Objective: To evaluate the malignancy or benignity of breast lesions through MRI, and its confirmation through biopsy. Materials and methods: Retrospective unicentric study. 166 patients were selected, however only 72 were included in the study, as their exams had a histopathological report compatible with a specific laboratory, responsible for analyzing the biopsies. Patients underwent MRI for breast investigation with injection of paramagnetic contrast medium, submitted to biopsy and with histopathological results. All participants were categorized by Bi-RADS through MRI exams. For statistical analysis, the Wilcoxon, Pearson's correlation and Chi-square tests were used. The confidence level used in the analyses was 95%. Results: The most frequent histological type was Invasive Ductal Carcinoma. The BI-RADS category with the highest occurrence was 2 (two). According to the analysis of malignancy given by histopathological findings, 56.9% (41 patients) of the lesions were malignant and 43.1% (31) were benign. When we analyzed the frequency of BI-RADS, in the categories with possibility of malignancy, we observed a distribution, BI-RADS 3 (three) with 8.3% (6 patients), BI-RADS 4 (four) with 9.7% ( 7 patients) and BI-RADS 5 (five) with 15.3% (11 patients), adding up to a total of 33.3% of potentially malignant lesions. Conclusion: It was observed in this study that the correlation between BI-RADS and neoplasia in the study population was weak, being categorized with a probably benign BI-RADS result was not a factor that ruled out the occurrence of cancer. Biopsy and histopathological analysis proved to be useful, demonstrating the importance of always performing it, and correlated with image findings more comprehensively.