Magnetic resonance imaging in the preoperative assessment of patients with primary breast cancer: systematic review of diagnostic accuracy and meta-analysis

Abstract

To estimate the diagnostic accuracy of magnetic resonance imaging (MRI) in detecting additional lesions and contralateral cancer not identified using conventional imaging in primary breast cancer. We have conducted a systematic review and meta-analyses to estimate diagnostic accuracy indices and the impact of MRI on surgical management. Fifty articles were included (n = 10,811 women). MRI detected additional disease in 20% of women and in the contralateral breast in 5.5%. The summary PPV of ipsilateral additional disease was 67% (95% CI 59-74%). For contralateral breast, the PPV was 37% (95% CI 27-47%). For ipsilateral lesions, MRI devices ≥1.5 Tesla (T) had higher PPV (75%, 95% CI 64-83%) than MRI with <1.5 T (59%, 95% CI 53-71%). Similar results were found for contralateral cancer, PPV 40% (95% CI 29-53%) and 19% (95% CI 8-39%) for high- and low-field equipments, respectively. True positive MRI findings prompted conversion from wide local excision (WLE) to more extensive surgery in 12.8% of women while in 6.3% this conversion was inappropriate. MRI shows high diagnostic accuracy, but MRI findings should be pathologically verified because of the high FP rate. Future research on this emerging technology should focus on patient outcome as the primary end-point.