Background: The possibility of primary tumor resection (PTR) improving the survival of de-novo Stage IV breast cancer (dn-StIV BC) patients has been evaluated by several prospective studies but remains controversial. We designed this phase 3 trial (JCOG1017) comparing with/without primary dissection after initial systemic therapy based on clinical subtype in dn-StIV BC patients. Methods: Dn-StIV BC patients were enrolled in the first registration. All patients received systemic therapies according to clinical subtypes. The patients not showing refractory disease were randomized to systemic therapy alone (arm A) or PTR plus systemic therapy (Arm B). The same systemic therapy was continued after randomization as additional therapy, for as long as possible. The primary endpoint was overall survival (OS). Secondary endpoints included the proportion of patients without progression at metastatic sites after initial systemic therapies for 3 months, local relapse-free survival (LRFS), primary tumor resection-free survival, and incidence of local ulcer/local bleeding and adverse events. The median overall survival time (MST) after initial systemic therapy for patients with dn-StIV BC was 20 months, on average, and a clinically relevant prolongation of the MST of Arm B was considered to be 6.0 months or longer (hazard ratio: 0.77). The required number of events was 359, to obtain a statistical power of 80% with a one-sided significance level of 0.05. Thus, the planned sample size was 410 patients for the second registration, assuming an accrual period of 7 years and a follow-up time of 4 years. Results: 570 patients were enrolled between 11/5/2011 and 31/5/2018 in the first registration. Of these, 407 eligible patients were randomized to either Arm A (N = 205) or Arm B (N = 202). The patient characteristics were well balanced between the two arms. The MST of randomized patients was 70 months, with 221 deaths. The difference in OS was not statistically significant (HR 0.857, 90% CI 0.686-1.072, one-sided p = 0.1283). MST was 69 months in Arm A and 75 in Arm B. The proportions of patients without progression at metastatic sites in Arms A and B were 81.5% and 67.3%, respectively (p = 0.0014). LRFS in Arm B was significantly longer than that in Arm A (median LRFS 20 vs 63 months: HR 0.415, 95% CI 0.327-0.527, p < 0.0001). In Arm B, patients with incomplete resection had poorer outcomes than those in whom resection was complete (94 vs. 61 months, HR 1.971 (1.161-3.347) p = 0.0120). In the subgroup analysis, PTR improved survival in patients with ER-positive tumors, pre-menopausal status or single-organ metastasis. Conclusions: PTR is not recommended for all dn-StIV BC patients but can control local disease and is acceptable in a select population of patients because of the clear improvement in local control. Clinical trial information: UMIN000005586.