San AntonioMagnetic 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. Its 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
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.
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 MRIs 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 dont know.