Every few months, another study reports that many breast cancers are being “overdiagnosed”—that is, detected and treated even though they would never cause problems if they were left alone. In one article, epidemiologists in Norway estimated that 15% to 25% of breast cancers found by mammograms were being treated unnecessarily.
The study in the Annals of Internal Medicine in April calculated that for every 2,500 women offered mammograms over 10 years, one breast-cancer death was averted, but six to 10 women were subjected to surgery, radiation and/or chemotherapy unnecessarily.
The researchers in Norway compared breast-cancer rates in counties where a government mammogram program had begun with those without such screening, as well as with past years. They found that detecting and treating many early-stage breast cancers reduced the number of late-stage cancers and deaths only slightly, prompting them to conclude that much of the treatment was unnecessary—in some cases because the cancers wouldn’t have progressed, and in some cases because they were fatal despite being treated early. Other studies have estimated that the overdiagnosis rate falls in a wide range, anywhere from 2% to 52%.
There is currently no way to tell which patients diagnosed with breast cancer—200,000 a year in the U.S.—could safely forgo treatment, breast-cancer specialists say. “When you can tell me which cancers need to be treated and which don’t, then I will consider this argument” about overdiagnosis, says Clifford Hudis, chief of the Breast Cancer Medicine Service at Memorial Sloan-Kettering Cancer Center in New York City.
Clinicians say leaving breast cancer untreated is a gamble they can’t take. “I don’t know anyone who offers women the option of doing nothing,” says Eric Winer, director of the breast cancer program at Dana-Farber Cancer Institute in Boston. “On the one hand, we are aware of the overtreatment, all of us. On the other hand, there are still 40,000 women every year who die of breast cancer.”
Otis Brawley, an epidemiologist and breast-cancer specialist who heads the American Cancer Society, notes such estimates are all statistical presumptions. “Even if we overdiagnose 1 in 5, we have numerous studies showing that by treating all these women, we save a bunch of lives,” he says.
A 2011 Cochrane review of seven trials in which 600,000 women were randomly assigned to get mammograms or not estimated that while 30% were overdiagnosed, breast cancer deaths were reduced by 15%.
Even in the precancerous stage, called ductal carcinoma in situ (DCIS) when abnormal cells are confined to a milk duct, physicians almost always advise women to have a lumpectomy or mastectomy along with radiation, because about 20% of the 65,000 cases of DCIS found every year in the U.S. become invasive cancer.
A few women do opt to monitor their DCIS to see if it progresses. A study in the journal The Breast last year of 14 such women, who took hormone-blocking drugs, found that after two years, eight decided to have surgery, with five of them having progressed to stage 1 invasive breast cancer. Six remained on active surveillance with no evidence their DCIS had worsened.
The authors, from the University of California, San Francisco, noted that even when DCIS becomes an invasive cancer, it’s treatable when found at an early stage: The odds that a 60-year old woman with a 7 millimeter breast tumor, would die from it in the next 10 years are less than 3%, half the risk she faces of dying from another cause during that time.
Physicians say it would be far riskier to leave invasive breast cancers untreated. “At this point, any breast cancer does need to be removed,” says Bhuvaneswari Ramaswamy, a breast cancer researcher at the Ohio State University Comprehensive Cancer Center in Columbus, Ohio. “We do more than we need to because we don’t know how to do less.”
Scientists have made progress in analyzing individual breast cancers and tailoring treatment accordingly. About two-thirds of tumors have estrogen receptors that make them vulnerable to hormone-blocking medications. About one-third test positive for a protein called HER2 that makes cancers particularly aggressive, but susceptible to the drug Herceptin.
A new wave of tests can predict how tumors will behave based on their genetic profile. The most commonly used test is Oncotype DX, which analyzes 21 genes. Cells from about half the breast tumors in the U.S. are now sent to Genomic Health Inc. in Redwood City, Calif., which developed the test. The company’s technicians determine how likely the cancer is to recur in 10 years and whether the patient would benefit from chemotherapy as well as radiation and surgery. Company officials say the test has reduced the number of U.S. breast-cancer patients on chemotherapy by 20% since it became available in 2004.
A new Oncotype DX test can predict whether a patient with DCIS would benefit from radiation in addition to surgery. But there is no test that can determine whether a breast tumor can be left untreated. “That’s a dream that we would all have for the future,” says Steven Shak, co-founder of Genomic Health, who led the development of Herceptin.
The company is testing a version of Oncotype DX that may be able to tell which prostate cancers don’t need treatment. More than 60% of prostate cancers are thought to be so slow-growing that they would never be life-threatening, but as of now, there is no way to tell those apart from the fast-growing ones that kill 28,000 men in the U.S. every year. About 20% of prostate cancer patients opt for “active surveillance” rather than immediate treatment—in part because the side effects of radiation and surgery can be severe, including impotence and incontinence.
Biostatisticians argue that fewer mammograms would reduce the number of small, early cancers found, as well as the rate of false positives that require additional scans and biopsies. That’s the rationale the U.S. Preventive Service Task Force used in 2009 when it recommended that women get mammograms every two years starting at age 50, instead of annually starting at age 40.
“I’m certainly not asking anyone to stop getting mammograms. I am asking my profession to tell women the truth about the [overdiagnosis] deal,” says H. Gilbert Welch, professor of medicine at Dartmouth Institute for Health Policy & Clinical Practice.
Patients are seldom even told about the overdiagnosis issue and there is very little data on the long-term side effects of cancer treatments, which can include chronic pain, debilitating fatigue, “chemo brain” or foggy thinking, and additional cancers linked to radiation treatments, says Fran Visco, president of the National Breast Cancer Coalition, a nonprofit advocacy group. “We shouldn’t not tell them the truth because we are worried they’ll be confused.”
Many physicians who treat breast cancer patients are loath to stop looking so hard for cancer, and hope for additional tests that can better predict which breast tumors will stay harmless. Meanwhile, says Memorial Sloan-Kettering’s Dr. Hudis, “we are quibbling over whether everyone benefits or only some people. Let’s not lose sight of the fact that we are saving lives.”
A look at data about the disease in the U.S. and the typical treatments.