Overall Survival Associated with Real-World Treatment Sequences in Patients with CLL/SLL in the United States

Introduction: Treatment guidelines suggest the preferred use of covalent Bruton tyrosine kinase inhibitors (cBTKi) and/or B-cell lymphoma 2 inhibitors (BCL2i) in earlier lines of therapy (LoT), with or without an anti-CD20 antibody (aCD20ab) in patients with chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL). However, the optimal sequencing of these and other regimens for CLL/SLL remain unknown. This study compared overall survival (OS) associated with frequently observed real-world treatment sequences in patients with CLL/SLL treated in the United States.

Methods: The nationwide Flatiron Health electronic health record (EHR)-derived de-identified database (cutoff Nov 2023) was used to select adult CLL/SLL patients who initiated systemic therapy on or after 2016 and had received at least two LoT. Treatment regimens were grouped into pre-defined, mutually exclusive categories based on drug class. Most common unique groups (n≥50) based on the sequence of treatments received in first (1LoT) and second (2LoT) LoT were included. All infrequent sequences including BCL2i + aCD20ab followed by (→) cBTKi monotherapy (n=12; 1LoT → 2LoT) were excluded. OS from the initiation of 1LoT, associated with each treatment sequence, was compared between sequences using two multivariable cox proportional hazard models, with significance level of 0.05. Model1 (M1) adjusted for baseline factors including age, sex, race, ethnicity, socioeconomic status, practice type, disease subtype, RAI stage, ECOG performance status (PS), del(11q), del(13q), tri12, del(17p)/TP53, IGHV, year of initiation of 1LoT, and time from initial diagnosis to 1LoT. Model2 (M2) also considered the cumulative number of LoT as a time-dependent covariate to account for potential differences in OS due to treatments received beyond the first two LoT. The Kaplan-Meier method was used to calculate median OS.

Results: Of 2,354 patients who met eligibility, n=1,711 (73%) received the 16 most frequent treatment sequences where the median age was 71 (IQR; 63–78) years, 63% were male, 74% non-Hispanic White, and had a median follow-up of 47 (26–69) months. Among patients with available data, 11% (148/1,365) had del(17p)/TP53 mutation, 62% (406/653) had unmutated IGHV, 91% (1,119/1,224) had ECOG PS 0/1, and 63% (651/1,039) were initially diagnosed with RAI stage 0/I. Key sequences of treatments patients received included chemoimmunotherapy (CIT) → cBTKi monotherapy (n=350; 20.5%); aCD20ab monotherapy → cBTKi monotherapy (n=201; 11.7%); aCD20ab monotherapy → BCL2i + aCD20ab (n=174; 10.2%); cBTKi monotherapy → aCD20ab monotherapy (n=108; 6.3%); aCD20ab monotherapy → CIT (n=99; 5.8%); CIT → CIT (n=94; 5.5%); cBTKi monotherapy → BCL2i + aCD20ab (n=83; 4.9%); cBTKi monotherapy → CIT (n=82; 4.8%); cBTKi monotherapy → BCL2i monotherapy (n=73; 4.3%); cBTKi monotherapy → cBTKi + aCD20ab (n=70; 4.1%); CIT → BCL2i + aCD20ab (n=65; 3.8%); and CIT → aCD20ab monotherapy (n=62; 3.6%). Compared to patients who received cBTKi monotherapy → BCL2i + aCD20ab (reference), treatment with the following sequences was associated with significantly worse OS in M1 and M2: cBTKi monotherapy → aCD20ab monotherapy (adjusted hazard ratio [aHR]: 2.47 [95%CI: 1.32–4.63]); aCD20ab monotherapy → CIT (2.68 [1.42–5.07]); and CIT → CIT (2.87 [1.54–5.34]). All other treatment sequences (vs reference) were either not significantly different in M1 and M2 or were only significantly different in M1. Median (95% CI) OS was not estimable for all sequences except for patients that received cBTKi monotherapy → aCD20ab monotherapy (63 [55–74] months).

Conclusions: Sequencing of CIT → CIT or aCD20ab, and cBTKi monotherapy → aCD20ab in earlier LoT were associated with worse OS compared to patients treated with cBTKi monotherapy → BCL2i + aCD20ab. Despite limiting the cohort to 2016 and later, many of the most frequent sequences evaluated in this study did not include use of targeted therapies within the first two LoT. As this study found no significant differences in outcomes for certain targeted therapy-based sequences (vs. reference), and as newer treatments like non-covalent BTKi and CAR-T are adopted, optimizing treatment sequencing will continue to remain an important area of future research in this patient population.

Keywords: treatment sequencing, CLL/SLL, real world data, targeted therapy, line of therapy