Mortality impact, risks, and benefits of general population screening for ovarian cancer: the UKCTOCS randomised controlled trial

Article history

The contractual start date for this research was in January 2017. This article began editorial review in January 2022 and was accepted for publication in August 2022. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The Health Technology Assessment editors and publisher have tried to ensure the accuracy of the authors’ article and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this article.

Copyright statement

Copyright © 2023 Menon et al. This work was produced by Menon et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – Journals Library, and the DOI of the publication must be cited.

2023 Menon et al.

Setting

The trial was set up within 27 participating primary care trusts (PCTs) (including local health boards in Wales) in the catchment area of 13 NHS Trusts in England, Wales and Northern Ireland between 2001 and 2005. Pre-identified Scottish centres were unable to participate due to a variety of logistical reasons (lack of space, retirement of potential trial-centre leads, unwillingness of NHS Trust management to commit to a 10-year trial, involvement in other ovarian cancer screening trials). The co-ordinating centre (CC) for the main trial was located at Barts and the London Medical School, Queen Mary University London, between January 2001 and March 2004 and moved to University College London (UCL) in April 2004. In UCL, from April 2004 to July 2018, the CC was based in the Department of Women’s Cancer, Elizabeth Garrett Anderson (EGA) UCL Institute for Women’s Health. It then relocated (August 2018 to January 2021) to the Medical Research Council Clinical Trials Unit (MRC CTU at UCL) at the Institute of Clinical Trials & Methodology, where all the trial data are archived. The bioresource created during the trial, UKCTOCS Longitudinal Women’s Cohort (UKLWC), is hosted and managed by the MRC CTU at UCL27 with biological samples located at the NIHR National Biosample Centre (UK Biocentre). 28 The co-ordinating centre for the psychosocial study was located at Sussex Health Outcomes Research and Education in Cancer (SHORE-C), Brighton and Sussex Medical School, University of Sussex.

Participants

Between April 2001 and October 2005, women aged 50–74 were randomly invited from age-sex registers of the 27 participating PCTs. Women were eligible if they were aged between 50 and 74 and were postmenopausal (defined as either >12 months amenorrhoea following a natural menopause or hysterectomy, or >12 months of hormone replacement therapy commenced for menopausal symptoms). Exclusion criteria were bilateral oophorectomy, previous ovarian malignancy, increased risk of familial ovarian cancer, active non-ovarian malignancy, and participation in other ovarian cancer screening trials (Project documentation A).

Approval was obtained from the Caldicott guardians (data controllers) of the 27 PCTs to access contact details of eligible women aged 50–74 resident in their jurisdiction (Figure 1). Specialised software was commissioned from the NHS which allowed 2000 to 10,000 women aged 50–74 to be randomly selected on a regular (usually 3-monthly) basis from the age/sex registers held by the participating PCTs. Their contact details, date of birth, NHS number and general practitioner (GP) details were forwarded as an electronic file to the CC. The software ensured that the women selected were flagged on the PCT register so that their details were not included in future downloads. These files were uploaded into a bespoke online trial management system (TMS) specifically developed for UKCTOCS. Women were then sent personal invitations (Project documentation B) to join the trial along with a brochure outlining the objectives, design and inclusion/exclusion criteria (Project documentation C). Some PCTs were unwilling to transfer contact details. For them, we negotiated an alternative method – the PCTs were sent standard invitation letters without a recipient’s name in sealed franked envelopes which they forwarded after pasting address labels. All women who wished to participate returned a tear-away slip in a Freepost envelope to the CC. If they fulfilled eligibility criteria, they were sent a detailed trial information sheet (Project documentation D) and an appointment to attend for recruitment at their local regional centre (RC). A dedicated telephone reception at the CC ensured that women were able to reschedule appointments easily.

FIGURE 1.

Invitation, recruitment and randomisation. (a) Recruitment in which primary care trusts (PCTs) provided electronic details of women; (b) recruitment in which PCTs did not allow access to contact details; TMS, trial management system; CC, co-ordinating centre.

At the recruitment appointment, women viewed an information video and participated in a group discussion. They completed an 18-item recruitment questionnaire (Project documentation E) which included information regarding eligibility. They then had an individual appointment with the study nurse, who went through the patient information and ensured that their questions were answered prior to signing consent. The nurse also checked through their completed recruitment questionnaire to ensure that the information was complete and accurate. 29 Women who were willing to participate in the psychosocial study were given a set of baseline quality-of-life questionnaires to complete and return to the researchers in a pre-paid envelope. 30 All volunteers provided a baseline serum sample.

The GPs of individual women were kept informed at all stages of recruitment and randomisation of their patients into the trial.

Consent

Separate written consent was obtained for the main trial, serum bank (Project documentation F) and the psychosocial study (Project documentation G). Each individual woman was also asked to sign a data protection form (Project documentation H). A copy of the consent form was sent to the CC, where it was checked and any exceptions to future use of data or samples noted on the TMS.

Processing and biobanking of blood samples

Blood was collected in two 8-ml gel separation serum tubes (VACUETTE^®; Greiner Bio-One, Stonehouse, UK) at the trial centres at recruitment. Additionally, in women randomised to the multimodal group, annual and repeat blood samples as per protocol were collected between 2001 and 2011. The barcoded blood tubes were scanned into the data management system at venepuncture and transported daily at ambient temperature to the central laboratory (Tumour Marker Laboratory) at the CC. Samples were scanned when they were received at the central laboratory, centrifuged at 1500×g for 10 minutes. Exact times for interval to spin (median 24 hours) were recorded for each sample. If it exceeded 56 hours, the sample was discarded, and the TMS sent women an appointment for a repeat blood draw. The separated serum was aliquoted into 10 500-μl straws (company – IMV) using a semi-automated MAPI platform (IMV). The straws were stored in liquid-nitrogen tanks at the central laboratory for 2001–2004. The tanks once full were transferred to a HTA licensed commercial cryofacility (Fisher Bioservices, UK) from 2004 onwards. The biobank was relocated to the UK Biocentre in March 2018.

Confirmation of eligibility

The completed questionnaires were sent to the CC, where they were scanned electronically using computerised intelligent character and optical mark reading software (Cardiff Software Inc, Teleform Elite version 8.1.1) which allowed rapid and accurate data entry. Any inconsistency or information not recognised by the data-capture software was verified manually by trained data-entry staff, who validated the computer-interpreted data. The TMS checked for completeness of eligibility data. When data were incomplete, women were placed ‘on hold’ and letters requesting further information were automatically sent. If the volunteer was placed ‘on hold’ because 12 months had not elapsed since her last menstrual period or start of HRT, she was informed and included in the randomisation process when 12 months had elapsed from the relevant date. The TMS generated lists of women who had a family history suggestive of ‘increased risk of ovarian cancer’. Such women were individually contacted and eligibility was confirmed. If they fulfilled criteria which put them at ‘increased risk’ of familial ovarian cancer, their GP was sent a letter requesting that the individual be referred to the Clinical Genetics department for risk assessment.

Randomisation and blinding

Following confirmation of eligibility, the TMS randomised women to no screening (control or C group) or to annual screening with serum CA125 (Multimodal or MMS group) or TVS (Ultrasound or USS group) in a 2:1:1 ratio using a computer-generated random-number algorithm. Randomisation was carried out by the TMS allocating a set of 32 random numbers to each RC with the lowest 8 allocated to the MMS group, the next 8 to the USS group and the remaining 16 to the control group. Each successive volunteer within the RC was randomly allocated one of the random numbers and subsequently randomised into a group. When all 32 random numbers had been used up a further set of 32 was generated. The randomisation was accomplished by using the Visual Basic randomisation statement and the SQL Rnd function.

Blinding of participants and trial staff was not possible given the nature of the interventions. However, the members of the outcome review committee were blinded to randomisation group.

The trial design, including details of recruitment, randomisation and screening, is summarised below and has been detailed in the protocol (Project documentation A).

Intervention – screening, clinical assessment, trial surgery

The women were invited for annual screening from randomisation (2001–2005) to 31 December 2011.

Follow-up

All participants were linked using their UK NHS number (surname and date of birth used where NHS number was incorrect) initially to the Office of National Statistics and then to NHS Digital (formerly Health and Social Care Information Centre, HSCIC) in England and Wales and Business Services Organisation (BSO) (formerly Central Services Agency) and Northern Ireland Cancer Registry in Northern Ireland for data on deaths and cancer registrations. In addition, women residing in England and Wales were also linked to the hospital administrative databases – Hospital Episodes Statistics (HES) and Patient Episode Database for Wales (PEDW) respectively. Data on cancers for 2001–2010 were also received from the National Disease Registration Service (NCRAS, formerly NCIN) for English participants. The dates when data were last received from the various agencies are detailed in Table 1.

In addition, all participants were sent two postal questionnaires (Project documentation I and J); the first 3–5 years post randomisation and the second in April 2014. A third questionnaire (Project documentation K) was sent to a subset of participants in June 2020 who had either exited the national registries or for whom it was not possible from HES data to ascertain if both ovaries had been removed. We also received ad hoc direct communication from trial participants, their families and clinicians.

Cancer site and cause of death review

Nineteen International Classification of Diseases (ICD-10) codes (Table 2) were identified at start of trial to require independent review to confirm or rule out a diagnosis of ovarian or tubal cancer. Throughout the trial, all the data sources were interrogated to identify women reported to be diagnosed with any of these 19 codes post randomisation. 33 Copies of medical notes were retrieved from hospitals where the women were treated. 33,34 The only exception was women with malignant neoplasm of uncertain origin (ICD-10 C80) who also had another non-ovarian cancer registration. Notes, with any reference to randomisation group redacted, were reviewed by the outcomes review (OR) committee consisting of gynaecological pathologists and oncologists. The OR committee assigned cancer site (using a previously audited pre-specified algorithm; Project documentation L),33 International Federation of Gynaecology and Obstetrics (FIGO) 2014 stage, grade, histotype, type of ovarian cancer and cause of death (Project documentation M). Ovarian and tubal cancer was defined using the revised WHO 2014 classification. 35,36 Death due to ovarian and tubal cancer was based on disease progression as evidenced by appearance of new lesions or increase in size of previously documented lesions with imaging, clinical worsening, or rising biomarker concentrations.

Additional data for secondary outcomes

For analyses related to ovarian and tubal cancer incidence and compliance with screening, it was essential to identify women who had bilateral (or unilateral if the other adnexa had been previously removed) salpingo-oophorectomy outside the trial. We did this by identifying all women who post randomisation had undergone any ovarian/adnexal surgery outside the trial by searching for the relevant procedure codes in the HES records of women resident in England and self-reported text in the questionnaires. Where possible and particularly if this was based on self-reporting, the procedure was confirmed by contacting the relevant hospital for copies of surgery and histology notes.

To collect data on any ovarian cancer screening undergone by the women in the C group (contamination), a series of questions related to screening for ovarian cancer outside UKCTOCS and reasons (symptoms, screening, at the woman’s own request or for other reason) for any CA125 or ovarian ultrasound performed after trial recruitment were included in the second postal follow-up questionnaire (Project documentation J) sent in 2014, three years after end of screening.

To ensure risks related to screening were properly recorded, RC teams were asked to report all complications that resulted from screening as well as trial surgery resulting in benign pathology findings. In addition, they were asked to report potential suspected unexpected serious adverse reactions (SUSAR) to a designated safety officer at the CC. The safety officer also reviewed operative and other relevant pages of the hospital notes of all women who underwent trial surgery and had false-positive benign pathology for any unreported complications. A designated senior gynaecological oncologist blinded to the randomisation group reviewed the medical notes and confirmed all surgical complications. This individual classified the complications as major or minor using a standard review form.

To assess whether the trial resulted in a cost-effective reduction in ovarian and tubal cancer deaths in the screen groups:

1. For each trial participant in the MMS and USS groups, all NHS resource usage relative to screening was captured. For those referred for clinical assessment, data related to hospital clinic visits, additional imaging, blood tests and trial surgery were captured through review of medical notes. The majority of the participants had treatment in the NHS, with a small minority treated privately, with the private health-care resource usage also captured.

2. Between February 2017 and March 2020 changes in quality of life were captured through the NICE preferred quality-of-life instrument, the EuroQoL EQ-5D-5L37 instrument (www.nice.org.uk/process/pmg9/chapter/foreword), and the disease-specific FACT-O questionnaires. These were sent to all the women diagnosed with ovarian and tubal cancer whose diagnosis was confirmed by the OR Committee, after ensuring that they were still alive. Those for whom the current address was not known to the team or who had previously requested ‘no further contact’ were excluded. The EuroQol instrument included EQ-5D-5L (five-question instrument using five domains) as well as the EQ-5D-5L Visual Analogue Scale (a numerical representation of QoL).

3. Additional EQ-5D-3L data on ovarian cancer patients – In view of the limited number of questionnaires received by October 2019, a decision was made to supplement the UKCTOCS QoL data with those collected in the course of the International Collaboration on Ovarian Neoplasms (ICON8) trial. The ICON8 trial had a population with considerable overlap with the UKCTOCS population with respect to treatment for ovarian cancer. The ICON8 EQ-5D-3L data on participants recruited from England to the standard-care 3-weekly carboplatin and paclitaxel arm of ICON8 were therefore requested as they map to the treatment the women in UKCTOCS received. Having these additional data from women with all Stages (I–IV) from the ICON8 cohort (collected at baseline, 9 months, 18 months, 5 years) would have enabled greater accuracy in the adjustment for overall quality of life and subsequent quality-of-life tariffs used in the QALY calculation for our cost-effectiveness/utility analysis.

Psychosocial study

All consented women completed psychosocial baseline questionnaires at randomisation. These include an examination of their knowledge, beliefs and attitudes to ovarian cancer screening and standardised questionnaires including the General Health Questionnaire 12 (GHQ-12), Speilberger Trait Anxiety Inventory (STAI) and Fallowfield’s Sexual Activity Questionnaire (SAQ). Additionally, as detailed in Figure 4, a longitudinal study was undertaken in two groups:

1. Random group About 1300 women were randomly chosen from the C, MMS and USS groups. Baseline data were entered into a database and all women received annual questionnaires plus additional assessments if results from annual screens indicated a need for repeat testing. The women from the C group, who did not receive any form of screening, received annual questionnaires only.

2. Events group This consisted of all participants in the MMS and USS groups who were recalled for repeat tests following their annual screen or underwent trial surgery. They were sent the GHQ-12, STAI and SAQ questionnaires for completion and return within 2 weeks of the repeat tests in the screening episode. If they were diagnosed with ovarian cancer, they were followed longitudinally at 6 months post treatment start. 30

FIGURE 4.

Design of psychosocial study.

Outcomes

All outcome data were kept confidential until unblinding. For most outcomes, subgroup analysis was undertaken for invasive epithelial ovarian and tubal cancer.

Changes to trial design

The trial spanned 20 years (April 2001–June 2021). During this period, there were a number of changes, with the most significant ones related to external growing evidence on the origins and natural history of ovarian cancer and on cancer screening. All are detailed below.

Stopping guidelines and interim analysis

There were stopping guidelines for safety specified in terms of excess morbidity resulting from screening or ‘false positive’ trial surgery that resulted in a diagnosis of benign pathology. This was monitored annually by the independent DMEC, as were the operating characteristics of each screening strategy. We undertook an interim analysis of the primary outcome when approximately half the number of initially planned ovarian and tubal cancer deaths had occurred in C group. There were no stopping guidelines for futility and the critical significance level guideline for a stopping for benefit for either screening strategy was very small (α = 0.001).

Sample size and power calculation

In 2000, the sample size was set at 200,000 women to be randomised 2:1:1 (C, MMS, USS groups), with six annual screens and follow-up of seven years. A 4% annual attrition rate was assumed. This was estimated to provide 80% power at a two-sided 5% significance level for a difference in ovarian and tubal cancer mortality of 25% (and over 90% power for a difference of 30%) between the combined screening groups and the C group and 70% power to detect a reduction of 30% in MMS versus C or USS versus C comparisons.

Power was recalculated following the extension of trial duration to 31 December 2014. It was estimated to be 80% at a two-sided 5% significance level to detect a reduction of 30% in MMS versus C or USS versus C comparisons.

At initial censorship on 31 December 2014, there were 358 ovarian and tubal cancer deaths in the C group. Compared to the C group, the ‘average’ estimated relative reduction in deaths was 11% (Cox model p = 0.240) MMS and 9% (Cox model p = 0.32) USS. Any mortality reduction was only apparent about seven years after randomisation. 45% (162/358) of the deaths in the C group during 2001–2014 had occurred before seven years. In 2015, for the no screening versus MMS or USS comparisons, we estimated that an additional 233 C group events would give 80% power at a two-sided 5% significance level for a difference in relative mortality of 25% during long-term (2015–2020) follow-up conditional on the observed mortality reduction of 11%. This translated to a target sample size of 591 overall events in the C group. All 233 new and 73% (431/591) of total C group events would occur beyond seven years.

Statistical analysis

Uptake and recruitment rates were described using descriptive statistics. A Consolidated Standards of Reporting Trials (CONSORT) diagram was constructed for the analysis of primary outcome (Figure 5). Descriptive statistics by randomisation group were calculated for baseline characteristics of all eligible participants. The trial consisted of three phases:

FIGURE 5.

CONSORT diagram. a, Events occurred before recruitment, but were discovered after randomisation.

1. Recruitment and randomisation phase between 2001 and 2005.

2. Screening phase, which ran concurrent with recruitment and continued till 2012, one year after the end of screening in 2011. For individual trial participants, it was defined as the period from randomisation to the anniversary of the randomisation date in 2012.

3. Follow-up with no screening from individual randomisation dates in 2012 to 30 June 2020 (final censorship date).

All ovarian and tubal cancers confirmed by the OR committee were classified by randomisation group as appropriate into:

1. screen detected (screen positives)

2. those not detected by screening (clinically diagnosed); the latter consisted of:

   1. screen negatives diagnosed within one year of last test in a screen episode (the false negatives used for calculating performance characteristics of screening strategies)

   2. screen negatives diagnosed more than one year after last test of screen episode during the screening phase

   3. diagnosed greater than one year after screening phase had ended

   4. diagnosed in women who never attended screening.

Details of stage, histotype, morphology and type of ovarian and tubal cancers were tabulated as were deaths due to the disease.

Baseline characteristics of eligible women by group and overall

The median age of participants at recruitment was 60.6 years. The majority were white, parous with two children, and more than half had used the oral contraceptive pill. About 40% of the oldest birth cohort (1925–9) reported no formal educational qualification, in contrast to 19% of women in the youngest birth cohort (1950–5). 56 Six per cent had a personal history of cancer and 1.6% reported maternal history of ovarian cancer. The baseline characteristics were balanced between the study groups (Table 3).

Follow-up information was available until death or censorship (30 June 2020) in 192,478 (95.0%) women (48,110 [95.0%] in the MMS group, 48,022 [94.9%] in the USS group and 96,276 [95.0%] in the C group). Overall, this amounted to 3.16 million women-years. Median follow-up of a participant was 16.3 years (IQR 15.1–17.3).

At censorship (30 June 2020), the OR committee reviewed 4482 women identified to have one of the 19 prespecified ICD-10 codes for possible ovarian or tubal cancer. Of them, 2055 (45.9%) women were confirmed to have ovarian or tubal cancer. The numbers diagnosed by years from randomisation are detailed in Table 4. This translated to 522 (11.6%) in the MMS group, 517 (11.5%) cancers in the USS group and 1016 (22.7%) in the C group.

Overall, 1805 (87.8%) of 2055 women had invasive epithelial ovarian or tubal cancers: 452 (0.9%) in the MMS group, 445 (0.9%) in the USS group and 905 (0.9%) in the no-screening group (Table 5).

Complications related to screening tests

Less than 1% of women reported screening complications – 30 (<1%) in the MMS group and 61 (<1%) in the USS group. The screening-related complication rate was 8.6 per 100,000 in the MMS group and 18.6 per 100,000 in the USS group. The most common complication reported in the MMS group was bruising (13 of the 30 women), and pain (20/61) and cystitis (11/61) in the USS group (Table 14).

Among women who were screen-positive and had trial surgery, 488 (1.0% of all randomised to the group) in the MMS group and 1634 (3.2% of those randomised to the group) in the USS group were found to have benign adnexal pathology or normal adnexa. In these women, there was a complication rate of 3.1% (95% CI 1.7 to 5.0; 15 of 488) in the MMS group and 3.5% (95% CI 2.7 to 4.5; 57 of 1634) in the USS group. 38

Principal findings

UKCTOCS, the largest ovarian cancer screening trial to date, did not find any evidence of a reduction in deaths from ovarian and tubal cancer compared to no screening at a median follow-up of 16.3 years from randomisation, with either the multimodal or the ultrasound screening strategy investigated in the trial. The suggestion following the 2014 censorship of a trend to delayed mortality reduction38 was not borne out. This was despite a stage shift with an increased incidence of Stage I (47%) and decreased incidence of Stage IV disease (24.5%) in the MMS compared to the C group that persisted even at nine and a half years after the end of screening. However, it must be noted that the overall decrease in incidence of advanced (Stage III and IV) disease was only 10%. There was no difference in stage between the USS group and the C group. With both strategies, there was no increase in incidence of ovarian and tubal cancer, suggesting that overdiagnosis was not a major issue. The trial findings reinforce the current guidance that women in the general population should not be offered ovarian cancer screening outside trials.

The UKCTOCS result should not be directly extrapolated to ovarian and tubal cancer screening of women at increased risk as they were specifically excluded from UKCTOCS. Previous single-arm screening studies in this population exploring three or four monthly screens using a MMS strategy alongside encouraging risk-reducing surgery found a reduction in number of women diagnosed with advanced disease. 27 One also needs to consider that women with BRCA gene mutations respond better to current treatments compared to the general population.

The UKCTOCS participants were predominantly white. Given that there is evidence of ethnic differences in histotype, with an increased incidence of clear-cell cancer in Asian populations, the mortality impact of population screening may be different in these populations. However, it is important to note that the major contributor to mortality in all populations is high-grade serous ovarian cancer. A future publication is in draft where we explore in detail the stage shift and mortality impact by histotype.

In the MMS group, earlier diagnosis of invasive epithelial ovarian and tubal cancer in a proportion of mainly asymptomatic women did not translate into lives saved. This suggests that it is likely that earlier detection in the symptomatic population may not translate to reduced mortality. However, this should not be considered as a reason to abandon the significant symptom-awareness efforts that are in place. As detailed above, there has been significant progress in the treatment of advanced ovarian cancer since 2011 when screening ended. These advances together with earlier diagnosis could contribute to better quality of life and improved outcomes. In addition, once symptomatic, a rapid diagnosis is invaluable to women and their families, as is avoiding emergency presentation to secondary care.

The detailed implications are discussed below using the UK National Screening Committee (NSC) criteria. 57

The condition

Ovarian cancer remains an important health problem because of the poor outcome (10-year survival of 35%), which has not changed dramatically over the past four decades. At the start of the trial, there was little understanding of the natural history of the disease. For many decades, the prevailing thought was that majority of the ovarian cancers arose from the surface epithelium of the ovary. It has only become widely accepted in the last decade (after screening had ended in UKCTOCS) that the majority of high-grade serous ovarian cancers arise from serous tubal intraepithelial carcinoma (STIC) lesions at the fimbrial end of the fallopian tube and spread when the adnexal lesion is very small. 58 Work is still under way to understand how the cancer develops from this latent stage to clinical presentation.

There are no primary prevention strategies available on a population basis. However, based on the above understanding there is increasing adoption of removal of tubes during hysterectomy for benign disease58,59 and as a family planning measure (in place of tubal ligation). 60 In women at increased risk of ovarian cancer, removal of tubes and ovaries is recommended after completion of family to reduce the risk of developing ovarian cancer. In this population, trials (PROTECTOR UK,61,62 TUBA Netherlands,63 Fimbriectomy USA64) are under way to explore the possibility of risk-reducing salpingectomy with delayed oophorectomy to mitigate the effect of early menopause.

The test

In the trial, two screening strategies were evaluated.

MMS strategy: this involved an annual CA125 blood test. Venepuncture is widely used in the clinical setting. Its safety in UKCTOCS is highlighted by the fact that complications were reported in only 1 per 10,000 annual screens. Serum CA125 has been used in the clinical setting since its discovery in 1981 for diagnosing and monitoring ovarian cancer. It is routinely performed in NHS laboratories, on multiple platforms with an established national quality assurance programme (NEQAS). 65 In UKCTOCS, all CA125 measurements were done centrally in a single laboratory that was enrolled in the NEQAS programme. The UKCTOCS laboratory at UCL was accredited for clinical work by the United Kingdom Accreditation Service (UKAS,66 formerly known as Clinical Pathology Accreditation Ltd).

The CA125 results were automatically uploaded from the Laboratory Information Management System (LIMS) into the UKCTOCS Trial Management System. Instead of using a cut-off, as in the clinical setting, in UKCTOCS serial CA125 was interpreted using a bespoke risk of ovarian cancer algorithm (ROCA) to triage women based on risk. This had already been piloted in a prospective trial of over 13,000 women with three rounds of annual screening to define risk cut-off levels for triage. 26 While the detailed MMS screening protocol was complex, it was built into the Trial Management System with automated classification of risk, follow-up actions and appointment scheduling. As a result, despite performing over 345,000 annual MMS screens over 11 years there were no instances of misclassification or errors in triaging, which suggests that it should be possible to run such a complex intervention as part of a screening programme. The only resource-intensive aspects were manual checking of all letters to women with elevated risk and regular monitoring of multiple steps of the blood test to results process.

USS strategy: women underwent a transvaginal ultrasound annually. The TVS was similar to that routinely performed in the NHS. However, it differed in that the majority of women had normal tubes and ovaries with no adnexal pathology. In addition, the subjective element of ultrasound scanning and age-related atrophy of adnexal structures made their visualisation more difficult. To address this, a training and accreditation programme in ultrasound screening of postmenopausal ovaries was developed for trial ultrasonographers by the Ultrasound Sub-Committee. 32 Alongside this, a quality assurance programme involving regular monitoring of individual ultrasonographers was developed and implemented with day-to-day oversight by a senior trial lead ultrasonographer. Despite this, a retrospective analysis by a single expert of static images of normal ovaries suggested a significant discrepancy between sonographer and expert. 67

Both screening procedures, venepuncture and TVS were only associated with minor complications with low complication rates. The psychosocial study found that screening generally did not increase anxiety. However, attending more intense screening for elevated risk (Level II) was associated with some general psychological morbidity (a mixture of depressive thoughts and worry, reflected by change in GHQ-12 scores) compared with repeat CA125 or TVS for intermediate risk. A meta-analysis of 12 randomised trials of screening (cancer and other diseases) that used similar measures to evaluate anxiety and psychological morbidity found comparable findings. In those receiving positive test results compared with those receiving negative results, there was no evidence of short (<1 month) or longer-term (beyond 1 month) anxiety. 68

UKCTOCS involved 7–11 annual screens with half the women undergoing >8 annual screens. Compliance was high, with almost 80% of women attending eligible screening appointments. The only other trials with six or more years of screening are the UK Age trial69 (seven annual screens) and the PLCO ovarian70 (six annual screens) and prostate cancer71 (six annual screens) screening trials. In the UK Age trial, participation in screening was 68.1% at the first screen and 69.1% subsequently. 69 During the initial four rounds of screening in the ovarian arm of PLCO, the compliance for combined screening with both CA125 and TVS decreased slightly from the 83% at the first to 78% at the fourth screen. When analysed separately, for CA125 alone, compliance in PLCO was 84% and 79%, and for TVS alone it was 83% and 78% for the first and fourth screen, respectively. 70 Equivalent data from UKCTOCS for MMS were 97% and 83% and for USS 98% and 84% for the first and fourth screen, respectively. In the prostate arm of PLCO, the compliance with routine PSA testing was much lower at 33% in year 1, increasing to 46% in year 5. 71 In UKCTOCS, despite the length of the trial, adherence with annual screening remained high, with the observed versus predicted compliance ratios 0.99 for the MMS and 0.98 for the USS group for the eleventh annual screen.

There was a clear diagnostic pathway for women found to have positive results on screening. They were assessed by the trial clinician in dedicated trial clinics (see methods). Options included conservative management if risk was equivocal or referral for trial surgery. All decisions were made in discussion with participants. Where appropriate, the women were referred to the NHS gynaecological oncology team using the two-week urgent cancer referral pathway. The diagnostic work-up undertaken in the trial is now part of the 2011 NICE 122 Guidance Ovarian cancer: recognition and initial management. 72

The sensitivity of the MMS strategy, following over 340,000 women-years of screening, was similar for detection of all (84%) and the subgroup of invasive epithelial (86%) ovarian and tubal cancers diagnosed within one year of the last test in a screening episode. Of the screen-detected invasive epithelial cancers, 38% were early Stage (I–II). Sensitivity of over 80% for ovarian cancer (83%; 6 of 7 detected) has been reported in the recent feasibility study, DETECT-A (Detecting cancers Earlier Through Elective mutation-based blood Collection and Testing) of 10,000 women aged 65–75 using a multimodal multi-cancer strategy incorporating the early version of the CancerSEEK blood test as first-line and diagnostic PET-CT as the second-line test. Of note, only 16% (1 of 6) of the screen-detected ovarian cancers in DETECT-A were early Stage (I–II), which suggests that significant improvements in test performance are necessary before the test can be used for ovarian cancer screening. 73 Encouraging performance has been reported in clinical sample sets using the targeted methylation of plasma cell-free DNA (cfDNA) test74 with the multi-cancer early detection (MCED) test in the Circulating Cell-free Genome Atlas study and using a micro-RNA signature in the OCaMIR test. 75 It is noteworthy that at the annual screen, half the women with screen-detected invasive epithelial ovarian and tubal cancer had CA125 below the normal cut-off (35 IU/ml)26 which was used in PLCO21 and SCSOCS76 trials. The results provide evidence of the significant improvement that longitudinal biomarker algorithms bring to early detection. This was further confirmed by use of the trial CA125 data to build and compare the performance of the available longitudinal algorithms. 77

For each woman detected with ovarian or tubal cancer, 2.5 additional women underwent surgery. This translated to 14 unnecessary (benign pathology or normal adnexa) false-positive operations per 10,000 screens in the MMS group. This was considerably fewer than the additional women who underwent surgery per cancer detected in the PLCO (18.5)70 and the SCSOCS (32)76 RCTs. However, they are still significant. While women participating in the trial consented to even higher false-positive rates (10 operations per cancer detected), it needs to be noted that they represented only 16.3% of the invited population. Additionally, the burden on the NHS would be considerable. Based on the projected mid-2020 UK population of 10.23 million women aged 50–75 and an 80% uptake of MMS annual screening, this would mean 11,455 false-positive operations per year. Surgical morbidity in these women is an additional major concern, especially with increasing comorbidity with age. The serious-complications rate (3.1%) in women who had benign pathology or normal adnexa, most of whom underwent laparoscopic bilateral salpingo-oophorectomy, was again significantly lower than that reported in the PLCO21 trial (15%). Rates were similar to that reported in a contemporary UK78 series of risk-reducing salpingo-oophorectomy (3.9%) in high-risk women.

The longitudinal algorithm used to interpret CA125 levels was innovative and forward-thinking. Such algorithms are now being explored for early-detection strategies in other cancers and are being improved through use of machine learning and Artificial Intelligence (AI). The latest UK health research programme ‘Our Future Health’,79 aimed to prevent, detect and treat disease, incorporates longitudinal collection of biospecimens to facilitate such approaches.

The sensitivity (73%) of USS screening was significantly higher than the sensitivity of TVS alone (44.6%; 33/74) noted during four rounds of ovarian cancer screening in the PLCO trial. 70 It must be noted that a combined strategy with first-line CA125 and TVS was used in the PLCO trial, so overall sensitivity was higher. Despite the improved sensitivity, performance characteristics of USS screening in UKCTOCS suggest that TVS is not an effective first-line test for population screening. The sensitivity for invasive epithelial ovarian cancer (63%) was low and 16 additional operations were undertaken per invasive cancer detected. It needs to be noted that when all (including non-epithelial and borderline) ovarian and tubal cancers were considered, in the USS group 10 additional operations were undertaken per cancer detected. This was below the 15 operations per cancer detected that was stated in the patient information leaflet (Project documentation D) based on estimates at the start of the trial in 2001. Despite a similar proportion (39%) of screen-detected invasive cancers detected at early stage in the USS group, nearly all (95%) missed cancers (clinically diagnosed) were advanced (III/IV) unlike in the MMS (69%) group. This together with the lower sensitivity of USS resulted in no evidence of a difference in incidence of early-stage (I–II) cancer between USS and C groups on the final intention-to-screen analysis.

A major limitation was the length of the trial – it spanned two decades. During this period, knowledge about this cancer evolved massively. By the end of the screening phase, it was widely accepted that a significant number of ‘ovarian’ cancers arose in the tube. 80,81 In 2009, modelling based on occult cancers found in risk-reducing surgery in BRCA gene mutation carriers suggested that the majority of Stage I high-grade serous cancers are 0.4–1.3 cm in diameter and therefore difficult to reliably image. 82 If this had been known prior to trial design, an ultrasound strategy using TVS as a first-line test might not have been included in the design. There would also have been less reliance on visualising an abnormal adnexal mass on TVS prior to recommending trial surgery in the MMS group. The MMS screening strategy did include trial surgery for the small minority of women for whom the ROCA indicated ‘severe’ risk due to rising biomarker levels but repeated imaging was normal. 83 However, it was challenging to convince clinicians to operate soon when faced with this scenario given the lack of understanding of disease evolution. In retrospect, second-line tests in the MMS strategy should have been further optimised so that the interval from annual intermediate risk screens to cancer diagnosis was reduced. 41 If we were planning a new trial, we would consider a second-line test such as circulating tumour DNA (ctDNA)84 or microRNA (miRNA)75 alongside TVS.

The intervention

The worldwide accepted standard of care for most women diagnosed with ovarian and tubal cancer (Stage IC–IV) is complete cytoreductive surgery and platinum-based chemotherapy. The two available options are (1) primary surgery followed by adjuvant chemotherapy and (2) neo-adjuvant chemotherapy with interval debulking surgery (three cycles, surgery, further three cycles). The standards are based on evidence from randomised controlled trials and are enshrined in numerous evidence-based recommendations. Those spanning the trial include ESMO 2001,85 200886 and 2010. 87 There are also nationally agreed treatment protocols for the United Kingdom (NICE, 201164). All women detected with ovarian cancer in the course of UKCTOCS, with few exceptions, were treated using these standard protocols at cancer centres within the NHS, ensuring that there was equity and balance between the groups.

However, the length of follow-up does mean that majority of screen-detected women were diagnosed and treated more than a decade ago (2001–2011). Since then, there have been further advances in clinical management (better staging, widespread use of ultraradical surgery, earlier treatment modulation based on better prognostic indicators, targeted therapies). The latter, if available during screening, may have impacted on trial outcomes.

Since end of screening on the trial, genetic testing for mutations in BRCA1/2 genes has been introduced for all women diagnosed with invasive epithelial ovarian and tubal cancer. It is likely that about 15% will be found to have germline mutations. 88 This would magnify the impact of any future ovarian cancer screening strategy through cascade testing in families of individuals who are mutation positive.

The screening programme

UKCTOCS did not find any evidence of a reduction in deaths from ovarian and tubal cancer compared to no screening at a median follow-up of 16.3 years from randomisation, with either the multimodal or the ultrasound screening strategy investigated in the trial. The only other RCT to report on the mortality impact of ovarian cancer screening was the PLCO cancer screening trial. Similar to UKCTOCS, there was no evidence of a reduction in ovarian cancer deaths between the screen and no-screen arms, either at initial (median follow-up of 12.5 years)21 or extended (14.8 years)89 follow-up. However, unlike in the UKCTOCS MMS group, there was no evidence of a difference in stage distribution between the screened and non-screened arms in the PLCO trial. In both trials there were varying rates of harm due to unnecessary surgery with its attendant complications. The trial findings reinforce the current guidance that women in the general population should not be offered ovarian cancer screening outside trials.

There is the possibility of dilution of the screening effect during the 9.5 years interval from end of screening (end 2011) to censorship (mid 2020). However, in most screening trials, long-term follow-up is the norm. In the European Randomized Study of Screening for Prostate Cancer,90 this did not impact on mortality reduction. Also, the ovarian cancer mortality hazard ratio (HR) was similar at 15 and 18 years from randomisation in our trial. In UKCTOCS, in line with other screening trials, disease-specific rather than all-cause mortality is the primary outcome.

Finally, the late effect of screening was not anticipated in the statistical design. It was only after the initial analysis38 that growing evidence from other screening trials of delayed mortality reductions led to significant concerns that the test chosen (proportional hazards Cox-model) for comparing the primary outcome may not be the best approach. Based on consultations with experts in the field of trials and statistics, the constant-effect primary analysis approach was changed to the Versatile test, which allows for a delayed effect. A transparent process with publication49 of all details was adopted. Our experience strongly suggests the need for discussion and consensus-building on how best to design and analyse large-scale screening trials.

Over 1.24 million women were invited to the trial and 202,638 took part, which equated to 1 in 6 eligible women aged 50–74 being invited to UKCTOCS and 1 in 6 of them taking part. Of the women invited, 23% accepted the invitation and finally 16% were randomised. As a result, the women who accepted the invitation were a self-selected motivated cohort who were more health-aware than most in the general population. This is a well-acknowledged aspect of participants of population screening trials and also part of real-world screening. We have explored these issues of uptake and its impact in previous publications. 29,39

The acceptance rate varied between different parts of the country from 19% in East London to 33% in Bristol. It was similar to that seen in contemporaneous UK trials of colorectal cancer screening and HRT use after the menopause (WISDOM) trial. We had not explored recruitment via invitation from the NHS registers in our feasibility pilot RCT of 13,500 women. 26 We instead estimated uptake from the sparse literature that was available in 1999. Parkes et al. 91 reported an 82% uptake in their RCT of ovarian cancer screening of 8678 women aged 50–64, where women attending a breast cancer screening centre in the UK were invited. Based on this, we estimated a somewhat lower uptake rate of 66% given that we were inviting all women rather than those attending screening. This turned out to be an overestimate. It highlights the importance of piloting all aspects prior to undertaking the main trial. Since then, there have been large studies in the UK that have used NHS registers to recruit participants from the general population. Even in non-interventional studies using this mode of invitation acceptance rates are low. For example, in UK Biobank, the uptake of invitations was 5.45%. 92 Acceptance rates also vary widely for population screening programmes, despite established evidence of benefit. In London, breast screening uptake is 63%. 64 For the more recent colorectal cancer screening programmes, uptake ranges from 68% in the Netherlands, 52% in England to 16% in Canada. 93 There are now many efforts focused on interventions (messaging, community engagement, education, improving test acceptability) to increase screening uptake. 94

Women underwent up to 11 years of annual screening involving over 670,000 annual screening episodes with high compliance. Women consented at trial recruitment to false-positive surgery rates of 9 in 10. In the MMS group, this was lower, with three false-positive operations per ovarian cancer detected. Diagnostic work-up and treatment was undertaken in 13 NHS Trusts located in England, Northern Ireland and Wales. There were very few protocol deviations or complaints from the wider group of health professionals in the NHS or the trial participants. Throughout the trial there was huge support from the ovarian cancer charities both in the UK and in the USA. There was also interest from politicians, especially the All Party Parliamentary Group on Ovarian Cancer. This suggests that such strategies (test, diagnostic procedures, treatment/intervention), had they been effective, might have been clinically, socially and ethically acceptable to the wider public.

The cost-effectiveness analyses based on data accrued till December 2014 suggested that an NHS ovarian cancer screening programme using the MMS strategy extrapolated out to lifetime and after accounting for lead time could approach the NICE thresholds for cost-effectiveness. This was highly dependent on a 20% mortality reduction being confirmed on continued follow-up. 47 Other groups also modelled the UKCTOCS data (2001–2014) and found that lifetime cost-effectiveness of MMS-based screening was promising, subject to long-term effectiveness. 52 However, as there was no reduction in deaths in the MMS arm on follow-up, this is no longer applicable.

Implementation criteria

If future screening trials show benefit, then it would be important to ensure that clinical management of ovarian cancer is further standardised across the UK. Recent national audit of the disease has shown that in England there is significant variability in the use of surgery and chemotherapy standard protocols. 95 Significant variation in the utilisation of both have also been reported previously. 96–98 Meanwhile other options being investigated to improve outcomes in this disease include improved risk prediction through use of algorithms incorporating genetic and epidemiological data99,100 and prevention options based mainly on risk-reducing surgery after completion of family in high-risk women.

For the trial, a bespoke commercially built web-based trial management system was developed by the team in 2000 with automation of key processes, remote data entry and concurrent central monitoring to ensure strict adherence to protocol. Use of electronic health records data to invite women ensured timely recruitment despite the huge sample size; appropriate consent and linkage from the start of the trial to national registries and administrative databases ensured completeness of follow-up; an independent outcomes review committee masked to randomisation group and following a strict algorithm for site assignment and cause of death ensured accuracy and reproducibility over time. Furthermore, keeping with new insights, cancer site assignment was revised to reflect the WHO 2014 classification and cancers were re-staged using the FIGO 2014 criteria. Many of the above approaches that were part of UKCTOCS from the start of the trial in 2001 are now considered best practice for trial conduct. Much could have been adapted to use in a screening programme had the trial been successful.

However, an ultrasound-based strategy would have been challenging to implement given the need to train a larger number of sonographers nationally to deliver first-line screening as well as the constant need for monitoring quality. The latter would require a set-up similar to those used for cervical cytological screening. A blood-based longitudinal CA125 strategy using NHS laboratories that routinely perform CA125 measurements, additional quality assurance measures developed in the trial and a digital platform would have been easier to implement. The MMS strategy too would have required an expanded NHS ultrasound workforce, but smaller as less than 1% of women undergoing MMS screening required a transvaginal ultrasound. An equally challenging issue would be the need for timely surgery of screen positives. This would be required for any ovarian cancer screening programme irrespective of the strategy. In this respect it is significantly different from cervical, breast, prostate and bowel cancer screening where it is possible to diagnose cancer safely and accurately using a directed biopsy. In ovarian cancer screening, women with abnormal screening results would be required to undergo surgery under general anaesthesia with complete removal of the tubes and ovaries, preferably laparoscopically.

The trial used age 50 and above and lack of high-risk family history of ovarian cancer to establish eligibility for general population screening. It is likely that in the future, risk will be further personalised using the new risk-prediction algorithms, but this would only serve to move women into the high-risk-group category. There is little scope in widening the eligibility criteria based on age as the risk of ovarian cancer in women under the age of 50 is low. A screening interval shorter than a year on a population scale would not have been possible to implement.

Conclusions

The group has worked to reduce deaths from ovarian and tubal cancer since the 1980s. 16,17 UKCTOCS, which has spanned 20 years, followed on from a pilot randomised controlled trial in the mid-1990s. 26 The trial owes its successful completion to the over 200,000 women who gave generously of their time, unstinting support from many UK funding agencies, efforts of numerous NHS staff over many years both at the RC and more widely, and the backing of multiple UK and international cancer charities, societies and expert groups. The trial has delivered a clear and definitive answer and addressed comprehensively all the questions posed. Currently, general population screening for ovarian and tubal cancer cannot be recommended. The findings of our trial are applicable to women from the general population and may not apply to those at high risk.

Key insights include the lack of impact on mortality despite detecting the disease at an early stage. This highlights the need to detect the cancer in asymptomatic women even earlier during its natural history and in a larger proportion of women than was possible in UKCTOCS. It reinforces the importance of having deaths due to the disease as opposed to stage at diagnosis as the primary outcome in ovarian cancer screening trials in the general population. It also sets benchmarks for surrogate clinical utility outcomes such as performance characteristics, unnecessary surgery rates, and compliance that need to be surpassed.

The women were exemplary in their support of the trial. In 2012, following the end of screening, we individually thanked all 202,638 participants who were alive and had not had ovarian cancer. Artwork, a painting of an ovarian follicle that was then digitised, was commissioned. Each dot on the digital image represented eight women who took part in the trial. The image was used in the card that was sent to each participant (Figure 6). 64 Further details of our engagement with the participants are detailed in the ‘Dissemination to participants and related patient and public communities’ section.

FIGURE 6.

Thank You card.

The generosity of the women has enabled collaboration with researchers around the world to explore new biomarkers for early detection and risk prediction in ovarian as well as other common cancers and in cardiovascular disease. It has spurred the development and validation of longitudinal algorithms. 77,101 The screening data provide a unique opportunity to study the natural history of ovarian and tubal cancer. The team are committed to ensuring wide access and sharing of samples and data with researchers worldwide.

Contributions of authors

Lesley Fallowfield (https://orcid.org/0000-0003-0577-4518), Alistair J McGuire (https://orcid.org/0000-0002-5367-9841), Stuart Campbell (https://orcid.org/0000-0003-4943-155x) and Steven J Skates (https://orcid.org/0000-0001-9249-4316) were co-investigators.

Mahesh Parmar (https://orcid.org/0000-0003-0166-1700) was the trial statistician.

Ian J Jacobs (https://orcid.org/0000-0002-5005-2672) was CI from 2000 to 2014 and co-investigator from 2015 to 2020.

All contributed to trial concept and, together with Aleksandra Gentry-Maharaj, Matthew Burnell, Andy Ryan and Naveena Singh, to trial design. Usha Menon, Andy Ryan and Aleksandra Gentry-Maharaj drafted the manuscript. Andy Ryan, Matthew Burnell and Usha Menon prepared the figures and tables. All contributed to review and revision of the manuscript. All authors approved the report before submission.

Ethics approval

UKCTOCS was approved in June 2000 by the UK North West Multicentre Research Ethics Committees (Ref: North West MREC 00/8/34), currently NRES Committee North West – Haydock, with site-specific approval from the local regional ethics committees and the Caldicott guardians (data controllers) of the 27 participating PCTs. All women provided written consent. Approval for follow-up of the entire cohort is until 31 December 2027.

The psychosocial study was approved by the North West research ethics committee (MREC 00/8/34), and separate written consent was obtained from all women.

Trial registration: UKCTOCS was registered with ISRCTN number 22488978; ClinicalTrials.gov. number NCT00058032.

Ethics statement

UKCTOCS was approved by the UK North West Multicentre Research Ethics Committee, currently Haydock NRES (Ref: 00/8/34) on 21 June 2000 with site-specific approval from the local regional ethics committees and the Caldicott guardians (data controllers) of the PCTs.

Information governance statement

UCL is both the data controller and processor of all personal data handled throughout the UKCTOCS trial. As per the EU General Data Protection Regulation and the UK Data Protection Act 2018, all trial/study participants were provided with additional details regarding the legal basis that allows us to hold and process their personal data (http://ukctocs.mrcctu.ucl.ac.uk/participant-info/).

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that they are stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives. You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to anonymised data may be granted following review.

Funding statement

The long-term follow-up UKCTOCS (2015–2020) was supported by National Institute for Health and Care Research (NIHR HTA grant 16/46/01), Cancer Research UK, and The Eve Appeal. UKCTOCS (2001–2014) was funded by the Medical Research Council (G9901012 and G0801228), Cancer Research UK (C1479/A2884), and the UK Department of Health, with additional support from The Eve Appeal. Researchers at UCL were supported by the NIHR UCL Hospitals Biomedical Research Centre and by Medical Research Council Clinical Trials Unit at UCL core funding (MR_UU_12023). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the HTA programme or the Department of Health and Social Care.

This article

The contractual start date for this research was in January 2017. This article began editorial review in January 2022 and was accepted for publication in August 2022. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The Health Technology Assessment editors and publisher have tried to ensure the accuracy of the authors’ article and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this article.

Disclaimer

The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

Media

2021

BBC News

www.bbc.co.uk/news/health-57087477

BBC Radio 4 The Today Programme (skip to 2:08:03)

www.bbc.co.uk/sounds/play/m000vyn2

BBC Radio 2 News (skip to 1:32:38)

www.bbc.co.uk/sounds/play/m000vyg8

BBC World Service’s ‘The Newsroom’ (skip to 16:55)

www.bbc.co.uk/sounds/play/w172xyxfw79rfxh

Sky News

https://news.sky.com/story/major-ovarian-cancer-screening-trial-fails-to-show-any-reduction-in-deaths-12305017

Yahoo News

https://uk.news.yahoo.com/ovarian-cancer-screening-doesnt-save-lives-223016448.html

PA via Yahoo

https://uk.news.yahoo.com/ovarian-cancer-population-screening-did-223000849.html

Telegraph

www.telegraph.co.uk/news/2021/05/13/screening-women-ovarian-cancer-does-not-reduce-deaths-study/

Guardian

www.theguardian.com/society/2021/may/12/annual-screening-for-ovarian-cancer-does-not-save-lives-study-finds?CMP=Share_iOSApp_Other

Independent

www.independent.co.uk/news/health/ovarian-cancer-screening-deaths-women-b1846182.html

iNews

https://inews.co.uk/news/health/ovarian-cancer-screening-does-not-reduce-deaths-major-trial-finds-997661

Sun

www.thesun.co.uk/fabulous/14942639/7-ovarian-cancer-signs-screening-plans-delayed-decade/

BMJ

www.bmj.com/content/373/bmj.n1223.short?rss=1&utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A%20bmj/recent%20%28Latest%20from%20BMJ%29

CBC Radio (Canada)

www.cbc.ca/radio/asithappens/as-it-happens-thursday-edition-1.6025264/ovarian-cancer-researchers-disappointed-after-trial-finds-regular-screenings-fail-to-reduce-deaths-1.6025268

Irish News

www.irishnews.com/magazine/science/2021/05/12/news/ovarian-cancer-population-screening-did-not-cut-deaths-study-2320664/

ABC (Spain)

www.abc.es/salud/enfermedades/abci-estudio-concluye-cribado-cancer-ovario-no-salva-vidas-202105171037_noticia.html?ref=https://www.google.com/

2019

The Adelaide Advertiser

On behalf of Ovarian Cancer Australia for Ovarian Cancer Awareness on 4 February 2019.

Body and Soul publication

On behalf of Ovarian Cancer Australia about Ovarian Cancer Awareness month published on 16 January 2019.

2014–2015

BBC News

Blood test “boost” in ovarian cancer fight (Tuesday 5 May 2015)

The Guardian

New ovarian cancer test twice as effective as existing methods (Monday 4 May 2015)

BBC News

Skirt size increase linked to breast cancer risk, says study (Thursday 25 September 2014)

NHS News

Skirt size increase ups breast cancer risk (Thursday, 25 September 2014)

2001–2010

NHS News

Ovarian cancer screening (11 March 2009)

BBC News

Trials offer ovarian cancer hope (Wednesday 9 March 2009)

The Guardian

Ovarian cancer screening on trial (Wednesday 22 March 2000; p. 7)

BBC News

Cancer screening trials to begin (Tuesday 21 March 2000)

Websites

MRC CTU website

www.cancerresearchuk.org/about-cancer/find-a-clinical-trial/a-study-looking-at-screening-the-general-population-for-ovarian-cancer

Cancer Research UK website

www.cancerresearchuk.org/about-cancer/find-a-clinical-trial/a-study-looking-at-screening-the-general-population-for-ovarian-cancer

Invited presentations

Review articles

Alexander-Sefre F, Menon U, Jacobs IJ. Ovarian cancer screening. Hosp Med 2002;63(4):210–13.

Campbell S, Gentry-Maharaj A. The role of transvaginal ultrasound in screening for ovarian cancer. Climacteric 2018;21(3):221–6. https://doi.org/10.1080/13697137.2018.1433656. Epub 2018 March 1. PMID: 29490504.

Crosbie E, Menon U. Epithelial ovarian cancer and induction of ovulation. Rev Gynaecol Pract 2005;5:131–8.

Gentry-Maharaj A, Menon U. Screening for ovarian cancer in the general population. Best Pract Res Clin Obstet Gynaecol 2012;26(2):243–56. https://doi.org/10.1016/j.bpobgyn.2011.11.006. Epub 2011 December 17. PMID: 22182415.

Gentry-Maharaj A, Sharma A, Menon U. Ovarian cancer (review). Geriatr Med 2007;37(11):8–12.

Hamilton W, Menon U. Ovarian cancer. BMJ 2009;339:b4650. https://doi.org/10.1136/bmj.b4650.

Jacobs I, Menon U. Can ovarian cancer screening save lives? The question remains unanswered. Obstet Gynecol 2011;118(6):1209–11.

Jacobs I, Menon U. The sine qua non of discovering novel biomarkers for early detection of ovarian cancer: carefully selected preclinical samples. Cancer Prev Res (Phila) 2011;4(3):299–302.

Jacobs IJ, Menon U. Progress challenges in screening for early detection of ovarian cancer. Mol Cell Proteomics 2004;3(4):355-66.

Kalsi J, Manchanda R, Menon U. Screening for gynecologic cancers. Exp Rev Obstet Gynecol 2013;8(2):143–60.

Lewis S, Menon U. Screening for ovarian cancer. Exp Rev Anticancer Ther 2003;3(1):55–62.

Menon U, Gentry-Maharaj A, Jacobs I. Ovarian cancer screening and mortality. JAMA 2011;306(14):1544.

Menon U, Griffin M, Gentry-Maharaj A. Ovarian cancer screening – current status, future directions. Gynecol Oncol. 2014;132(2):490–5. https://doi.org/10.1016/j.ygyno.2013.11.030. Epub 3 December 2013. PMID: 24316306; PMCID: PMC3991859.

Menon U, Jacobs IJ. Ovarian cancer screening in the general population – an update. Curr Opin Obstet Gynaecol 2001;13(1):61–4.

Menon U, Jacobs IJ. Ovarian cancer screening in the general population – current status. Int J Gynecol Cancer 2001;11(S1):3–6.

Menon U, Karpinskyj C, Gentry-Maharaj A. Ovarian cancer prevention and screening. Obstet Gynecol 2018;131(5):909–27. https://doi.org/10.1097/AOG.0000000000002580. PMID: 29630008.

Menon U. Ovarian cancer screening has no effect on disease-specific mortality. Evid Based Med 2012;17(2):47–8.

Menon U. Ovarian cancer screening. CMAJ 2004;171(4):323–4.

Menon U. Ovarian cancer: challenges of early detection. Nat Clin Pract Oncol 2007;4(9):498–9.

Menon U. Steps towards effective gynaecological cancer screening. Nat Rev Clin Oncol 2018;15(9):538–40. https://doi.org/10.1038/s41571-018-0049-4. PMID: 29855609.

Nash Z, Menon U. Ovarian cancer screening: current status and future directions. Best Pract Res Clin Obstet Gynaecol 2020;65:32–45. https://doi.org/10.1016/j.bpobgyn.2020.02.010. Epub 2020 Mar 3. PMID: 32273169.

Rosenthal AN, Menon U, Jacobs IJ. Screening for ovarian cancer. Clin Obstet Gynecol 2006;49(3):433–47.

Rufford B, Menon U, Jacobs IJ. Ovarian Cancer Screening. Proceedings of the 2nd World Congress on Controversies in Obstetrics, Gynaecology and Infertility, September 2001, Paris, France.

Sharma A, Menon U. Screening for gynaecological cancers. Eur J Surg Oncol 2006;32(8):818–24, ISSN: 0748-7983.

Sharma A, Menon U. The value of ovarian cancer screening. Br J Hosp Med 2006;67(6):314–17.

Widschwendter M, Menon U. Circulating methylated DNA: a new generation of tumor markers. Clin Cancer Res 2006;12(24):7205–8.

Chapters

Alexander Sefre F, Menon U, Jacobs IJ. Ovarian cancer screening. In Farquharson R, editor. Vignettes for the MRCOG. London: Quay Books; 2005. pp. 57–62.

Fourkala E-O, Gentry-Maharaj A, Menon U. MUC16, alias CA125 and ovarian cancer. In Taylor-Papadimitriou J, Burchell JM, editors. Mucins and Cancer. London, UK: Future Medicine; 2013. pp. 82–93.

Gentry-Maharaj A, Griffin M, Menon U. Ovarian cancer prevention and screening. In Eeles R, Berg C, Tobias J, editors. Cancer Prevention and Screening: Concepts, Principles and Controversies. Oxford: Wiley-Blackwell; 2018. pp. 331–48.

Gentry-Maharaj A, Jacobs I, Menon U. Development and identification of tumor serum markers. In Barakat RR, Berchuck A, Markman M, Randall ME, editors. Principles and Practice of Gynecologic Oncology. 6th edn. Baltimore: Lippincott Williams & Wilkins; 2013. pp. 89–114.

Gentry-Maharaj A, Jacobs I, Menon U. Development and identification of tumour markers. In Barakat RR, Markman M, Randall ME, editors. Principles and Practice of Gynecologic Oncology. 5th edn. USA: Wolter Kluwer Health; Lippincott, Williams & Wilkins; 2009. pp. 145–62.

Gentry-Maharaj A, Jacobs I, Menon U. Ovarian cancer tumor markers and screening. In Berek JS, Hacker NF, editors. Berek and Hacker’s Gynecologic Oncology. 6th edn. Lippincott Williams & Wilkins; 2014. pp. 443–63.

Gentry-Maharaj A, Kalsi J, Menon U. Screening for gynaecological cancers. In Patel HRH, Mould T, Joseph JV and Delaney CP, editors. Pelvic Cancer Surgery: Modern Breakthroughs and Future Advances. Chap 26, Springer; ISBN 978-1-4471-4257-7; 2015. pp. 267–82.

Gentry-Maharaj A, Kalsi J, Menon U. Screening for gynecologic cancers. In Patel H, Mould T, Joseph J, Delaney C, editors. Pelvic Cancer Surgery: Modern Breakthroughs & Future Advances. Berlin/Heidelberg, Germany: Springer Science; 2014. pp. 267–81.

Gentry-Maharaj A, Menon U. Development and identification of tumour markers. In Chi D, Berchuck A, Dizon DS, Yashar CM, editors. Principles and Practice of Gynecologic Oncology. 7th edn. Wolters Kluwer; 2017. pp.79–99.

Gentry-Maharaj A, Menon U. Early detection of hereditary gynaecological cancers. In Jacobs C, Robinson L, Webb P, editors. Genetics for Health Professionals in Cancer Care: from Principles to Practice. 1st edn. Oxford, UK: Oxford University Press; 2014. pp. 173–9.

Gentry-Maharaj A, Menon U. Screening for ovarian cancer in the general population. Best Pract Res Clin Obstet Gynaecol 2012;26(2):243–56.

Manchanda R, Jacobs IJ, Menon U. Tumour markers and screening. In Berek J, Hacker N, editors. Practical Gynecologic Oncology. 5th edn. Philadelphia: Lippincott, Williams & Wilkin; 2010. pp. 233–66.

Menon U, Brinkman DA, Skates SJ, Jacobs IJ. Screening: ovarian cancer. In Cuzick J, editor. Evidence Based Oncology Part 3. Amsterdam, Netherlands: Elsevier; 2003.

Menon U, Dilley J. Tumour markers. In Richard Smith J, Del Priore G, Coleman RL, Monaghan JM, editors. An Atlas of Gynecologic Oncology: Investigation and Surgery. 4th edn. Chap 7, CRC Press; ISBN 135114166X- 9781351141666; 2018.

Menon U, Jacobs IJ. CA125 and other tumour markers in screening and monitoring of ovarian cancer. In Altchek A, Deligdisch L, editors. Diagnosis and Management of Ovarian Disorders. New York: Academic Press; 2002.

Menon U, Jacobs IJ. Screening for ovarian cancer. In Arulkumaran S, editor. Baillière’s Best Practice & Research Clinical Obstetrics & Gynaecology. London; 2002. pp. 469–82.

Menon U, Jacobs IJ. Serum tumour markers. In Smith JR, Priore GD, Curtin J, Monaghan J, editors. An Atlas of Gynaecological Oncology. London: Martin Dunitz; 2005. pp. 21–43.

Menon U, Jacobs IJ. Serum tumour markers. In Smith JR, Priore GD, Curtin J, Monaghan J, editors. An Atlas of Gynaecological Oncology. London: Martin Dunitz; 2001. pp. 19–34.

Menon U, Jacobs IJ. The current status of screening for ovarian cancer. In Jacobs IJ, Shepherd JH, Oram DH, Blackett AD, Luesley D, Berchuck A, Hudson CN, editors. Ovarian Cancer. London: Oxford University Press; 2002. pp. 171–8.

Menon U, Jacobs IJ. Tumour markers and screening. In Berek J, Hacker N, editors. Practical Gynecologic Oncology. 4th edn. Philadelphia, PA: Lippincott Williams & Wilkins; 2004. pp. 43–66.

Reynolds K, Menon U, Jacobs IJ. Development and identification of tumour markers. In Hoskins WJ, Perez CA, Young RC, editors. Principles and Practice of Gynecologic Oncology. 4th edn. Philadelphia, PA: Lippincott Williams & Wilkins; 2005, pp. 157–78.

Rosenthal A, Menon U, Jacobs IJ. Ovarian cancer screening. In Reznick R, editor. Contemporary Issues in Cancer Imaging: A Multidisciplinary Approach. Cambridge: Cambridge University Press; 2006.

Rufford B, Menon U, Jacobs IJ. Ovarian cancer screening. Cancer Imaging 2001;1(2):14–16.

Rufford B, Menon U, Jacobs IJ. Screening for familial ovarian cancer. In Morrison PK, Hodgson SV, Haites NE, editors. Familial Breast and Ovarian Cancer Genetics, Screening and Management. Cambridge: Cambridge University Press; 2002. pp. 220–33.

Sharma A, Menon U, Jacobs IJ. Symptoms in the diagnosis of ovarian cancer. In Hillard T, editor. The Yearbook of Obstetrics and Gynaecology. London: RCOG Press; 2007. pp. 159–70.

Sharma A, Menon U, Ledermann J. Advanced Ovarian Cancer (Published online 3/10/2006). BMJ Clinical Evidence handbook.

Presentation to All-Parliamentary Group

Prof Menon gave evidence to the panel of the All-Party Parliamentary Group (APPG) on Ovarian Cancer in 2010 and answered questions regarding ovarian cancer screening which included two UK studies – UKCTOCS and UKFOCSS.

UKCTOCS committees and teams

Project documentation

1. Protocol Version 9.0

2. Invitation letter

3. Brochure

4. Patient Information Sheet

5. Recruitment Questionnaire

6. Consent for the Main Trial

7. Consent for the Psychosocial Study

8. Data Protection form

9. Follow Up Questionnaire 1

10. Follow Up Questionnaire 2

11. Follow Up Questionnaire 3

12. Outcomes Review Form – Cancer site

13. Outcomes Review Form – Cause of death