival and has become a standard therapy in premenopausal women with estrogen receptor-positive early breast cancer. The Chemotherapy / Hormonal therapy and Ovarian Ablation reports also provided evidence for the NICE guidelines on Early and Locally Advanced Breast Cancer update of 2009.

Under previous versions of this Agreement the University of Oxford received identifiable MRIS Members and Posting, Cause of Death, Flagging Current Status and Cohort Event Notification Reports for the purpose of a research project referred to as Early Breast Cancer Trialists Collaborative Group (EBCTCG).

All data received under previous iterations of this Agreement relates to women who have been diagnosed with operable breast cancer (or breast cancer which might become operable through the use of neo-adjuvant therapy) and enrolled in one of seven randomised trials (listed below) comparing treatments for breast cancer, with recurrence or death as a principal outcome.

RANDOMISED TRIALS:

• Ovarian Irradiation Trial, Part I: Entry JUN-1948 to DEC-1950; 189 entered; (J Faculty Radio 10: 175-80 1959)

• Addenbrooke's Stage II: Entry OCT-1958 to MAY-1965 233 entered; (Lancet 2:1086-7 1971)

• Cardiff Trial: Entry SEP-1967 to JUN-1973; 200 entered; (Ann Surg Oncol 6(5):455-60 1999)

• Lister Trial: Entry SEP-1969 to SEP-1976; 534 entered; (Ann R Coll Surg Engl 63(4)239-43 1981)

• Regional Breast Study 1 & 2: Entry MAR-1970 to OCT-1975; 1012 entered;; 279 participants still alive in 1990 flagged in 2001 (Br J Surg 69(12) 693-6 1995)

• King's/Cambridge Radiotherapy (C.R.C. 1): Entry MAY-1970 to JUN-1975; 2800 entered;; a number of UK participants were flagged in 1999 (World J Surg. 1994 Jan-Feb;18(1):117-22)

• Tamoxifen trial: Entry NOV-1976 to JUL-1982; 1005 entered; (Br J Cancer 62(Suppl 12):16 1990)

• CMF/Tamoxifen UK: Entry MAY-1978 to MAR-1984; 118 entered; (Lancet 2 (8412):1148-9 1984)

• C.R.C. Adjuvant Breast Trial (C.R.C. 2): Entry SEP-1980 to APR-1986; 2230 entered;; a number of UK participants were flagged in 1999 (Br J Cancer 57(6):604-7 1988)

• Conservation Trial: Entry OCT-1982 to DEC-1987; 708 entered; (Clin Oncol R Coll Radiol 5(5):278-83 1993)

The original cohort consisted of 9029 individuals, of which 1385 are still alive. Appropriate legal permissions are in place for the University of Oxford to retain data on the remaining 1385, however there is no legal basis for the retention of identifiable data relating to the remaining 7664 now deceased individuals. The University of Oxford will be required to securely destroy all identifying data received from NHS Digital currently held on the 7664 now deceased individuals and provide evidence of this destruction to NHS Digital.

The Early Breast Cancer Trialists' Collaborative Group (EBCTCG) was created in 1985 by researchers at the Clinical Trial Service Unit (CTSU), University of Oxford. The membership of EBCTCG consists of research groups which share their trial data for the purpose of the meta-analyses that assess the benefits and risks of treatments for early breast cancer. The EBCTCG Secretariat are members of staff at CTSU. They identify trials, contact collaborators to request data, process & analyse data and, with input from the collaborators, publish reports of the analyses.

Breast cancer is a common condition and if some widely practicable treatment could be shown to produce even a moderate reduction of 10 to 15% in mortality this could prevent the deaths from breast cancer of thousands of women each year. Hence, it is important to be able to distinguish between a treatment which produces a modest but real effect on mortality and one that has little or no effect. If such differences are to be assessed reliably, both moderate random errors and moderate biases must be avoided. Systematic overviews of all relevant randomised clinical trials can help in both respects, but to avoid introducing any bias it is important to get as complete follow up information as possible on all randomised patients in all such trials. This avoids undue emphasis on just the more promising (or just the less promising) results. Any possible long-term toxic effects of treatment can also be evaluated by looking at cause-specific mortality of patients who die from causes other than breast cancer and by comparing the incidence of second cancers at other sites in treated and untreated patients.

The most reliable way of ascertaining the vital status of lost patients, of obtaining certified causes of death and ensuring long term mortality follow up is to use central registries. Any data thus obtained can be used to update the results of the relevant trial, so that the overview analyses can be performed more accurately. Data on the incidence of second cancers can also be obtained through the cancer registries and used to monitor long term toxicity.

This study began in 1985, when individual patient data from many randomised trials worldwide was provided to the EBCTCG secretariat for a systematic overview of treatment effects. This first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five-year mortality and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patients. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.

Further cycles of the collaboration were needed to help discover whether these differences persist, to help discover which types of patient are most likely to benefit from such treatments (and which variants of such treatments are most promising), to help assess the various other types of treatment for early breast cancer that have been studied in randomised trials, and to evaluate any long term toxicity of treatments. Therefore, it was intended that the overview data would be updated every five years with data collection taking place in the periods: January 1989 to April 1990, January 1994 to April 1995, January 1999 to April 2000, etc.

The EBCTCG has informed the treatment of women with breast cancer worldwide and will continue to do so into the future. For instance, the results informed the National Institutes of Health (NIH) consensus development conference on the treatment of early breast cancer in 2000. Subsequently, the 2005 report on chemotherapy and endocrine therapy showed the substantial effects on 15-year survival of the chemotherapy regimens and hormonal regimens (such as at least 5 years of tamoxifen in women with oestrogen-positive (ER+) disease) that were tested in the 1980s. The 2005 report on surgery and radiotherapy showed that treatments that substantially improve local control have little effect on breast cancer mortality during the first few years, but definite beneficial effects by 15 years. Updates of these analyses in the 2010s have added further weight to the evidence and further informed national and international guidelines on therapy, including those from the National Institute of Health and Care Excellence (NICE) in the UK. In the current decade, overviews of bisphosphonates, aromatase inhibitors, neoadjuvant chemotherapy and dose-intense chemotherapy, as well as targeted, partial and axillary radiotherapy will provide guidance to healthcare systems globally.

Identifiable patient data have been supplied by NHS Digital to CTSU for an original set of flagged trials. The aggregate number of participants in these trials was about 9,000 in 11 randomised controlled trials; there are currently fewer than 1,385 survivors from 7 of these trials. These trials were undertaken between 1948 and 1987 and since that time the participants have been followed up through their health records. During this period, the CTSU has been acting under the instruction of the trialists; the trials into which these participants were enrolled ceased follow-up and the trialists then gave EBCTCG authorisation to continue to receive follow-up of their outcomes to include in the EBCTCG meta-analyses.

Once received by CTSU these identifiable data have been kept separate from the EBCTCG database by confining their use to the CTSU Clinical Trials Follow-up Service for EBCTCG (CTSU-ctfs ). The CTSU-ctfs is a separate service provided by CTSU with the purpose of processing personally identifiable information received from NHS-Digital. The CTSU-ctfs carries out the linkage activities between data from NHS-Digital and the original trial data to provide de-identified data to EBCTCG, holding any identifiable information in a separate, secure file store.

The data received from NHS Digital has been combined with an international cohort of women in meta-analyses of randomised controlled trials in the fields of surgery (18,000), radiotherapy versus surgery (51,000), targeted radiotherapy (26,000) and endocrine therapy (8000), this meta-analyses has been carried out by the EBCTCG at the University of Oxford. The wider EBCTCG project includes many more contemporary treatments and targeted therapies, and overall will collect data from more than 725,000 women for around 80 distinct comparisons. Most of the trialists in the EBCTCG collaboration will provide follow-up of individual patient mortality and cancer events directly to the EBCTCG every 5 years until they stop collecting this data. Where registry data is available all possible follow-up information will be sought.

The overarching purpose of this Agreement, and it previous versions, is to provide outcomes by linkage with NHS Digital data for UK clinical trials so that the EBCTCG can perform individual participant data meta-analyses of the comparative effects of different treatments for women with breast cancer or death from breast cancer. The primary outcome measure of the EBCTCG analyses is time to death from breast cancer. The secondary outcome measures are time to recurrence of breast cancer, time to death from any cause, time to death from non-breast-cancer causes, time to second cancer and type of second cancer.

The sole data controller is the University of Oxford, and all data are processed in the Nuffield Department of Population Health (NDPH) at the University of Oxford. No other organisation processes the data.

The Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) Secretariat at NDPH, University of Oxford receives de-identified data from CTSU-ctfs and analyses them for inclusion in the EBCTCG meta-analyses. The EBCTCG comprises a further 213 organisations from around the world, which contribute data to the overview and are involved with the production of EBCTCG publications. None of these organisations make decisions in how NHS Digital data are processed, and do not have access to NHS Digital data.

The lawful basis for processing the data is GDPR article 6 (1) (e): ‘Processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller’. The EBCTCG overview and meta-analyses report on the efficacy and safety of trial therapies and, through dissemination via treatment guidelines to physicians, inform future treatments for individuals with early breast cancer. Therefore processing the data is in the public interest.

The application also falls under Article 9 (2) (j) of EU GDPR, which details that processing is necessary for scientific and research purposes, subject to appropriate safeguards. Processing data from the participants of trials in the treatment of breast cancer is in the public interest and is necessary for scientific and research purposes in order to provide secondary analyses of the harms and benefits of breast cancer treatments. This makes full use of the available information for the possible benefit of future breast cancer patients, minimising research waste, while not identifying participants in the trials beyond what is essential for processing.

It is not considered that there is a risk of potential harm to the public from the dissemination of this information. Security measures in place at the University of Oxford ensure data privacy so that the likelihood of a breach of the participant’s privacy is extremely low. This has been confirmed by a data protection impact assessment at the University of Oxford.

Reports detailing the EBCTCG project’s findings will be published, subject to acceptance, in high impact peer-reviewed journals such as The Lancet or Journal of Clinical Oncology. The EBCTCG publications have been cited many thousands of times. The most recent reports published by the collaboration are listed at

https://www.ctsu.ox.ac.uk/research/the-early-breast-cancer-trialists-collaborative-group-ebctcg/ebctcg-publications

Reports of early findings and final results of the analyses will be presented in seminars at the University of Oxford and closed meetings between members of the collaborative group, including the 5-yearly EBCTCG main meeting and the annual EBCTCG Steering Committee meeting, which takes place in September each year.

Findings from the project will be presented at the San Antonio Breast Cancer Symposium (SABCS) in December 2021, and the American Society of Oncology (ASCO) meeting in June each year, and other national and international breast cancer and general cancer symposia.

The EBCTCG and Oxford University web pages will be updated with findings from the EBCTCG project.

All outputs to publications, presentations and web pages will be of aggregate data only with small numbers suppressed in line with HES Analysis Guidance.

Dissemination of project results/outputs is achieved primarily by the EBCTCG in the form of reports and presentations. On submission of a manuscript to a peer-reviewed journal, the funders, Cancer Research UK (CRUK) are informed so they can draft a press release if needed. The readership of EBCTCG reports consists mainly of academic researchers and clinicians, although material is also of interest to members of the public. Among the EBCTCG and its steering committee are leaders in the field of breast cancer who are instrumental in updating and disseminating treatment guidelines for early breast cancer. The EBCTCG Steering committee has three patient representatives who disseminate the findings throughout the international community of breast cancer patients.

All Intellectual Property arising from the analyses and the results of the analyses are owned by the University of Oxford.

In order to increase the impact and accessibility of the findings and in line with NDPH publications policy all EBCTCG results are published with Open Access.

Each research question is updated at approximately 5-year intervals. EBCTCG research questions address the benefits and risks of treating early breast cancer with chemotherapies, surgery, radiotherapy, endocrine therapies and therapies targeted towards molecular pathways and cancer dissemination. Research questions specific to the trials currently flagged by NHS Digital data are:

Axilliary radiotherapy vs not

Accelerated partial breast irradiation (APBI) to whole breast vs tumour bed only

Axillary biopsy with/without axillary radiotherapy

Axillary dissection vs not

Axillary radiotherapy vs axiillary dissection

Chemotherapy with / without tamoxifen

More vs less surgery to axilla

Nodal surgery vs radiotherapy

Ovarian irradiation vs not

Perioperative chemotherapy vs not

Perioperative chemotherapy vs tamoxifen

Radical mastectomy vs modified radical mastectomy

Tamoxifen vs not

Tamoxifen vs ovarian irradiation / oophorectomy

Target dates are not stated for the EBCTCG publications since they are subject to the availability of clinical trial data from many sources. Publication might be delayed while waiting for data in the interests of producing meaningful reports.

New EBCTCG publications are widely reported via press releases, websites, TV and radio.

EBCTCG Publications that use NHS Digital data, currently in progress and in planning, are:

1. A meta-analysis of trials of surgery with or without tamoxifen versus tamoxifen alone in women over the age of 70 will be submitted to the Journal of Clinical Oncology in 2021

2. A meta-analyses on endocrine therapy.

3. A report on ovarian suppression therapy in early breast cancer is in progress

4. A meta-analysis is being finalised of radiotherapy to the regional nodes (internal mammary chain, supraclavicular region and axilla) to obtain as precise an understanding as possible of the potential benefits and harms to evaluate the balance of disease control against late effects of radiotherapy.

5. Meta-analyses are being conducted of different radiotherapy treatments to the axilla, following clinical staging or sentinel lymph node biopsy (SLNB).

6. Data are being collected for a meta-analysis of whole versus partial breast radiotherapy.

Publications which are referenced in the Yielded Benefits below are:

1. Early Breast Cancer Trialists’ Collaborative Group (2014). Effect of radiotherapy after mastectomy and axillary surgery on 10-year recurrence and 20-year breast cancer mortality: meta-analysis of individual patient data for 8135 women in 22 randomised trials.

Lancet. 383:2127-2135. DOI: 10.1016/S0140-6736(14)60488-8.

2. Early Breast Cancer Trialists' Collaborative Group (2015). Aromatase inhibitors versus tamoxifen in early breast cancer: patient-level meta-analysis of the randomised trials. Lancet 386:1341-52. DOI: 10.1016/s0140-6736(15)61074-1

3. Early Breast Cancer Trialists' Collaborative Group (2015). Adjuvant bisphosphonate treatment in early breast cancer: meta-analyses of individual patient data from randomised trials. Lancet 386:1353-61. DOI: 10.1016/S0140-6736(15)60908-4

EBCTCG papers published between January 2008 and May 2017 were primary references in important updates to each of the major international guidelines for the management of early breast cancer. These include those produced by the National Institute for Health and Care Excellence (NICE), the UK’s Royal College of Radiologists (RCR) and the Scottish Intercollegiate Guidelines Network (SIGN); the European Society for Medical Oncology (ESO-ESMO), the Japanese Breast Cancer Society and the St Gallen international consensus on breast cancer; the American National Comprehensive Cancer Network (NCCN), Cancer Care Ontario and the American Society of Clinical Oncology (ASCO), and the American Society for Radiation Oncology (ASTRO). These updated guidelines brought in new recommendations (detailed in the Yielded Benefits) based on EBCTCG evidence, to improve the survival of breast cancer patients.

EBCTCG reports have been published in major medical journals and by September 2020 the 23 published reports (1985-2019) had been cited more than 27,000 times. Twelve of these reports, on ovarian ablation, chemotherapy, endocrine therapy (tamoxifen), surgery and radiotherapy, include data from NHS Digital (reference list).

The 1996 Lancet report on effects of ovarian ablation on recurrence and death showed that 15-year survival was significantly improved in premenopausal women receiving ovarian ablation in surgery or irradiation in trials that began before 1980. The planned update of this report for 2021 will include trials that began between 1987 and 2009.

The 2005 Lancet report on systemic therapies updated reports from 1998 and earlier to show the substantial effects on 15-year survival of the chemotherapy regimens (eg. About 6 months of anthracycline-based chemotherapy) and endocrine regimens (eg.5 years of tamoxifen) that were being tested in the 1980s.

The 2012 Lancet chemotherapy report updated evidence from 2005, bringing together data from 100,000 women in 123 randomised trials of chemotherapy, showing that chemotherapy can reduce breast cancer mortality not only in ER-negative but also in ER-positive disease, and showing benefits of taxane-based over standard anthracycline chemotherapy.

The 2005 Lancet Report on surgery and radiotherapy updated reports from 1995 and 2000 to show that regimens that substantially reduce 5-year local recurrence rates have little effect on 5-year breast cancer mortality, but moderately reduce 15-year breast cancer mortality.

The 2011 Lancet report on 20,000 women in 20 randomised trials of about 5 years of tamoxifen versus no tamoxifen, showed a highly significant reduction of about a third in breast cancer mortality not only during years 0-4 and 5-9 after starting treatment but also during years 10-14. Tamoxifen is effective whether or not chemotherapy has been given and, importantly, even in weakly ER positive disease.

The long-term nature of the treatment benefits found in the above reports underlines the importance for EBCTCG of obtaining extended follow-up from older trials. For example, surgery and radiotherapy trial follow-up ending at five years would not have detected the 15-year reduction in mortality. Also, if 5-year tamoxifen trial follow-up had ended at 10 years the improvement in survival with 5 years of tamoxifen after 15 years would be missed. The update planned for 2021 of the ovarian suppression meta-analysis, which includes some of the oldest trials flagged by EBCTCG, is likely to provide important information about the long-term effects of this treatment.

The impact of the EBCTCG meta-analyses has been the dissemination of evidence of several moderate treatment effects that together have almost halved breast cancer mortality rates for women aged 35-69 since 1990. In addition, by demonstrating the need for big data in randomised trials the EBCTCG has fostered respect for the value of conducting much larger randomised controlled trials than was customary, in order to assess the effects of treatment properly.

Through conducting large-scale population-level studies, EBCTCG researchers have quantified the long-term benefits and risks of different breast cancer treatments, providing conclusive evidence where there was previously uncertainty. These discoveries have already changed clinical practice, ultimately saving patient lives and improving secondary outcomes. A series of meta-analyses of breast cancer-related clinical trials worldwide generated robust evidence into the effectiveness of treatments. These discoveries have been translated into clinical practice, ultimately benefitting the millions of women who are diagnosed with breast cancer worldwide each year. In particular, these studies identified a more effective treatment (aromatase inhibitors) than the standard endocrine therapy; identified new sub-groups of patients that would benefit from radiotherapy treatment; and demonstrated that bisphosphonates reduce the risk of bone metastasis and death from breast cancer.

The Early Breast Cancer Trialists' Collaborative Group (EBCTCG), based at the University of Oxford has since the 1980s been using individual patient data meta-analysis to assess the long-term benefits and side-effects of different treatment options for early breast cancer, collaborating with breast cancer trialists worldwide to collect long-term outcome data. This puts it in a unique position to assess early and late effects of treatment (either beneficial or harmful), which are not assessed as reliably within the original clinical trial reports. These large datasets allow definitive estimates of treatment effects overall and exploration of any differences in the effects of treatments in different subgroups of women, or between treatments within the same class, analyses which cannot be performed reliably using individual clinical trial publications. The EBCTCG notably provided the first conclusive evidence of the effectiveness of tamoxifen in the mid 1980s and has since produced evidence on other treatment modalities. Notable recent findings from the EBCTCG are:

a) A meta-analysis of 8,000 women in 22 trials of radiotherapy after mastectomy (published 2014) showed that radiotherapy reduced breast cancer recurrence and mortality not only in women whose breast cancer had spread to many lymph nodes but also in those with spread to only 1-3 axillary lymph nodes.

b) A meta-analysis (published 2015) using individual data on almost 32,000 women determined conclusively that aromatase inhibitors (which block oestrogen production) are a more effective treatment for breast cancer than tamoxifen in postmenopausal women. Specifically, aromatase inhibitors were found to reduce recurrence rates by 30% and 10-year mortality rates by 15% compared with tamoxifen.

c) A meta-analysis of 24 trials of the use of bisphosphonate therapy (19,000 women) (published 2015) demonstrated that bisphosphonates reduce the risk of bone metastasis and death from breast cancer in postmenopausal women. Furthermore, bisphosphonates also strengthen bones and effectively reduce damage from osteoporosis caused as a side effect by aromatase inhibitors, therefore adding to their clinical benefit.

In these studies it is important to weigh up benefits and risks which may not occur at the same time: in particular, radiotherapy and treatment with drugs such as anthracyclines can have potential late harms; whereas any benefit of long term treatment with endocrine therapy (such as aromatase inhibitors) is likely to persist may years after surgery. The data from original trial publications may only tell part of the story, and long-term flagging is essential to provide a fully balanced and nuanced exploration of benefits and risks, and their relative timings.

Changes in clinical practice to improve breast cancer treatment

Speaking of the 2014 update to the NICE guidelines on Early and locally advanced breast cancer: diagnosis and management, the Director for Guidelines for NICE wrote:

Widespread adoption of aromatase inhibitors:

Following the publication of a meta-analysis of aromatase inhibitors versus tamoxifen aromatase inhibitors have been adopted as standard-of-care: as a result, starting endocrine therapy with an aromatase inhibitor is now recommended for postmenopausal women with breast cancer whereas previously they were managed with either tamoxifen or an aromatase inhibitor. A study in 2019 found the majority of older women (55+) received aromatase inhibitors [G].

Extension of radiotherapy treatment:

Life-saving radiotherapy treatment is now recommended to a wider range of breast cancer patients, including those whose cancer spreads to 1-3 axillary lymph nodes.

Reducing deaths and adverse side-effects using bisphosphonates

The results of the meta-analysis prompted an immediate policy change, with bisphosphonates now being globally recommended for breast cancer patients, when they were not before. An increase in bisphosphonates use was confirmed by a survey in March 2020 by the charity Breast Cancer Now, finding that 94% of NHS Trusts who have a breast cancer service were routinely prescribing them and a further 3% in process of making them available.

The results presented in the Lancet were recognised by policy makers and used to lobby for improved patient access to bisphosphonates. For instance, during a 2017 debate in the House of Commons on Breast Cancer Drugs, Baroness Blackwood argued in favour of bisphosphonates being licensed as treatment for breast cancer patients, saying ‘research in The Lancet in 2015...found that bisphosphonates can be used to help women who are being treated for early breast cancer after the menopause by reducing the risk of the breast cancer spreading to the bone by 28%’. At least 35,700 postmenopausal women are diagnosed with primary breast cancer in the UK each year. Routine treatment with bisphosphonates prevents 1,180 women from dying from breast cancer annually: equivalent to one in ten breast cancer deaths with an overall net NHS saving of £5.09m per annual cohort of patients.

Since 13/10/2018 the University of Oxford have been permitted to hold, but not process, the data provided by NHS Digital. For this reason, there are no more recent yielded benefits.

Mortality and Cancer data were supplied to the Clinical Trial Service Unit (CTSU) at the University of Oxford by ONS and subsequently the Health and Social Care Information Centre (which has since become NHS Digital) for the purpose of a research project referred to as ‘Early Breast Cancer Trialists Collaborative Group’.

This Data Sharing Agreement permits the retention of the data for an interim period but no other processing of the data is permitted. For clarity, this restriction on the purpose for processing applies to any data supplied under this Agreement and any versions of the Agreement it supersedes, including copies of that data which have been incorporated into a larger dataset.

Permission to retain the data for the interim period is a practical step to enable the study to comply with the necessary legal and ethical requirements. If, for any reason, it is not possible for the study to meet the necessary requirements, this Agreement will be terminated and destruction of the data will be required.

The following provides background information on the purpose of the original study:

Breast cancer is a common condition and if some widely practicable treatment could be shown to produce even a moderate reduction of 10 to 15% in mortality this could prevent the deaths from breast cancer of thousands of women each year. Hence, it is important to be able to distinguish between a treatment which produces a modest but real effect on mortality and one that has little or no effect. If such differences are to be assessed reliably, both moderate random errors and moderate biases must be avoided. Systematic overviews of all relevant randomised clinical trials can help in both respects; but to avoid introducing any bias, it is important to get as complete follow up information as possible on all randomised patients in all such trials. This avoids undue emphasis on just the more promising (or just the less promising) results.

Any possible long term toxic effects of treatment can also be evaluated by looking at cause specific mortality of patients who die from causes other than breast cancer and by comparing the incident of second cancers at other sites in treated and untreated patients.

Individual patient data from a large number of randomised trials worldwide was provided to the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) secretariat for a systematic overview of treatment effects in 1985. This first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five year mortality, and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patient. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.

Further cycles of the collaboration were needed to help discover whether these differences persist, to help discover which types of patient are most likely to benefit from such treatments (and which variants of such treatments are most promising), to help assess the various other types of treatment for early breast cancer that have been studied in randomised trials, and to evaluate any long term toxicity of treatments. Therefore it was intended that the overview data would be updated every five years with data collection taking place in the periods: January 1989 to April 1990, January 1994 to April 1995, January 1999 to April 2000, etc.

The most reliable way of ascertaining the vital status of lost patients, of obtaining certified causes of death and ensuring long term mortality follow up is to use central registries. any data thus obtained and be used to update the results of the relevant trial, so that the overview analyses can be performed more accurately. Data on the incidence of second cancers can also be obtained through the cancer registries and used to monitor long term toxicity.

No new outputs will be produced under this Data Sharing Agreement.

In any future application, the applicant will be required to provide details of the outputs that were produced and disseminated by the study as well as details of any future outputs planned.

The first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five year mortality, and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patient. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.

Mortality and Cancer data were supplied to the Clinical Trial Service Unit (CTSU) at the University of Oxford by ONS and subsequently the Health and Social Care Information Centre (which has since become NHS Digital) for the purpose of a research project referred to as ‘Early Breast Cancer Trialists Collaborative Group’.

This Data Sharing Agreement permits the retention of the data for an interim period but no other processing of the data is permitted. For clarity, this restriction on the purpose for processing applies to any data supplied under this Agreement and any versions of the Agreement it supersedes, including copies of that data which have been incorporated into a larger dataset.

Permission to retain the data for the interim period is a practical step to enable the study to comply with the necessary legal and ethical requirements. If, for any reason, it is not possible for the study to meet the necessary requirements, this Agreement will be terminated and destruction of the data will be required.

The following provides background information on the purpose of the original study:

Breast cancer is a common condition and if some widely practicable treatment could be shown to produce even a moderate reduction of 10 to 15% in mortality this could prevent the deaths from breast cancer of thousands of women each year. Hence, it is important to be able to distinguish between a treatment which produces a modest but real effect on mortality and one that has little or no effect. If such differences are to be assessed reliably, both moderate random errors and moderate biases must be avoided. Systematic overviews of all relevant randomised clinical trials can help in both respects; but to avoid introducing any bias, it is important to get as complete follow up information as possible on all randomised patients in all such trials. This avoids undue emphasis on just the more promising (or just the less promising) results.

Any possible long term toxic effects of treatment can also be evaluated by looking at cause specific mortality of patients who die from causes other than breast cancer and by comparing the incident of second cancers at other sites in treated and untreated patients.

Individual patient data from a large number of randomised trials worldwide was provided to the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) secretariat for a systematic overview of treatment effects in 1985. This first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five year mortality, and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patient. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.

Further cycles of the collaboration were needed to help discover whether these differences persist, to help discover which types of patient are most likely to benefit from such treatments (and which variants of such treatments are most promising), to help assess the various other types of treatment for early breast cancer that have been studied in randomised trials, and to evaluate any long term toxicity of treatments. Therefore it was intended that the overview data would be updated every five years with data collection taking place in the periods: January 1989 to April 1990, January 1994 to April 1995, January 1999 to April 2000, etc.

The most reliable way of ascertaining the vital status of lost patients, of obtaining certified causes of death and ensuring long term mortality follow up is to use central registries. any data thus obtained an be used to update the results of the relevant trial, so that the overview analyses can be performed more accurately. Data on the incidence of second cancers can also be obtained through the cancer registries and used to monitor long term toxicity.

No new outputs will be produced under this Data Sharing Agreement.

In any future application, the applicant will be required to provide details of the outputs that were produced and disseminated by the study as well as details of any future outputs planned.

The first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five year mortality, and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patient. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.

Mortality and Cancer data were supplied to the Clinical Trial Service Unit (CTSU) at the University of Oxford by ONS and subsequently the Health and Social Care Information Centre (which has since become NHS Digital) for the purpose of a research project referred to as ‘Early Breast Cancer Trialists Collaborative Group’.

This Data Sharing Agreement permits the retention of the data for an interim period but no other processing of the data is permitted. For clarity, this restriction on the purpose for processing applies to any data supplied under this Agreement and any versions of the Agreement it supersedes, including copies of that data which have been incorporated into a larger dataset.

Permission to retain the data for the interim period is a practical step to enable the study to comply with the necessary legal and ethical requirements. If, for any reason, it is not possible for the study to meet the necessary requirements, this Agreement will be terminated and destruction of the data will be required.

The following provides background information on the purpose of the original study:

Breast cancer is a common condition and if some widely practicable treatment could be shown to produce even a moderate reduction of 10 to 15% in mortality this could prevent the deaths from breast cancer of thousands of women each year. Hence, it is important to be able to distinguish between a treatment which produces a modest but real effect on mortality and one that has little or no effect. If such differences are to be assessed reliably, both moderate random errors and moderate biases must be avoided. Systematic overviews of all relevant randomised clinical trials can help in both respects; but to avoid introducing any bias, it is important to get as complete follow up information as possible on all randomised patients in all such trials. This avoids undue emphasis on just the more promising (or just the less promising) results.

Any possible long term toxic effects of treatment can also be evaluated by looking at cause specific mortality of patients who die from causes other than breast cancer and by comparing the incident of second cancers at other sites in treated and untreated patients.

Individual patient data from a large number of randomised trials worldwide was provided to the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) secretariat for a systematic overview of treatment effects in 1985. This first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five year mortality, and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patient. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.

Further cycles of the collaboration were needed to help discover whether these differences persist, to help discover which types of patient are most likely to benefit from such treatments (and which variants of such treatments are most promising), to help assess the various other types of treatment for early breast cancer that have been studied in randomised trials, and to evaluate any long term toxicity of treatments. Therefore it was intended that the overview data would be updated every five years with data collection taking place in the periods: January 1989 to April 1990, January 1994 to April 1995, January 1999 to April 2000, etc.

The most reliable way of ascertaining the vital status of lost patients, of obtaining certified causes of death and ensuring long term mortality follow up is to use central registries. any data thus obtained an be used to update the results of the relevant trial, so that the overview analyses can be performed more accurately. Data on the incidence of second cancers can also be obtained through the cancer registries and used to monitor long term toxicity.

No new outputs will be produced under this Data Sharing Agreement.

In any future application, the applicant will be required to provide details of the outputs that were produced and disseminated by the study as well as details of any future outputs planned.

The first collaboration produced clear evidence of a modest but real effect, in at least some women, of adjuvant hormonal and cytotoxic therapy on five year mortality, and gave statistically stable evaluations of the effects of treatment on recurrence free survival in different types of patient. The results of this first cycle of the overview have already altered routine clinical practice in the UK and elsewhere.