Drs. Smith, Duffy, and Tabár provide an excellent history of breast cancer screening, including a summary of the randomized trials, the technical approaches, and the range of recommendations for screening. I would only add the perspective that our improved understanding of breast cancer biology should also inform our recommendations on screening frequency and improve our ability to discern where screening has the most benefit.
Advances in molecular biology have enabled us to clearly establish that breast cancer is not just one disease. Our treatments have dramatically changed and now focus on biologic type as much as stage. We need to approach screening with this same framework.
There are many ways to define breast cancer subtypes, such as the examples shown in the table. Tools such as the 70-gene profile separate tumors into low and high risk for recurrence, and help identify women with a low risk for recurrence in the first 5 years after diagnosis, in the absence of any systemic therapy. Importantly, there is the potential to use this tool to define an “ultra-low” risk threshold that might correspond to what has been termed IDLE (InDolent Lesions of Epithelial origin) tumors, which are tumors not destined to progress to metastatic disease. A study from the RASTER (MicroarRAy PrognoSTics in Breast CancER) screened cohort in the Netherlands, in which almost all tumors were molecularly profiled, suggests that the frequency of extremely-low-risk tumors may be as high as 30% of screen-detected cancers.
The standard of care today includes systemic treatment of cancers with hormonal therapy and chemotherapy, based on biologic features such as receptor status, proliferation, and other features as measured by multigene tests, based on their risk for systemic disease. Targeted treatments for HER2-positive cancers, for example, have had the greatest impact on lowering mortality from these tumor types. The impact of screening will necessarily be less because routine care now includes assessment of biology and the use of adjuvant systemic therapy, accordingly. Also, small tumor size will not necessarily preclude treatment, as biology trumps stage. The screening trials were done prior to standard adjuvant interventions, and it is estimated that two-thirds of the benefits previously ascribed to screening are due to the use of systemic therapy.[4,5] It is also important to remember that the tumors that pose the greatest risk for early recurrence often present as interval cancers, which arise between screens. Here, reduction in mortality will come from improved systemic therapy or from our ability to predict populations at risk and tailor prevention and screening accordingly.[6-9]
There is scant evidence, if any, that annual screening is better, as the RCTs were performed with biennial or triennial screening. Additionally, there may be evidence of harm. The biopsy rate is 50% higher with annual screening (increased chance of a false-positive result), and there is no statistically significant increase in the proportion of stage 2 or higher cancers.[10,11] Drs. Smith, Duffy, and Tabár provide the evidence that patients most likely to benefit from screening are those who have the slower-growing cancers with a trajectory for death after 10 years, noting that, “In fact, most of the breast cancer deaths prevented occurred 10 years or longer after the inception of screening.” These are clearly not the aggressive HER2-positive or triple-negative breast cancers, nor even the very-high-risk HR-positive ones, but the more slow-growing HR-positive tumors. Others have also shown that the types of tumors found by screening are more likely slower-growing tumors. That is why screening every other year or every 3 years, as was done in the Swedish trials, was effective. Indeed, it is true that young women are much more likely to have fast-growing tumors. But being in one’s 40s is not a sufficient risk factor, as the absolute risk of getting these cancers is quite low and thus will lower the positive predictive value of screening, as the authors point out. The additional question that remains is whether screening will have a significant impact on these tumors, even if screening is every 6 months.
The authors recognize the potential problem of overdiagnosis, but they doubt that it is truly a significant problem or that we have the capacity to discriminate indolent from more lethal tumors. I would submit that we have that capacity and need to put the effort into validating predictors already on the market or which are emerging as commercial tests. Importantly, we need to recognize that with age, the likelihood of finding a low-risk tumor increases dramatically, which is why screening and early detection may not make much difference in women over the age of 75, even if they do not have significant comorbidities.
Although screening techniques have improved, the ability to find more lesions has also increased. In the US especially, we have dramatically increased the frequency of performing biopsies and surfacing premalignant lesions of uncertain potential for progression.[14,15] Not all of these lesions merit a biopsy. Indeed, there is tremendous variation in the rates of biopsy, so clearly there is room for improvement. The scientific community must now step up to the plate to determine a better definition of cancer—not just one based on the presence of what appear to be cancerous cells based on a light microscopy evaluation but one based on likelihood of invasive or metastatic spread—and the time course for that risk. This type of approach will now be a major focus of activity for the National Cancer Institute, and it will improve screening by enabling radiologists to focus on detecting those lesions that truly have the potential for harm, thereby avoiding the anxiety, morbidity, and cost associated with biopsies that turn out to be either benign or of uncertain malignant potential over many years.
Screening has an impact today, but it is not as great as we would like and the cost and resources are very significant (in billions of dollars). Given the types of tumors that screening is most likely to impact, the United States Preventive Services Task Force (USPFTF) guidelines enable the most benefit with the least harm. This approach is standard in all countries outside the US.
I think we can all agree that the real opportunity going forward is to find a way to improve our approach to screening and to focus on prevention as well, based on susceptibility to breast cancer type. There is accumulating evidence that higher breast density, for example, appears to be a risk factor for higher-grade tumors, and a number of investigators have proposed to tailor screening frequency based on specific risk factors. Such approaches should be encouraged and studied. We need to work together, going forward, to determine where screening and prevention are likely to have the most impact, for which population, and at what frequency, and to understand how we can use emerging molecular tools to build better risk models and avoid harm from overtreatment.
Financial Disclosure: The author has no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.