Breast lesions classified as probably benign (BI-RADS 3) on magnetic resonance imaging: a systematic review and meta-analysis.

Purpose: To investigate prevalence, malignancy rates, imaging features, and follow-up intervals for probably benign (BI-RADS 3) lesions on breast magnetic resonance imaging (MRI).

Methods: A systematic database-review of articles published through 22/06/2016 was performed. Eligible studies reported BI-RADS 3 lesions on breast MRI. Two independent reviewers performed a literature review and data extraction. Data collection included study characteristics, number/type of BI-RADS 3 lesions, final diagnosis (histopathology and/or follow-up). Sources of bias (QUADAS-2) were assessed. Meta-analysis included data-pooling, heterogeneity testing, and meta-regression.

Results: Fifteen studies were included. Prevalence was reported in 11 studies (range: 1.2-24.3%). Malignancy rates ranged between 0.5-10.1% (pooled 61/2814, 1.6%, 95%-CI:0.9-2.3% (random-effects-model), I=53%, P=0.007). In a subgroup of 11 studies (2183 lesions), highest malignancy rates were observed in non-mass lesions (pooled 25/714, 2.3%, 95%-CI:0.8-3.9%, I=52%, P=0.021) followed by mass lesions (pooled 15/771, 1.5%, 95%-CI:0.7-2.4%, I=0%, P=0.929), and foci (pooled 10/698, 1%, 95%-CI:0.3-1.7%, I=0%, P=0.800). There was non-significant negative association between prevalence and malignancy rates (P=0.077). Malignant lesions were diagnosed at all follow-up time points.

Conclusion: While prevalence of MRI BI-RADS 3 lesions was strongly heterogeneous, pooled malignancy rates met BI-RADS benchmarks (<2%). Malignancy rates varied, exceeding 2% in non-mass lesions. Twenty-four-month surveillance is required to detect all malignant lesions.

Key Points: • Probably benign (BI-RADS 3) lesions showed a pooled malignancy-rate of 1.6% (95%-CI:0.9-2.3%). • Malignancy rates differ and are highest in non-mass lesions (2.3%, 95%-CI:0.8-3.9%). • The prevalence of BI-RADS 3 lesions on breast MRI ranged from 1.2-24.3%. • Malignant lesions were diagnosed at follow-up time points up to 24 months.