Estimating the population health impact of a multi-cancer early detection genomic blood test to complement existing screening in the US and UK

Like Comment
Read the paper

What if new multi-cancer early detection tests were widely available?

One of the questions I think about quite a bit as someone who works in both cancer research and health policy is how to best impact population health with limited healthcare resources.  In our recent BJC publication, we look at how new multi-cancer early detection (MCED) tests may help us utilize our resources more efficiently by complementing existing screening recommendations.

We examined two different healthcare systems, the United States and the United Kingdom, to estimate (by modelling) what current cancer screening achieves and then what impact adding an MCED test might have.  As MCED tests are a new screening method, there are no real world data available yet, so our modelling was based on an earlier version of the commercially available Galleri™ test.1

How many and what types of cancer can we detect with current and MCED screening?

Cancer screening is currently based on the principle of one test for one cancer type. Current single-cancer screening tests are recommended for only five cancers (breast, bowel, cervical, lung [US only], prostate), and prostate cancer screening is not recommended for population-wide testing in either the US or UK .2,3 Current screening tests are effective for those cancers that they target, but adding an MCED blood test could potentially have a huge impact because they can target many cancer types (see Figure 2 in manuscript).4 MCED tests have some major features:

  • they can find types of cancers that do not have any current screening options
  • they can find breast, bowel, cervical and lung cancers that were missed by current screening tests, or were among people (of screening age) who are not eligible for current screening

Using current screening, an estimated 189,498 cancers (breast, bowel, cervical and lung; true positive in Figure 2) are detected in the US each year. But when an MCED test is used alongside current screening, up to an additional 422,105 people with >50 types of cancers could be detected (Figure 1). In the UK, 24,888 cancers (breast, bowel and cervical) are estimated to be detected annually using recommended screening (Figure 2), but up to an additional 92,817 cancers could be found each year (Figure 1).

Figure 1. Additional cancers that may be detected with MCED screening

What are other potential benefits of MCED screening tests?

As shown in Figure 1, a major strength of MCED tests is that they can pick up many cancer types beyond those recommended for single-cancer screening, and therefore potentially thousands more people who have cancer. There are other important outcome measures. 

One is screening efficiency, that is the number of people who end up with a confirmed cancer diagnosis after a positive screening test (true positive) versus those who are not found to have cancer after a positive screen result (false positive). For current screening in the US, the ratio of true positives to false positives is 1:43 (Figure 2). This means for every 1 diagnosed cancer, 43 people did not have cancer but they had a positive screening test result, and many of these would be offered diagnostic investigations including scans and biopsies. In the UK, this ratio is 1:18 (Figure 2).  In comparison, an MCED test is projected to have a true to false signal ratio of only 1:2 in both countries. This represents a striking difference.

Another important outcome is resource utilization, i.e., amount of money spent on investigating people who have a positive screening test result (cancer signal). The cost per cancer diagnosed is considerably less with an MCED test than with single-cancer screening because it can detect many more cancer cases but with a relatively small number of false positives.  

Figure 2. Screening test outcome measures

For the US and the UK, the top rows of this figure show, from left to right, the total number of people with a positive screening test result with current screening measures, the additional results with MCED testing (assuming 100% uptake), and the sum of these. The second rows show the number of people confirmed to have cancer after they have positive screening test results (true positive, TP). The last rows (TP:FP ratio) depict how many people without cancer (false positive, FP; light blue -) may be flagged for diagnostic investigations at a hospital or clinic in order to detect one person with cancer (TP; dark blue +), when they have a positive screening test.
aCurrent screening in the US for breast, colon, cervical, and lung cancers.
bCurrent screening in the UK for all cancer types. 

The top row of this figure shows, from left to right, the total number of people with a positive screening test result with current screening measures, the additional results with MCED testing (assuming 100% uptake), and the sum of these. The second row shows the number of people confirmed to have cancer after they have positive screening test results (true positive, TP). The last row (TP:FP ratio) depicts how many people without cancer (false positive, FP; light blue -) may be flagged for diagnostic investigations at a hospital or clinic in order to detect one person with cancer (TP; dark blue +), when they have a positive screening test. aCurrent screening in the US for breast, colon, cervical, and lung cancers. bCurrent screening in the UK for all cancer types. 

This big impact on screening efficiency and costs is achieved when the MCED test has a very low false-positive rate (a sort of error rate: that is, flagging people for cancer investigations because they had a positive test but they do not have cancer). This will be an essential feature of a good MCED test because a test might be able to identify many cancers, but if too many people without cancer are flagged, this could lead to an inefficient and expensive screening policy.

Our analyses indicate that by combining current screening policies with MCED testing of cancers without recommended screening, you can greatly increase the number of cancers detected without a large increase in the false positive rate, and at a reasonably low cost per cancer diagnosed. Given that the goal of oncology practitioners and public health is to give as many patients the best chance of surviving their cancer in the most efficient manner, this combination is likely to be the best approach.

Our BJC paper represents a preliminary first look at how an MCED test can be used on a population scale. Further research is needed to evaluate how well MCED tests detect advanced stage cancers and also whether screening can lead to fewer deaths from cancer. This information would then be used for more comprehensive health economic assessments.

What does an ideal test look like?

Figure 3. Criterion of a successful MCED

We cannot discuss cancer screening without acknowledging that healthcare resources are often not equitably distributed.  Several diseases that are treatable when caught early continue to be lethal in those without access to quality healthcare, and cancer is no exception. In the list of characteristics that comprise a successful MCED screening test (Figure 3), the first four are technical, but one critical point is that no matter how well the test performs once you have the blood sample, without wide uptake, you are not maximizing the potential of the test.  Tests that can inexpensively screen for multiple cancers (especially those cancers that are often lethal at late stage and do not currently have any screening tests available) with no more patient inconvenience than a simple blood draw, could help bridge the healthcare inequality gap in cancer screening. In fact, a test that has all of these attributes could have a major positive impact on the health of the populations of many countries. While we estimated potential outcomes for two high-income countries (US and UK) in our BJC paper, MCED tests may be even more impactful in middle-income countries where expensive cancer treatments put a greater strain on limited resources, and access to current screening programmes is variable because it requires people to attend a specialist clinical facility (except bowel cancer screening using a faecal sample test, and in the near future cervical cancer screening using self-sampling kits).

Healthcare providers and public health policy makers need the best data possible to make decisions about how to best utilize cancer screening tests.  The UK NHS, for example, has set a goal of having 75% of cancers diagnosed at an early stage by 2028.5 Innovations such as multi-cancer screening are critical for national goals like this.  For any new MCED test, a comprehensive evaluation will come from clinical trials and real world studies. To do this for the Galleri test, which is currently available in the United States, there is a large UK National Health Services randomised trial underway.6 We look forward to talking to you about those results when they are in!

REFERENCES

    1. Liu MC, Oxnard GR, Klein EA, Swanton C, Seiden MV, CCGA Consortium. Sensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA. Ann Oncol Off J Eur Soc Med Oncol. 2020;31(6):745-759. doi:10.1016/j.annonc.2020.02.011
    2. United States Preventive Services Task Force. USPSTF A and B Recommendations. Published 2019. https://www.uspreventiveservicestaskforce.org/Page/Name/uspstf-a-and-b-recommendations/
    3. UK National Health Service (NHS). NHS Screening. Published Updated 2020. https://www.nhs.uk/conditions/nhs-screening/
    4. Hackshaw A. Estimating the population health impact of a multi-cancer early detection genomic blood test to complement existing screening in the US and UK. Br J Cancer.:13.
    5. NHS England » Diagnosing cancer earlier and faster. Accessed July 28, 2021. https://www.england.nhs.uk/cancer/early-diagnosis/
    6. NHS England » NHS to pilot potentially revolutionary blood test that detects more than 50 cancers. Accessed August 19, 2021. https://www.england.nhs.uk/2020/11/nhs-to-pilot-potentially-revolutionary-blood-test/

Allan Hackshaw

Professor/ Director, University College London