Objective ·To investigate the efficacy and safety of weekly paclitaxel in combination with cisplatin (DP) as first-line regimen compared with other treatment of physician's choices (TPC) for metastatic breast cancer (MBC). Methods ·The clinical data of 117 MBC patients who were hospitalized in the Department of Oncology, Renji Hospital, Shanghai Jiao Tong University School of Medicine from July 2008 to June 2021 were retrospectively analyzed. Fifty-five patients were divided into DP group and 62 patients were divided into TPC group, according to whether the DP regimen was used as first-line chemotherapy. The DP group received paclitaxel 80 mg/m² combined with cisplatin 25 mg/m², administered on the 1st, 8th and 15th day, respectively, every 28 d as a cycle, a total of 6 cycles of treatment. TPC group received other first-linechemotherapy regimens, including single or combined regimens of capecitabine, gemcitabine, docetaxel and other regimens.Clinicopathologic characteristics including age, menopausal status, pathological results and molecular subtypes were collected and analyzed. Objective response rate (ORR) and clinical benefit rate (CBR) were used to evaluate the efficacy of different regimens based on Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. The adverse effects (AEs) were assessed in accordance with National Cancer Institute's criteria for common adverse effect 5.0 (NCI CTC 5.0), and NCI Ⅲ/Ⅳ AEs were analyzed as severe AEs. The numbers and sites of MBC relapses were analyzed as progression pattern. The differences of clinicopathologic characteristics, efficacy and severe AEs between the two groups were compared by χ² test. Kaplan-Meier survival analysis was used to compare the difference of first-line progression free survival (PFS1) between the two groups, and the factors with probable influence on PFS1, including clinicopathologic characteristics, treatments and other factors were analyzed. The univariate Cox regression analysis was further applied to analysis of the possible influencing factors screened above. Results ·Except the progesterone receptor (PR) expression (P=0.048), there was no significant difference in clinicopathologic characteristics between the two groups (P>0.05). The ORR (47.3% vs 22.6%, P=0.009) and CBR (78.2% vs 41.9%, P=0.002) in the DP group were significantly higher than those in the TPC group. The median PFS1 was 12.0 months (95%CI 10.0?15.0 months) in the DP group, while the median PFS1 was 6.0 months (95%CI 5.0?9.0 month) in the TPC group, demonstrating significant difference (Log-rank P=0.000). Only the factor "receiving DP regimen as first-line chemotherapy" was recognized as a protective factor for MBC progression by univariate Cox analysis (P=0.000, HR=0.419, 95%CI 0.271?0.649). There was no significant difference (P>0.05) in the progression pattern between the two groups except for fewer bone metastasis (P=0.006). The AEs of the two groups were tolerable, and there was no significant difference in the proportion of severe AEs between the two groups (P>0.05). Conclusion ·The first-line DP regimen demonstrates better efficacy and prolonged PFS1 than TPC regimens. It could be an effective selection and should be considered for MBC according to the clinical situations.