PURPOSE: Activating T-cell costimulatory receptors is a promising approach for cancer immunotherapy. In preclinical work, adding an OX40 agonist to in situ vaccination with SD101, a TLR9 agonist, was curative in a mouse model of lymphoma. We sought to test this combination in a phase I clinical trial for patients with low-grade B-cell lymphoma. PATIENTS AND METHODS: We treated 14 patients with low-dose radiation, intratumoral SD101, and intratumoral and intravenous BMS986178, an agonistic anti-OX40 antibody. The primary outcome was safety. Secondary outcomes included overall response rate and progression-free survival. RESULTS: Adverse events were consistent with prior experience with low-dose radiation and SD101. No synergistic or dose-limiting toxicities were observed. One patient had a partial response, and nine patients had stable disease, a result inferior to our experience with TLR9 agonism and low-dose radiation alone. Flow cytometry and single-cell RNA sequencing of serial tumor biopsies revealed that T and NK cells were activated after treatment. However, high baseline OX40 expression in T follicular helper and T regulatory type 1 cells, as well as high posttreatment soluble OX40, shed from these T cells upon activation, associated with progression-free survival of less than 6 months. CONCLUSIONS: Clinical results of T-cell costimulatory receptor agonism have now repeatedly been inferior to the motivating preclinical results. Our study highlights potential barriers to clinical translation, particularly differences in preclinical and clinical reagents and the complex biology of these coreceptors in heterogeneous T cell subpopulations, some of which may antagonize immunotherapy.