Evaluation of outcomes by line of therapy in marginal zone lymphoma: Real-world evidence from 10 US academic centers

Abstract

Abstract Introduction: Marginal zone lymphoma (MZL) is an indolent B-cell non-Hodgkin lymphoma characterized by relapses over patients' (pts) lifetime and often necessitates multiple lines of therapy (LOT) with long intervals between new treatments (tx). In follicular lymphoma, successive LOT have been associated with decreasing overall response rates (ORR), progression-free survival (PFS), and overall survival (OS) (Casulo C, Lancet Haematol 2022). However, it is unclear if the sequential worsening of outcomes holds for MZL. Hence, we sought to evaluate treatment outcomes across successive LOTs in MZL. Methods: This multicenter retrospective study included adults with splenic (SMZL), nodal (NMZL), or extranodal (EMZL) MZL treated at 10 U.S. academic centers between 2010–2022. Patients must have received ≥1 LOT (systemic or local), and those with transformation at diagnosis were excluded. LOT was determined by investigators, and changes in therapy due to toxicity, maintenance therapy, or other non-relapse events were not considered as a new LOT. Outcomes assessed per LOT included ORR, complete response rate (CRR), PFS, and OS. Response rates were assessed using Lugano criteria. Sequential changes in ORR/CRR were compared in hierarchical ordinal logistic models, and PFS/OS in proportional hazard models accounting for intra-patient correlation. Results: The study included 914 pts receiving at least 1L of therapy for MZL. Median age was 63 years, with 53% women, 57% EMZL, 22% NMZL, 21% SMZL. At diagnosis, 81% had ECOG of 0-1, 57% had stage 3/4 MZL, and 21% had elevated serum LDH. With median follow-up of 5.1 years from diagnosis, 28% of pts received 2L tx, 8% received 3L tx, 3% received 4L tx, and 1% received 5L tx. 1L tx included rituximab (R) monotherapy (41%), R+chemotherapy (R+chemo, 29%), radiation therapy (RT, 18%), surgery (8%), and other tx (5%). The most common 2L tx (n=253) included anti-CD20 monoclonal antibody (mAb, 27%), R+chemo (28%), BTK inhibitors (BTKi, 18%), RT (7%), lenalidomide (2%), Pi3K inhibitors (Pi3Ki, 1%), and other tx (17%, largely cytotoxic chemo without R). The most common 3L tx (n=75) was BTKi (39%) followed by other tx (21%), anti-CD20 mAb (13%), R+chemo (12%), lenalidomide (9%), RT (3%), Pi3Ki (1%), and autologous stem cell transplant (1%). We observed no significant differences in ORR/CRR/PFS/OS outcomes for pts receiving 3rd vs 4th vs 5th LOT, so we aggregated these LOTs into one category (“3L+”) for reporting. ORR sequentially decreased between 1L (88%), 2L (79%) and 3L+ (68%) tx (p<0.001 for sequential decrease). Similarly, the CRR significantly decreased from 69% in 1L, 51% in 2L, to 37% in 3L+(p<0.001 for sequential decrease). The 5-year PFS estimate in 1L vs 2L vs 3L+ was 55%, 38%, and 39%, respectively, with median PFS 69 months (mo), 39 mo, and 24 mo, respectively. The PFS was significantly shorter with increasing LOT (HR=1.16 [95%CI 1.30-1.99] for 2L, and HR=2.21 [95%CI 1.60-3.06] for 3L+). We observed no significant difference in PFS after 2L tx (“PFS2”) depending on the type of 1L tx when accounting for baseline stage (p=0.30). Pts receiving 1L RT had a prolonged PFS2 (88% at 3 years), but 79% of those pts had initially early stage MZL. Pts receiving retreatment with anti-CD20 mAb as 2L tx upon relapse after 1L R had median PFS2 of 44 mo. The 5-year OS estimate for pts starting 1L, 2L, and 3L+ tx was 87%, 72%, and 60%, respectively. We further compared outcomes of pts receiving BTKi in 2L (n=45), 3L (n=29), and 4L (n=14), The ORR and CRR did not significantly differ regardless of the timing of the BTKi use (ORR in 2L/3L/4L: 75%, 71%, and 69%, respectively, p=0.64; CRR: 27%, 21%, and 31%, respectively, p=0.92). Similarly, median PFS on BTKi did not significantly differ by LOT (28 mo, 26 mo, and 23 mo, respectively, p=0.89). Discussion: In this large, real-world cohort, treatment efficacy in MZL declined across early LOTs, with significant drops in response rates and PFS between 1L, 2L, and 3L+ therapy. However, outcomes appeared to plateau beyond 3L, and notably, BTKi demonstrated consistent efficacy regardless of line of use. These findings highlight the importance of optimizing early-line therapies and support flexible sequencing of BTKi in relapsed MZL. Clinical trial efforts should prioritize the first and second lines of treatment, where the potential for long-term disease control is greatest.