Breast MRI slowly reaches consensus on indications

Over the last several years, imagers and patients have witnessed passionate debate, sometimes even legal action, involving breast MR imaging. Dissent has finally given way to consensus over a few key points: Breast MRI is meant to complement, not replace, mammography or sonography, and it's a handy tool to guide biopsies, assess disease before surgery, and evaluate treatment response.

Although fervent, the process leading to the standardization of breast MR usage has not been chaotic so much as it has been slow. It took policymakers, lobbyists, activists, and domestic and international research institutions almost a decade to come up with sound performance standards for the procedure. The American College of Radiology published the first edition of the Breast Imaging Reporting and Data System for MR (BIRADS MRI) in 2003 and the first set of practice guidelines in late 2004.

The debate seems far from over, however, about when breast MRI makes sense and what breast imagers need to know to use it optimally.

ONGOING CONTROVERSY

The learning curve for reading breast MRIs tops radiologists' list of concerns regarding the procedure. Differences with other breast cancer specialists-pathologists, oncologists, and surgeons-regarding preoperative staging usually stem from the necessity of distinguishing clinically relevant lesions occult on mammography and other tests from irrelevant ones. Numerous studies have shown that contrast-enhanced MRI is highly sensitive to breast cancers, as all malignancies enhance. Specificity is low, however, due to false-positive enhancement of benign lesions.

"The high false-positive rate and subsequent increase in the number of unnecessary benign biopsies have swayed many surgeons," said Dr. Kristina A. Siddall, a radiologist in the department of imaging sciences at the University of Rochester Medical Center in New York.

Overutilization of breast MRI leads to more aggressive surgical treatment at many breast cancers. Oncologists question whether staging of these occult lesions on MR can eliminate or reduce recurrence and have a real impact on survival, Siddall said.

Radiologists assume that breast conservation therapy leaves cancer behind, which should be taken care of by postoperative radiation. That's the source of another argument for and against preoperative staging, said Dr. D. David Dershaw, director of breast imaging at Memorial Sloan-Kettering Cancer Center in New York City.

"What's the point of knowing how much cancer those patients have if the radiation will wipe it out?" he said.

Radiation doesn't succeed in all patients, however. Mapping out the extent of tumor before lumpectomy, including positive margins and residual calcifications, could reduce the likelihood that a patient will need to return to the operating room. Unfortunately, for obvious ethical reasons, no data can support this assertion. Researchers cannot knowingly leave behind a trace of disease just to see how it reacts under radiation therapy, but radiologists suspect that patients with unseen residual tumor are the ones who fail radiation treatment. Staging thus appears to be a reasonable way to boost treatment success.

"From what we know about the biology of radiation in breast cancer, one would suspect that large residual disease left in the breast is not a good thing," Dershaw said.

MR screening for breast cancer turned into a major controversy a few years ago. Growing public interest plus misconceived perceptions about the modality fed by mainstream media ultimately led to questionable marketing schemes. AmeriScan, a nationwide company that offered screening services, was forced to close down in 2003 after facing a wave of lawsuits and action from California's attorney general for false advertising.

"Patients hear about the [MRI] and then don't want to have mammograms because they are not comfortable with the radiation and so on, and then they go to places where they can get an MR as a screening study. We certainly see that at our institution," said Dr. Haydee Ojeda-Fournier, an assistant professor of radiology at the University of Cincinnati College of Medicine.

Recent studies have shown that certain patients at high risk could possibly benefit from screening with MR. Though promising, that approach remains under investigation and criteria for categorizing patients as high risk are evolving, Ojeda-Fournier said.

SUITABLE INDICATIONS

Continued concerns over breast MRI's clinical cost-effectiveness can be fully addressed only if radiologists can reach a consensus on the appropriate indications and development of standard protocols, according to Siddall. She presented an educational exhibit at the 2006 RSNA meeting on the topic.

Acceptance of breast MRI is limited by cost issues, variability of equipment, lack of clarity over the technique and interpretation criteria, inability to reliably detect microcalcifications and ductal carcinoma in situ, and a higher false-positive rate compared with mammography. MR-guided biopsies are still not widely available outside of academic institutions and tertiary-care centers.

The ACR is developing criteria for individual program accreditation that focuses on many of these concerns. As the list of appropriate indications is solidified, radiologists and surgeons will have a better understanding of how best to use breast MRI, Siddall said.

Breast MRI is most commonly used for evaluating patients with a known cancer. Accepted indications include the following:

• defining tumor size, extent, multicentricity, and multifocality;

• screening the contralateral breast for cancer; and

• following treatment, determining response to chemotherapy, evaluating residual tumor after surgery, and assessing possible tumor recurrence.

Evaluation for occult breast primary tumor in the setting of axillary node metastasis is a less common but widely accepted indication.

Protocol and indication differences have limited the reproducibility of research results, and technical factors such as motion artifacts, incomplete fat suppression, and sequence selection have affected the accuracy of individual studies. In patients with follow-up MRI exams, the time of menstrual cycle has added another variable.

Patient selection criteria have also directly influenced the scientific data. In a patient population with a low suspicion of disease, for example, the sensitivity of breast MRI is likely to lead to several false-positive findings and biopsies, according to Siddall.

Mr. Abella is associate editor of Diagnostic Imaging.

Philip Ward contributed to this report.