Executive Summary

Centre of Research Excellence in Precision Public Health Approaches to Breast Cancer Screening, Early Detection and Mortality Reduction

Aim

The aim of this Centre of Research Excellence (CRE) is to apply a precision public health risk-based approach to address multiple aspects of breast cancer screening, so as to enable greater reductions in the morbidity and mortality arising from breast cancer.

Specifically, we plan to:

  1. develop, apply and teach advanced analytics including AI applied to Big Datasets comprising integrated state-of-the-art epidemiological and genetic studies using digital mammograms, family history and genomic testing, to better stratify risk of breast cancer – and better detect cancers
  2. co-design novel personalised screening pathways that translate our new models and BRISK into improved breast cancer screening by working with radiologists (BreastScreen and private), general practitioners, clinicians and women using qualitative interviews, focus groups and stakeholder engagement, and study the health economic consequences of these models
  3. create widely accessible educational material to support implementation
  4. use technological applications to provide an interface between women and clinicians by creating a simple automated decision support tool (BRISK) for women and health professionals
  5. build a new generation of up-skilled early and mid-career research and clinical leaders with opportunities for training, mentorship, career development and international exchange.

Introduction

The CRE is implementing a fundamental change in breast cancer control for Australia and worldwide by tailoring breast cancer screening precisely to risk – a concept gaining acceptance internationally. It is creating a screening model that can better support younger women, women with family history and Aboriginal and Torres Strait Islander women (who have lower participation and higher mortality rates). The COVID-19 pandemic amplified this need, as all BreastScreen Australia services were shut-down for a period and risk-based decisions are now required to prioritise scarce screening resources.

We are determining optimal personalised mammographic screening programs and improving automated cancer detection. We are working with others towards a new and more cost-effective personalised screening model that could be used by both BreastScreen and private providers. We are continuing to inform pathobiology research on mammogram-based causes of breast cancer.

We are building on the successful translational and organisational strategies we developed and implemented through two previous successful CREs for bowel cancer and twins’ research. The highly innovative CRE team is demonstrating how analytics, big data and technology can be used in clinical practice and public health decision-making, by combining human and artificial intelligence (AI).

Breast cancer is a major health problem

For Australian women, breast cancer is the most common cancer (20,000 new cases p.a.) and the second most common cause of cancer-related deaths (3,000 deaths p.a.). One in seven women will be diagnosed with breast cancer by age 85 years (average life expectancy). Women vary greatly in their underlying risk [1, 2], but current risk assessment tools lack precision and can be cumbersome to utilize, limiting routine use by clinicians [3]. While early detection through population screening is associated with survival benefits, current early strategies are yet to fully benefit from the recent surge in development and application of AI and other analytical techniques to screening.

Breast cancer screening in Australia

Mammography – which has been digital in Australia for the last decade –  has been at the front line of breast cancer control in Australia for more than two decades due to its ability to detect many cancers at an early stage, with the aim of reducing mortality from breast cancer [1,4]. Almost 1 million Australian women are screened each year by BreastScreen, which is targeted at women based on age. BreastScreen Australia is a government-funded program, established in 1991, providing biennial, free screening mammograms. The target age-range is 50-74 years however women are eligible to attend from the age of 40 years. The aim is to reduce mortality and morbidity by detecting breast cancer early before symptoms are noticed and when treatment is likely to be more successful. A two-view screening mammogram is performed at fixed and mobile screening units. Following independent, double reading by radiologists (with arbitration by an independent third reader if needed), women whose images are suspicious for breast cancer are recalled for further investigation by a multidisciplinary team at an assessment service. Further investigation can include clinical examination, further mammograms, ultrasound and biopsies.

Breast cancer screening has significant known challenges

The historical challenges for mammographic screening are:

  • One-size-fits-all approach: Other than for the small proportion of women with a strong family history or previous breast biopsy of an atypical/borderline lesion, all women are screened every 2 years – irrespective of risk. This might be suboptimal for most women if lower-risk women are screened too frequently, and higher-risk women are not screened frequently enough.
  • Lack of sensitivity and specificity: Over 34,000 Australian women each year experience a false positive outcome at screening (called back for further investigation) which results in anxiety, unnecessary interventions and costs [1,5,6]. False negatives (i.e. missed cancers) delay detection until cancers are larger and may have spread to the lymph nodes, with more intense treatment needed, poorer outcomes and increased mortality. This challenge is strongly driven by dense breasts.
  • Delays in service delivery: Currently there is a 14-day turnaround time for all-clear results, and 28-day turnaround for recalled assessments. Is there a cost-effective way this can be reduced?
  • Lack of efficiency: screening costs are rising with an ageing population.

Dense Breasts

Our CRE is devoting much-needed attention to ‘dense breasts’, a major challenge for mammography increasingly recognised worldwide. The white or bright (mammographically but not necessarily physically dense) areas on a mammogram, also known as breast density, make it difficult for radiologists to detect existing cancers [7]. This ‘masking’ compromises the effectiveness of screening. However, women are not routinely informed about this, an issue made prominent by the late consumer-activist Nancy Cappello [8]. Following strong consumer pressure, most US states and some Canadian provinces mandated that women identified as having dense breasts at screening should be notified and encouraged to discuss supplemental screening and concerns with their healthcare provider. Breast density notification typically depends solely on Breast Imaging Reporting and Data System’s (BI-RADS) classifications with no consideration of personal risk [9]. Approximately 40% of women in the US aged 40 to 74 years have dense breasts [10, 11]. However, there is little evidence about whether other screening modalities are optimal for these women [12].

In Australia, there has been increasing demand from women to know their breast density [13, 14]. In its position statement on breast density [15], BreastScreen Australia recognises that more information is needed on: (i) how density changes with age and time; (ii) how it interacts with other risk factors on breast cancer risk, (iii) the reporting and implementation of breast density notification, (iv) the optimum, cost-effective screening protocol according to breast density level, and (v) the impact of breast density notification on mental and physical health outcomes for women. This CRE will address issues (i), (ii) and (iv), but on a wider scale.

Young women and screening policy based on risk

Younger Australian women diagnosed with breast cancer have a far greater mortality [16], and far greater impact in terms of morbidity and loss of years of productive life.

Current screening programs are problematic for younger women. The best time to have a first screen is a balance between risk and expected benefit and cost; and is a debated topic. Younger adult women are more vulnerable to having cancers missed by mammography as they have on average mammographically-denser breasts. But in terms of relative risk, they differ much more in familial, genetic and mammography-based risk[17]. Risk stratification will be more effective at younger ages.

There is a growing recognition of the need to move beyond age-based criteria, and one-size-fits-all models for breast cancer screening. Clinical trials are examining screening tailored to a woman’s risk based on mammographic density, family history, and/or gene-based measures (see [18]).

The more a woman is at risk, the more important her other risk factors become

This critical relationship underlies our approach to find out how women can benefit from a tailored screening strategy that combines mammogram-based, genetic and other familial risks.

We are engaging with all Australian groups working in related aspects of risk-stratified breast screening to support the development of nationally agreed models of implementation. These include but not be restricted to Cancer Australia, Cancer Councils, BCNA, BreastScreen. We are bringing international perspectives on the implementation of risk stratified screening through our strong connections to the WISDOM and other trials. This CRE is providing the hard evidence and tools, and the experience and expertise, to make a major contribution to an inevitable change in practice.

We are addressing the key challenges to the implementation of cancer risk prediction models into clinical practice [82]:

  1. Choosing the risk model that can be implemented in the proposed clinical setting. This CRE will generate new breast cancer risk prediction models and calculate their discrimination and calibration. In choosing which variables to include we will take into account the clinical setting, practical constraints of implementation such as the cost of collecting different information, and the relative benefits, harms and costs of missed diagnoses and over-diagnoses.
  2. Choosing when and where risk should be predicted. We will consider several clinical settings where a risk prediction tool could be implemented including general practice, private radiology and BreastScreen clinics. We will determine the most suitable age group of women in which risk assessment should occur and how to identify them.
  3. Understanding and overcoming barriers to use. We will identify potential barriers to use of risk prediction tools depending on the specific clinical setting and the design of the tool by conducting a series of Feasibility and Optimisation studies designed to explore these barriers, identify ways to overcome them, and co-design the risk assessment and decision support tool (BRISK).
  4. Communicating risk. The formats in which risk information is presented is a key aspect. We will use randomised vignette studies comparing different risk presentation formats (icon arrays, expected frequency trees, comparative graphs, etc.). We will apply the same methods we have used successfully to examine the communication of risks and benefits of bowel cancer screening [83-85] and chemo-prevention for breast and bowel cancer. We will also conduct qualitative studies of women to understand their perceptions about risk-stratified breast cancer screening and approaches to communicate information that could make receiving less screening acceptable.
  5. Evaluating the effectiveness and cost-effectiveness of using a risk assessment and decision support tool to tailor breast cancer screening. The CRE findings will be used to model the potential health and economic benefits of alternative personalised screening models and risk-assisted cancer detection, and support the pre-trial Development, Feasibility and Optimisation phases of future research. This will underpin the design and application for additional funding to conduct effectiveness RCTs in clinical settings (e.g. general practice, radiology clinics) to assess the potential real-life impacts of using such tools.

References

  1. Australian Institute of Health and Welfare, BreastScreen Australia monitoring report. Cancer series no. 129, CAT. NO: 135. 2020, Australian Government: ACT.
  2. Hopper, J.L., Odds per adjusted standard deviation: Comparing strengths of associations for risk factors measured on different scales and across diseases and populations. Am J Epidemiol, 2015. 182(10): p. 863-7.
  3. Phillips, K.-A., et al., Accuracy of Risk Estimates from the iPrevent Breast Cancer Risk Assessment and Management Tool. JNCI Cancer Spectrum, 2019. 3(4).
  4. Roder, D., et al., Population screening and intensity of screening are associated with reduced breast cancer mortality: evidence of efficacy of mammography screening in Australia. Breast Cancer Res Treat, 2008. 108(3): p. 409-16.
  5. Long, H., et al., How do women experience a false-positive test result from breast screening? A systematic review and thematic synthesis of qualitative studies. Br J Cancer, 2019. 121(4): p. 351-358.
  6. Salz, T., A.R. Richman, and N.T. Brewer, Meta-analyses of the effect of false-positive mammograms on generic and specific psychosocial outcomes. Psychooncology, 2010. 19(10): p. 1026-34.
  7. Holm, J., et al., Risk factors and tumor characteristics of interval cancers by mammographic density. J Clin Oncol, 2015. 33(9): p. 1030-7.
  8. Cappello, N.M., D. Richetelli, and C.I. Lee, The impact of breast density reporting laws on women’s awareness of density-associated risks and conversations regarding supplemental screening with providers. J Am Coll Radiol, 2019. 16(2): p. 139-146.
  9. Burnside, E.S., et al., The ACR BI-RADS experience: learning from history. J Am Coll Radiol, 2009. 6(12): p. 851-60.
  10. Sprague, B.L., et al., Trends in clinical breast density assessment from the breast cancer surveillance consortium. J Natl Cancer Inst, 2019. 111(6): p. 629-632.
  11. Sprague, B.L., et al., Prevalence of mammographically dense breasts in the United States. J Natl Cancer Inst, 2014. 106(10).
  12. Melnikow, J., et al., Supplemental Screening for Breast Cancer in Women With Dense Breasts: A Systematic Review for the U.S. Preventive Services Task Force. Ann Intern Med, 2016. 164(4): p. 268-78.
  13. Darcey, E., et al., Post-mammographic screening behaviour: A survey investigating what women do after being told they have dense breasts. Health Promot J Austr, 2020.
  14. Barter, K. Breast Density: The state of play in Australia. 2018 [cited 2020 25 November]; Pink Hope is a preventative health hub that allows every individual the necessary tools to assess, manage and reduce their risk of breast and ovarian cancer, while providing personalised support for at risk women.]. Available from: https://pinkhope.org.au/breast-density-the-state-of-play-in-australia/.
  15. Position Statement on Breast Density and Screening within the BreastScreen Australia Program: A joint Australian, State and Territory Government Program. 2020, Australian Government: Canberra, ACT.
  16. Jayasekara, H., et al., Mortality after breast cancer as a function of time since diagnosis by estrogen receptor status and age at diagnosis. Int J Cancer, 2019. 145(12): p. 3207-3217.
  17. Hopper, J.L., et al., Going beyond conventional mammographic density to discover novel mammogram-based predictors of breast cancer risk. J Clin Med, 2020. 9(3).
  18. Boyd, N.F., et al., Mammographic density and the risk and detection of breast cancer. N Engl J Med, 2007. 356(3): p. 227-36.
  19. Usher-Smith, J., et al., Risk prediction tools for cancer in primary care. British Journal of Cancer, 2015. 113(12): p. 1645-1650.