The purpose of the study was to evaluate the outcome of initially only magnetic resonance mammography (MRM)-detected breast lesions as a function of radiologic features, history of breast cancer and lesion size. We evaluated core needle biopsy (CNB) (148) and follow-up (25) results of 173 initially only MRM-detected lesions-142 with and 31 without "second-look" correlate, as a function of (1) radiologic features (sonographic correlate, MRI BI-RADS category); (2) history of breast cancer; (3) MRM indication in case of history of breast neoplasm; (4) side and size of synchronous cancer; (5) lesion diameter. (1) Overall malignancy rate was 28.3% (49/173); significantly higher among lesions with a sonographic correlate (46/142), than among those without (3/31) (p=0.014). Frequencies of malignancy for MRI BI-RADS categories 2, 3, 4 and 5, were 0% (0/1), 5.4% (4/73), 26.1% (17/65) and 82.3% (28/34), respectively. (2) Malignancy rate was significantly higher in case of history of breast carcinoma (40/118 versus 9/55; p=0.027); in particular, of 42 MRI BI-RADS category 3 lesions in women with history of breast cancer and of 31 in patients without history, 3 (7%) and 1 (3%) proved to be malignant, respectively (non-significant). (3) Malignancy was more frequent when MRM was performed for pre-operative assessment than for follow-up (30/78 versus 10/40; non-significant). (4) Malignancy rate increased in presence of ipsilateral (19/35 versus 11/43; p=0.018), large (cut-off 6 mm: 30/75 versus 0/3, non-significant; 11 mm: 28/61 versus 2/17, p=0.011; 16 mm: 24/48 versus 6/30, p=0.015; 21 mm: 14/21 versus 16/57, p=0.004) primary tumors. (5) The frequency of malignancy was significantly higher in lesions equal to or larger than 6, 11 and 16 mm, compared with smaller lesions (6 mm: 45/136 versus 4/37, p=0.007; 11 mm: 21/51 versus 28/122, p=0.025; 16 mm: 12/24 versus 37/149, p=0.021). Radiologic features, history of breast cancer and large diameter are associated with high likelihood of malignancy in case of initially only MRM-detected lesions. Nevertheless, biopsy might be spared just for MRI BI-RADS 3 lesions in patients without history of breast carcinoma.