Purpose The optimal design of radiation fields for patients with positive sentinel lymph nodes (SLNs) who do not undergo axillary dissection is unknown. We have previously shown that modified breast tangent fields can include most axillary Level I-II lymph nodes. We have also reported that irradiation of the axillary apex/supraclavicular fossa is indicated for patients with four or more positive axillary lymph nodes. To determine the optimal arrangement for patients with positive SLNs, we studied what factors predicted for having four or more positive lymph nodes. Methods and materials We reviewed the records of 339 consecutive patients with one to three positive SLNs who underwent complete axillary dissection at our institution between 1995 and 2002. We separately analyzed the outcome for those initially treated with surgery ( n = 265) and those receiving neoadjuvant chemotherapy ( n = 74). A logistic regression model was used to identify independent factors predictive for four or more positive lymph nodes. Results A total of 28 of 265 patients in the initial surgery group and 20 of 74 patients in the neoadjuvant group had four or more positive lymph nodes. In the initial surgery group, the independent factors associated with four or more positive lymph nodes were no drainage seen on lymphoscintigraphy (rate, 38%, odds ratio [OR] = 5.4, p = 0.03), more than one positive SLN (rate, 24–42%, OR = 2.9, p = 0.02), and lymphovascular space invasion (LVSI; rate, 25%, OR = 4.8, p = 0.01). Of the 106 patients without any of these factors, only 2 had four or more positive lymph nodes. For the patients treated with neoadjuvant chemotherapy, the independent factors were clinical Stage III (rate, 48%, OR = 3.1, p = 0.03), more than one positive SLN (rate, 37–67%, OR = 4.8, p = 0.03), and LVSI (rate, 62%, OR = 8.1, p = 0.02). Of the 28 patients without any of these factors, only 1 had four or more positive lymph nodes. Conclusion It is reasonable to treat with modified tangents fields that include most axillary Level I-II nodes for patients with one positive SLN who do not undergo axillary dissection if drainage is seen on lymphoscintigraphy and no LVSI is present. This approach is also reasonable for patients treated with neoadjuvant chemotherapy who have Stage II disease, no LVSI, and only one positive SLN. The remaining patients have a greater risk of having four or more positive lymph nodes, and, therefore, the high axilla/supraclavicular fossa should also be included in the radiation fields.