Radiology society’s PR message about untreated breast cancer: Sound evidence or scare-mongering?

Kathlyn Stone is an associate editor with She tweets at @KatKStone.

The American College of Radiology recently issued a fairly terse 183-word news release on a new study published in its journal on the “natural history of untreated screen-detected breast cancer.” The report is subtitled “An argument against delayed screening or increasing the interval between screening” and is based on a small survey of Society of Breast Imaging members.  

While no one questions that the release writers had the best interests of women in mind, the messages strike us as unnecessary fear-mongering built on very limited evidence, the drawbacks of which were not mentioned in the release. There is also a notable lack of context accompanying the release’s rather ominous take-home conclusion: “Medical science cannot yet determine which cancers will or will not advance to kill the person afflicted.”

Sidestepping the dilemmas posed by DCIS

Is the premise of the release – that all undetected and untreated breast cancers are a potential death sentence – a sound one? I asked some of our regular expert contributors to weigh in on the short document, paragraph by paragraph.

Paragraph 1: “Breast cancers detected by mammography screening do not spontaneously disappear or regress if left untreated, according to a new study published in the Journal of the American College of Radiology (JACR®). The scientific findings contradict claims that many cancers found via mammography may simply “go away” if left undiscovered or untreated.”

There are many who question the value of mammography performed at the intervals recommended by the American College of Radiology, but the basis for that skepticism is not primarily that cancers may simply “go away” — even though there is some evidence that this may occur. The bigger issue is whether some cancers might be so slow-growing (as with many prostate cancers) as to never cause symptoms of disease.

“There are several studies supporting spontaneous tumor regression,” says Deanna Attai. MD, a breast cancer surgeon and assistant clinical professor of surgery at UCLA, “and we know that DCIS [short for ductal carcinoma in-situ] may not develop into invasive disease. I think an important point to make is not whether or not tumors regress, but do they ever become clinically significant without treatment? It is clear that some tumors will not become clinically significant, but we do not yet have the ability to accurately predict this (with the 100 percent certainty that a patient and her physician would want) for an individual patient.”

It’s important to note that some experts are loath to even call DCIS a type of “cancer.” About 20 percent of mammography detected cancers are DCIS, and as we reported earlier this year: “Some experts estimate that up to 80% of women with DCIS found via screening may not need any treatment at all–and instead should just keep an eye on things.”

Conclusions based on “best estimates” and “recollection of individual anecdotes”

That brings us to Paragraph 2 and an incomplete discussion of the study itself:

“Among all practices involved in the study, 25,281 screen-detected invasive breast cancers and 9,360 cases of screen-detected ductal carcinoma in situ were reported over the 10-year study period. Among these cancers, there were 240 cases of untreated invasive breast cancer and 239 cases of untreated ductal carcinoma in situ. None were reported to have spontaneously disappeared or regressed at next mammography.”

The release doesn’t mention where those numbers come from. When we looked up the study, we learned it was based on a survey of 42 members of a society devoted to breast imaging and includes either audit data or “best estimates” from participants’ mammography practices. According to the “limitations” section of the study:

“Limitations of this study include those common to all surveys, including selection bias with respect to responses, the possibility of incorrect responses (especially when some survey respondents answered with best estimates instead of audit data), and that it depends upon the recollection of individual anecdotes. Related to the latter, no data were collected regarding patient demographics, tumor histology, or clinical progression of the disease; although these additional data would be pertinent to a better understanding of the frequency and clinical importance of overdiagnosis and would be helpful in providing guidance about who might be managed with “watchful waiting” and who should not, this was beyond the scope of our study.” 

The study authors also note, “Furthermore we did not determine the exact reasons why some patients had no treatment.” In short, this study had serious and multiple limitations, and the release presented data from this very small survey, with some of the data based on guesses, and used it as a strong-arm message to women that they should “do this or risk dying.” That is not a helpful message.

No insight into why some women waited

Why did some patients forego treatment? Neither the release or study gives us that answer, but there are many reasons women may choose to watch and wait rather than be treated immediately. Mandy Stahre, PhD, an epidemiologist and breast cancer survivor, says “It’s not just patient demographics, but health of the patient and co-morbid conditions which could affect whether a patient seeks treatment. Age is also an important factor. If an 75 year old woman has a mammogram with a positive tumor, are you going to treat them or wait?”

When women are armed with good information and counsel from their physicians the best course for some might be to watch and wait, particularly if their screen suggests DCIS. The release should have acknowledged this.

Paragraph 3: “An unknown percentage of these 479 cases represent overdiagnosis, but because all untreated screen-detected cancers were visible and suspicious for malignancy at next mammographic examination, delaying the onset of screening (from age 40 to ages 45 or 50) or increasing the interval between screenings (biennial vs annual) should not reduce the frequency of overdiagnosis.”

Paragraph 4: “Medical science cannot yet determine which cancers will or will not advance to kill the person afflicted.”

Paragraphs 3 and 4 dismiss very real concerns about over-diagnosis. Over-diagnosis leads to over-treatment. The unnecessary surgery, radiation and chemotherapy that some women undergo for non-life-threatening cancer are a harm that should be discussed more openly. Estimates of over-diagnosis vary greatly, from 20 percent to more than 80 percent. These paragraphs both seem to suggest that there’s nothing to be done about over-diagnosis and that women must simply accept it as the cost of eliminating potentially deadly cancers. But the reality is there are trade-offs that fully informed women can and should be able to weigh before making any decision about breast screening and treatment. Active surveillance is an option that some women should discuss with their doctors. That’s what media messages should be encouraging — not ominous warnings that fail to give people the full picture.

You might also like

Comments (7)

Please note, comments are no longer published through this website. All previously made comments are still archived and available for viewing through select posts.

Donna Pinto

May 10, 2017 at 12:13 am

THANK YOU for tackling this issue in such a fair way so that women can truly be informed. It is shameful that radiologists keep important information from women about overdiagnosis and downplay the harms of mammography. I was most likely over-diagnosed and over-treated for low grade DCIS until I said NO MORE despite heavy pressures for mastectomy and/or weeks of radiation and years of tamoxifen. I started a blog (DCIS 411) to help educate women and to provide support for women who choose active surveillance. This was not available to me in 2010 and I may have had 3 unnecessary surgeries — all to discover I did not have invasive cancer…yet the mammogram and docs had scared me that I did. Please keep writing about this topic! Thank you!!

Marie Dagosta

May 10, 2017 at 9:46 pm

This is spot on! When I was first diagnosed with DCIS, I was not given the choice of active surveillance. It was surgery, radiation, and 5 years of tamoxifen. OR a mastectomy!! Had I not done intensive research on DCIS, I would have gone ahead with standard protocol. Then I found Donna Pinto’s blog and that helped me make a decision to go with lumpectomies and a yearly MRI instead. So far, 5 years and no regrets!

Dorothy Pugh

May 11, 2017 at 11:12 am

This excellent article raises many important issues, e.g., the lack of mention of cases that were stable, i.e., were indolent cancers (especially in older women), the exclusion of quality-of-life considerations, and the lack of objective lab data. And how many subjects got second opinions? Another issue is the very small number of untreated patients (2%); the article’s point that their characteristics were not examined was well-taken. I was dismayed to notice that none of the authors was a professional statistician or epidemiologist, which is actually quite unusual in clinical studies, and crucial, in my opinion, for observational ones.

Brian Casey

May 11, 2017 at 3:52 pm

I’m afraid the author of this column is off base on several points. The point of the ACR release was exactly what it said — to refute a controversial theory that some breast cancers just “spontaneously regress,” or just go away. Her statement “even though there is some evidence that this may occur” is actually not true, there is no evidence of this. It’s telling that neither she nor Dr. Attai provide any references to back their contention..
The debate over DCIS and overdiagnosis is entirely different from that over spontaneous regression, and as such these topics were not the subject of the ACR press release and the survey on regression. The breast imaging community itself does not fully agree on how to handle DCIS, and overdiagnosis is an acknowledged problem.
To conflate an unproven theory like spontaneous regression with legitimate concerns regarding DCIS and overdiagnosis is missing the point.

    Kathlyn Stone

    May 12, 2017 at 7:02 am

    Hi Brian. Thanks for your comment. I did not set out to argue for or against “spontaneous regression” of breast cancer. Instead, the focus of the article was on the many holes in the news release, the important unanswered question about why some women may not seek treatment and the weak survey the report was based on.

    I also sought to point out that the release is misleading. In arguing against spontaneous regression, the release makes a leap to claiming that “Medical science cannot yet determine which cancers will or will not advance to kill the person afflicted.” That’s an incomplete message. Medical science knows that certain types of DCIS are unlikely to progress, and that’s information that readers of this release should have been made aware of.

    Finally, there is in fact evidence that that breast cancers spontaneously regress, from a study that followed the results of breast cancer screening over several years ( The ACR release, by contrast, describes a study that examined no patient data whatsoever. A survey of 42 screening specialists – who at times relied on recall and did not even try to assess why women put off treatment — is not strong evidence.

    Kathlyn Stone
    Associate Editor

      Brian Casey

      May 15, 2017 at 2:25 pm

      I’m afraid you’re mistaken about the Zahl paper offering evidence for spontaneous regression. In actuality, Zahl et al found a statistical anomaly in their data, and in order to explain it they came up with the theory that cancers had spontaneously disappeared. That’s not evidence, at least in my book. Have any other researchers duplicated this finding? I’d love to see that data.
      Again, I don’t think it’s fair to criticize the ACR for not addressing a subject — DCIS and overdiagnosis — that was simply beyond the limited purview of that particular press release. DCIS is a major and controversial subject within radiology and breast imaging, and if you look at their overall body of work the ACR has been quite active in addressing it.

Kathlyn Stone

May 16, 2017 at 8:27 am

You asked for evidence – which was provided – and it’s your choice to dismiss it. I never said that the evidence was definitive. In any case, my point was that this is a bigger topic than a discussion of spontaneous regression. If the purview of this release was limited to the issue of spontaneous regression, why did it conclude with a claim that “Medical science cannot yet determine which cancers will or will not advance to kill the person afflicted?” That statement goes far beyond spontaneous regression and raises issues that were appropriately addressed in my post. The release was full of holes, lacked context, and drew sweeping conclusions that aren’t supported by the weak evidence provided.