Checkpoint inhibitors barely move the needle in unselected advanced soft-tissue sarcoma — objective responses hover near 2%. The TLS-positive cohort of the phase 2 PEMBROSARC trial tested a different approach: enrolling only patients whose tumors already contained tertiary lymphoid structures. In these 30 patients, pembrolizumab plus low-dose cyclophosphamide produced a 30% objective response rate and met its primary endpoint. Spatial transcriptomics and multiplex imaging then traced response to intratumoral plasma cells and dendritic cells — pointing toward a biomarker-driven selection strategy for sarcoma immunotherapy.
Published in Nature Medicine, this PEMBROSARC trial cohort was led by Prof. Antoine Italiano (Institut Bergonié, University of Bordeaux) — with Alban Bessede of Explicyte as a co–first author — and drew on the French Sarcoma Group together with the tertiary-lymphoid-structure biology group of Catherine Sautès-Fridman and Wolf Hervé Fridman at the Centre de Recherche des Cordeliers. The work was supported by the RHU CONDOR program and by MSD, France’s Agence Nationale de la Recherche, INSERM, the Ligue contre le Cancer, and the French National Cancer Institute. Explicyte contributed the exploratory biomarker workup that explains who responds: spatial transcriptomic profiling on the NanoString GeoMx whole-transcriptome atlas and multiplex immunofluorescence panels for TLS imaged on the PhenoImager HT, applied to responder and non-responder tumors.
Selecting sarcoma patients by TLS status converts a treatment with negligible activity into one with meaningful, durable responses — and the immune profiling nominates plasma cells as a refinement beyond TLS status alone.
For sponsors developing immunotherapy in sarcoma, TLS status offers a prospectively validated enrichment strategy that can rescue trials otherwise undermined by low unselected response rates. The plasma-cell and regulatory-T-cell signals point toward a next-generation biomarker panel and rational combinations — pairing anti–PD-1 with Treg- or CTLA-4–directed agents. Two randomized CONDOR-associated trials are already testing TLS-based selection in the neoadjuvant and metastatic settings.