Clinical efficacy and biomarker analysis of neoadjuvant atezolizumab in operable urothelial carcinoma in the ABACUS trial

Thomas Powles, Mark Kockx, Alejo Rodriguez-Vida, Ignacio Duran, Simon J. Crabb, Michiel S. van der Heijden, Bernadett Szabados, Albert Font Pous, Gwenaelle Gravis, Urbano Anido Herranz, Andrew Protheroe, Alain Ravaud, Denis Maillet, Maria Jose Mendez, Cristina Suarez, Mark Linch, Aaron Prendergast, Pieter-Jan van Dam, Diana Stanoeva, Sofie DaelemansSanjeev Mariathasan, Joy S. Tea, Kelly Mousa, Romain Banchereau, Daniel Castellano

Research output: Contribution to journalComment/Letter to the editorAcademic

393 Citations (Scopus)

Abstract

Antibodies targeting PD-1 or its ligand 1 PD-L1 such as atezolizumab, have great efficacy in a proportion of metastatic urothelial cancers1,2. Biomarkers may facilitate identification of these responding tumors3. Neoadjuvant use of these agents is associated with pathological complete response in a spectrum of tumors, including urothelial cancer4–7. Sequential tissue sampling from these studies allowed for detailed on-treatment biomarker analysis. Here, we present a single-arm phase 2 study, investigating two cycles of atezolizumab before cystectomy in 95 patients with muscle-invasive urothelial cancer (ClinicalTrials.gov identifier: NCT02662309). Pathological complete response was the primary endpoint. Secondary endpoints focused on safety, relapse-free survival and biomarker analysis. The pathological complete response rate was 31% (95% confidence interval: 21–41%), achieving the primary efficacy endpoint. Baseline biomarkers showed that the presence of preexisting activated T cells was more prominent than expected and correlated with outcome. Other established biomarkers, such as tumor mutational burden, did not predict outcome, differentiating this from the metastatic setting. Dynamic changes to gene expression signatures and protein biomarkers occurred with therapy, whereas changes in DNA alterations with treatment were uncommon. Responding tumors showed predominant expression of genes related to tissue repair after treatment, making tumor biomarker interpretation challenging in this group. Stromal factors such as transforming growth factor-β and fibroblast activation protein were linked to resistance, as was high expression of cell cycle gene signatures after treatment.
Original languageEnglish
Pages (from-to)1706-1714
JournalNature medicine
Volume25
Issue number11
DOIs
Publication statusPublished - Nov 2019

Cite this