A Randomized Phase 2, Investigator-Led Trial of Glofitamab-R-CHOP or Glofitamab-Polatuzumab Vedotin-R-CHP (COALITION) in Younger Patients with High Burden, High-Risk Large B-Cell Lymphoma Demonstrates Safety, Uncompromised Chemotherapy Intensity, a High Rate of Durable Remissions, and Unique FDG-PET Response Characteristics

Background:

Patients (pts) with high-risk (HR), IPI≥3/NCCN-IPI≥4 large B-cell lymphoma (LBCL) treated with R-CHOP have 1-yr PFS<75% and <50% are cured. They are underrepresented in trials partly due to the demands of screening when rapid treatment is needed. Previous studies adding novel drugs to R-CHOP often increased toxicity and reduced chemotherapy delivery intensity without improving outcomes. Polatuzumab-R-CHP, in contrast, improves PFS over R-CHOP in pts with IPI≥2 (Tilly, NEJM 2022), but the rate of relapse in HR pts remains an area of need. For fit younger HR pts many centers use intensified chemotherapy (e.g. DA-EPOCH-R) despite added toxicity, cost, and uncertain benefit. Glofitamab, a CD20xCD3 bispecific antibody, is approved in relapsed/refractory LBCL. In a multi-center investigator-led trial designed to address this need, we evaluated the safety and efficacy of glofit-R-CHOP and glofit-pola-R-CHP in younger pts with HR LBCL.

Methods:

Pts aged 18-65 yrs with untreated LBCL and ≥1 HR feature were enrolled prior to or after one cycle of R-CHOP. HR was defined as IPI ≥3, NCCN-IPI ≥4, or proven HGBL with MYC and BCL2 and/or BCL6 rearrangements (HGBL-DH). ECOG <4 at baseline or <2 at cycle 2 was permitted. All pts received 1 cycle of R-CHOP (C1) and were randomized to receive 5 cycles of glofit (2.5 mg & 10 mg step-up in C2, 30 mg in C3-6) with either R-CHOP (Arm A) or pola-R-CHP (Arm B), followed by 2 cycles of glofit (C7-8).

Primary endpoints were safety and deliverability, measured by relative dose intensity (RDI) and rate of early discontinuation. Secondary endpoints included ORR, CRR, PFS, OS. Responses were evaluated per Lugano 2014 by FDG-PET after cycles 2, 4 and 6 and then every 3-6 months. Adverse events (AEs) were CTCAE V5.0 graded, except CRS and ICANS (ASTCT). Exploratory endpoints included PET parameters, genomic profiling and MRD monitoring by cell free DNA (cfDNA).

Results:

80 pts were enrolled and completed end of induction (EOI) at the 15 Feb 2024 data cut-off, with median follow up (FU) of 14.6 months. Baseline pt characteristics were similar between the arms. In total, 71 (89%) pts had de novo DLBCL/HGBL, 9 (11%) had transformed indolent lymphoma; 30/80 (38%) were non-GCB subtype. Median age was 58 years (range, 24-65), and the median time from diagnosis to first dose of R-CHOP was 14 days (IQR 8-20). HR features were: IPI ≥3 in 83%, NCCN-IPI ≥4 in 85%, and HGBL-DH in 11%. 97% had stage III-IV disease. The median metabolic tumour volume was 842 cm³ (IQR 408-1342 cm³).

Gr≥3 AEs occurred in 22/40 (55%) in each arm, including 1 Gr 5 AE each (cardiomyopathy and AML). In Arm A and B, febrile neutropenia was observed in 0/40 (0%) and 6/40 (15%) and CRS Gr 1 in 7/40 (18%) and 8/40 (20%), respectively. One episode of Gr 2 CRS was reported in each Arm and no Gr 3-5 CRS. Peripheral neuropathy was limited to Gr 1-2: 18/40 (45%) Arm A and 17/40 (42%) Arm B. No ICANS was reported.

For > 90% of pts, the RDI was ≥90% for all individual components except vincristine (78%). There were 2 dose interruptions of doxorubicin (AMI [n=1], and poor performance status [n=1]), and 5 interruptions of glofit, (infection [n=3], hepatotoxicity [n=1], and rash [n=1]). One pt ceased treatment after 5 cycles and remains in CMR at 13 months FU.

Best overall response (BOR) rate was 100% (CMR/PMR 92%/8%). At EOI, rates of CMR/PMR were 70%/29% in Arm A, and 80%/18% in Arm B. There were 6 PFS events: in Arm A, 3 PD and 1 death in remission (cardiomyopathy) and in Arm B, 1 PD and 1 death in remission (AML). BOR in pts with PD were PMR in 2 and CMR in 2 (time from first CMR to PD of 5 and 11 months).

The PFS/OS at 12 months is 88%/96% in Arm A and 95%/97% in Arm B. Of the 18 pts in PMR at EOI, none has progressed, while 14 converted to CMR. Given this finding, analysis of cfDNA was performed (KAPA NHL; Roche Dx) and 80% (8/10) of pts with available samples were negative for MRD at EOI in the context of PMR. Genomic evaluation and extended clinical follow up will be reported at the meeting.

Conclusion:

Glofit-R-CHOP or glofit-pola-R-CHP each resulted in high rates of deep and durable responses in a very high disease burden, HR pt population. The study design enabled rapid treatment initiation, minimal new toxicity, and uncompromised chemotherapy delivery intensity. With these regimens, PMR at EOI does not necessarily represent treatment failure, and cfDNA may be useful to further investigate outcomes in this setting. CRR and preliminary PFS is promising compared to historical outcomes in pts with HR LBCL.