ASCOThree studies of MRI screening for women at high risk of breast cancer, presented at the 39th Annual meeting of the American Society of Clinical Oncology, show high sensitivity and the ability to detect cancers missed on mammography or ultrasound, but in two of the studies, the technique had lower specificity than mammography, resulting in unnecessary biopsies. The better specificity seen in the study from Germany may stem from the greater experience of the physicians involved in that study.
The researchers agreed that MRI is not recommended for breast cancer screening in the general population, but should be considered in high-risk women as a complement to mammography, in an attempt to find breast cancers early in these women and reduce the need for prophylactic mastectomies. The German researchers, however, suggested that MRI could replace mammography screening in women at high risk of developing the disease.
When MRI was compared with mammography and ultrasound in a German study of 462 women with proven or suspected hereditary breast cancer (abstract 4), it "was the most powerful tool to detect cancer, classify lesions, and prevent unnecessary biopsies," said Christiane Kuhl, MD, associate professor of radiology and neuroradiology, University of Bonn. The study included women who had or were likely to have BRCA mutations based on genetic testing or pedigree analysis. Women in the study underwent annual two-view mammography, high-resolution ultrasound, and MRI. Dr. Kuhl reported on imaging findings after the first 5 years (1996 to 2001) of the ongoing study.
The first midterm results of this prospective clinical trial showed that MRI was more sensitive in detecting breast cancer than either of the other imaging modalities. MRI had a sensitivity of 95% vs 34% for mammography and 42% for ultrasound. MRI depicted all but 2 of a total of 51 cancers detected in the study population, while mammography and ultrasound found less than half of the cancers, Dr. Kuhl said.
"High sensitivity was not achieved at the expense of specificity," she said. Specificity was 95%, 94%, and 88% for MRI, mammography, and ultrasound, respectively. Moreover, MRI was associated with the lowest rate of unnecessary biopsies.
MRI also had a higher positive predictive value (PPV) (54%) than mammography (26%) or ultrasound (16%), and it found 20 of 51 cases of multicentric disease. Dr. Kuhl said that the higher specificity and higher PPV in this study are due to the team’s 12 years of experience in reading MRIs.
In one woman who had ultrasound and MRI scanning on the same day, ultrasound identified a lesion in the left breast and no abnormalities in the right breast. MRI clearly demonstrated that the suspicious area in the left breast was scar tissue, and it revealed an enhancing area in the right breast that turned out to be a 4 mm invasive carcinoma.
In addition to its diagnostic ability, MRI may be safer than mammography for women with a BRCA mutation, Dr. Kuhl said. Because the BRCA gene deficit is thought to affect tumor suppression, in theory, women who harbor the genetic defect may not be able to fight off the mutagenic effects of ionizing radiation. "MRI is not done with ionizing radiation, so there is no harm to the patient with a BRCA mutation," Dr. Kuhl said.
Because the breast parenchyma in carriers of BRCA gene mutations may be sensitive to radiation and mammography has a lower diagnostic yield than MRI, Dr. Kuhl suggested that MRI should replace mammography when screening women with familial breast cancer.
Although Jan Klijn, MD, PhD, chair of the Rotterdam Family Cancer Center, The Netherlands, agreed that MRI is needed to detect small tumors in high-risk women, he stressed that "it is too early to leave out mammography."
Dr. Klijn reported on the Dutch MRI Screening Study (MRISC), the largest prospective, nonrandomized, multicenter trial of MRI and mammography in women with a mutation in the BRCA1 or BRCA2 gene or with a family history of breast cancer (abstract 5).
In the study, women underwent mammography and contrast-enhanced MRI once a year as well as clinical breast examination twice a year. A total of 1,874 women were eligible for study analysis.
After a median follow-up of 2.1 years, a total of 41 tumors have been detected: 34 invasive breast carcinomas, one lobular carcinoma in situ, five ductal carcinoma in situ (DCIS), and one lymphoma. Sensitivity for MRI was 71% vs 36% for mammography. Specificity was 88% for MRI vs 95% for mammography.
Dr. Klijn noted that 42% of tumors found by MRI were less than 1 cm in size, compared with 13% detected by mam-mography. Further, 76% of cases found by MRI were node negative and 52% were grade 1. These findings suggest that by detecting early-stage cancers, MRI screening may be able to reduce the number of deaths due to breast cancer, he said.
The overall event rate was 1% a year, which is seven times more than expected in a normal patient population of the same age (average 40 years).
The overall breast cancer detection rate was 10.5/1,000 woman-years. The rate of cancer detection per 1,000 woman-years varied by risk category: 25.9 in women with a 50% to 55% risk of developing breast cancer in their lifetime, 7.0 in those with a 30% to 50% lifetime risk of breast cancer, and 9.1 in women with a 15% to 30% breast cancer risk.
For detecting invasive carcinoma, MRI was more sensitive than mammography: 83% vs 37%. Dr. Klijn explained that, for every 10 patients with invasive cancer, MRI will detect at least 8 tumors and mammography will detect fewer than 4. For finding DCIS, however, the opposite is true. Of five cases of DCIS, four were detected by mammography. In this case, mammography is more sensitive than MRI, he concluded.
The US study, from Memorial Sloan-Kettering Cancer Center (abstract 362), looked only at women with known BRCA1 or BRCA2 mutations. The 54 BRCA-positive women were part of a larger breast cancer surveillance program at the center. Mark E. Robson, MD, assistant attending physician, reported the results.
The women underwent 97 screening breast MRIs in addition to their routine mammograms between July 1998 and April 2003, Dr. Robson said. Another 28 MRIs were performed as short-term follow-up approximately 6 months after the previous MRI had detected a minor abnormality. Four more MRI examinations were performed either at the time of a contralateral breast cancer diagnosis or before a prophylactic mastectomy.
During the study period, 2 of the 54 women developed DCIS, and 1 developed invasive breast cancer, all detected by MRI. All three of these women had reportedly normal mammograms within 6 months of their MRI. No woman on the study has yet developed cancer with 12 months of a normal MRI.
MRI was 100% sensitive for breast cancer detection, with 83% specificity. A significant number of women had false-positive exams. Only 3 (14.3%) of 21 biopsies recommended after the 129 MRI studies resulted in a cancer diagnosis.
"MRI should not be recommended to the general population at the present time because its specificity is not yet adequate, and the majority of abnormalities detected in average-risk women would turn out to be false positives," Dr. Robson said. High-risk women considering breast MRI screening, he said, must be counseled about the pros and cons of the test.
The discussant for the two plenary session studies (abstracts 4 and 5), Elizabeth Morris, MD, assistant attending physician, Memorial Sloan-Kettering Cancer Center, said that while MRI was superior in finding early breast cancers in high-risk populations, "at this time, it cannot replace mammography as a primary breast cancer screening tool."
Dr. Morris, a coauthor of the Memorial Sloan-Kettering study, said that for high-risk women, "by adding another layer of screening with breast MRI, we may be able to show which patients have cancer so that more aggressive therapies, such as prophylactic mastectomy or chemo-prevention, are not needed."