T-cell dysregulation informs radiotherapy-immunotherapy response in B-cell lymphoma: results from a phase 1 trial

Abstract

Diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL) are highly radiosensitive with immune-driven abscopal responses reported. Programmed cell death 1/programmed cell death ligand 1 (PD-L1) inhibitors are relatively ineffective in DLBCL/FL; however, evidence suggests synergy with radiotherapy (RT), but no clear biomarkers. This phase 1 study examined the safety of escalating RT dose and treated volumes with durvalumab (PD-L1 inhibitor) in 34 adults with relapsed/refractory DLBCL and relapsed/refractory FL, and the role of immune-cell subsets on outcomes. Patients received external-beam RT (2.5-30 Gray [Gy], 5 or 10 fractions up to 3 target sites) plus durvalumab from RT day 2, until progression. Novel positron emission tomography (PET) biodistribution studies of 89Zr-durvalumab and CD8 T-cell minibody-89Zr-Df-crefmirlimab were incorporated. The RT recommended phase 2 dose was 10 Gy/5 fractions and 30 Gy/10 fractions to 3 sites for FL and DLBCL, respectively. The most common grade 3 to 4 toxicities included anemia (9%), neutropenia (11%), and liver dysfunction (5%). Overall response was 60% in FL (3/5; complete response, 40% [2/5]), and 14% in DLBCL (4/27; complete response, 7% [2/27]). Distinct peripheral blood and tumor T-cell features, including CD8 PET-determined intratumoral CD8 T-cells, correlated with response (P < .05). RT-durvalumab with 30 Gy/10 fractions of RT to 3 disease sites is safe, and offers promising responses in FL. Intratumoral and peripheral blood CD8 T-cell dysregulation correlate with treatment response. This trial was registered at www.clinicaltrials.gov as #NCT03610061.

Graphical abstract

Figure 1.

Treatment outcomes and response. (A) Consort diagram. (B) Treatment duration and follow-up (F/U) postcompletion according to disease and RT dose level. Durvalumab treatment indicated by colored solid bars. Triangles represent best response, circle represents PD. Black arrows illustrate participants who were alive at data lock. Participants in the PET substudy are denoted by colored arrows according to tracer. (C) TMTV change in lesions according to RT dose level and in-treated volume vs out-of-treated volume responses. X represents participants who did not have a F/U PET/CT. CNS, central nervous system; MRI, magnetic resonance imaging; CMR, complete metabolic response; PMR, partial metabolic response; SD, stable disease; PD, progressive disease.

Figure 2.

Biomarker analysis. Gene expression profiling in tumor biopsies was examined, and differentially expressed genes evaluated based on best response. (A) Heat map of T cells and M2 macrophage-related genes demonstrating differential gene expression according to best response. Participants who underwent CD8 PET are highlighted by color-coded boxes that correspond to Figure 3D results. (B-D) Expression of T-cell activation score (as described in “Methods”), immune ratio, and TCF7 expression based on best response are represented by panels B, C, and D, respectively. (E) Matched tumor biopsies from baseline (BL), C1D8-15 and progression were evaluated for TCF7 expression (3 CMR in green; 2 PMR in blue; and 7 PD in red). (F-G) Flow cytometry analysis of immune checkpoints (TIGIT and PD-1) and TCF7 was evaluated in CD4⁺ and CD8⁺ T cells at BL and C1D8 (CMR, n = 3; PD, n = 7) for panels F and G, respectively. (H) TCF7 expression was examined in PD1/TIGIT dual-positive cells for CD4⁺ and CD8⁺ cells (CMR, n = 3; PD, n = 7). ∗P < .05; ∗∗P < .01; ∗∗∗P = .009.

Figure 3.

Baseline ⁸⁹Zr-Df-crefmirlimab (CD8 T-cell) uptake correlates with posttreatment tumor FDG-PET-CT response. (A-C) Representative baseline (BL) CD8 T-cell and paired FDG-PET imaging (and F/U scans ∼D50). (A) A 50-year-old female patient with stage II FL (at diagnosis): whole-body ⁸⁹Zr-Df-crefmirlimab scan 24 hours postinjection (p.i.) (left) at BL and day 51 on treatment with paired FDG-PET (right). Axial slices of left axillary mass demonstrate intense uptake at BL and CMR on treatment. RT fields are illustrated on the FDG-PET image. (B) A 86-year-old male with DLBCL: whole-body ⁸⁹Zr-Df-crefmirlimab scan 24 hours p.i. (left) at BL and F/U on treatment with paired FDG-PET (right). Axial slice of bilateral neck lesions demonstrate intense uptake at BL and CMR on treatment. RT fields are illustrated on the FDG-PET image. (C) A 76-year-old female with DLBCL: whole-body ⁸⁹Zr-Df-crefmirlimab scan 24 hours p.i. (left) at BL and F/U on treatment with paired FDG-PET (right). On BL scans, axial slices of right external iliac mass demonstrate intense uptake of FDG pretreatment, but no BL ⁸⁹Zr-Df-crefmirlimab uptake in tumor sites; however, some marrow uptake plus excretion of catabolized ⁸⁹Zr in bowel is seen. Although the patient had a CMR in RT-treated volume, they had extensive PD out-of-treated volume on treatment. Radiotherapy fields are illustrated on the FDG-PET image. (D) Tumor uptake (SUVmax) of ⁸⁹Zr-Df-crefmirlimab 24 hours p.i. at baseline was significantly higher in patients who achieved a CMR on study (P = .003; adjusted P values ∗P = .01, ∗∗P = .002). D51, 51 days after treatment commencement; ns, not significant; PD, progressive disease; PMR, partial metabolic response.