The Paradox of “Precision Medicine” When Applied to Breast Cancer Screening

The term “precision medicine” is so overused today that it suffers the same fate as any buzz phrase that is tossed about carelessly – loss of impact. Whether you call it “precision medicine” or “personalized medicine,” the intent is the same – to customize medical care for the individual. Effective treatments are applied only to those who benefit. Who can argue with that?

In its purest sense, it’s the ideal approach to medicine. Imagine a patient newly diagnosed with cancer whose tumor is analyzed at the molecular level, whereupon the exact combination of therapeutic agents designed for the tumor’s particular profile is utilized, and the cancer is eradicated. Great strides are being made in that direction.

Paradoxically, however, when it comes to breast cancer screening, the use of precision medicine to justify doing less benefits society at the expense of the individual. This is not the alleged intent of precision medicine; it’s the exact opposite.  Precision medicine is supposed to benefit the individual.  Let’s see how so-called precision works when it’s applied to the justification to do less screening for the early diagnosis of breast cancer. This is not hypothetical, by the way. The term “precision medicine” is often used in the defense of new breast cancer screening guidelines where only those women at increased risk are to consider mammography in their 40s.

The primary care physician, caught in the storm of controversy, is currently being pressured to say this: “Since you don’t have a family history for breast cancer, I’m going to recommend waiting until you’re age 50 to begin mammographic screening.” Or, this: “Since your mother had breast cancer, I’m going to recommend that you begin screening at age 40.” It sounds perfectly reasonable, but the logic is deeply flawed if one understands the numbers behind these recommendations.

For the decade of the 40s, the general incidence of breast cancer is only 1.5% over the course of these 10 years. That is, only 15 women out of 1,000 are going to develop breast cancer during their 40s (compared to 25/1,000 in their 50s, and 35/1,000 in their 60s). In order to improve cancer detection rates, the recommendation has been made to selectively screen only those women with risk factors.

However, the majority of women who develop breast cancer have no identifiable risk factors, so this logic is flawed from the git-go, even before we get to actual numbers.

This selective approach for breast cancer is to be sharply distinguished from lung cancer where “precision” screening using chest CT works quite well because of the tight correlation of smoking to the disease being addressed through screening. In lung cancer, 80% of patients have a smoking history. But more importantly, the level of lung cancer risk is 20-fold in smokers over the general population. Thus, smokers are advised to screen, while non-smokers are advised to do less, that is, nothing. Inevitably, there are those who believe we should adopt the same approach for all cancer types – screen only those at risk.

But breast cancer is not even in the same ballpark. Instead of 80% who have risk factors as in lung cancer, only 20-25% of newly diagnosed breast cancer patients have a positive family history. So, our target population is far weaker than what is seen in lung cancer. But it gets worse – the power of risk due to family history in breast cancer is often negligible. Rather than the 20-fold risk of smoking and lung cancer, the usual risk seen for breast cancer (with one first-degree relative with the disease) is more in the range of 2-fold, or one-tenth the power of smoking and lung cancer risk.

With the exception of patients who have multiple family members affected by breast cancer (with or without testing positive for a breast cancer predisposition gene), the usual “high risk” patient for breast cancer is not strikingly elevated above and beyond baseline risk. Again, unlike lung cancer where the baseline risk is low (in non-smokers), the baseline risk for breast cancer is high for all women.


Here’s how the numbers work in the decade of the 40s:

Risk of breast cancer over 10 years if there are no identifiable risk factors – 1%

Risk of breast cancer over 10 years in the general population of women – 1.5%

Risk of breast cancer over 10 years with one first-degree relative with breast cancer – 2%


Do we really limit screening to the group with 2% 10-year risk and ignore the women at 1%, understanding that this maneuver will exclude the majority of women with breast cancer? Certainly, the yield improves when you limit screening to higher risk patients, but “cancer yields” don’t tell you how many women were excluded to get the boost from 1% to 2%. Cancer Detection Rates (CDRs), or cancer yields, can be very misleading.

The cost savings with “precision medicine” in this instance is staggering. If we take 100,000 women in the general population at baseline risk, we will need to perform one million mammograms over the course of 10 years to identify 1,500 cancers. But if we limit screening to only those women with a positive family history, we will need to perform only 200,000 mammograms over 10 years rather than 1,000,000. And, our cancer yields go up! 2% rather than 1.5%. Great. But we will only find 400 cancers (2% of 200,000), while missing the majority of breast cancers with this “precision” approach. Had we simply screened everyone, “without” precision, we would have encountered 1,500 cancers, instead of only 400.

Actual breast cancer deaths are calculated differently, and require taking into account cancers missed by mammography, but a reasonable estimate is that there will be an additional 1,000 breast cancer deaths every year in the U.S. if we restrict screening to a starting age of 50.

In a roundtable committee on screening, I once stated that women in their 30s were already disenfranchised from early diagnosis of breast cancer (5% of eventual breast cancer patients), but to start screening at 50, we would disenfranchise a much larger group of women in their 40s (an additional 20% of eventual breast cancer patients). I was challenged by a colleague who said, “These women are not disenfranchised, they have physical exam.” Wow. Clinical exam usually detects Stage II disease (or worse) in this population, and there’s not a shred of evidence that lives are saved. Yet, the evidence is strong that mammographic screening lowers mortality in the 40s, even if the benefit is not as great as later in life (even the Task Force admits to a 15% relative reduction in mortality for women screening in their 40s).

When it comes to breast cancer screening for women in their 40s, the admonition to do less in the name of precision medicine benefits society by lowering health care costs, but does little for the individual and could cause substantial harm, the exact opposite of what “precision medicine” claims to be.

For a detailed look at the fallacy of risk-based screening as used to justify doing less, refer to my book — Mammography and Early Breast Cancer Detection, Chapter 20 — Risk-based Screening — It Feels So Right, But Wait…