A new comprehensive gene-based breast cancer prediction device


  • CanRisk is a new online gene-based health-risk evaluation algorithm for detecting breast cancer
  • It identifies people with different levels of risk of breast cancer, not just those at high risk
  • As the infotech and biotech revolutions merge expect authority in medicine to be transferred to algorithms
  • CanRisk has the potential to provide a cheap, rapid, non-invasive, highly sensitive and accurate diagnosis before symptoms present
  • Breast cancer is the most common cancer in women worldwide and is the 5th most common cause of death from cancer in women
  • Currently mammography screening, which has a sensitivity between 72% and 87%, is the gold standard for preventing and controlling breast cancer
  • For every death from breast cancer that is prevented by screening, it is estimated there will be three false-positive cases that are detected and treated unnecessarily
  • Lack of resources do not support breast cancer screening in many regions of the world where the incidence rates of the disease are rapidly increasing
  • In the near-term expect interest in the CanRisk algorithm to increase
 
 A new comprehensive gene-based breast cancer prediction device

 
A new online gene-based health-risk evaluation device called CanRisk has the potential to identify women with different levels of risk of breast cancer; not just women who are at high risk. Predicated on a comprehensive algorithm, CanRisk is one of several innovations currently in development, which include novel methods for predicting the recurrence of breast cancer, a new class of molecules that aim to halt or destroy breast cancer, and liquid biopsies, which determine the presence and recurrent risk of the disease through the detection of tumour cells in peoples’ blood.
 
Although over the past two decades there have been significant improvements in the detection and treatment of breast cancer, the disease remains the most common cancer in women worldwide, with some 1.7m new cases diagnosed each year, which account for about 25% of all cancers in women and it is the fifth most common cause of death from cancer in women, with over 0.52m deaths each year.
 
Game changer for breast cancer
 
Findings of CanRisk were reported in the January 2019 edition of Genetics in Medicine. Findings of a less comprehensive version of the device’s algorithm were published in the July 2016 edition of the same journal. Commenting on the 2019 study, Antonis Antoniou, Professor of Cancer Risk Prediction at the University of Cambridge and lead author of the two studies said: "This is the first time that anyone has combined so many elements into one breast cancer prediction tool. It could be a game changer for breast cancer and help doctors to tailor the care they provide depending on their patients' level of risk”.
 
When fully developed and approved, CanRisk will be well positioned to provide a cheap, rapid, non-invasive, highly sensitive and accurate diagnostic test to detect breast cancer early in people with diverse levels of risk. This might be expected to provide an alternative to the current gold standard population-based mammography screening and assist in making a significant dent in the vast and escalating global burden of the disease.
 
In this Commentary
 
This Commentary describes the algorithm that drives CanRisk, which benefits from the increasing availability of vast and growing amounts of genomic and other personal data and significant advances in genomic sequencing technologies. The confluence of these two phenomena facilitates and enhances the quality and speed of data analysis and drives the development of new and innovative diagnostic and prognostic cancer technologies. The fact that CanRisk is based on UK data and its algorithm is available to researchers globally, presents a potential  opportunity for medical research organizations in emerging regions of the world where the burden of breast cancer is increasing. The Commentary briefly describes the heterogeneous nature of breast cancer and highlights some of its complexities and risk factors. Originally perceived as a Western disease, breast cancer is growing rapidly in Asia and other regions of the world where it tends to be detected late and managed less effectively. Developed economies prevent and manage breast cancer through well-established population-based mammography screening programs. Because of  the lack of resources,  such screening programs are not widely available in low to middle income countries (LMIC). As the infotech and biotech revolutions merge expect authority in medicine to be transferred to Big Data algorithms such as CanRisk. This not only could provide an alternative to gold standard mammography screening, but also provide a cheap and effective device for use in developing nations where the burden of breast cancer is significant and increasing.
 
CanRisk: a world first
 
CanRisk, developed by members of the Centre for Cancer Genetic Epidemiology at the University of Cambridge, UK, takes advantage of discoveries in both cancer genomics and epidemiology and aims to become a popular device used by primary care physicians, in consultation with their patients, to effectively assess patients’ diverse levels of risk of developing breast cancer. The device is predicated on an algorithm called BOADICEA (the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm). This is the world’s first polygenic breast cancer risk model and the only one to-date, which is available to the international research community. Also, it is the first breast cancer risk model to incorporate pathology data and population-specific cancer incidences in risk calculations. The algorithm accounts for over 300 genetic risk factors, including BRCA1, [BReast CAncer gene] BRCA2PALB2CHEK2, and ATM, which are genes that have been found to impact a person’s chances of developing breast cancer. The device uses a Polygenic Risk Score (PRS) based on 313 single-nucleotide polymorphisms (SNPs), [SNPs, pronounced ‘snips’, are the most common types of genetic variation in people. Each SNP represents a difference in a single DNA building block and is called a nucleotide] which explains 20% of breast cancer polygenic variance. CanRisk also includes a residual polygenic component, which accounts for other genetic/familial effects; known lifestyle/hormonal/reproductive risk factors and mammographic density [Dense breast tissue can make it harder to evaluate mammographic results and may also be associated with an increased risk of breast cancer].

 

Authority increasingly being transferred to algorithms
 
Over the past two decades we have increasingly learnt to accept the authority of Big Data algorithms. For example, without question we expect algorithms to give us directions, tell us what movies to watch, who to date, what clothes to wear, where to go on holiday, what flight to take, what hotel to stay in and where to eat. We are  comfortable with algorithms assigning us our credit rating, limiting our overdraft and capping our payments. Furthermore, we are beginning to accept the authority of algorithms in medicine. For example, we are gradually replacing the authority of primary care doctors with algorithms that can diagnose common diseases more accurately and more cost effectively.


You might also be interested in:

China’s rising MedTech industry and the dilemma facing Western companies

In December 2018, for the first time in history, the US FDA approved an algorithm to diagnose patients without a doctor’s interpretation. The algorithm, called IDx-DR, detects diabetic retinopathy by analysing images of the back of the eye. Indeed, we are living on the cusp of history when the twin revolutions of information technology and biotechnology are merging and providing the basis for us to transfer authority in medicine to algorithms. In the next two decades, it seems reasonable to assume that it will become common practice to accept the authority of algorithms such as CanRisk, which will inform us that we are suffering from a medical condition long before we present any signs or symptoms.
 
Increasing supply of data
 
CanRisk takes advantage of the fact that genetic and other risk factor data are becoming more easily available in clinical practice through electronic health records, biometric sensors that convert biological processes into electronic information, which computers can store and analyse, cost-effective high speed, high capacity genomic sequencing technologies, and efforts such as the 100,000 Genomes ProjectA UK Government sponsored initiative completed in December 2018, which collected, stored and analysed data from the genomes and medical records of 85,000 NHS England patients affected by cancer or rare disease. Genomics Englandwhich is wholly owned by the UK’s Department of Health, was set up in 2003 to deliver the project. Because CanRisk solely is based on UK population data, its findings are likely to be more applicable to similarly developed Western populations, and less so to populations in other regions of the world. This provides a potential opportunity for international organizations interested in early breast cancer diagnosis. 
 
International sequencing projects
 
The UK’s genomes project is part of a much larger rapidly growing and dynamic global genomics market comprised of data and gene sequencing technologies. 100,000 genomes have been the goal of several other nations interested in improving their healthcare - and lowering costs  - by carrying out precision medicine based on insights from sequencing data. Currently the global genomics market is estimated to be about US$19bn and projected to reach US$41bn by 2025. The market is driven by increasing government funding, the consequent rise  in the number of genomics projects, decreasing gene sequencing costs, growing application areas of genomics and the entry and fast growth of commercial players.

China has become the world’s leader in genomic sequencing. In 2010, the Beijing Genomics Institute (BGI) in Shenzhen was understood to be hosting a higher sequencing capacity than that of the entire US. While most government projects aim to sequence 100,000 genomes, China’s sequencing program is set to sequence 1m human genomes, which include subgroups of 50,000 people, each with specific conditions such as cancer or metabolic disease. The data will also include cohorts from different regions of China, which will facilitate “the analysis of different genetic backgrounds of subpopulations”.
 

Revolution in genome sequencing
 
The first human genome project began in 1990, took 13 years and about US$1bn to complete. The last two decades have seen a revolution in genome sequencing with dramatic increases in its speed and efficiency coupled with massive reductions in cost. Genomic sequencing has proved its usefulness as a diagnostic and prognostic tool. Today it is possible to get your genome sequenced for around US$1,000 in a few days and delivered by  post from firms such as Dante Labs and 24 Genetics in Europe, and Veritas Genetics and Sure Genomics in the US.
 
Breast cancer
 
Returning to breast cancer. It is important to note that the disease is not one, but  a group of conditions that manifest themselves with maladies in the same organ. Breasts are comprised of three main parts: lobules, which produce milk; ducts, which carry milk to the nipples; and fibrous and fatty connective tissue, which hold everything together. The type of breast cancer depends on which cells in the breast mutate, but most breast cancers begin in the ducts or lobules. Some mutated cells in the breast may never spread, however, most breast cancers tend to be invasive and may present with a number of different characteristics in terms of hardness and shape, which can provide some indication of their likely progression. Breast cancer can spread outside the breast through blood and lymph vessels. Further, there are significant differences in breast cancer at the genetic level. A study published in the April 2012 edition of Nature compared the genetic makeup of breast cancer tumour samples with their other characteristics for some 2,000 women, for whom information about the tumour characteristics had been meticulously recorded; and identified at least 10 distinct sub-types of breast cancer, each with its own unique characteristics. Although the study contributed to how breast cancer is diagnosed, classified and treated, in practice certain characteristics of these tumours were already known and tested for: most notably cellular receptors for estrogen, and progesterone, which are the two most significant steroid hormones responsible for various female characteristics. Their presence or absence generally suggests the potential utility of additional medication to accompany surgery, radiotherapy and chemotherapy.

 
Despite population screening and advanced therapies breast cancer remains a killer disease
 
Let us briefly consider breast cancer in the world’s most advanced and wealthiest nation: the US. Although there have been significant improvements in the detection and treatment of breast cancer in the US; still about 1 in 8 American women will develop an invasive type of the disease over the course of her lifetime. In 2019, an estimated 268,600 new cases of invasive breast cancer are expected to be diagnosed in the US, along with 62,930 new cases of non-invasive (in situ) breast cancer. Breast cancer death rates for women in the US are higher than those for any other cancer, besides lung cancer. As of January 2019, there were more than 3.1m women with a history of breast cancer in the US. Although breast cancer death rates in the US have been decreasing over the past three decades and women under 50 have experienced larger decreases, still some 41,760 are expected to die in 2019 from the disease. About 2,670 new cases of invasive breast cancer are expected to be diagnosed in men in the US in 2019 where a man’s lifetime risk of breast cancer is about 1 in 883.
 
Breast cancer challenges in Singapore
 
There are also breast cancer challenges in wealthy non-Western developed economies such as Singapore. Over the past four decades, the incidence of breast cancer in Singapore has more than doubled: from 25 to 65 per 100,000 women. Breast cancer is not just the most common cancer for Singaporean women, accounting for one in three cancers in women, but it is also the top killer. Data reported in the country’s Cancer Registry showed that 2,105 women died of the disease between 2011 and 2015. Notwithstanding, Singapore has extensive awareness-raising programs; population-wide mammography screening; excellent, multi-disciplinary primary and long-term care and improving palliative care, which have contributed to a significant increase in the survival rates of breast cancer patients. However, a substantial proportion of Singaporean women still appear to have a patchy knowledge of aspects of the disease, which leads to comparatively low participation rates in the nation’s breast cancer screening services, and this contributes to late presentation of the disease when it is more difficult to cure and more challenging to treat.

You might also be interested in:

Breast cancer and harmful BRCA gene mutations


AstraZeneca’s strategy to target early cancer
Breast cancer growing rapidly in Asia
 
Breast cancer was once largely confined to developed Western countries and Australasia, but it has now become the most common cancer in Asia. Although Asian data on breast cancer are patchy, an Economist Intelligence Unit report, suggests that, “since the 1990s, increases in the incidence of breast cancer in Asia, as measured by age-standardised rates (ASRs), is four to eight times that of the global average”. Indeed, as younger cohorts of Asian women age and adopt Western diets and lifestyles (particularly fertility patterns, such as later first childbirth and shorter breast feeding), breast cancer incidence rates in Asia look set to converge with the much higher ones in the West.
 Further, in LMIC breast cancer is increasing at a more rapid rate than in the West and has become a significant healthcare challenge: 50% of breast cancer cases and 58% of deaths from the disease occur in LMIC.
 The significance of early detection
 
The good news is that if caught in its early stages, breast cancer can be treated effectively, with high survival rates. The average 5-year survival rate for women with invasive breast cancer is 90%. The average 10-year survival rate is 83%. If the cancer is located only in the breast, the 5-year survival rate of women with breast cancer is 99%. In all types of the disease early detection is the cornerstone of breast cancer control.
 
 Gold standard breast cancer mammography screening
 
The current gold standard for preventing and controlling breast cancer is population-based mammography screening. This is a non-invasive process that uses an x-ray of the breast to look for disease in women who do not have symptoms. The method has reasonable sensitivity (72%–87%) that increases with age and allows for the early detection of breast cancer, which helps increase survival, especially in women between 50 and 70. Notwithstanding, mammograms are not pleasant as the breast is squashed between two metal plates and further some women may find mammograms embarrassing.
 
Success of population-based mammography screening
 
Following a landmark Swedish study that began in 1977 mammography screening has been adopted in more than 26 developed countries worldwide. Findings of the study, reported in a 1989 edition of the Journal of Epidemiology and Community Health, suggested that mortality from breast cancer dropped 31% after screening of women aged 39 to 74. More recent findings of the UK screening program published in the June 2013 edition of the British Journal of Cancer, suggested mortality rates from breast cancer were reduced by 20% in the screened group compared to the unscreened group across all age groups. A study published in 2018 in Cancer, which tracked 52,438 Swedish women aged 40-69 from 1977 to 2015, suggested that regular mammograms contributed to a 60% decrease in breast cancer death during the first 10-years of diagnosis, and a 47% reduced risk within 20-years. Research has shown that mammography has relatively little benefit for women under 50.
 
Diverging views about mammography screening
 
Despite evidence to support the benefits of population-based mammography screening, there are diverging views among healthcare professionals about the impact of several decades of high levels of screening. Some argue that traditional mammography screening stretches finite resources and is not cost-effective because the majority of people who undergo screening do not have cancer and may never go on to develop it. Others suggest that there are significant uncertainties about the magnitude of the harms from mammography screening especially associated with false positives (a test result, which wrongly indicates that breast cancer is present).

Challenges of mammography screening
 
The sensitivity of mammography is between 72% and 87%, but is higher in women over 50 and in women with fatty rather than dense breasts. Dense breast tissue can make it harder to evaluate results of a mammogram. According to the Marmot review, for every death from breast cancer that is prevented by screening, it is estimated there will be three over-diagnosed or false-positive cases that are detected and treated unnecessarily. The chance of having a false positive result after one mammogram ranges from 7% to 12%, depending on age (younger women are more likely to have false positive results). After 10 yearly mammograms, the chance of having a false positive is about 50-60%. The more mammograms a woman has, the more likely it is she will have a false positive result. This makes it difficult for doctors to weigh and communicate the benefits and risks of mammography screening programs and fuels interest in innovations such as CanRisk.
 
Takeaways
 
Mammography screening for breast cancer is not 100% accurate. Further, knowhow, trained healthcare professionals and significant resources are required to effectively implement and manage a well-organized and sustainable breast cancer screening program that targets the right population group and ensures effective coordination and quality of actions across the whole continuum of care. These attributes tend to exist only in developed wealthy countries. CanRisk, and other innovative breast cancer early diagnostic devices under development, offer the potential for cheap, rapid, reliable and exquisitely accurate diagnosis that can be easily used in primary care settings throughout the world. In time, as authority in medicine passes to algorithms, expect these new and innovative devices to replace mammography screening in wealthy countries and quickly become devices of choice in developing economies and significantly dent the vast and rapidly growing global burden of breast cancer.

Comments