BACKGROUND The impact of locoregional treatment (LRT) on overall survival (OS) in de novo metastatic breast cancer (dnMBC) is still under debate, with very few data available regarding exclusive radiotherapy (ERT) as a therapeutic modality. METHODS We evaluated the impact of ERT, exclusive surgery, or a combination of surgery plus radiotherapy (bimodality therapy, BMT) on survival outcomes in a national real-life dnMBC cohort. The primary and secondary end points were OS and progression free survival (PFS) according to LRT (ERT, exclusive surgery, BMT) and no LRT. Sensitivity analyses were performed using propensity score matched analyses. RESULTS From 2008 to 2014, 4507 dnMBC patients were identified. Only patients alive and free from progression under systemic therapy at least 1 year after diagnosis were included (n = 1965). Forty-five percent of patients (891/1965) underwent LRT: 41.1% (n = 366) ERT, 13.7% (n = 122) exclusive surgery, and 45.2% (n = 403) BMT. OS adjusted for major prognostic factors was significantly longer in the ERT and BMT group compared with no-LRT group, but not exclusive surgery (hazard ratio (HR) = 0.63, 95% confidence interval (CI) [0.49, 0.80], p < 0.001, HR = 0.61, 95%CI [0.47, 0.78], p < 0.001 and HR = 0.87, 95%CI [0.61, 1.26], p = 0.466 respectively). Results were similar after matching on a propensity score. ERT, surgery and BMT were all associated with a significantly better PFS in multivariable analysis. CONCLUSION ERT was significantly associated with better OS in dnMBC, in the same magnitude as BMT, compared with no-LRT. However, even with statistical models adjusted for known prognostic factors and propensity score analysis, selection biases cannot be eliminated from observational studies.

中文翻译：

---

回顾性多中心队列分析中原发肿瘤的放射治疗新发转移性乳腺癌和总生存期

背景 局部治疗 (LRT) 对新发转移性乳腺癌 (dnMBC) 总生存期 (OS) 的影响仍存在争议，关于作为一种治疗方式的排他性放疗 (ERT) 的数据很少。方法 我们在全国真实 dnMBC 队列中评估了 ERT、独家手术或手术加放疗的组合（双峰疗法，BMT）对生存结果的影响。主要和次要终点是根据 LRT（ERT，独家手术，BMT）和无 LRT 的 OS 和无进展生存期 (PFS)。使用倾向评分匹配分析进行敏感性分析。结果 从 2008 年到 2014 年，确定了 4507 名 dnMBC 患者。仅包括在诊断后至少 1 年在全身治疗下存活且无进展的患者（n = 1965）。45% 的患者 (891/1965) 接受了 LRT：41.1% (n = 366) ERT、13.7% (n = 122) 独家手术和 45.2% (n = 403) BMT。与非 LRT 组相比，ERT 和 BMT 组在调整主要预后因素后的 OS 显着更长，但不排除手术（风险比 (HR) = 0.63，95% 置信区间 (CI) [0.49, 0.80]，p < 0.001, HR = 0.61, 95%CI [0.47, 0.78], p < 0.001 和 HR = 0.87, 95%CI [0.61, 1.26], p = 0.466)。匹配倾向得分后的结果相似。在多变量分析中，ERT、手术和 BMT 都与显着更好的 PFS 相关。结论 与无 LRT 相比，ERT 与 dnMBC 中更好的 OS 显着相关，与 BMT 的幅度相同。然而，即使根据已知的预后因素和倾向评分分析调整了统计模型，