Prospective or retrospective test accuracy studies that evaluated clinically suspicious breast lesions using MRI (according to the American College of Radiology practice guidelines) were eligible for inclusion.
Eligible studies were conducted in women with mammographic and sonographic abnormalities in whom physical examination findings were inconclusive. MRI was performed using high-field strength commercially available 0.5 or 3 T systems, with a dedicated surface breast coil. A number of different criteria (details provided in the article) were used to indicate a positive MRI. The reference standard used to confirm diagnosis was histological analysis of standard paraffin-embedded sections after surgery. Studies had to include at least two of the following diagnoses: benign, high risk lesions and breast cancer.
Studies of patients who were known carriers of a BRCA 1 or BRCA 2 mutation, had a first-degree family member known to carry a BRCA 1 or BRCA 2 mutation, or had a personal or family history highly suggestive of BRCA 1 or BRCA 2 mutation, were excluded. Studies were excluded if participants had a 15 to 20% lifetime risk for breast cancer on the basis of a personal history of breast or ovarian cancer, biopsy proven lobular neoplasia, or atypical ductal hyperplasia, and/or had received radiation therapy to the chest before the age of 30 years and at least eight years previously. Studies which did not report sufficient data to construct 2x2 contingency tables (numbers of true positive, false negative, false positive and true negative test results) were excluded.
The mean age of study participants ranged from 45 to 58.5 years (where reported). Most included studies were conducted in Europe or the USA.
Studies were assessed for inclusion by six reviewers and any disagreements were resolved by consensus, or (where this was not possible) by the arbitrarily decision of one reviewer.