Division of Oncology, Huntsman Cancer Institute, University of Utah, Salt Lake City, Utah.
Department of Population Health Sciences, University of Utah, Salt Lake City, Utah.
J Urol. 2021 Mar;205(3):709-717. doi: 10.1097/JU.0000000000001412. Epub 2020 Oct 20.
Five programmed cell death protein 1 or its ligand (L1) inhibitors are approved for treatment of platinum refractory, locally advanced/unresectable or metastatic urothelial carcinoma. However, their comparative effectiveness is unknown. We compared time to initiation of third therapy or death, and overall survival with different programmed cell death protein 1/L1 inhibitors in patients with platinum refractory metastatic urothelial carcinoma.
Patientlevel data were extracted from a real-world de-identified database. Comparative effectiveness was inferred via Cox proportional hazards model, weighted by matching weights. Each patient's propensity for each treatment was modeled via random forest, based on potential drivers of treatment selection. A propensity score for each therapy was used to calculate a matching weight, targeting the same estimand as 1:1 matching of treatment groups with balance among potential confounders. Eligibility criteria included diagnosis of metastatic urothelial carcinoma, receipt of first line treatment with a platinum based chemotherapy, followed by initiation of single agent programmed cell death protein 1/L1 inhibitor after disease progression from August 1, 2016 through May 1, 2019.
Overall, 609 patients were eligible for analysis. Median time to initiation of third therapy or death with atezolizumab, nivolumab and pembrolizumab was 4.2, 5.3 and 4.5 months, respectively, and median overall survival was 6.4, 8.0 and 8.3 months, respectively. Matching weighted analyses did not show strong evidence of differences among programmed cell death protein 1/L1 inhibitors in terms of time to initiation of third therapy or death and overall survival.
In this large real-world cohort, effectiveness in terms of time to initiation of third therapy or death and overall survival with programmed cell death protein 1/L1 inhibitors in patients with platinum refractory locally advanced/unresectable or metastatic urothelial carcinoma was similar.
目前已有 5 种程序性死亡蛋白 1 或其配体（L1）抑制剂获批用于治疗铂类耐药、局部晚期/不可切除或转移性尿路上皮癌。然而，这些药物的相对疗效尚不清楚。本研究比较了不同 PD-1/L1 抑制剂在铂类耐药转移性尿路上皮癌患者中的起始三线治疗时间或死亡时间以及总生存期。
从真实世界的去标识数据库中提取患者水平数据。采用 Cox 比例风险模型进行比较有效性分析，权重通过匹配权重加权。基于治疗选择的潜在驱动因素，通过随机森林对每位患者的每种治疗方法的倾向性进行建模。针对每种治疗方法的倾向评分用于计算匹配权重，目标是在平衡潜在混杂因素的情况下，对治疗组进行 1:1 匹配，达到相同的估计值。入选标准包括诊断为转移性尿路上皮癌、接受一线铂类化疗治疗，随后在疾病进展后于 2016 年 8 月 1 日至 2019 年 5 月 1 日开始使用单药 PD-1/L1 抑制剂。
共有 609 名患者符合分析条件。阿替利珠单抗、纳武利尤单抗和帕博利珠单抗的起始三线治疗时间或死亡时间的中位值分别为 4.2、5.3 和 4.5 个月，中位总生存期分别为 6.4、8.0 和 8.3 个月。匹配加权分析未显示 PD-1/L1 抑制剂在起始三线治疗时间或死亡时间和总生存期方面存在显著差异的有力证据。
在这项大型真实世界队列研究中，铂类耐药局部晚期/不可切除或转移性尿路上皮癌患者使用 PD-1/L1 抑制剂在起始三线治疗时间或死亡时间和总生存期方面的疗效相似。
