Supplemental Breast Cancer Screening: Unmasking the Truth
This article contains several of my own opinions to foster interest and further learning on the important topic of supplemental breast cancer screening. I link to several full-text manuscripts for further reference that may be behind firewalls or otherwise require paid access to medical journals. This is a summary of my own thoughts on supplemental breast cancer screening and nearly all references cited comprise my own publications as a primary author from peer-reviewed publications. Many other useful references exist. I encourage the reader to seek these out on PubMed or a similar database.
Do you prefer to listen to this as a podcast episode? If so, click here.
The purpose of screening mammography is to lower breast cancer mortality by detecting non-palpable breast cancers at an early stage. Through this mechanism screening mammography lowers breast cancer mortality by about half. While laudable, I remain disconcerted that the mortality reduction is not much higher.
Short of a silver bullet cure for metastatic breast cancer, can we achieve close to 100% reduction in breast cancer mortality using current screening technologies? A careful re-assessment and re-implementation of current screening approaches may get us much closer. If this topic interests you, read on.
Challenges with current Mammography screening paradigms
To understand how we can improve, we first must understand some of the primary limitations of our current beast cancer screening approaches.
Dense Breast Tissue
A screening mammogram does not benefit every person equally. The success of a screening mammogram to detect invasive breast cancer early—meaning cancer that is confined to the breast and small—mostly varies based on breast density. As the density of breast tissue increases, the likelihood that a mammogram will detect cancer decreases.
The good news is that mammography performs very well in breasts that are non-dense. About 60% of the population undergoing screening mammography has non-dense breast tissue. On the other hand, mammography has lower sensitivity for cancer detection for the other 40% of the population who have dense breast tissue. Furthermore, these 40% of individuals with dense breast tissue are also more likely to develop breast cancer compared to individuals with non-dense breasts.
This creates the unfortunate situation where individuals with dense breast tissue are more likely to develop breast cancer and are also more likely to have the cancer go undetected on a screening mammogram.
To see illustrations, and learn more about this issue—including a pictorial example of how breast density can obscure a cancer (created by yours truly)—see this page at mydensitymatters.org.
If you have dense breast tissue and your mammogram is reported as “negative” this means that no breast cancer was seen on the mammogram, but does not mean that no breast cancer is present.
Radiologist Apathy
Radiologists performing breast imaging are simply taking too long to collectively accept that too many breast cancers are going undetected in dense breasts on screening mammography. Despite unequivocal evidence that mammography frequently fails to detect breast cancers in dense breasts many radiologists continue to believe that mammography is sufficient as a stand-alone test to screen individuals with dense breast tissue.
If you think this is illogical, I would agree.
Although radiologists in breast imaging want to save as many lives as possible through screening, too many nonetheless oppose or otherwise fail to implement changes to historical screening paradigms to make this a reality. It is sadly true that screening techniques other than mammography continue to meet resistance in the breast imaging community. That any radiologist would limit access to robust breast cancer screening regardless of dense breast tissue (whether intentional or not) is very sad and reflects poorly on my field.
Population Disparities from Screening Mammography
If you are sitting in a room with 100 women or others who have had a recent screening mammogram, those approximately 60 with non-dense breasts will fare better in terms of beast cancer reduction in comparison to the other 40 with dense breasts who are at risk to have a breast cancer go undetected despite participating in screening. Given this 60/40 split between non-dense and dense breasts, you personally know many people in both groups.
This meets the definition of a healthcare disparity: some people fare better than others based on health care systems and procedures. Some win and others lose. And the distribution of winners and losers is typically not random. Given that breast density varies by age (younger women are more likely to have dense breast tissue) and race (dense breast tissue is more common in certain racial groups including people who are African American and Asian women). It is sad but true that certain age and racial groups are more likely to have a breast cancer go undiagnosed on a screening mammogram. This is an ethically troublesome issue that deserves quick action to resolve.
The Magnitude of the Problem
How many cancers are we missing on mammography due to dense breast tissue? I estimated in a recent publication that 267,000 breast cancers likely went undetected on screening mammography in 2021 in the U.S. due to dense breast tissue on mammography (click for article). This is roughly equivalent to the total number of breast cancers that were detected in that same year. In essence, we are possibly detecting as many breast cancers through screening mammography as we are missing. To me, this hints at a possible major reason why screening mammography has lowered breast cancer mortality by around 50%.
As a breast radiologist myself, I do not understand why my field has been slow to solve the issue of dense breast tissue limiting detection of breast cancer on screening mammography. Certainly, progress has been made as a result of the individual or collective effort of many breast radiologists. There are also many breast radiologists who have long been passionate about this issue and are dedicating their careers to solve this problem. Nonetheless, our field has done far too little at this point to implement actionable solutions to this problem. Many of the changes that have been enacted to tackle this issue started and continue to be mostly driven by individuals and groups external to breast radiology. More on this later.
I cringe whenever I hear a breast imager say something like “out of all the screening options available, if you have extremely dense breast tissue, and are not at high risk of breast cancer, mammography with tomosynthesis remains your best option”. The assertion that mammography is the best stand-alone screening option for individuals with extremely dense breast tissue not at high-lifetime risk of breast cancer is scientifically questionable in my opinion. Indeed, an objective observer of available science is unlikely to reach the conclusion that a mammogram is the best stand-alone screening test for a person with extremely dense breast tissue. Nonetheless, even the American College of Radiology Appropriateness Criteria continues to voice the opinion that tomosynthesis (3D mammography) is the most appropriate exam for supplemental screening of individuals with dense breast tissue not at high-lifetime risk of breast cancer. I have been a vocal critic of this assertion (click here to read more).
When a large problem with a status quo medical approach becomes manifest, I believe physicians should earnestly look for new or alternative approaches to solve the problem. Various potential solutions already exist to detect the breast cancers in dense breasts that mammography misses. These solutions include policy changes for dense breast tissue screening, application of alternative advanced screening technologies, and use of artificial intelligence for risk prediction and aiding imaging interpretation.
Potential Solution: Legislative Action
Dense Breast Tissue Notification
Nancy Capello, PhD, was a prominent advocate for breast density notification and breast cancer awareness. She became an influential figure in the breast cancer community following her personal experience with missed breast cancer from screening mammography due to dense breast tissue and her subsequent efforts to raise awareness about breast density.
In 2004, Nancy Capello founded the organization "Are You Dense?" with the goal of educating society about breast density and advocating for legislation that would require healthcare providers to inform individuals participating in screening mammography about their breast density following screening mammography. The organization aimed to raise awareness about the importance of supplemental screening for individuals with dense breasts and to empower them with information to make informed decisions about their breast health.
Sadly, Nancy Capello passed away in 2019, but her legacy continues through the work of "Are You Dense?" and other organizations that strive to improve breast cancer awareness and detection. Through her work, she has inspired countless individuals to take charge of their breast health and push for improvements in breast cancer screening practices. Indeed, many additional champions of dense breast tissue legislation have been found through grass-roots efforts in local communities. Although the concept of dense breast tissue notification initially found strong resistance from the radiology community, many radiologists now advocate for legislative adoption of notification laws, which now exist in nearly all U.S. states.
These laws require healthcare providers to inform individuals participating in screening mammography if they have dense breast tissue and to provide information about additional screening options that may be more effective for detecting breast cancer in dense breasts. It is important to note that the specific details and requirements of dense breast tissue notification laws can vary between jurisdictions. Federal law will soon establish dense-breast notification standards throughout the United States.
That such laws exist at all speaks to the failure of the breast imaging community to adequately address this problem on their own. Indeed, after Dr. Capello’s initial breast cancer diagnosis, she was dismayed to learn that radiologists understood that mammography had limited sensitivity for cancer detection for people with dense breasts, yet did not routinely explain this to those with dense breasts undergoing screening mammography. This effectively deprived individuals such as Dr. Capello of the chance to pursue additional screening options until it was too late, and cancer had already spread.
Finding that the breast imaging community was unwilling to embrace increased transparency and patient education regarding this issue, Dr. Capello turned to her state legislature to enforce disclosure to empower others with dense breasts to seek additional screening options. This legislation was initially implemented to essentially force radiologists to do the right thing and be transparent about the limitations of screening mammography for those with dense breasts.
While breast density notification is an important step toward awareness of the limitations of screening mammography for those with dense breasts, notification itself does not detect cancer or lower breast cancer mortality. No additional lives will be saved unless other screening strategies are made available to those with dense breasts.
Therefore, notification and awareness of the limitations of screening mammography for those with dense breasts is only step one. Step two requires having robust supplemental screening options available and accessible for those with dense breast tissue. This is currently not happening in many communities.
Mandatory Insurance Coverage for Supplemental Screening:
Various U.S. states have enacted laws that private and public insurers must cover certain supplemental screening tests which most commonly include whole breast ultrasound and breast MRI.
Currently, the Find it Early Act has been proposed to require insurance coverage nationwide for supplemental screening examinations. If enacted, this Find it Early Act could provide an enormous boost to supplemental screening access and subsequent use throughout the United States.
Potential Solution: Supplemental Screening with Advanced Imaging
Background
Too many breast imaging practices continue to offer only mammography with tomosynthesis to screen individuals with dense breasts. Even if woman with dense breasts were to request additional screening beyond tomosynthesis, many breast imaging practices simply do not have other technologies to offer. This needs to change if we are to ever solve this problem.
Imaging technologies, such as contrast-enhanced breast MRI, molecular breast imaging (MBI), contrast-enhanced mammography (CEM), contrast-enhanced breast CT, and whole breast ultrasound (hand-held and automated) can detect cancers in dense breast tissue that mammography cannot. Use of these technologies has been termed supplemental screening, as they supplement the performance of standard mammography to boost cancer detection.
There is wide variation in use and availability of supplemental screening technologies at breast imaging centers throughout the U.S. Various breast imagers, breast imaging practices, and even particular geographic regions may embrace one or more of these technologies as their preferred method(s) and are thereafter quick to point out the shortfalls of imaging techniques they are not using. I have personally been at fault of this technological bias.
I am trying to personally leave this technology bias behind because physicians should be objective observers of scientific data. This requires that we set biases and preconceived notions aside. This is difficult to do.
Informed readers may notice that I omitted tomosynthesis (3D mammography) as a supplemental screening option. I hope this is not a result of my own bias, but rather based on objective review of published scientific data. I think tomosynthesis is a good primary screening study (meaning your initial screening mammogram), but I question tomosynthesis for supplemental screening. Let me explain why.
Incremental Cancer Detection Rates
The incremental cancer detection rate (ICDR) is a metric that is helpful to compare supplemental screening options. The ICDR assesses the effectiveness of a specific screening modality or technique to detect additional cancers beyond what was already detected by standard mammography alone. The ICDR is an important metric because it helps healthcare providers and policymakers assess the value of adding supplemental screening methods to the standard mammography approach. A higher ICDR suggests that the supplemental screening modality is more effective in detecting additional cancers and may be beneficial for screening dense breasts.
The ICDR is calculated by comparing the number of cancers detected by a supplemental screening method (such as whole breast ultrasound or MRI) in addition to those detected by mammography alone, divided by the number of people screened. It represents the incremental number of cancers identified per 1,000 individuals screened using a specific supplemental screening method. For example, if 2 additional cancers were detected using ultrasound in a group of 1,000 individuals who had already undergone mammography, the ICDR would be 2 per 1,000, indicating that two more cancers were detected beyond what mammography alone would have found.
It is worth noting that the ICDR can vary depending on factors such as the population being screened, the characteristics of the screening modality, and the expertise of the healthcare providers performing the screening. Therefore, the ICDR should be interpreted in conjunction with other factors, including cost-effectiveness, potential harms or false positives, and individual patient preferences when determining the optimal breast cancer screening strategy for an individual or population.
Although ranges and variability in estimates exist, these numbers are representative ICDRs for current leading supplemental screening options:
Whole breast ultrasound: 3 cancers per 1,000 screened
Molecular breast imaging: 9 cancers per 1,000 screened
Contrast-enhanced mammography: 13 cancers per 1,000 screened
Contrast-enhanced breast MRI: 16 cancers per 1,000 screened
and
Tomosynthsis: 1 cancer per 1,000 screened
Tomosynthesis: a supplemental screening exam?
If one considered tomosynthesis as a supplemental screening method (which I personally do not) the ICDR would be about 1 per 1,000 screened, recently confirmed in this study of over over 1 million women. However, current ACR Appropriateness Criteria Recommendations state that tomosynthesis is the ONLY examination that is “usually appropriate” for dense breast supplemental screening for individuals with dense breast tissue not at high-risk for breast cancer (link here).
The primary problem in considering tomosynthesis as a supplemental screening method, is that tomosynthesis leaves far too many cancers undetected because it, like standard 2D mammography, struggles to identify breast cancers that are masked in dense breast tissue. That is the entire point of supplemental screening—to find those cancers that mammography cannot. Therefore, although tomosynthesis may detect one additional cancer per 1,000 exams, it will not detect the other 15 cancers predicted to exist in this same group of 1,000 women based on the ICDR’s of breast MRI. Therefore, tomosynthesis as a stand-alone supplemental screening exam is simply not sufficient.
I agree that tomosynthesis is the preferred method for initial screening mammography for most individuals, but tomosynthesis has an unacceptably low ICDR to be considered appropriate for supplemental screening. In other words, a good screening strategy includes tomosynthesis at time of initial screening mammography to be followed by a robust supplemental screening option beyond tomosynthesis for those with dense breasts.
Calling tomosynthesis a supplemental screening exam provides false assurance to patients and providers that robust supplemental screening has been provided if tomosynthesis was performed at initial screening. This is clearly not the case when one considers the extremely low ICDR of tomosynthesis in comparison to other imaging technologies for supplemental screening. In summary, tomosynthesis is a powerful primary screening exam, especially for those without dene breast tissue, but frequently struggles, similar to standard mammography, to identify breast cancers in those with dense breast tissue.
Whole Breast Ultrasound:
The most widely available test for supplemental screening (assuming my assertion that tomosynthesis is not a supplemental screening exam) is currently whole breast ultrasound, whether hand-held or automated. From an ICDR perspective this is disappointing because ICDRs are much lower than other options such as CEM, MBI and breast MRI.
Data shows that whole breast ultrasound will detect fewer than half of all breast cancers that may be present in those with dense breast tissue and normal mammography. Whole breast ultrasound is likely to detect approximately one-fourth as many breast cancers as breast MRI.
What are advantages of whole breast ultrasound for supplemental screening? The biggest advantage is availability and access. Others would include that ultrasound is an extremely safe imaging study. Ultrasound is a very safe technology to implement in that it does not directly harm bodily tissues through imaging. However, I would caution the reader to not conclude that ultrasound is the safest supplemental screening method. Ultrasound will miss many breast cancers that are detectable by other methods, and this carries risks of morbidity and mortality that must be considered.
My current opinion on whole breast ultrasound may be summed up as follows: Whole breast ultrasound is better than no whole breast ultrasound in terms of improving breast cancer detection in those with dense breasts. However, if alternative supplemental screening options with higher ICDRs are available, why would you not choose those instead?
Molecular Breast Imaging:
Molecular breast imaging (MBI) is frequently undervalued by breast imagers, as well as the American College of Radiology, but has tremendous potential for supplemental breast screening. Many (if not most) breast imagers view this exam as unsafe due to radiation exposure. Many are also ignorant of the literature that exists for MBI and take community hearsay as truth. Existing research shows that MBI is likely one of the safest of all supplemental screening options we have in terms of averting deaths from breast cancer—even when accounting for the maximal possible theoretical risks from radiation exposure.
MBI involves the intravenous (IV) injection of a radiotracer called Technetium-99m sestamibi for subsequent imaging with a specialized nuclear medicine breast imaging camera. This IV injection of radiotracer circulates through the body and accumulates in areas that have a high density of mitochondria and increased blood flow—both of which are commonly found at sites of breast cancer. This radioactivity can then be detected by specialized, highly sensitive detectors. MBI is not impaired by dense breast tissue and is therefore adept at revealing breast cancers that mammography cannot. Current MBI scanners require a small fraction of the radiation that was necessary for prior generation breast-specific gamma imaging (BSGI).
To evaluate whether an exam is worth the radiation exposure from current techniques, one can calculate benefit-to-radiation risk ratios which I have published with Dr. Brown. These calculations look at how many deaths would be averted by using a technology of interest, compares this to how many deaths could result from radiation exposure received at imaging, and thereafter informs whether the benefit of imaging exceeds, equals, or falls below the risks of radiation.
From our calculations—and those performed by other investigators previously—MBI has a net benefit that exceeds risk by 5-times or greater (depending on age). My calculations with Dr. Brown estimate that MBI would essentially double the net-deaths averted compared to screening mammography alone, even when accounting for radiation risks (See Figure from our peer-reviewed publication in Radiology:Imaging Cancer by clicking here).
Knowing that all published benefit-to-risk ratios show net-benefit from use of MBI, I continue to be astonished when breast radiologists advocate that this technology should not be used because the radiation risks are too high. We should seriously consider using any technology wherein the net deaths averted is positive. A positive benefit-to-risk ratio literally means that more people would be alive if the technology were used versus not using that technology.
Many advocate using whole breast ultrasound as the “safer” alternative supplemental screening. However, whole breast ultrasound detects around 3 additional cancers per 1,000 studies compared to about 9 for MBI. How is non-detection of 6 very real cancers by choosing ultrasound instead of MBI safer? Even when considering small theoretical risks of future cancer induction from MBI that may or may not exist (we aren’t sure these actually exist at MBI exposure levels), I favor detecting more cancers now compared to avoiding fewer theoretical cancers later.
However, in attempt to avoid a smaller number of theoretical cancers that may or may not exist decades down the road from MBI radiation exposure, breast imagers and the American College of Radiology have effectively discouraged adoption of MBI—a powerful tool in our fight for early detection of breast cancer. This has been a great personal source of frustration to me. Read my various letters to the Journal of the American College of Radiology and the Journal of Breast Imaging on this topic by clicking here, here, and here.
It is not acceptable to fail to detect a breast cancer that exists today based on our current understanding of MBI radiation risks wherein benefit exceeds risk. It is also not acceptable to judge any imaging test merely by its radiation risks. If radiation exposure, regardless of diagnostic benefit, were the primary metric to judge the value of an imaging test, a chest x-ray would be considered a generally superior examination to a chest CT which is preposterous for many clinical uses.
What are the primary limitations of MBI? First and foremost is limited accessibility of MBI for patient use. The exam is also lengthy to perform though generally well tolerated as a patient may sit during imaging and only light breast compression is utilized.
While imaging acquisition is slightly longer than a whole breast ultrasound study and comparable to many contrast-enhanced breast MRI protocols, imaging interpretation is frequently very rapid and relatively simple. Most cancers detected by MBI tend to be obvious on imaging, and are invasive node-negative cancers which are the cancers we most want to detect to avoid morbidity and mortality. Finally, the negative predictive value of MBI is very high. Therefore, a negative MBI study not only means that no cancer was seen, but also that no cancer is likely to be present.
Contrast Enhanced Mammography:
Contrast enhanced mammography (CEM) is one of the most promising exams for supplemental breast cancer screening. Given that CEM has been FDA approved for around a decade for diagnostic breast imaging, it is frustrating that CEM use for supplemental screening is currently limited. It is a lost opportunity for breast imaging to have not implemented CEM much earlier for use as a supplemental screening exam.
The rationale for CEM is this: Mammogram + Contrast = Improved Cancer Detection. We have known for decades that intravenous contrast aids cancer detection on CT scans. The same principle holds true for CEM. CEM also uses a dual-energy subtraction technique that removes dense breast tissue from images. Therefore, the expanded formula for CEM: Mammogram + Contrast - Dense Breast Tissue = Robust Cancer Detection. CEM makes many cancers extremely easy to see even if one has extremely dense breast tissue.
CEM and tomosynthesis were both approved by the FDA around the same time. Tomosynthesis was paraded down the Main Street of breast imaging to great applause, spurred on by effective marketing and industry investment. CEM was essentially left parked in the garage, remaining largely out of public view or clinical use.
One notable exception is at the Mayo Clinic in Arizona where I completed breast imaging fellowship. Credit to the Mayo Arizona breast imaging faculty that saw the promise of CEM early because they have reaped the benefits of frequent utilization of CEM ever since. Since leaving Mayo Arizona many years ago I have helped in the implementation of CEM at two institutions, and I have lectured multiple times on CEM implementation at RSNA. (For disclosure, I previously consulted for Hologic for CEM, but have no current financial relationship with Hologic; my institution does have research support for CEM through Fuji). For those interested in CEM implementation, click here and here for my detailed thoughts on this topic.
If I were to select the top reason why CEM has yet to find widespread adoption in the United States it is reimbursement. In short, CEM is currently not a money maker for radiology practices. Despite imaging capabilities that far exceed standard mammography, CEM typically reimburses at the rate of a standard diagnostic mammogram plus around $35 for contrast. Compare this with breast MRI that reimburses at least four-times this amount at most centers. For CEM to be reimbursed commensurate to its clinical value, it likely needs its own unique CPT billing code, and higher negotiated reimbursement rates. Until this happens, CEM use may remain limited to niche applications such as providing contrast-enhanced imaging for individuals who can’t complete a breast MRI.
What are other downsides of CEM? Having to add IV placement to the mammography workflow is a limitation that is real but not insurmountable. Having breast imaging staff on site capable of managing rare but sometimes serious iodinated contrast reactions is also a notable yet not insurmountable limitation. CEM is associated with increased radiation exposure compared to standard mammography. Although formal benefit-to-risk calculations on CEM have not been performed according to my knowledge, the benefit of increased cancer detection from CEM is highly likely to exceed radiation risks for appropriately selected patients. There is also a surprising lack of research on CEM for screening use. The studies that exist are promising but, in my view, not as robust as data that currently exists for MBI. More research on CEM for screening is sorely needed.
One concern regarding CEM for screening that is sometimes discussed is whether individuals with dense breast tissue would be willing to get an IV placed for a CEM exam. Existing data suggest that they are. Indeed, when you consider what people are willing to go through to get a screening colonoscopy, an IV for a screening mammogram seems like a small ask.
A unique advantage for CEM is that, in theory, this single exam could count as both the primary screening exam (via low energy images) and the supplemental screening exam (via recombined contrast-enhanced images). Imagine a future breast imaging workflow wherein individuals with dense breast tissue simply undergo screening with CEM instead of a standard mammogram. This should provide highly sensitive breast cancer screening for those with dense breast tissue in a single setting, effectively removing a major disparity from current breast cancer screening practice between those whose breasts are non-dense versus dense.
Contrast-Enhanced Breast MRI:
Contrast-enhanced breast MRI is currently considered the most sensitive imaging test for breast cancer detection. So why can’t we simply provide breast MRI for all individuals with dense breast tissue? First, a reasonable proportion of people cannot undergo or tolerate MRI due to claustrophobia, non-MRI compatible medical implants, or other reasons. Second, MRI is currently expensive, and may not be readily available to all who may need this for screening, particularly in rural areas. Additionally, some fraction of gadolinium (a potentially toxic metal used for MRI contrast) can deposit and remain in tissues in the body such as portions of the brain. More longitudinal data is necessary to know whether this is truly a significant problem. When considering that people with dense breasts may undergo an annual breast MRI over a lifetime of screening, that equates to a lot of gadolinium exposure and potential tissue deposition for which long-term consequences are currently unclear.
Potential advances in MRI include abbreviated protocols—sometimes termed fast or even ultra-fast MRI—that speed up the exam by skipping potentially non-essential MRI sequences. Abbreviated MRI is supposed to be a slightly cheaper, slightly faster, and therefore more accessible MRI exam that approximates the accuracy of a standard full protocol breast MRI. Will abbreviated MRI protocols allow breast MRI to be fast and cheap enough to allow population-wide supplemental screening for those with dense breast tissue? I do not know.
Non-contrast breast MRI techniques using diffusion weighted imaging (DWI) instead of intravenous contrast, and other non-contrast techniques that may aid cancer detection are under study. It is unclear whether DWI or other MRI techniques can provide the same sensitivity for cancer detection as contrast-enhanced breast MRI protocols.
The high cost of purchasing and maintaining an MRI in comparison to other screening options that are simply less expensive and cumbersome to operate is an additional consideration.
Nonetheless, if the goal for supplemental screening is to have the most sensitive imaging test available for this purpose, contrast-enhanced breast MRI is the current best bet.
Other Imaging Technologies:
Breast CT: There may be significant potential here, but robust studies are still forthcoming. Breast CT offers the possibility of getting detailed x-ray-based images without the necessity of breast compression which is an advantage from a patient comfort perspective. Contrast-enhanced breast CT images should boost the sensitivity of the study and better reveal cancers in dense breast tissue that standard mammography cannot. Compelling use cases for why one would choose breast CT over CEM need to be clearly elucidated. Stay tuned because the jury is still out on this one. CEM is already gaining a rapid foothold in breast imaging practices, so breast CT has a lot of work to do to try to catch up and prove its value.
Breast Elastography: Breast elastography for breast cancer screening typically refers to a hand-held device that emits ultrasound waves for measurement of breast stiffness but does not provide ultrasound images for interpretation. The device simply alerts a positive result if it runs over an area of the breast with increased tissue stiffness. This should not be confused with breast ultrasound elastography which does provide images for interpretation that potentially help radiologists make decisions such as which masses with borderline imaging features require a biopsy versus imaging follow-up.
I think this non-imaging screening device is potentially promising to augment the accuracy of a clinical breast exam for cancer detection. Imagine a primary care provider using this portable elastography device to screen the breast during an annual physical exam. If a positive result is found over a certain area of the beast, this would then trigger subsequent workup with diagnostic mammogram and ultrasound.
I think handheld non-image-producing breast elastography devices have potential as an assistive device for breast cancer screening, particularly for individuals for whom mammography may not yet be indicated, or for screening in countries outside of the U.S. where formal mammography screening programs are not yet implemented.
Some ask whether breast elastography could be used to triage those with dense breast tissue to supplemental screening depending on whether the elastography study is positive (indicating a need for additional supplemental imaging) or negative (indicating mammography-alone is sufficient). Elastography could potentially fill this triage role, but do we want to add an intermediary step in the screening pathway? In a world where CEM could provide primary and supplemental screening at the same initial imaging session, it is hard for me to believe a superior workflow would include a screening mammogram to be followed by elastography and, if positive, additional screening with ultrasound, CEM, MBI, or MRI.
Breast PET: This technology is rare and mostly exists for research. Various names have been applied for PET imaging of the breast including dedicated breast PET (dbPET) and positron emission mammography (PEM). Breast PET is not in current clinical use for supplemental screening at any center, to my knowledge. I think breast PET is better suited, based on existing data, for breast cancer staging and treatment evaluation than for breast cancer screening.
Contrast-enhanced breast ultrasound: I do not know why more research or discussion on contrast-enhanced breast ultrasound for breast cancer screening exists. The contrast used for ultrasound appears to be safe, does not have tissue deposition concerns such as gadolinium for MRI, and should increase the sensitivity of ultrasound for breast cancer detection which is a great thing. However, many more studies on this technology are needed.
Which supplemental screening exam is best?
In a pragmatic sense, whatever supplemental screening study an individual can access given financial and geographic constraints is the best option for that individual. Generally, the most accessible supplemental screening examination (though not universal) is whole breast ultrasound. More cancers will be detected early with whole breast ultrasound than with mammography or tomosynthesis alone. However, if ANY other supplemental screening study (CEM, MBI, MRI) is available, these detect many more cancers than ultrasound and would be my preferred method of supplemental screening.
The current supplemental screening options that I consider the best are, in no specific order, MBI, CEM, and MRI. Unlike breast ultrasound and tomosynthesis, these three exams are predicted to detect the majority of breast cancers that mammography cannot, which is the entire point of supplemental screening.
Potential Solution: Artificial Intelligence
Artificial intelligence has shown promise to improve breast cancer detection on mammography and has primarily been directed toward screening mammography. However, I worry that the marketing or hype around AI in breast imaging is that it will eventually allow us to detect all existing breast cancers using mammography. It is likely that AI will boost cancer detection for breast cancers that are mammographically visible, but I do not think AI will be able to detect most breast cancers that are hidden by dense breast tissue. The problem will not be on AI’s end, but rather that mammography simply cannot depict cancers that are masked by dense breast tissue. It is like asking AI to find Waldo in a Where’s Waldo image, except that Waldo is not on the page. It simply won’t happen.
Great promise may exist in AI risk-prediction models, particularly those that include mammogram images in addition to factors like family history into the risk estimation. If AI can tell which patients with dense breasts are more likely than others to benefit from a supplemental screening exam, that could allow us to target supplemental screening to those most likely to benefit and spare the expense and bother for those who are unlikely to benefit.
Additionally, not all dense breasts are equal in terms of potential masking effects or complexity that can contribute to a cancer going undetected. Some breast tissue that is technically dense is nonetheless fairly bland and less distracting on a mammogram. Other dense or non-dense breast tissue has an appearance the mirrors that of architectural distortion, or is very nodular mimicking masses, or has diffuse benign microcalcifications that may make it harder to perceive suspicious microcalcifications. Perhaps AI can help us identify which mammograms are more likely to not show a cancer based not only on the density of breast tissue, but also as a function of breast complexity. Intelligent AI-based triage could allow us to tailor alternative screening approaches to such individuals.
If AI can tell which patients are likely to do well with screening mammography alone, a combination of mammogram plus MRI (for example), or a screening strategy that omits mammography entirely—wouldn’t that be superior to the current “mammography for all” approach? Imagine a future where AI not only helps us interpret medical images, but also helps us intelligently create personalized screening strategies.
AI may also be able to assign ideal personalized screening intervals based on risk-prediction models. What if AI can tell that one person will be well-screened on an annual basis, another person on an 18-month basis, and another person on a 9-month basis? The notion that either 12-months or 24-months are the absolute best intervals for screening may be incorrect. When one considers that these intervals have more to do with the time it takes the earth to orbit the sun, and less to do with the biology of breast cancer, how fast breast cancer cells replicate, or individual risk factors for developing breast cancer, this truth becomes evident.
AI that considers a person’s mammographic features (including density, since it’s the single highest risk factor) as well as family, genetic, social, and environmental factors to provide tailored screening recommendations on the optimal type of imaging to perform, and the frequency at which screening should occur, could be greatly impactful.
By way of caution, we must also be cognizant of ever-present bias, including the potential that AI can bias a human radiologist against making the right call on a breast imaging study. We must not blindly follow AI until (or if) it becomes truly worthy of our trust and is shown to benefit breast cancer detection.
Potential Solution: Improved Medical Society Recommendations
The American College of Radiology is among the strongest advocates for robust primary breast cancer screening. The ACR held strong that screening should start at age 40 with mammography on an annual basis. This is a more aggressive approach than was advocated by most other medical bodies. This includes the U.S. Preventive Services Taskforce (USPSTF) which until very recently recommended starting mammography screening at age 50 on an every-other-year basis. The USPSTF now has released draft recommendations that mammography screening should start at age 40, though on an every-other-year basis. This appears to be a partial admission by the USPSTF that the ACR was correct all along, and that more lives truly will be saved if one starts screening at 40 instead of 50. The ACR also strongly advocates that those at high-risk for breast cancer (meaning those with a 20% or greater lifetime risk of developing breast cancer) should add annual breast MRI screening on top of mammography.
However, the ACR has strangely NOT been a strong advocate for supplemental breast cancer screening. This has been very frustrating for many supplemental screening advocates including myself. Many in the supplemental screening community view the ACR’s current recommendations for supplemental screening for those with dense breast tissue otherwise not at high risk as a huge obstacle that prevents those with dense breast tissue to access additional screening beyond tomosynthesis. Supplemental screening availability is often tied to insurance coverage, and insurance coverage hugely hinges on things like ACR Appropriate Use Criteria recommendations. Therefore, the ACR’s unwillingness to endorse any imaging study as usually appropriate other than mammography with tomosynthesis for dense breast tissue supplemental screening effectively limits access to life-saving supplemental screening for most individuals with dense breasts.
Why the ACR continues to assert that tomosynthesis is the most appropriate supplemental screening exam for those with dense breast tissue is confusing. The ACR’s assertion that tomosynthesis is the most appropriate supplemental screening exam for dense breasts defies logic and is contrary to published evidence on this topic. Indeed, even the ACR Appropriateness Use Criteria summaries indicate that tomosynthesis has been found in several studies to provide no statistically significant benefit compared to standard mammography for individuals with extremely dense breast tissue which is the entire point of supplemental screening.
If the ACR would endorse as “usually appropriate” supplemental screening options other than tomosynthesis, this would be a major step forward to saving more lives from breast cancer. It is time for the ACR—and other medical societies and policy groups such as the USPSTF— to support robust supplemental screening for every individual with dense breast tissue.
Screening recommendations that are one-size-fits-all will always introduce disparities by being better tailored to some patient groups than others. More individualized screening recommendations are something we need to see in the future. And policy bodies need to be composed of various experts, including those directly involved in the care of individuals with breast cancer, for which the USPSTF has historically fallen short. Read more of my thoughts on this topic published on KevinMD by clicking here.
Potential Solution: Circulating Tumor DNA and other fluid-based Testing
It is possible that powerful breast cancer screening could be obtained from a blood draw that checks for presence of breast cancer DNA in the bloodstream, such as a circulating tumor DNA (ctDNA) test. It is also possible that image-based breast cancer screening will be partially or fully replaced in the future by such serum-based assays. These blood tests may prove to be a simpler and cheaper way to accurately check for breast cancer and may become equally or even more efficacious in comparison to image-based screening approaches. Perhaps future breast imaging for cancer detection will be performed only if blood markers are first abnormal.
There are already a variety of ctDNA and other DNA-based assays that exist, and many others are in development. The rate of progress is on ctDNA testing is relatively fast, with more sensitive testing methods always on the horizon. The impact that ctDNA and similar serum-based cancer detection assays can have on medical imaging should not be understated. These tumor-detecting serum tests could disrupt the entire breast cancer screening landscape.
If not used for primary breast cancer screening, ctDNA and similar assays could also be considered for supplemental screening. Perhaps mammography alone is ideal for those with non-dense breasts, and additional screening with ctDNA may be indicated specifically those with dense breast tissue in lieu of other image-based supplemental screening tests. We currently do not know optimal screening strategies with ctDNA assays. However, I believe these blood assays will play an important role in breast cancer screening in the not-to-distant future. Nonetheless, additional research is necessary before these tests are ready for primetime, and these do not currently stand as replacements for FDA approved screening such as mammography.
Breast cancer screening could also potentially be performed through testing other bodily fluids such as tears or saliva for which studies are currently proposed and ongoing.
Stay tuned because this is exciting technology that could prove of great value to cancer detection and monitoring.
The Way Forward
Randomized controlled trials that prove mortality reduction from screening are the holy grail to prove screening works. It is not likely that such trials will be available for supplemental screening any time soon, if ever. The sheer cost of conducting such a trial is nearly prohibitive, and I do not know of any organization shy of national governments who have the money or patience necessary to support such a study. Unfortunately, the U.S. has recently invested heavily on funding the extremely expensive TMIST trial to evaluate tomosynthesis for screening which several of our breast imaging giants say is a poor use of resources that could have been spent on a more impactful study (click for representative article). I agree that the TMIST trial is potentially a huge disappointment because we only get so many of these studies in a career or lifetime, so we need to prioritize our biggest and most expensive studies on the most impactful science possible. I believe that if the roughly $100 million U.S. dollars spent on TMIST were instead used to prospectively study CEM, MRI, and MBI for dense breast screening, hundreds of thousands of lives could be spared from breast cancer mortality as a direct result of this science over my remaining professional career.
I believe we have two paths forward. The first is to get the best science we can, short of randomized clinical trials with mortality reduction as the endpoint. We can perform less expensive studies proving beyond doubt whether CEM, MBI, and MRI have ICDRs and other performance metrics that justify use of these technologies in addition to standard mammography. We can perform comparative-effectiveness research to evaluate the costs and benefits of national supplemental screening programs. We can study how to best implement these technologies in communities throughout the U.S. that are historically under-screened for breast cancer. There is so much science we can do that is desperately needed to assure that payers of all varieties have the data requisite to provide widespread payment for these studies.
Alternatively, continued state and national legislation can force insurers to pay for these exams, and even require states to implement supplemental screening programs for their constituents. This solution to the problem would again stem from citizens and lawmakers being fed up that breast radiologists and medical societies such as the ACR are doing too little to promote and advance technology beyond mammography for dense breast screening, as well as frustration with insurers’ reluctance to pay for these exams. Such appears to be the origin of the Find It Early Act that is currently proposed national legislation to mandate insurance coverage for supplemental screening exams.
There are two more barriers to implementation of effective supplemental screening that I have not yet addressed. First, is public health policy that appears to me to be too frequently anti-screening. This has always astounded me as effective screening widely applied is one of the best public health interventions that I can imagine in our fight against breast cancer. However, we are likely losing lives to metastatic breast cancer because of well-intentioned experts who worry that individuals participating in screening cannot tolerate false-positive results from imaging studies or false-positive percutaneous biopsies. To avoid this anxiety, public health experts and governing bodies have sometimes proposed less intensive screening. Fewer screening exams would mitigate false-positive findings from screening, lower patient anxiety as a result, but also allow too many breast cancers to become metastatic and incurable.
Is anxiety over a false positive imaging result or a minimally invasive biopsy sufficient that we should limit screening in the U.S.? For me, the reality that 40,000 deaths from breast cancer continue to occur each year in the U.S. should heavily weigh into this decision. The fastest way to drive this number of breast cancer deaths to as close to zero as possible, as far as I am aware, is to widely implement effective breast cancer screening.
One price we have to pay as a society to lower breast cancer mortality through screening is acceptance that false-positive screening results will occur. False-positive results should be reduced as low as possible, but it is an error in judgement, and even ethics, to implement policies that prioritize reduction of false-positive imaging results at the expense of lives lost to metastatic breast cancer given the asymmetric risks. When considering the same anxiety-reduction argument, coping with metastatic breast cancer would certainly cause vastly more anxiety than the anxiety that results from a false-positive mammogram or needle biopsy. Therefore, anxiety over false-positive screening findings, while real and concerning and something that deserves attention, is NOT a reason to forego effective breast cancer screening.
Critics of breast cancer screening also frequently speak of overdiagnosis of breast cancers as a reason to screen less aggressively for breast cancer. An overdiagnosed breast cancer is described as a breast cancer that is diagnosed in a person’s lifetime but will cause that person no physical harm. A person with an overdiagnosed breast cancer will die of another cause before the cancer advances and interferes with quality of life. Such an overdiagnosed cancer truly may harm a patient due to anxiety regarding the cancer diagnosis and will introduce morbidity from cancer treatment. This is a problem. However, an overdiagnosed breast cancer is, by definition, non-lethal. An undiagnosed breast cancer can be lethal. Therefore, which of these cancers should we allow to continue?
To me, the answer is clear: lives spared from metastatic breast cancer through supplemental screening, and avoidance of the morbidity and financial devastation that can accompany metastatic breast cancer therapy, is likely worth current and even potentially somewhat higher rates of false-positive imaging findings, minimally invasive biopsies, and cancer overdiagnosis.
If we implement less aggressive screening regimens, and continue to rely on mammography alone with or without tomosynthesis or ultrasound for supplemental screening, we will continue to have tens of thousands of deaths from breast cancer annually in the U.S. That is unless significant progress is made in treating metastatic breast cancer which currently is not curable. Therefore, until metastatic breast cancer is curable, the best path forward is to lower rates of metastatic breast cancer through effective breast cancer screening for all, widely applied.
If every person with non-dense breasts participated in screening mammography, and every person with dense breast tissue had screening mammography plus MBI, CEM, or MRI, nearly all breast cancers would be detected early while not only treatable, but also curable. Future breast cancer deaths would plumet to the lowest levels we have ever seen.
The technology to perform effective breast cancer screening for every individual is already here. We simply need to target these technologies better to those who need them most.
As a breast imager, it is excruciatingly frustrating and sad whenever I see a patient present with a palpable lump or other clinical change that reveals an advanced breast cancer that was not seen on screening mammography due to dense breast tissue, and no other screening option was utilized to detect this cancer. This occurs far too frequently. To know the technology to find these cancers exists, yet is not finding these patients in time is terrible. I want to see this problem solved during my career. I hope enough people will understand the importance of this problem, and are able to see the technological way forward that already exists, to make this possible.
How long will it take us to overcome current barriers to implementation and enter a new era of effective breast cancer screening for all? How many more missed opportunities of early breast cancer diagnosis will it take for us to approach screening paradigms with fresh eyes?
The time to act was yesterday. So, what will we do today?
To learn more and help spread awareness of dense breast tissue and supplemental screening, here are some organizations to check out:
Selected References and Articles of Interest from Dr. Covington on this Topic:
Check out other articles from The Radiology Review Insider by clicking here.
Do you want to contribute to The Radiology Review Insider?
The article submission process is simple: email your proposed article to theradiologyreview@gmail.com. Include with your article your name and professional affiliation. Submission of all articles is appreciated but submission does not guarantee publication.
The Radiology Review may receive referral fees from purchases made with links on this page.
