Supplementary Material for: Real-World Treatment Patterns, Survival, and Prediction of CNS Progression in ALK-Positive Non-Small-Cell Lung Cancer Patients Treated with First-Line Crizotinib in Latin America Oncology Practices

Objective: This study describes the real-world characteristics, treatment sequencing, and outcomes among Hispanic patients with locally advanced/metastatic ALK-positive non-small-cell lung cancer (NSCLC) treated with crizotinib. Methods: A retrospective patient review was conducted for several centers in Latin America. Clinicians identified ALK-positive NSCLC patients who received crizotinib and reported their clinical characteristics, treatments, and survival. Overall survival and progression-free survival (PFS) were described. A Random Forest Tree (RFT) model was constructed to predict brain progression. Results: A total of 73 patients were included; median age at diagnosis was 58 years, 60.3% were female, and 93.2% had adenocarcinoma. Eighty-nine percent of patients were never smokers/former smokers, 71.1% had ≥2 sites of metastasis, and 20.5% had brain metastases at diagnosis. The median PFS on first-line crizotinib was 7.07 months (95% CI 3.77–12.37) and the overall response rate was 52%. Of those who discontinued crizotinib, 55.9% progressed in the central nervous system (CNS). The RFT model reached a sensitivity of 100% and a specificity of 88% for prediction of CNS progression. Conclusions: The overall response rate and the PFS observed in Hispanic patients with ALK-positive NSCLC treated with first-line crizotinib were similar to those in previous reports. An RFT model is helpful in predicting CNS progression and can help clinicians tailor treatments in a resource-limited practice.研究目的：本研究旨在描述接受克唑替尼（crizotinib）治疗的局部晚期/转移性间变性淋巴瘤激酶（anaplastic lymphoma kinase, ALK）阳性非小细胞肺癌（non-small-cell lung cancer, NSCLC）西班牙裔患者的真实世界临床特征、治疗序贯方案及预后转归。 方法：本研究针对拉丁美洲多家医疗中心开展回顾性病例研究。由临床医师筛选接受克唑替尼治疗的ALK阳性NSCLC患者，收集其临床特征、治疗方案及生存数据，并分析总生存期（overall survival, OS）与无进展生存期（progression-free survival, PFS）。此外，构建随机森林树（random forest tree, RFT）模型以预测中枢神经系统（central nervous system, CNS）进展。 结果：本研究共纳入73例患者；诊断时中位年龄为58岁，其中60.3%为女性，93.2%的患者病理类型为腺癌。89%的患者为从不吸烟者或既往吸烟者，71.1%存在≥2个转移部位，诊断时即合并脑转移者占20.5%。一线克唑替尼治疗的中位PFS为7.07个月（95%置信区间3.77~12.37），总体缓解率为52%。在停用克唑替尼的患者中，55.9%出现CNS进展。本研究所构建的RFT模型预测CNS进展的灵敏度达100%，特异度为88%。 结论：本研究中接受一线克唑替尼治疗的西班牙裔ALK阳性NSCLC患者，其总体缓解率与PFS与既往相关报道结果相近。RFT模型可有效预测CNS进展，有助于临床医师在资源受限的医疗环境中制定个体化治疗策略。