Saturday, December 13, 2008

Jane Zones, Board Member

MRI and Breast Cancer

Monica Morrow is a surgical oncologist, who heads Memorial Sloan-Kettering Cancer Center’s Breast Service. Her major research interest is the application of knowledge from clinical trials to daily surgical practice, and her talk this morning was a beautiful example of this. Although she is co-author of Breast Cancer for Dummies, Dr. Morrow’s examination of the role of magnetic resonance imaging (MRI) in breast cancer was for smarties, and perhaps those more savvy about cost-benefit ratios than about molecular underpinnings of cancer biology, which is the main topic at this conference. This was a really smart presentation.

MRI is used in detecting breast cancer in asymptomatic women (screening) and in providing information to improve patient outcome in women with breast cancer (diagnosis). Dr. Morrow only addressed MRI’s use as a diagnostic tool. The potential benefits of MRI in diagnosis are to refine decisions about breast conservation therapy, to determine the extent of the tumor, to identify potential contralateral cancer, and to decrease the risk of local recurrence.

A meta-analysis of MRI detection of multifocal/multicentric breast cancer showed that MRI detected about 16% additional disease (of which over a third turned out to be false positive), resulting in a parallel increase in the proportion of mastectomies or wide local excisions (WLE) rather than breast conserving surgery (BCS).

In general, selecting a patient for BCS is not problematic in stages 0, I or II. Some are excluded from the possibility of BCS. Contraindications include pregnancy up to 6 months gestation (because of the risk to the fetus from radiation), two or more tumors in separate quadrants of the breast, diffuse suspicious calcifications, a prior history of chest irradiation, or repeated positive margins. About 8% of women who receive BCS are later found to require mastectomy.

In a range of studies, the total number of mastectomies is persistently double in women who have MRI. Furthermore, having MRI delays surgery on average three weeks. Diagnosis-related MRI studies have been retrospective and not randomized. Women who undergo MRI are on average 6 years younger and selected for imaging because they are more likely to benefit, which would result in more favorable research outcomes for MRI. Even so, no advantage has been shown for such imaging.

Dr. Morrow summed it up by saying that MRI finds more cancer but that what is found is not clinically relevant. She showed a great deal of data that indicate outcomes in the long term—particularly local recurrence—does not have significant effects upon patients. Using MRI results in increased rates of mastectomy that prove to be unnecessary. Neither short term surgical outcomes, nor long term local control or contralateral cancer rates are improved with MRI. Because of this, she recommends MRI only for BRCA1 and 2 carriers, those who present with axillary disease, those who are being assessed for neoadjuvant therapy, or those whose diagnosis is not resolved by physical exam, mammogram and ultrasound.

Dr. Morrow ended her lecture by saying that “the routine use of MRI in cancer patients requires some evidence of clinical benefit. To date, this does not exist.” In her last slide, she itemized concepts in breast cancer that were intuitively obvious, but that proved incorrect:

  • high dose chemo + bone marrow transplant is superior to conventional chemotherapy;
  • Endocrine therapy is inferior to chemotherapy and will not result in a survival advantage;
  • Treatment of breast cancer with less than mastectomy is dangerous;
  • local therapy does not influence survival.

And she added to that list: MRI finds cancer not found by other modalities. It must be useful.

Dr. Morrow’s presentation was followed by a report on the first and only prospective study of MRI, the COMICE trial, which was sponsored by the research arm of the British National Health Service. (England and Canada sponsor significant research on actual effectiveness as a means of cost containment). COMICE randomized 1623 women into two groups: those who received MRI prior to BCS, and those who did not. The primary endpoint was the rate of re-operation (those who had to have further excision, and those who had a subsequent mastectomy). There were not significant differences between the two study groups in re-operation rates, which substantiated Morrow’s perspective.