CR REPORTS FROM THE SABCS

SATURDAY, DEC. 13, 2008

The Shape of Progress in Breast Cancer:
Use and overuse of diagnostic MRI

Posted by Musa Mayer

What is the shape of progress in breast cancer? In the United States, where our collective romance with technology and innovation often colludes with a lack of evidence-based health insurance coverage, it follows a predictable curve. A new approach is put forth, often against much resistance, but then is widely adopted. In time, this leads to indiscriminate overuse and off-label prescribing, often fueled by marketing strategies. Finally, new studies emerge that question this overuse, demonstrating its harm and costs. New methods are proposed to better select patients who will actually benefit, and spare those who will not.

We have seen this happen again and again in breast cancer. Extensive, mutilating surgical approaches, like the radical mastectomy, eventually gave way to simple mastectomy and then to lumpectomy. Radiation to the breast became conformal, and today, more localized radiation to the tumor bed alone is being studied. Complete axillary dissection of lymph nodes has been succeeded by node sampling, and then by sentinel node biopsy. Widespread use of adjuvant chemotherapy led to the premature adoption of high-dose chemotherapy for metastatic and locally advanced breast cancer and eventually to more targeted therapies and the development of multi-gene assays to determine which patients actually need chemotherapy.

In this regard, one of the more interesting plenary sessions at this year’s San Antonio Breast Cancer Symposium was presented by surgical oncologist Monica Morrow, of Memorial Sloan-Kettering Cancer Center, who offered an overview of the role of magnetic resonance imaging (MRI) in newly diagnosed breast cancer patients (see abstract P3-1). Her talk was focused on the use of diagnostic MRI imaging to improve patient outcomes (not  on the use of screening MRI to detect breast cancer in high-risk women). [Listen to the related podcast.]

In the clinic, MRI has been increasingly used for a number of purposes, Morrow said: to assess the extent of disease; to detect an unknown primary when breast cancer has first manifested though a metastatic site, like an axillary lymph node; and to look at treatment response before and after neoadjuvant treatment. It has been widely supposed that MRI would assist in determining which patients could have breast-conserving surgery, thus reducing the number of mastectomies. By determining the extent of disease within the breast, it was assumed that MRI could also help in reducing the number of re-excisions, in which a surgeon has to operate again to get clean margins. This, it was hoped, would also decrease local recurrence rates. And by finding “mirror” disease in the contralateral (unaffected) breast, MRI might find second breast cancers early enough to improve outcomes.

Morrow reviewed the many studies that have examined these uses. These randomized trials and retrospective analyses found that MRI did not improve surgical outcomes by preventing re-operation, and did not reduce the number of mastectomies, she said. In fact, MRI use served to increase the number of mastectomies. In many cases, the use of MRI also tended to delay surgery, which other studies have suggested may be of real concern. In lobular cancers, which are more difficult to image, MRI was not able to reduce either the need for mastectomy or the number of re-excisions. The COMICE study, presented immediately following Morrow’s plenary talk, is a U.K. trial that examined MRI planning for breast-conserving treatment. It, too, found no significant difference in reoperation rates.

As far as the detection of cancer in the other breast goes, while MRI found more cancers, this detection failed to improve outcome. Morrow is not the first to offer the observation that many of these “occult” breast cancers don’t need treatment, and will never grow or metastasize. This was another instance of a rarely discussed harm related to mammography screening: the overdiagnosis and overtreatment of indolent cancers. [Read more about this topic in the Spring 2007 CR magazine or listen to a related podcast.] A truism of all testing in cancer is that the more you look, the more you find—and the more you tend to treat or make some other sort of medical intervention in response to that finding. We often assume this must be beneficial, but in fact this has rarely been studied.

“How is it possible,” Morrow asked, “that finding additional cancer is not important?” In this context, she cited a number of well-conducted pathological studies of occult multifocal disease that support this surprising observation.

Summing up, Morrow said that it is clear that MRI finds two to three times more disease than is found without MRI, resulting in an increased mastectomy rate, but that the benefit is questionable. “Neither short-term surgical outcomes,” she concluded, “nor long-term local control or contralateral cancer rates are improved with MRI.” Although it might seem “intuitively obvious,” she said, the routine use of MRI ought to require evidence of clinical benefit, and to date, this has been lacking. In fact, research shows that only a few uncommon uses of diagnostic MRI in newly diagnosed breast cancer patients have proven beneficial.

This sobering message suggests that the path to progress in breast cancer treatment does not always lie in the direction of embracing new tests and new treatments. Sometimes, progress means stopping and asking: Is this really necessary? Does this really help? And what is the evidence?