Background: R/R DLBCL patients (pts) who fail multiple lines of treatment or are ineligible for novel treatments, including CAR T-cell therapy, have poor clinical outcomes. The combination of the anti-CD47 monoclonal antibody (mAb) magrolimab and rituximab enhances macrophage-mediated antibody-dependent cellular phagocytosis in DLBCL (Advani et al. NEJM 2019). We hypothesized that the Bruton's tyrosine kinase inhibitor acalabrutinib would further enhance clinical activity in R/R DLBCL. We present the final analysis of Arm 3 of the Phase 1 PRISM platform study of acalabrutinib with magrolimab and rituximab in all molecular subtypes of R/R DLBCL (NCT03527147). Methods:Adults with DLBCL and ECOG PS ≤2 who had failed ≥1 line of prior chemoimmunotherapy or were ineligible for high-dose chemotherapy with stem-cell rescue and/or CAR T-cell therapy were eligible. Pts with prior CAR T-cell therapy were excluded from the dose-limiting toxicity (DLT) evaluable pts. Pts received magrolimab 1 mg/kg IV priming dose on cycle 1 day 1 (C1D1) of a 28-day cycle, then 30 mg/kg IV on D8, D15, and D22, on D1, D8, D15, and D22 of C2, and on D1 and D15 thereafter. Rituximab 375 mg/m2 was given on D8, D15, and D22 of C1, D1 of C2-6 and every other cycle from C8 onwards. Acalabrutinib was given continuously at 100 mg twice daily. Combination treatment was given until disease progression or for a maximum of 2 years. DLTs were evaluated from first treatment administration until the end of C1. The primary objective was safety of the combination regimen. Secondary objectives included response rate, pharmacokinetics (PK), and immunogenicity. Exploratory analyses included baseline tumor samples, baseline and longitudinal immunophenotyping of peripheral blood, and cytokine and mutation analysis of plasma samples. Results:By 31 March 2021, 7 pts with DLBCL (including 3 with transformed disease) were enrolled. Median age was 72 years (51-84). Pts had received a median of 4 prior regimens (range 1-6). Three (42.9%) pts were refractory to their most recent regimen. No DLTs were observed in 6 DLT-evaluable pts. Five pts discontinued treatment due to progressive disease (PD), 1 due to an adverse event (AE), and 1 due to study termination. Grade ≥3 AEs regardless of causality reported in ≥10% of pts included anemia (42.9%), thrombocytopenia (28.6%), bilirubin elevation (14.3%), vomiting (14.3%), and ALT elevation (14.3%). Two pts experienced serious AEs (1 pneumonitis, 1 septic shock). Overall, the safety profile was consistent with magrolimab monotherapy studies. No treatment-related deaths were reported. Two (28.6%) pts had an objective response per Response Evaluation Criteria in Lymphoma (RECIL) 2017; both were complete responses (CR). At a median follow-up of 8.3 months, response durations for these pts were 0.03 and 7.36 months. One CR pt had a non-germinal center B-cell (GCB) DLBCL; the other was unknown. Of the 2 PD pts, 1 had non-GCB DLBCL with a BN2/A53 genetic subtype and 1 had GCB DLBCL with an EZB/A53 genetic subtype. Fluctuations in cell-free DNA (cfDNA) and cytokines tended to correlate with response by imaging. Both PD pts had higher plasma cfDNA concentrations than other pts (1 pt at screening, C1D1, and C1D8 and the other at C1D15, C2D1, and end of treatment). CR pts had lower mutational burden and lower variant allele frequencies than PD pts at screening and longitudinally. The allele frequencies of copy number alterations in tissue samples at screening were also higher in stable disease and PD pts. PD pts had higher baseline levels of ICAM1, IL-10, IL-15, IL-6, MCP-1, MIP1a, MIP1b, TNF-b, VEGF-A, and VEGF-D than non-PD pts. PK data were consistent with monotherapy results and no antibodies to magrolimab were detected. Conclusions:Acalabrutinib combined with magrolimab and rituximab was safe and tolerable in pts with advanced R/R DLBCL and no prior exposure to CAR T-cell therapy. The safety profile of the combination was consistent with the known safety profiles of the individual agents. However, due to the lack of obvious clinical synergy of the triplet over acalabrutinib monotherapy or the combination of magrolimab and rituximab, the study was stopped prematurely. Conclusions from cfDNA and cytokine analyses are limited by small pt numbers, but such analyses should be included in future studies of targeted combinations because mounting evidence suggests they tend to correlate with response.