Introduction: Breast cancer (BC) is a disease that affects about 2.2 million people worldwide. The prognosis and treatment of these patients depend on clinical and histopathologic staging, in which more aggressive cancers need a less conservative therapeutic approach. Previous studies showed that patients with BC have an increased frequency of systemic microvesicles (MVs) that are associated with invasion, progression, and metastasis, which can be used in liquid biopsy to predict the therapeutic response in individualized treatment. Objective: This study proposes the development of a minimally invasive BC diagnostic panel and follow-up biomarkers as a complementary method to screen patients. Methods: The quantification of circulating MVs in 48 healthy women and 100 BC patients who attended the Mario Penna Institute between 2019 and 2022 was performed by flow cytometry. In addition, the MVs of BC patients were analyzed before treatment and 6, 12, and 24 months post-treatment. Machine learning approaches were employed to determine the performance of MVs to identify BC and to propose BC classifier algorithms. Results: Patients with BC had more neutrophil- and endothelial cell-derived MVs than controls before treatment. After treatment, all MV populations were decreased compared to pre-treatment, but leukocyte- and erythrocyte-derived MVs were increased at 12 months after treatment, before decreasing again at 24 months. Conclusions: Performance analyses and machine learning approaches pointed out that MVs from neutrophils and endothelial cells are the best candidates for BC diagnostic biomarkers. Neutrophil- and endothelial cell-derived MVs are putative candidates for BC biomarkers to be employed as screening tests for BC diagnosis.