Abstract Background: CART and bispecific antibodies (BsAb) are both effective treatments for patients (pts) with relapsed/refractory (R/R) large B-cell lymphoma (LBCL). In prospective trials, BsAbs demonstrate similar response rates regardless of prior CART exposure. However, the impact of BsAb exposure on subsequent CART efficacy is unknown and the appropriate sequencing of these therapies has not been studied. In this real-world analysis, we compare outcomes in pts who received CART and/or BsAb for R/R LBCL to understand the impact of sequencing on survival. Methods: We identified 1031 adult pts with R/R LBCL treated with a CD20-directed BsAb and/or CD19-CART at 15 US cancer centers from 2015-2024. De novo DLBCL, transformed FL, and other LBCL histologies were included. Pts were grouped according to sequence of therapy received. Baseline characteristics were compared with the Pearson Chi-squared test and included: demographics (age, race, ethnicity, sex), disease features cell of origin (COO), histology, double hit lymphoma (DHL), LDH at treatment, and treatment characteristics primary refractory disease (PRD) defined as progression on or relapse within 12 months of start of frontline therapy, BsAb or CART product, line of therapy, prior ASCT or bendamustine. Median duration of complete response (mDOCR), progression free survival (mPFS) 0.05 was significant. Results: We compared BsAb pts with prior CART (group 1, n=121) vs those without prior CART (group 2, n=109): Glofitamab was the most common BsAb in both group 1 and group 2 (56% vs 45%). Cohorts did not differ by sex, race, histology, bendamustine exposure, or drive time to center. For group 1, median time from CART relapse to BsAb administration was 43 days. Pts in group 2 were older ( age 60, 69 vs 56%, p=0.04), less fit (ECOG PS 0-1, 74 vs 91%, p0.001), more often GCB subtype (70 vs 48%, p=0.002), more often with elevated LDH (67 vs 53%, p=0.03), and received fewer prior lines of therapy (BsAb 3rd vs 5th line, p0.001). We examined outcomes following BsAb between groups. There was no difference in rate of best response, including CR rate (32 vs 38%, p=0.3), mDOCR (34.4 mo vs NR, p=0.2), time to relapse (4.3 vs 4.9 mo, p=0.7), or rates/severity of CRS or neurotoxicity (NT). Pts with DHL had superior mDOCR in group 2 (21.7 vs 5.1 mo, p=0.025). There was no difference in mPFS (3.4 vs 4.3 mo, p=0.7) or mOS (13.2 vs 10.4 mo, p=0.3), including among patients age 60 or with PRD, DHL, or elevated LDH. BsAb product had no impact on mPFS or mOS. For group 1, time from CART relapse to BsAb treatment ( 90 days, 90-180 days, 180 days), did not impact mPFS or mOS. We compared CART pts with prior BsAb (group 3, n=12) vs those without prior BsAb (group 4, n=922): Groups did not differ by sex, race, histology, DHL, LDH, stage, extra-nodal disease, prior bendamustine, or prior ASCT. Pts in group 4 were younger ( age 60, 56 vs 91%, p=0.03), more fit (PS 0-1, 90 vs 64%, p=0.02), had lower baseline IPI (IPI 3-4, 51 vs 90%, p=0.04), more non-GCB subtype (45 vs 0%, p=0.002), and fewer prior lines of therapy (3rd vs 4h line, p=0.005). We examined outcomes following CART between groups. There was no difference in rate of best response, including CR rate (58 vs 61%, p=0.3), mDOCR (43 mo vs NR, p0.9), time to relapse (11.2 vs 7.9 mo, p=0.6), or rates/severity of CRS or NT. There was no overall difference in mPFS (2.6 vs 2.1 mo, p=0.4) or mOS (30.4 vs 15.2 mo, p=0.15), including among pts age 60, with DHL, or with elevated LDH. However, mOS among PRD pts was superior in group 4 (25.4 vs 5.9 mo, p=0.03). Conclusions: In pts with LBCL, those treated with BsAb without prior CART are often older and have worse performance status. Among PRD pts, CART without prior BsAb offers a survival advantage to CART with prior BsAb exposure. Otherwise, sequencing of CART and BsAb does not appear to impact mPFS and mOS following each therapy, including among older, DHL, and high LDH subsets. In DHL pts, mDOCR with BsAb may be improved if not preceded by CART. Results are skewed by retrospective assessment, including unbalanced cohort characteristics and ineligibility for subsequent therapy due to worsening clinical status or death. As CART is potentially curative, based on this analysis, we favor CART over BsAb, particularly as BsAb appears similarly effective following CART.
Thomas et al. (2025) studied this question.