PO.CL07.05 · 临床研究
晚期黑色素瘤中免疫检查点抑制剂与BRAF/MEK抑制剂的序贯用药：基于Kaplan-Meier重建个体水平数据的荟萃分析
Sequencing immune checkpoint inhibitors and BRAF/MEK inhibitors in advanced melanoma: A meta-analysis of Kaplan-Meier-reconstructed individual level data
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
目的：对于BRAFV600突变型晚期黑色素瘤，免疫检查点抑制剂（ICI）与BRAF抑制剂（BRAFi）的最佳序贯方案仍不明确。我们开展了一项基于Kaplan-Meier（KM）曲线重建的个体患者数据（IPD）的荟萃分析，比较两种临床常用序贯方案：一线ICI后接BRAFi（ICI-BRAFi）与一线BRAFi后接ICI（BRAFi-ICI）。
方法：我们检索了MEDLINE、Embase和Cochrane数据库，纳入报告了III/IV期黑色素瘤患者（接受ICI-BRAFi或BRAFi-ICI治疗）KM曲线的II/III期研究。结局指标为无进展生存期（PFS）和总生存期（OS）。次要终点为毒性和无脑转移生存期（BMFS）。重建的IPD采用Cox比例风险（PH）模型进行分析。我们还估算了截至24个月（m）的限制性平均生存时间（RMST）（Grambsch-Therneau检验，p<0.05）。
结果：共纳入6项研究、832例患者，其中426例接受ICI-BRAFi，406例接受BRAFi-ICI。两组的中位PFS均约为12 m（p=0.1，log-rank检验）。ICI-BRAFi和BRAFi-ICI在24 m时的PFS分别为42.8和29 m。24 m时PFS的RMST差异为0.32 m（95% CI：-1.1至1.8个月；p=0.66）。ICI-BRAFi的中位OS为58.2 m，BRAFi-ICI为40.3 m。ICI-BRAFi和BRAFi-ICI组24 m时的OS分别为65.9%和60.7%（p=0.4；log-rank检验）。RMST显示OS存在轻微且无统计学意义的差异，为-0.49 m（95% CI：-1.64至0.67个月；p=0.41）。在仅纳入两项随机研究SECOMBIT和DREAMseq的汇总分析中，我们发现ICI的中位PFS为31.5 m，BRAFi为13.1 m（p=0.003）；ICI和BRAFi的中位OS分别为NA和32.7 m（p=0.005）。
结论：仅评估随机临床试验时，ICI-BRAFi较BRAFi-ICI带来显著更长的PFS和OS，支持ICI优先的序贯方案。
$$table_{CF128EEB-D4BD-4130-95C0-74E89761CA49}$$
ICI-BRAFi与BRAFi-ICI队列的生存数据 ICI–BRAFi BRAFi–ICI p值 总体中位PFS 12.5个月（9.6-23.3） 12（10.7-14） 0.1 24个月时PFS 42.8个月（36.9-49.7） 29个月（24.2-34.4） - 仅RCT的中位PFS 31.5个月（13.9-44.1） 13.1个月（11.3-16.7） 0.003 总体中位OS 58.2个月（37.5-NA） 40.3个月（32.4-NA） 0.4 24个月时OS 65.9 %（60.2-72.1 %） 60.7 %（55.6-66.2 %） - 仅RCT的中位OS NA（55.5-NA） 32.7（23.1-59.7） 0.005
查看英文原文 English abstract
Objectives: Optimal sequencing of immune checkpoint inhibitors (ICI) and BRAF inhibitors (BRAFi) for BRAFV600-mutant advanced melanoma remains uncertain. We conducted a meta-analysis of Kaplan-Meier (KM)-reconstructed individual-patient data (IPD) comparing two clinically used sequences: First-line ICI followed by BRAFi (ICI-BRAFi) versus first-line BRAFi followed by ICI (BRAFi-ICI).
Methods: We searched MEDLINE, Embase, and Cochrane for phase II/III studies reporting KM curves for stage III/IV melanoma patients treated with either ICI-BRAFi or BRAFi-ICI. Outcomes were progression-free survival (PFS) and overall survival (OS). Secondary endpoints were toxicity and brain-metastasis-free survival (BMFS). Reconstructed IPD were analyzed with Cox proportional hazards (PH) models. We additionally estimated restricted mean survival time (RMST) up to 24 months (m) (Grambsch-Therneau test, p<0.05).
Results: Six studies with 832 patients were included, of which 426 patients received ICI-BRAFi, and 406 BRAFi-ICI. Median PFS was approximately 12 m for both groups (p=0.1, log-rank). PFS at 24 m was 42.8 and 29 m for ICI-BRAFi and BRAFi-ICI, respectively. The PFS RMST difference at 24 m was 0.32 m (95% CI: -1.1 - 1.8 months; p=0.66). Median OS was 58.2 m for ICI-BRAFi and 40.3 m for BRAFi-ICI. OS at 24 m was 65.9% and 60.7% for ICI-BRAFi and BRAFi-ICI groups, respectively (p=0.4; log-rank). The RMST revealed a slight and nonsignificant OS difference of -0.49 m (95% CI: -1.64 - 0.67 months; p=0.41). In a pooled analysis including the only two randomized studies SECOMBIT and DREAMseq, we found a median PFS of 31.5 m for ICI and 13.1 m for BRAFi (p=0.003); and median OS was NA and 32.7 m for ICI and BRAFi (p=0.005).
Conclusion: ICI-BRAFi led to significantly longer PFS and OS than BRAFi-ICI when evaluating only randomized clinical trials, supporting ICI-first sequencing.
$$table_{CF128EEB-D4BD-4130-95C0-74E89761CA49}$$
Survival data in ICI-BRAFi vs BRAFi-ICI cohorts ICI–BRAFi BRAFi–ICI p value Overall median PFS 12.5 months (9.6-23.3) 12 (10.7-14) 0.1 PFS at 24 months 42.8 months (36.9-49.7) 29 months (24.2-34.4) - Median PFS with RCT only 31.5 months (13.9-44.1) 13.1 months (11.3-16.7) 0.003 Overall median OS 58.2 months (37.5-NA) 40.3 months (32.4-NA) 0.4 Os at 24 months 65.9 % (60.2-72.1 %) 60.7 % (55.6-66.2 %) - Median OS with RCT only NA (55.5-NA) 32.7 (23.1-59.7) 0.005
利益披露 Disclosure
A. R. Simoes, None..
I. Michelon, None..
W. Neto, None..
M. Ciampricotti, None..
L. Cavalcante, None.