MRI vs. Mammography: Which is Better for Screening Women at High Risk of Breast Cancer?
Key Words: Breast cancer, prevention, screening, magnetic resonance imaging (MRI), ultrasound, mammography, clinical breast exam, BRCA1, BRCA2. (Definitions of many terms related to cancer can be found in the Cancer.gov Dictionary.)
Summary Three studies examined the value of magnetic resonance imaging (MRI) in screening women at high risk for breast cancer. The studies concluded MRI was a more sensitive screening tool in finding cancers than mammography, ultrasound, or clinical breast exams (CBEs). However, results also showed that MRI gives a higher number of “false positives.” False positive results can lead to unnecessary additional tests, biopsies, and increased patient anxiety.
Source American Society of Clinical Oncology (ASCO) annual meeting, Chicago, June 2, 2003.
Background: Women are considered to be at high risk for breast cancer if they have a strong family history of the disease (i.e., multiple affected first-degree relatives, especially at an early age), or if they have documented or suspected BRCA1 or BRCA2 gene mutations, which confer a greater than 50 percent chance of developing breast cancer. Three studies presented at the 2003 annual meeting of the American Society of Clinical Oncology (ASCO) examined whether MRI might be a more effective screening tool than mammography for such high-risk women.
The studies considered two measures of screening effectiveness: sensitivity (how well the tool detects a cancer when one is present) and specificity (how well the tool avoids “false positives” – suggesting a tumor when there isn’t one). In the studies, if screening suggested the presence of cancer, a biopsy was performed to confirm or deny the finding.
Study 1 From November 1999 to August 2002, researchers with the Dutch MRI Screening Study evaluated 1,911 high-risk women in a non-randomized, prospective, multi-center study. The women received a twice-yearly breast exam by a health professional (called a clinical breast exam, or CBE), a yearly mammography, and a yearly MRI. Researchers evaluated each woman’s mammography results independently of her MRI results, so if one imaging modality suggested the presence of cancer the evaluator would not be biased towards expecting to see signs of cancer in the other modality, and vice versa.
Study 1 ResultsEach woman in the study was followed for about two years. During this time, researchers found invasive breast cancers or non-invasive tumors (such as ductal carcinoma in situ, or DCIS) in 40 of the women. Forty-six percent of the tumors were small (1 centimeter or less) and 77 percent were confined (localized) to the breast. While the clinical breast exam detected 16 percent of the tumors and mammography found 36 percent, MRI sensitivity was found to be 71 percent. MRI sensitivity was even more pronounced in cases of invasive cancer (spread beyond the layer of tissue in which it developed), with 20 percent found by CBE, 26 percent by mammography, and 83 percent by MRI.
However, MRI was less specific than the other two modalities; that is, it was more likely to produce “false positive” results – to incorrectly suggest the presence of cancer. Twelve percent of the time, MRI suggested that there was cancer when there wasn’t, compared to 5 percent for mammography and 3 percent for CBE.
Breast cancer oncologist William J. Gradishar, M.D., of Northwestern University in Evanston, Ill., who moderated the presentation of these data at an ASCO press conference on June 2, 2003, said the results of the Dutch study suggest that “the use of MRI for high-risk woman is an appropriate and promising way of identifying tumors.” Nonetheless, he said, the ultimate value of the approach “requires more study.”
Study 2A second clinical trial presented at ASCO came from the University of Bonn. Researchers in this trial screened 462 high-risk women by yearly CBE, mammography, ultrasound, and MRI. Over a five-year period, 51 breast cancers were detected in 45 of the women. This study is ongoing and will continue another five years.
Study 2 ResultsThe results so far suggest that MRI offers the greatest sensitivity – 96 percent – for correctly diagnosing breast cancer in this high-risk group. In contrast, the sensitivity rate for detecting breast cancer was 25 percent for CBE, 43 percent for mammography, and 47 percent for ultrasound. Unlike the Dutch trial (Study 1), this trial found that MRI actually resulted in fewer cases of false positives than the other forms of screening, leading the German researchers to suggest that MRI should replace mammography as a screening tool for high-risk women. (Note: these data were subsequently published in the Nov. 20, 2005, issue of the Journal of Clinical Oncology; see the journal abstract.)
Study 3 In a third study that did not directly compare MRI with other forms of breast cancer screening, researchers with Memorial Sloan-Kettering Cancer Center in New York reviewed medical and radiology reports for 54 women with BRCA mutations who had 115 MRI exams between 1998 and 2002.
Study 3 ResultsMRI was 100 percent sensitive for correctly detecting breast cancer – finding a tumor when there was one – but only 83 percent specific. That is, 17 percent of MRI’s positive findings were false alarms. Researchers with this study concluded that MRI sensitivity is encouraging but that the high false-positive rate limits its use as a routine practice.
Limitations of the Three StudiesWhile the sensitivity of MRI in detecting breast cancer among high-risk women is encouraging, the significant false-positive rate indicates that further research is warranted before recommending that MRI be commonly used in early detection even among this at-risk group.
Also, the sensitivity is likely to be lower than the figures reported, since biopsies were only performed for abnormal tests. Some cancers were not seen by any available test and so they were not biopsied.
All the researchers emphasized that MRI is inappropriate for women not considered at very high genetic or familial risk for breast cancer, due to the cost of MRI and the rate of false positives.
 Back to Top |
