研究背景：ROS1 基因重排是非小细胞肺癌治疗的重要靶点。在中国人群中，关于 ROS1检测方法、治疗选择和治疗结果的数据尚不充分。因此，本研究旨在分析并总结中国晚期 ROS1 阳性非小细胞肺癌患者的 ROS1 检测方法、治疗方案和临床结果，以及探寻其进展模式和可能的耐药机制。方法：对 2011 年 7 月至 2021 年 11 月诊断为 ROS1 阳性晚期非小细胞肺癌的患者进行了单中心回顾性收集，以记录这些患者接受的 ROS1 检测方法、治疗选择、疗效和对ROS1 抑制剂的耐药性。结果：本研究共纳入了 40 名患者。ROS1 检测方法和初始治疗选择与时代显著相关（FISH 是 2019 年之前最常用的检测方法：67%，NGS 是 2019 年之后最常见的检测方法：96.3%，P <0.001；2019 年之前 60.0%的患者选择 ROS1 抑制剂作为一线治疗，而 2019 年之后 92.0%的患者选择 ROS1 抑制剂进行一线治疗，P=0.041）。绝大多数患者（90.0%）选择克唑替尼作为初始 ROS1 抑制剂，ORR（针对有靶病变的患者）为 82.8%（95%CI: 68.1-97.9%），PFS 为 18.7 月（95%CI: 8.9-28.4 月）。体力情况是克唑替尼疗效的一个影响因素（ECOG 评分 0-1：35.9 月 vs ECOG 评分≥2：11.1 个月，P=0.02）。中枢神经系统是克唑替尼最常见的进展部位（60%，13/26，包括 11 例单独的颅内进展）。对于 ROS1 抑制剂耐药的患者，48.3%的患者在整个疾病过程中接受了再次活检，其中 G2032R 突变是最常见的继发耐药突变类型（7/8）。结论：随着诊疗技术的革新和医保目录范围的扩大，越来越多患者从新技术和靶向药物中获益。克唑替尼虽然为 ROS1 阳性患者带来了优异的治疗数据，但对于 ROS1 阳性患者仍应在未来探索更佳的脑保护策略。此外，真实世界中 ROS1 阳性患者再活检比例偏低，应在未来提高重视。

Background: ROS1 gene rearrangement is an important target for NSCLC treatment. There is not yet sufficient real-world data on ROS1 diagnostic methods, treatment selection and clinical outcomes in the Chinese population. Therefore, this study was intended to retrospectively analyze the ROS1 detection methods, treatment options and clinical outcomes of advanced ROS1 positive NSCLC patients in China, as well as the progression pattern and possible drug resistance mechanism. Methods: A single-center retrospective collection of patients with diagnosis of ROS1-positive advanced NSCLC from July 2011 to November 2021 was performed to document the method of ROS1 testing, treatment option, efficacy and resistance to ROS1 inhibitors in these patients. Results: The method of ROS1 testing and initial treatment selection were significantly correlated with time. ROS1 testing shifted from FISH (67% pre-2019) to NGS (96.3% post-2019; P <0.001). First-line ROS1-TKI use increased from 60.0% to 92.0% (P=0.041). The vast majority of patients (90.0%) chose crizotinib as the initial ROS1 inhibitor, with objective response rates (for patients with target lesions) and median progression-free survival of 82.8% (95% CI: 68.1%–97.9%) and 18.7months (95% CI: 8.9–28.4months), respectively. CNS was the most common site of progression for crizotinib (60%, 13/26, including 11 intracranial progressions alone). Compared to patients who received a chemotherapy-based regimen (n=8) as first-line therapy, patients who received ROS1-TKI (n=32) as first-line therapy had significantly longer median PFS (18.3months vs. 3.7months, P <0.001). For ROS1 inhibitor-resistant patients, 48.3% of patients underwent rebiopsy throughout the course of their disease, with G2032R being the most common secondary ROS1 mutation (7/8). Conclusion: With the innovation of diagnostic and therapeutic methods and the expansion of the scope of health insurance catalog, more and more patients are benefiting from new technologies and targeted drugs. Although crizotinib has brought excellent therapeutic data for ROS1-positive patients, better brain protection strategies should be explored for ROS1-positive patients in the future. In addition, the low rate of rebiopsy in real-world ROS1-positive patients should also be emphasized in the clinical practice.