Sequential immunotherapy and targeted therapy for metastatic BRAF V600 mutated melanoma: 4-year survival and biomarkers evaluation from the phase II SECOMBIT trial

Abstract

No prospective data were available prior to 2021 to inform selection between combination BRAF and MEK inhibition versus dual blockade of programmed cell death protein-1 (PD-1) and cytotoxic T lymphocyte antigen-4 (CTLA-4) as first-line treatment options for BRAFV600-mutant melanoma. SECOMBIT (NCT02631447) was a randomized, three-arm, noncomparative phase II trial in which patients were randomized to one of two sequences with immunotherapy or targeted therapy first, with a third arm in which an 8-week induction course of targeted therapy followed by a planned switch to immunotherapy was the first treatment. BRAF/MEK inhibitors were encorafenib plus binimetinib and checkpoint inhibitors ipilimumab plus nivolumab. Primary outcome of overall survival was previously reported, demonstrating improved survival with immunotherapy administered until progression and followed by BRAF/MEK inhibition. Here we report 4-year survival outcomes, confirming long-term benefit with first-line immunotherapy. We also describe preliminary results of predefined biomarkers analyses that identify a trend toward improved 4-year overall survival and total progression-free survival in patients with loss-of-function mutations affecting JAK or low baseline levels of serum interferon gamma (IFNy). These long-term survival outcomes confirm immunotherapy as the preferred first-line treatment approach for most patients with BRAFV600-mutant metastatic melanoma, and the biomarker analyses are hypothesis-generating for future investigations of predictors of durable benefit with dual checkpoint blockade and targeted therapy.

Fig. 1. Kaplan–Meier survival curves with 3-year and 4-year rates for Arm A (blue), Arm B (green), and Arm C (red).

A Total progression free survival; B Overall survival. Source data are provided as a Source Data file.

Fig. 2. HR according to biomarkers analysis.

Forest plot representing HR for Overall Survival according to Jak mutations, IFN gamma expression and TMB in Arm A (targeted therapy followed by immunotherapy), Arm B (immunotherapy followed by targeted therapy), and Arm C (a course of targeted therapy preceding immunotherapy and targeted therapy) of SECOMBIT. Source data are provided as a Source Data file.

Fig. 3. Overall survival by JAK mutation status.

Kaplan–Meier survival curves for patients with wild type JAK1/2 (blue) and deleterious mutations in JAK1/2 (green) are shown, in Arm A (targeted therapy followed by immunotherapy), Arm B (immunotherapy followed by targeted therapy), and Arm C (a course of targeted therapy preceding immunotherapy and targeted therapy) of SECOMBIT. Dotted lines represent 95% confidence intervals. Source data are provided as a Source Data file.

Fig. 4. Overall survival by baseline serum interferon gamma (IFNy).

Kaplan–Meier survival curves for patients with low baseline serum IFNy (blue) and high baseline serum IFNy (green) in Arm A (targeted therapy followed by immunotherapy), Arm B (immunotherapy followed by targeted therapy), and Arm C (a course of targeted therapy preceding immunotherapy and targeted therapy) of SECOMBIT. Dotted lines represent 95% confidence intervals. Source data are provided as a Source Data file.

Fig. 5. Heatmaps.

Panel A (n = 89 patients) shows hierarchical co-clustering of clinical variables and baseline serum cytokine levels in the overall cohort. Levels of correlation of each cytokine with the other ones in arm A (panel B, n = 27 patients), in arm B (panel C, n = 28 patients), in arm C (panel D, n = 34 patients). The setting for the visible lower and upper scale bounds is two standard deviations (ANOVA was used for comparison of groups). *shows unavailable cytokines.

Fig. 6. Network interaction map depicting regulatory interactions between the JAK/STAT signaling pathway and the PD-1(PDCD1)/PD-L1(CD274) axis.

Line thickness represents the strength of data. STRING (https://string-db.org/) uses a spring model to generate the network images. Nodes are modeled as masses and edges as springs; the final position of the nodes in the image is computed by minimizing the ‘energy’ of the system^,.