Women at high risk for breast cancer undergoing their first MRI screening have higher false-positive rates and require significantly more follow-up, showing that MRI’s notoriously low specificity improves as women accumulate prior MR screening histories, a study published Feb. 1 in Radiology discovered.
MRI is an increasingly common modality for screening women at high-risk for breast cancer, particularly women with family or prior histories of breast cancer. “MRI is a very sensitive exam, but the downside is that it is less specific—has a higher false-positive rate—than mammography, which keeps some doctors from ordering and some women from undergoing MRI,” Martha B. Mainiero, MD, of the department of diagnostic imaging, at Warren Alpert Medical School of Brown University, Rhode Island Hospital in Providence told Health Imaging News.
Researchers have found that as women accrue more mammograms, false-positive rates drop because radiologists become better able to monitor abnormal lesions over time to help determine their benignancy or malignancy. “We hypothesized that similar to mammography, if you’ve had past MRIs, the false-positive rate would go down because you would have something to compare the findings to,” Mainiero continued. Such a finding would bode well for improving screening rates and specificity in high-risk women who are wary of the costs and anxiety that accompany false-positive screenings, indicated Mainiero and co-author Gil Abramovici, MD, also of Rhode Island Hospital’s department of diagnostic imaging.
Mainiero and Abramovici reviewed 650 breast MRIs, 307 baseline studies and 343 from women with prior breast MRIs, to compare the false-positive rates of each. All women were at high-risk for breast cancer, with a mean age of 51 years. Among the 343 women with past images, the average interval between the study image and their most recent prior screening was 14 months.
“We found that women who had a prior MRI for radiologists to compare had a much lower false-positive rate,” Mainiero said. The authors observed significantly more BI-RADS category 3 masses in the baseline study group (10.1 percent) compared with women with prior MRIs (2.6 percent).
The positive predictive value of a BI-RADS category 4 or 5 interpretation was 11.1 percent for the baseline group and 18.8 percent for the group with previous images. This difference was not significant, however, due to the small number of category 4 and 5 interpretations, the authors argued.
“These findings should be encouraging to women at high risk for breast cancer,” Mainiero insisted. “While yes, at first, MRI does have a higher risk of false-positive results, but as you get your annual MRIs this becomes less of a problem.” Mainiero and Abramovici hoped that this “information may provide some high-risk patients reassurance when they consider whether to undergo screening breast MR imaging.”
“This study confirms our clinical impressions and what makes sense. The findings are definitive regarding the fact that the false-positive rate goes down as women continue to undergo annual breast MRI screenings.”
The authors also found that despite the disparity in false-positive rates, the cancer detection rates for both groups remained equivalent. Mainiero said that additional studies would be necessary to confirm this finding due to the small number of malignant lesions observed in their sample.
“When you don’t have prior images for comparison, and you see an abnormality that you think is unlikely to be cancer, you’re more likely to recommend a follow-up biopsy,” Mainero described, from the point of view of reading radiologists. “But if you have prior images from last year or the year before, and you can see that the lesion does not change, then you’re no longer worried about it being malignant.”
Mainiero and Abramovici acknowledged that their loss of 10 out of 40 BI-RADS category 3 patients could have biased their results. Moreover, the authors said that the two groups being composed of different patients could limit their findings, though Mainiero said that because the groups did not differ substantially, any effects would be minimal.
“We found our hypothesis to be true, that women who had prior MRIs for radiologists to compare had much lower false-positive rates,” Mainero said. “I think this should encourage women at high-risk for breast cancer and make them less fearful about getting