Trial-level meta-analysis of minimal residual disease and progression-free survival in follicular lymphoma Free

Background: Follicular lymphoma (FL) is an indolent but incurable B-cell lymphoma. Frontline therapy for FL induces remission in large majority of patients. However, a significant proportion will experience relapse and experience progressively shortened progression-free (PFS) and overall survival (OS) with subsequent therapies. Identification of patients at risk for progression early during the FL treatment may support optimized therapeutic interventions and lead to improvement in long-term FL outcomes. Measurable residual disease (MRD) has emerged as a prognostic biomarker in FL. Detection of persistent MRD may identify patients at risk of progression or death, highlighting MRD negativity as a potentially useful intermediate endpoint. Although multiple studies have demonstrated an association between MRD negativity and improved PFS, the strength and consistency of this relationship across trial level remain uncertain. To better characterize the relationship between these endpoints, we conducted a trial-level meta-analysis evaluating the association between MRD negativity rates and median PFS, which may support the use of MRD as an intermediate or complementary endpoint in FL drug development.

Methods: A comprehensive literature search was conducted , based on the PubMed and major conference proceeding (ASH, ASCO, EHA) search, we identified randomized controlled trials (RCTs) and single-arm studies conducted between May 2000 and July 2022 that reported MRD status and PFS in patients with FL. Reviews, non–English-language articles, and studies with insufficient description of MRD were excluded. Studies were included if MRD negativity was assessed in peripheral blood or bone marrow using standardized polymerase chain reaction or next generation sequencing assays (sensitivity <10^-4), and if PFS was defined as the time from randomization or treatment initiation to disease progression or death. For each study cohort, we extracted the MRD negativity rate and corresponding median PFS. If only landmark PFS results were available, the model based median PFS was estimated based on exponential distribution, considering only RCTs enrolling more than 500 patients per arm. Trial-level associations between MRD negativity rates and median PFS were evaluated using weighted Pearson correlation coefficient method, with weights proportional to cohort sample size to account for varying study sizes. The primary measure of association was the Pearson correlation coefficient (R²) along with its 95% confidence interval (CI). To assess the robustness of the observed relationship, a sensitivity analysis using the Spearman correlation coefficient was conducted to assess consistency in rank-order relationships. An R² value of ≥0.8 indicates a very strong association between the MRD and median PFS, suggesting that MRD predicts a large proportion of the variability in the median PFS. This level of association may support the use of MRD as a surrogate endpoint for the clinical outcome.

Results: We identified five randomized, two-arm trials and two single-arm trials, yielding a total of twelve evaluable cohorts with a total of 3799 enrolled patients. Among the included studies, a moderate trial-level correlation between MRD negativity rates and median PFS was observed. The weighted Pearson regression yielded an R² of 0.67, suggesting that approximately 70% of the variability in median PFS could be explained by differences in MRD negativity rates across studies. The Spearman method further supported the robustness of this association with a rank-based R² of 0.64.

Conclusions: This trial-level meta-analysis demonstrates a moderate correlation between MRD negativity and median PFS in FL, supporting further exploration of MRD and its potential as an intermediate efficacy endpoint for the intention of accelerated approval. Inclusion of more studies in future trial-level meta-analysis may be necessary to further validate the MRD as a surrogate endpoint for PFS and OS. These studies should aim to address the impact of variability in MRD detection platforms, timing of MRD assessment, baseline disease burden, and provide longer term follow-up, all of which may contribute to observed heterogeneity.