Background/Objectives: Spinal metastases are a common and severe complication of lung cancer, particularly in small cell lung cancer (SCLC), and are associated with poor survival. Despite advancements in treatment, optimal management strategies remain unclear, with significant differences between non-small cell lung cancer (NSCLC) and SCLC. This study evaluates treatment patterns, survival outcomes, and prognostic factors in lung cancer patients with spinal metastases, integrating deep learning survival prediction models. Methods: This retrospective cohort study analyzed the National Cancer Database (NCDB) to identify NSCLC and SCLC patients diagnosed with spinal metastases. Demographics and treatment modalities were analyzed and adjusted for age, sex, and comorbidities. Kaplan–Meier analysis and Cox proportional hazards models assessed overall survival (OS). Five advanced survival prediction models estimated 1-year and 10-year mortality, with feature importance determined via permutation analysis. Results: Among 428,919 lung cancer patients, 5.1% developed spinal metastases, with a significantly higher incidence in SCLC (13.6%) than in NSCLC (5.1%). SCLC patients had poorer OS. Radiation therapy alone was the predominant treatment, and stereotactic body radiation therapy (SBRT) predicted better short- and long-term survival compared to other radiation techniques. High-dose radiation (71–150 Gy BED) improved OS in NSCLC, while reirradiation benefited NSCLC but had a limited impact in SCLC. SurvTrace demonstrated the highest predictive accuracy for 1-year and 10-year mortality, identifying age, radiation dose, reirradiation, and race as key prognostic factors. Conclusions: The management of spinal metastases requires a histology-specific approach. Radiation remains the primary treatment, with SBRT predicting better short- and long-term survival. High-dose radiation and reirradiation should be considered for NSCLC, while the benefits are limited in SCLC. These findings support histology-specific treatment strategies to improve survival of patients with metastatic lung cancer to the spine.
背景/目的：脊柱转移是肺癌常见且严重的并发症，尤其在小细胞肺癌（SCLC）中更为突出，并与不良生存预后相关。尽管治疗手段不断进步，但最佳管理策略仍不明确，且非小细胞肺癌（NSCLC）与SCLC之间存在显著差异。本研究结合深度学习生存预测模型，评估肺癌脊柱转移患者的治疗模式、生存结局及预后因素。方法：本回顾性队列研究通过分析美国国家癌症数据库（NCDB），识别诊断为脊柱转移的NSCLC和SCLC患者。对人口统计学特征和治疗方式进行分析，并对年龄、性别及合并症进行校正。采用Kaplan-Meier分析和Cox比例风险模型评估总生存期（OS）。通过五种先进的生存预测模型估算1年及10年死亡率，并利用置换分析确定特征重要性。结果：在428,919例肺癌患者中，5.1%发生脊柱转移，其中SCLC患者发生率（13.6%）显著高于NSCLC（5.1%）。SCLC患者总生存期更差。单纯放疗是主要治疗方式，与其他放疗技术相比，立体定向体部放疗（SBRT）预示更好的短期及长期生存。高剂量放疗（71-150 Gy BED）可改善NSCLC患者OS，再程放疗对NSCLC有益但对SCLC影响有限。SurvTrace模型在预测1年及10年死亡率方面表现出最高的准确性，确定年龄、放疗剂量、再程放疗和种族为关键预后因素。结论：脊柱转移的治疗需采用组织学特异性策略。放疗仍是主要治疗手段，其中SBRT预示更好的短期及长期生存。NSCLC患者应考虑高剂量放疗和再程放疗，而SCLC患者获益有限。这些发现支持采用组织学特异性治疗策略以改善肺癌脊柱转移患者的生存预后。
肿瘤(癌症)患者之家