MRI Screening of BRCA Mutation Carriers: Comparison of Standard Protocol and Abbreviated Protocols With and Without T2-Weighted Images

Increasing evidence supports the role of abbreviated MRI protocols for breast cancer detection. However, abbreviated protocols have been poorly studied in patients who are or mutation carriers. Furthermore, the need for T2-weighted sequences in abbreviated protocols remains controversial. The purpose of this study was to compare, in the evaluation of patients with mutations, the diagnostic performance of a standard full breast MRI protocol with the performance of abbreviated protocols that included and did not include a T2-weighted sequence. This retrospective study included 292 patients (mean age, 47.9 years) who were or mutation carriers who underwent 427 screening breast MRI examinations according to a standard full protocol who could be classified as having benign ( = 407) or malignant ( = 20) findings based on histopathology or imaging follow-up. Four readers independently assessed examinations in three separate sessions (theoretic abbreviated protocol, which included the first postcontrast acquisition; theoretic abbreviated protocol with addition of a T2-weighted sequence; and the standard full protocol) and assigned BI-RADS categories. Categories 3-5 were considered to represent positive examinations. Interreader agreement was assessed, and diagnostic performance was compared by use of pooled reader data. Interreader agreement on BI-RADS category, expressed as kappa values, was 0.55 for the standard, 0.45 for the abbreviated, and 0.57 for the abbreviated plus T2-weighted protocols. Pooled sensitivity was 94% for the standard, 92% for the abbreviated, and 90% for the abbreviated plus T2-weighted protocols (all > .001). Pooled specificity was 80% for the standard, 71% for the abbreviated, and 83% for the abbreviated plus T2-weighted protocols ( < .001 for abbreviated plus T2-weighted compared with both standard and abbreviated). Pooled PPV was 19% for the standard, 14% for the abbreviated, and 20% for the abbreviated plus T2-weighted protocols ( < .001 for abbreviated compared with both standard and abbreviated). Pooled NPV was 100% for the standard, 99% for the abbreviated, and 99% for the abbreviated plus T2-weighted (all > .001) protocols. Pooled accuracy was 80% for the standard, 73% for the abbreviated, and 83% for the abbreviated plus T2-weighted protocols ( < .001 for abbreviated compared with both standard and abbreviated plus T2-weighted). The abbreviated protocol without T2-weighted imaging had suboptimal performance. However, addition of th