Circulating tumor DNA (ctDNA) by clonoSEQ to monitor residual disease after axicabtagene ciloleucel (axi-cel) in large B-cell lymphoma (LBCL).

Abstract

7547 Background: Most patients (pts) with LBCL respond to axi-cel, but many will eventually experience disease progression (Jacobson et al. JCO. 2020). Monitoring of ctDNA from blood, a minimally invasive diagnostic tool, has been used to assess measurable residual disease (MRD) with prognostic value, including in pts with diffuse LBCL treated with axi-cel in the third line of therapy (3L; Frank et al. JCO. 2019). Here, we explored the value of ctDNA to predict outcomes in LBCL after axi-cel across lines of treatment. Methods: Pts with LBCL from ZUMA-12 (Phase 2 axi-cel in 1L), ZUMA-7 (Phase 3 axi-cel or standard of care [SOC] in 2L), and ZUMA-14 (Phase 2 axi-cel + rituximab in 3L) were included. All studies included initial MRD testing ≈30 days post axi-cel infusion. ClonoSEQ MRD assay (limit of detection≈0.0001%-0.00001%) was used to define the lymphoma B-cell clonotype in formalin-fixed, paraffin-embedded (FFPE) biopsy tissue prior to axi-cel infusion and to track ctDNA in blood after treatment. Positive predictive value (PPV; MRD+ pts who relapsed or were nonresponders/total MRD+ pts ×100) and negative predictive value (NPV; MRD- pts in ongoing response/total MRD- pts ×100) were assessed at Day 28, Month (Mo) 3, and Mo 5 for 3L; Days 50, 100, 150, Mo 9,and Mo 24 for 2L; and Day 28, Mo 3, and Mo 6 for 1L. Results: In 3L, the PPV at Day 28 was 88% (7/8) and NPV was 83% (10/12). NPV remained 83% at Mo 3 (10/12) and increased to 100% (8/8) at Mo 5. Overall, 83% (5/6) of relapsed pts had MRD detected at any time; of those 5, 100% (5/5) had MRD detected at any time prior to or at progression with a median detection of 43 days prior to progression. In 2L, the MRD+ detection rate among evaluable pre-infusion samples was only 69% (11/16). At Day 50, PPV was 100% (7/7) in the SOC arm, whereas it was only 57% (4/7) in the axi-cel arm. PPV increased over time in the axi-cel arm, reaching 100% by Mo 9 (2/2). At Day 50, NPV was 53% (8/15) in the axi-cel arm and 38% (5/13) in the SOC arm. Overall, 47% (9/19) of relapsed pts on the axi-cel arm had MRD detected at any time; of those 9, 78% (7/9) had MRD detected prior to or at progression with a median of 35 days prior to progression. In 1L, most pts (11/14) remained in ongoing response by data cut off. The PPV at Day 28 was 50% (1/2): of the two pts who had detectable MRD at Day 28, one subsequently became MRD negative at Mo 3 yet relapsed just after Mo 3, while the second became negative by Mo 3 and had ongoing response up to Mo 24. The NPV at Day 28 was 88% (7/8). Conclusions: The prognostic value of MRD assessment by clonoSEQ varied across lines of therapy in pts with LBCL treated with axi-cel. A relatively high rate of undetectable MRD at baseline in the 2L setting and prior to relapse in the 2L and 1L settings warrants exploration of more sensitive ctDNA monitoring methods. Clinical trial information: NCT03391466 NCT03761056 NCT04002401 .