You Get What You Pay For: Breast MRI Screening of Women With Dense Breasts Is Cost-effective

In this issue of the Journal, Geuzinge and colleagues demonstrate that in women with dense breasts, screening by magnetic resonance imaging (MRI) every 3 to 4 years is cost-effective compared with the current standard (ie, biannual mammographic screening) (1). The authors provide critical evidence for a simple fact: A more sensitive and more expensive diagnostic test is cost-effective, even if it is used for screening, compared with a less sensitive and less expensive test. Their publication thus provides long-awaited evidence on the economic practicability of the concept of risk-adjusted screening—and is therefore of pivotal importance.

Why is this finding so important? In the absence of effective methods of primary prevention (avoiding the development of breast cancer), secondary prevention (ie, early diagnosis) is the best means of improving outcomes for women with breast cancer. Screening for breast cancer helps by identifying breast cancer while it is still in an early stage (ie, before metastatic spread and/or clonal diversification challenge curative treatment) (2). Screening means to sieve, though: It is the classic case of the search for a needle in a large haystack. Although breast cancer is the most frequent type of cancer in women (3), at the time of a given screening round, only a small number of women—usually about 4 to 7 per 1000 women—will be diagnosed with cancer; the remaining 993 to 996 will be diagnosed as cancer free (4). Now, there is a simple truth regarding screening: Only the small fraction of individuals with the target condition—in our case, those with breast cancer—can, in principle, benefit from screening. All others—the vast majority of women who do not have breast cancer—cannot benefit but can expect only a disadvantage, mainly through false-positive screening findings. This situation is unlike that in therapeutic settings, where treatment is delivered to patients who all have the target condition. In therapeutic settings, the number of individuals who can, in principle, benefit from an intervention is therefore by several orders of magnitude larger when we compare it with screening. Screening is thus far from being a “cost-effective” intervention.

In this regard, the concept of screening is similar to that of many insurances. A home insurance policy, for instance, does not prevent a house from being destroyed by fire or flood—just as screening does not prevent breast cancer. In the unlikely event of one’s home burning down (just as in the unlikely event of a woman developing breast cancer), however, having paid for a home insurance policy (having undergone an efficient screening test for early diagnosis) is an effective way to alleviate the disastrous consequences of that event. People buy home insurance policies in full awareness of the fact that, in all likelihood, their payments will be in vain because their house will never burn down. Similarly, women may decide to invest in screening fully aware that, hopefully and likely, she will never develop breast cancer. In other words, people pay for insurance (or screening) even though it will likely not be cost-effective.

Unfortunately, the net cost-effectiveness of screening is further reduced for similar reasons that modulate cost-effectiveness of therapeutic interventions: Treatment may not work in an individual patient. In fact, the treatment may even cause additional morbidity rather than providing the expected benefit. Similarly, not every woman of the already small fraction of women who do have breast cancer at the time of the screening examination will necessarily benefit from it. For one, not all cancers that are detected by mammographic screening should be diagnosed (“overdiagnosis”) (5). For another, not all cancers that should be diagnosed are detected (“underdiagnosis”) (4,6).

Overdiagnosis means true-positive detection of cancer but of cancers that, if left undiagnosed and thus untreated, would not cause clinical symptoms or premature death. About 12% of screen-detected malignancies, mainly ductal carcinoma in situ, belong to this group (7). Treatment of “overdiagnosed disease” may result in costs and side effects without providing benefit. To avoid such overtreatment, or at least to mitigate its negative consequences, cancer biology is now used to guide the treatment of every woman who has breast cancer, and substantial research efforts are directed to further deescalate treatment in women who have less aggressive disease.

Underdiagnosis means failure to find life-threatening cancer early enough to prevent progression to an uncurable stage and thus to suffering and premature death. In quality-assured mammographic screening programs, between 16% and 47% of cancers are diagnosed as so-called “interval cancer” in the time after a negative screening round (4). Plus, another 20% to 45% of cancers that are screen detected are already in a locally advanced or metastatic stage (Union for International Cancer Control stage II or higher) (8,9). Hence, well over half and up to two-thirds of women who do have breast cancer at the time of screening will not benefit from it because their cancer is either not detected at all or not detected early enough. The substantial degree of underdiagnosis associated with mammographic screening may contribute to the fact that breast cancer is still a or the major cause of cancer death in women, despite mammographic screening programs being established and well attended for several decades now.

The bottom line is that only women with breast cancer can benefit from screening, and even these few women may not benefit because of over- and underdiagnosis. Does this mean that we should abandon screening?

In the absence of effective methods to cure metastatic breast cancer—in particular, metastatic luminal cancer—likely not. It does, however, indicate that current breast cancer screening methods are far from perfect.

Given the high rate of underdiagnosis associated with mammographic screening and in view of the severe medical, societal, and financial implications of underdiagnosis, further research must concentrate on avoiding underdiagnosis. We need screening methods that keep their “value proposition,” which means that they should reliably identify prognostically relevant cancer early enough. This is what breast MRI does in women with dense breasts. The DENSE trial provided level 1 evidence that MRI screening effectively avoids underdiagnosis, as measured by interval cancer rates observed in women who underwent MRI vs mammographic screening alone (10).

Thus, the DENSE trial concentrated on women known to be underserved by current mammographic screening—that is, women with extremely dense breasts. This is paradigmatic: To combat underdiagnosis of breast cancer, we must move beyond the current one-size-fits-all approach, where everybody undergoes the same test (mammography) from the same age onwards in the same intervals until the same age. Instead, methods for risk prediction should be used to tailor screening to the individual woman’s risk (10). Such risk-adjusted, tailored screening should replace the current scattergun approach. Such risk-adjusted approaches would also enable us to open up resources in the subpopulation of women who may go along with less or even no screening at all and reallocate the resources to those who need them.

Offering preventive measures, such as cancer screening, at no cost is a societal decision, and it is a matter of political prioritizing of health care resources to decide how large this investment should be. As explained earlier, the investment has a return not so much in monetary terms but in the well-being of individuals and families. Most wealthy countries, therefore, offer breast cancer screening either through publicly organized screening programs based on invitation or by systematic education of women (2).

Despite the cost-effectiveness established by the current paper by Geuzinge et al. (1), many societies will not be able or willing to offer breast MRI for women with dense breasts. Where public health systems do not cover the additional cost for breast MRI, women who are at risk of being underserved by mammographic screening may decide to pay the costs for MRI screening out of their own pocket. Based on the results by Geuzinge et al., this would amount to direct costs of about €68 per year. Despite being a limited amount of money, this may increase current health care disparities.

Does that mean that we should not recommend MRI for screening until it is covered by public health care systems?

Health care priorities differ between countries and societies as well as between individuals. Fair and objective information about a woman’s personal risk and screening options is essential to enable individual women to make their own, educated choices. For women at risk of being underserved by mammographic screening, greater investment in breast cancer screening can be considered a useful investment in one’s bodily health.

A screening test that predictably will not provide the expected diagnostic information is not only painfully disappointing for those who participated and still receive a late diagnosis of cancer but is a waste of money. In women with dense breasts, undergoing mammography for screening is similar to having a home insurance policy that, in case of claims, would not pay. To put it with the “common law of business balance” frequently attributed to J. Ruskin, “When you pay too much, you lose a little money—that is all. When you pay too little, you sometimes lose everything, because the thing you bought was incapable of doing the thing it was bought to do.”

Funding

None.

Notes

Role of the funder: Not applicable.

Disclosures: The authors have no disclosures.

Author contributions: Writing—original draft: CK; writing—review and editing: CK, PB.

Data Availability

Not applicable.