DP004 | RESPONSE PREDICTION AND MUTATIONAL PROFILING BY CIRCULATING TUMOR DNA OF RELAPSED/REFRACTORY LARGE B-CELL LYMPHOMA PATIENTS RECEIVING GLOFITAMAB

Glofitamab is a CD20:CD3 bispecific antibody that has been approved for treating R/R LBCL. The pivotal phase 2 trial (NCT03075696) using fixed-duration Glofitamab in R/R LBCL demonstrated high complete response (CR) rates, clinically meaningful durability of responses, and a manageable safety profile. However, a subset of patients experiences disease relapse or progression. To address this issue, we assessed ctDNA to explore response biomarkers and resistance mechanisms to Glofitamab and optimize risk stratification for R/R LBCL patients. Blood samples from 57 patients treated with Glofitamab in phase 1-2 trial or the Italian expanded access program were collected at: baseline, cycle 3 (C3), cycle 6, end of treatment (EOT), or at progression. Cell-free DNA extracted from plasma was sequenced using the CAPP-Seq strategy. Following somatic analysis, the ctDNA load was calculated as haploid genome equivalent per ml (hGE/ml) of plasma. Copy number (CN) and fusion analyses were performed using CNVkit and FACTERA, respectively. At baseline, the median ctDNA load in 57 patients was 292 hGE/ml (range 27-14,804). Higher ctDNA levels were associated with bulky disease (p=0.02) and high IPI (p=0.03). ctDNA was assessed longitudinally in 43 patients who received ≥10 mg of Glofitamab. At C3, 18/43 (42%) patients were Minimal Residual Disease positive (MRD+), and all progressed during treatment, including five cases that were in metabolic CR or partial response (PR) at C3. In contrast, 25/43 (58%) were MRD negative (MRD-) at C3. Among the MRD-, 19/25 experienced a persistent CR, whereas 6/25 progressed during treatment, as reflected by a simultaneous reappearance of MRD. Moreover, MRD- patients had a significantly longer median progression-free survival (PFS) compared to MRD+ (11.8 vs 2.5 months; p<.0.0001). At baseline, we observed that patients with TP53 disruptions [mutations (n=22), deletions (n=4)] and wild-type cases had comparable CR rates (42% vs 36%) and median PFS (4.6 vs 4 months). Our study suggests that the detection of C3 MRD in R/R LBCL patients receiving Glofitamab may effectively predict progression. Additionally, mutational analysis allowed us to deeply characterize our cohort's genetic profile and further investigate the mechanisms of resistance to Glofitamab. Despite the limited sample size, these findings highlight the potential role of ctDNA in disease monitoring and optimization of risk stratification in R/R LBCL patients.