Background. Due to the high false negative rate (FNR) associated with sentinel lymph node biopsy (SLNB) after neoadjuvant systemic therapy (NAST), the standard surgical treatment for patients with an initially positive axilla and indicated for NAST is axillary lymph node dissection (ALND). To avoid unnecessary ALND, this multicenter, prospective, observational study aimed to determine the effectiveness and ease of using magnetic seeds (Magseed®) for targeted axillary dissection (TAD) when the seeds are placed before or after NAST. Materials and Methods. We recruited 81 patients diagnosed with T1-T3 breast cancer, with clinically/radiologically positive nodal involvement (cN1, 75 patients with 1–3 nodes suspected nodes and 6 patients with up to 4 suspected nodes) prior to NAST. Positive nodes detected by fine-needle aspiration biopsy or core needle biopsy were marked with a stainless steel marker coil and after NAST with Magseed® prior to surgery (Post-NAST group), or directly with Magseed® before NAST (Pre-NAST group). The correlation between lymph nodes marked with Magseed® (MLNs) and sentinel lymph nodes (SLNs) was calculated based on pathologic assessment with the OSNA assay (Sysmex Corporation, Kobe) or conventional sectioning and staining techniques according to the standard protocols of each center. Results. All magnetic seeds were successfully identified and retrieved in just over 10 minutes of surgery, guided by the Sentimag® magnetometer system. The overall concordance rate between MLNs and SLNs was 81.5%, and the concordance between MLNs and SLNs with metastasis was 93.8%. Metastasis was detected in 54.3% of the MLNs compared with 48.1% of SLNs. In cases that presented negative MLN and negative SLN (negative TAD), the FNR was 0%. No significant differences were found between the Post-NAST and Pre-NAST groups. Conclusions. Our results validate the use of Magseed® for long-term marking of axillary lymph nodes and show that when used in combination with SLNB for TAD, a FNR of 0% can be achieved, avoiding unnecessary ALND.