Introduction: Immune checkpoint inhibitor (ICPis) re-challenge could be an attractive therapeutic option considering its good safety profile. However, little data is available regarding anti-PD-1/anti-PD-L1 retreatment. We conducted a meta-analysis focusing on outcomes of solid cancer patients performing this strategy.
Methods: Fourteen full papers involving 74 patients were included. Individual data about best response or progression-free survival (PFS) upon the first and second course of anti-PD-1/anti-PD-L1 were collected.
Results: Non-small-cell lung cancer (53%) and melanoma (34%) were the most represented cancers. Higher objective response (46% versus 24%, P=4.10-4) and disease control rates (73% versus 52%, P=7.10-3) were obtained upon the first ICPi course compared to re-challenge. No association between responses obtained with the two ICPis courses was found (P=3.10-1). The PFS upon the first ICPi (PFS1) was longer than after re-challenge (PFSR) (6.6 versus 2.8 months, hazard ratio (HR) 0.57, P=2.10-3). A longer PFSR was obtained in patients with a longer PFS1 (P=6.10-3), in those who discontinued the first ICPi due to toxicity or per protocol (8.8 versus 2.1 months if disease progression occurs, P=2.10-3), and in those not receiving intercalated treatment between the two ICPis (6.6 versus 2.1 months for the treated ones, P=1.10-3).
Discussion: Anti-PD-1/anti-PD-L1 re-challenge showed interesting clinical activity in selected patients, mainly in those achieving a long-term response upon the first ICPi course, that do not discontinue therapy because of disease progression, or that are able to keep a treatment-free period.
Keywords: Anti-programmed cell death ligand 1; Anti-programmed cell death protein 1; Immune checkpoint inhibitor; Meta-analysis; Re-challenge; Solid cancer.
Copyright © 2020 Société Française du Cancer. Published by Elsevier Masson SAS. All rights reserved.