Lymphopenia Confers Poorer Prognosis in Myelodysplastic Neoplasms (MDS) with Very Low and Low IPSS-M

Abstract

Introduction: In addition to MDS cell-intrinsic genetic lesions, we and others found increasing evidence of dynamic alterations in both adaptive and innate immunity in the disease genesis and/or progression (Boy et al., Nat Commun 2023;14:588). Although a few reports have suggested a prognostic impact of absolute lymphocyte count (ALC) in both del(5q) and non-del(5q) MDS, lymphopenia is not captured by current prognostic scores. Moreover, the association between lymphopenia and somatic mutations has not been reported. In this work, we describe the prevalence of lymphopenia in MDS patients, the interaction with molecular features, and the impact of lymphopenia on patient outcomes. Methods: We retrospectively included all MDS patients followed in our center (Saint-Louis Hospital, Paris, France) between January 2015 and January 2022 with a molecular evaluation. According to the local laboratory threshold, lymphopenia was defined as ALC <1.5 G/L. Low (LR), intermediate (IR), and high-risk (HR) MDS were defined according to IPSS-M (LR: IPSS-M Low and Very Low, scores < -0.5; IR: IPSS-M Moderate Low and Moderate High, scores -0.5 to 0.5; and HR: IPSS-M High and Very High, scores >0.5 respectively).Survival analyses were estimated between MDS diagnosis until death from any cause (OS) or last follow-up, or until AML transformation (LFS) or last follow-up. Allogeneic stem cell transplantation was considered a censoring event. Results: Our cohort included 372 MDS patients (45% females, median age 72 IQR[65-79]) with 230 (62%), 70 (19%), and 72 (19%) patients with LR, IR and HR-MDS respectively. Compared to IPSS-R, molecular IPSS reclassified 175 patients (47%, with 97 (26%) upstaged and 78 (21%) downstaged). The median ALC in the whole cohort was 1.20 IQR[0.87-1.60] G/L. Lymphopenia was highly prevalent in all MDS subgroups (68%, 79% and 75% in LR, IR and HR-MDS respectively), and ALC inversely correlated with higher IPSS-M (Spearman rho=-0.13, p=0.01). Lymphopenia was associated with lower ANC (median 1.90 IQR[1.1-3.4] versus 2.70 IQR[1.6-4.2], p=0.01) and monocyte count (median 0.31 IQR[0.17-0.46] versus 0.42 IQR[0.25-0.78], p=0.04). Lymphopenia was not significantly associated with an excess of marrow blasts (78/267, 29% versus 21/105, 21% in patients with or without lymphopenia respectively, p=0.09). However, its prevalence was significantly lower in MDS with ringed sideroblasts (RS) (27/49 (55%) versus 240/267 (90%), p<0.01). Consistently, SF3B1 mutations were more prevalent in MDS patients without lymphopenia (29/107 (27%) versus 27/265 (10%), p<0.01), while the rest of the somatic mutational landscape did not differ between the 2 groups. The median follow-up of the entire cohort was 4.1 IQR[3.7-4.6] years. Among IPSS-M categories, lymphopenia adversely impacted OS and LFS in LR-MDS patients (median 8.8 years vs not reached for OS, p=0.02; median 12.3 years versus not reached, p=0.02 for LFS) but not in IR (p=0.34 and p=0.49 for OS and LFS respectively) or in HR-MDS (p=0.82 and p=0.83) patients ( Figure 1). In SF3B1 mutated LR-MDS patients (n=46), the adverse prognostic impact of lymphopenia at diagnosis was still significant for both OS (p=0.03) and LFS (p=0.03). Conclusion: In this single-center retrospective analysis, lymphopenia was highly prevalent at MDS diagnosis in all IPSS-M subgroups and ALC negatively correlated with disease severity. The presence of lymphopenia adversely impacted OS and LFS in the LR-MDS subgroup, potentially suggesting a significant role of host immunity in early disease phase and possibly adding a prognostic parameter to IPSS-M.