A low lymphocyte-to-monocyte ratio (LMR) predicts PFS, POD24 and OS in previously untreated, high tumor burden follicular lymphoma (FL): an analysis from the RELEVANCE trial
Résumé
Introduction: The peripheral blood LMR has been postulated as an accessible piece of information about the composition of the tumor microenvironment. In a single-center, retrospective, unselected cohort of FL, a low LMR was shown to be associated with an older age and higher tumor burden and to predict for a shorter progression-free and overall survival (PFS, OS) and a higher risk of histological transformation and second primary malignancies (Mozas, Leuk & Lymph, 2020). We explored the impact of the LMR in patients included in the phase III RELEVANCE trial (Morschhauser, NEJM, 2018 and JCO, 2022), which compared rituximab-chemotherapy (R-chemo) with rituximab-lenalidomide (R2) in patients with previously untreated, high tumor burden FL.
Methods: Xtile and the maxstat package of R software were used to find the best LMR cutoff based on PFS data and then validated using a truncated power basis spline method. Baseline characteristics, PFS per investigator assessment, early relapse (POD24) and OS were compared between LMR risk groups. Multivariable Cox regression models including the FLIPI score and the treatment arm were built for survival analyses and uni/multivariable logistic regression was used for POD24 analysis.
Results: Among the 1030 patients included in the RELEVANCE study, 1018 had LMR data available. The median LMR was 2.5 (range, 0–93) and a LMR cutoff of 2 was found to best predict PFS. Patients with a LMR ≤2 (n = 372, 37%) were older and displayed higher-risk features (Figure A). In the global cohort, a LMR ≤2 was predictive of a shorter PFS (HR = 1.39 Figure B and C) and OS (HR = 1.44), but its negative impact in the multivariable model remained statistically significant solely for PFS (HR for LMR = 1.31). Likewise, a LMR ≤2 was associated with a higher risk of POD24 (univariable OR = 1.84; multivariable OR = 1.71). No significant interaction was observed in the PFS analysis between treatment arms and the LMR, despite the fact that the LMR was significantly associated to PFS only in the R-chemo arm (P = 0.001) and not in the R2 arm (P = 0.08).
Conclusions: In this RELEVANCE subanalysis, we demonstrated that the LMR is a strong predictor of PFS and early relapse in high tumor burden FL patients. Whether treatment with immunomodulators such as lenalidomide may abrogate its negative prognostic impact needs to be further investigated.