Abstract
BACKGROUND: Although de novo metastatic breast cancer (dnMBC) is acknowledged as a heterogeneous disease, the current staging systems do not distinguish between patients within the M1 or stage IV category. This study aimed to refine the M1 category and prognostic staging for dnMBC to enhance prognosis prediction and guide the choice of locoregional treatment. METHODS: We selected patients with dnMBC from the SEER database (2010-2019), grouping them into training (N = 8048) and internal validation (N = 3450) cohorts randomly at a 7:3 ratio. An independent external validation cohort (N = 660) was enrolled from dnMBC patients (2010-2023) treated in three hospitals. Nomogram-based risk stratification was employed to refine the M1 category and prognostic stage, incorporating T/N stage, histologic grade, subtypes, and the location and number of metastatic sites. Both internal and external validation sets were used for validation analyses. RESULTS: Brain, liver, or lung involvement and multiple metastases were independent prognostic factors for overall survival (OS). The nomogram-based stratification effectively divided M1 stage into three groups: M1a (bone-only involvement), M1b (liver or lung involvement only, with or without bone metastases), and M1c (brain metastasis or involvement of both liver and lung, regardless of other metastatic sites). Only subtype and M1 stage were included to define the final prognostic stage. Significant differences in OS were observed across M1 and prognostic subgroups. Patients with the M1c stage benefited less from primary tumor surgery in comparison with M1a stage. CONCLUSION: Subdivision of the M1 and prognostic stage could serve as a supplement to the current staging guidelines for dnMBC and guide locoregional treatment.