The Mystery of a Common Breast Cancer Statistic

Nick Mulcahy

August 18, 2015

UPDATED August 19, 2015 // A commonly cited breast cancer statistic — that 30% of all early-stage breast cancers will progress, despite treatment, to deadly metastatic disease — appears to have no strong contemporary evidence to back it up.

Nonetheless, the statistic appears widely. For example, it is cited in an academic report (J Intern Med. 2013;274:113-126), in a breast cancer charity report, in a pharmaceutical marketing piece, and on a major cancer center website.

In short, the 30% figure is conventional wisdom — despite the absence of an authoritative epidemiologic source.

But is that statistic accurate and reflective of current clinical reality? And should clinicians repeat it to patients? Perhaps more importantly, does the statistic really matter? After all, the treatment of women with early-stage disease will not change whatever the statistic is, correct?

Medscape Medical News went in search of answers to these questions and found angry patients, a clinician author trusted blindly by a lot of people, and special access to a common database that provides some measure of insight into the proportion of early-stage patients who progress to metastatic disease.

Our story begins with multiple women with metastatic breast cancer who are dismayed or angry about the fuzziness and mystery of the 30% statistic, and have said so online.

I would like to know the true stats of how many breast cancers come back no matter what the hell we do for treatment.

For example, in a 2013 post on the bulletin board, "SusansGarden" from Gig Harbor, Washington, wrote: "I would like to know the true stats of how many breast cancers come back no matter what the hell we do for treatment."

The topic has been discussed repeatedly by "metsers" for a few years, but a recent blog post got a lot of attention.

On July 21, metastatic breast cancer patient and blogger Ann Silberman, from Sacramento, California, examined the 30% statistic. For the individual patient, "none of this matters," she wrote. "You will relapse or you won't." But Silberman, who unsuccessfully looked for a credible source for the statistic for 7 months, added that "it's harmful to mis-state things, use scare tactics, and otherwise try to make a bad thing worse."

The post, with its reference to scare tactics by prominent breast cancer organizations, including Komen for the Cure, prompted a response from the Metastatic Breast Cancer Network (MBCN), a respected patient advocacy group.

"It is estimated that 20% to 30% of all breast cancer cases will become metastatic," said the MBCN in response, repeating a statistic from its own website.

The primary source for this declaration is a 2005 CME review on metastatic disease pulished in the Oncologist by prominent medical oncologist Joyce O'Shaughnessy, MD, from the Baylor University Medical Center in Dallas.

"Despite advances in the treatment of breast cancer, approximately 30% of women initially diagnosed with earlier stages of breast cancer eventually develop recurrent advanced or metastatic disease," Dr O'Shaughnessy wrote.

But there is no reference for this 30% claim.

Nevertheless, the O'Shaughnessy review appears to have become the mother lode for the 30% statistic, mined repeatedly by academics, nonprofit organizations, and industry.

For example, two faculty members of Harvard Medical School in Boston explained in their analysis published in 2013 in the Journal of Internal Medicine that "nearly 30% of women initially diagnosed with early-stage disease will ultimately develop metastatic lesions, often months or even years later." The reference? The unreferenced assertion in Dr O'Shaughnessy's review.

And two oncologists — one from Houston Methodist Hospital and one from the National Institute of Neoplastic Diseases in Peru — explained in their review published in 2010 in Breast Cancer Research that "approximately 30% of the women diagnosed with early-stage disease in turn progress to metastatic breast cancer, for which therapeutic options are limited." The reference? Again, the O'Shaughnessy review.

In addition, a brochure from Pfizer reported that "even when diagnosed at an early stage, nearly 30 percent of women with early breast cancer will eventually progress to metastatic disease." The reference? O'Shaughnessy.

Medscape Medical News reached out to Dr O'Shaughnessy to learn the source of her much-cited assertion, but received an out-of-office reply.

Still, the 30% statistic did not come out of nowhere.

In fact, in 2005, Laura J. van't Veer, PhD, who helped pioneer genetic testing for predicting breast cancer treatment outcomes and is a creator of the MammaPrint test, and her colleagues stated, in their review (Nat Rev Cancer. 2005;5:591-602), that "approximately one-third of women with breast tumors that have not spread to the lymph nodes develop distant metastases."

That statement has a reference — a 1989 study of 644 patients with stage I (T1N0M0) or stage II (T1N1M0) breast carcinoma, all treated with mastectomy (J Clin Oncol. 1989;7:1239-1251). During the median follow-up of 18 years, 148 patients (23%) died of recurrent breast carcinoma.

Dr van't Veer and her colleagues presumably rounded their figures up (from 23% to 33%) because the referenced population included only patients with stage I and II disease, and therefore did not comprise all early-stage disease.

According to the National Cancer Institute (NCI), the definition of early-stage breast cancer is that which has not spread beyond the breast or the axillary lymph nodes. The range includes stage I, stage IIA, stage IIB, and stage IIIA disease.

So this particular 30%-ish statement from Dr van't Veer and colleagues appears to be an estimate based on a clinical study that is not contemporary. In short, it is not strong evidence.

A claim similar to that of the MBCN was made in a study published in 2010 (J Clin Oncol. 2010;28:3271-3277): "Despite advances, 20% to 30% of patients with early breast cancers will experience relapse with distant metastatic disease." The reference for this statement is a 2005 meta-analysis conducted by the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) (Lancet. 2005;365:1687-1717).

But the EBCTCG data are from 194 clinical trials (and thus not epidemiologic) and are limited to 15 years of follow-up. According to experts, early breast cancers are known to metastasize at 20 years or beyond.

Not able to locate a strong source for the 30% statistic, Medscape Medical News turned to America's two most prominent cancer organizations: the National Cancer Institute (NCI) and the American Cancer Society (ACS).

The NCI was no help. According to an email from the NCI press office, the institute does no collect national data on progression from early-stage to late-stage breast cancer.

"What we don't count, we can't plan for," metastatic breast cancer advocate Musa Mayer has said in the past about this NCI omission.

The Surveillance, Epidemiology, and End Results (SEER) program of the NCI records only incidence, initial treatment,andmortality data. And most breast cancers do notpresent as metastatic.

"The cancer registry does not track recurrence, which is how the majority of people are thrust into the metastatic breast cancer ranks," according to the MBCN website, which has repeatedly criticized the limited statistical approach of the NCI and its SEER program.

However, the ACS took a stab at solving the mystery of the breast cancer progression statistic.

It turns out that the ACS has special access to SEER data.

28% of the women who died of breast cancer during that time period had localized disease at diagnosis.

The organization has an agreement with the SEER program that it will not identify individuals, said Otis Brawley, MD, chief medical officer at ACS in Atlanta. "So we get a larger look at SEER data," he explained.

Dr Brawley worked with two ACS epidemiologists to examine the issue. They looked at breast-cancer-specific mortality (as identified on death certificates) in 12 health districts in the United States from 2008 to 2012. They were surprised by the finding: "28% of the women who died of breast cancer during that time period had localized disease at diagnosis," said Dr Brawley.

We all thought 30% was too high.

The result was unexpected. "We all thought 30% was too high," said Dr Brawley.

He did not say whether the ACS would publish the data. But he did emphasize that, in general, he avoids discussing treatment outcomes and prognosis with statistics. "I will always avoid a precise number," he said.

Early in his career, Dr Brawley learned a lesson from a patient that he has never forgotten.

"About 20 years ago, a male patient with lymphoma told me: 'Doc, when you're talking to me, everything is 0% or 100%," Dr Brawley said, making the point that either an individual's disease progresses or not, and that averages are not the stuff of an individual.

He pointed out that treatment decisions are based on disease and patient characteristics.

Finally, Dr Brawley said he fully expected that if the same research approach used data from the 1990s, it would reveal higher rates of disease progression from early-stage to late-stage disease or death.

"I'm assuming there has been a drop" in the rate of progression from early- to late-stage disease, he explained, and that "the drop is largely due to treatment efficacy."

The biggest treatment improvements have been the related to the use of tamoxifen and trastuzumab (Herceptin, Roche/Genentech), he said.

Editor's note: A previous version of this story incorrectly suggested that the statistical analysis from the ACS represented a confirmation of the 30% statistic.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.