MRI Used as Surrogate for Response, and as Surveillance

San Antonio—Magnetic resonance imaging (MRI) continues to show promise as a means of diagnosing and evaluating breast cancer, according to study results reported at the San Antonio Breast Cancer Symposium.

Serial MRI assessment of 31 patients has shown good correlation between imaging results and pathologic findings, reported Laura Esserman, MD, director of the Comprehensive Breast Cancer Program, University of California, San Francisco. “Our objective is to see whether we can use MRI as a surrogate marker for response to therapy,” Dr. Esserman said at a poster session.

The study involves women with diagnosed stage III-IV breast cancer who are entered into a neoadjuvant chemotherapy protocol consisting of doxorubicin and cyclophosphamide for four cycles. Serial MRI breast scans are performed prior to chemotherapy, after the first cycle, and after the fourth cycle.

Dr. Esserman and her colleagues use high-resolution, 3D, T1-weighted, gradient echo MRI imaging with fat suppression and gadolinium enhancement. The scans take about 45 minutes to complete, and no patient has dropped out of the study because of aversion to MRI.

“Most patients love MRI unless they are claustrophobic,” Dr. Esserman said. “There is nothing like the kind of picture that MRI can provide to give a patient a better idea of what she actually has. It’s useful for helping patients make treatment decisions.”

Of the 31 patients enrolled thus far, size as determined by MRI has shown good correlation with actual tumor size (.7878). The results also indicate that MRI is a good surrogate marker for residual disease, compared with clinical and pathologic assessment.

In the study, a clinical response is defined by the change in the longest distance of the tumor. A complete response means a nonpalpable lesion. Partial response reflects a decrease in longest distance by at least 30%, which corresponds to at least a 50% reduction in volume. A minimal response reflects a change of less than 30%.

MRI response criteria are defined by automated measurement of volumetric change and signal enhancement ratios. A complete response is defined as a greater than 90% MRI change, partial response as a 50% to 90% change, and minimal response as a 10% to 50% change. Pathologic response is defined by the absence or size of residual tumor.

Of the first 31 patients, MRI showed 3 complete responses and 17 partial responses. Clinical assessment classified 9 responses as complete and 14 as partial. Pathologic assessment revealed 2 complete responses and 13 partial responses (see Table).

“We are starting a multicenter clinical trial to evaluate serial MRI scans and other possible surrogate markers to see whether we can come up with an algorithm to predict which patients are responders and which ones are not,” Dr. Esserman said. “Then we can begin to design trials and enroll patients who have less-than-optimal responses and design novel therapies for them, all without taking the patients to the operating room.” She noted that the multicenter trial will address the issue of technique variability by employing standardized MRI protocols.

MRI Surveillance

Canadian investigators have found MRI useful for surveillance of high-risk women with hereditary breast cancer. An ongoing study ultimately will involve 200 women aged 25 to 60 years who have a known BRCA1 or BRCA2 mutation or a greater than 25% probability of being a mutation carrier, Charles Catzavelos, MD, a pathologist at the University of Toronto, said at the symposium. The principal investigator for the study is Ellen Warner, MD, a medical oncologist at the University of Toronto.

The women have multimodality assessments twice a year, including a clinical breast examination, digital mammography, breast ultrasound, and breast MRI.

Thus far, 139 patients have completed the first round of imaging. The mean age of the patients is 44 years; 45% are mutation carriers, 8% are first-degree relatives of mutation carriers, and 47% have a strong family or personal history of breast cancer. Dr. Catzavelos reported that 53% of the women are premenopausal, and 30% have a history of breast cancer.

Six invasive breast cancers have been detected to date, and all the patients proved to be node negative. MRI identified all six cancers, compared with two identified by mammography and three by ultrasound. None of the tumors was palpable by breast examination.

Overall, MRI has identified 13 breast abnormalities, including the six tumors. Ultrasound showed 10 abnormalities, three of which proved to be cancer. Mammography identified three abnormalities. Dr. Catzavelos said that 97% of the women have been satisfied or very satisfied with their overall clinical experience, and discomfort scores for MRI and mammography have been identical.

Despite the favorable findings from the study, Dr. Catzavelos believes MRI will have a fairly limited role in the diagnosis and evaluation of breast cancer. “I think we will always have to define a subgroup of patients in whom MRI will be useful,” he said. “This particular group has hereditary breast cancer. There might be other subgroups for whom MRI will be useful in a particular clinical setting, but I doubt that MRI will have widespread use in breast cancer screening.”

Dr. Esserman offered a more optimistic assessment of MRI’s role in breast cancer management. She cited a paper she and her colleagues published last year, indicating that MRI can be a cost-effective initial imaging modality in patients who have multifocal disease.

“Because so many patients are multifocal and because the decisions associated with multifocal disease can be so wrenching, having MRI initially can actually be cost-effective,” she said. “For large tumors, in particular, I think MRI will play an enormous role in helping direct changes in therapy. Whether it becomes the standard for everybody, I don’t know.”