Immune Recovery Refines the Prognostic Impact of Minimal Residual Disease on Overall and Progression Free Survivals after Frontline Fludarabine, Cyclophosphamide and Rituximab (FCR) in CLL Patients

Introduction: in chronic lymphocytic leukemia (CLL), treatment with FCR is the standard of care for treatment-naive and fit enough patients (pts). MRD threshold assessed 3 months after end of treatment (EOT) is considered as a surrogate for both progression free survival (PFS) and overall survival (OS). On the other hand, antibiotic prophylaxis is often started and maintained until CD4>200/mm 3 . We hypothesized that a monitoring of MRD versus normal immune cells recovery kinetics during follow-up (FU) could help refine the predictive power of MRD EOT evaluation, and allow identification of patients with increased risk of either infection and/or relapse. Methods: we prospectively collected fresh peripheral blood (PB) samples of 162 pts from 2 institutions in France (2006-2017) at the end of treatment (EOT), then every 6 months until month 24 (M24). We used 4-color flow cytometry to quantify both MRD (classified as 3 clusters: low ( -4 ), intermediate (≥10 -4 but -2 ) or high (≥10 -2 ) level) and the reconstitution of normal lymphocyte subsets (CD4, CD8, NK). We plotted survival curves using Kaplan-Meier method and landmark analyses at different time points (9-15-21 months from day 1 cycle 1 FCR). Cox proportional model with time-dependent covariates (MRD, CD4, CD8 and NK counts) was also used to evaluate the correlation of covariates with outcomes. Results: one hundred and sixty-two pts (69% male, median age 58y) received a median of 6 courses of FCR. IGHV was mutated in 36.6%, and del13q was the most frequent cytogenetic abnormality (37.4%). Only 7 pts had del17p and/or mutation of TP53. After a median follow-up of 60.5 months, 46.3 % of pts have relapsed and/or died (median PFS estimation: 65.7 months, median OS not reached). At EOT, 65.3%, 27.2% and 7.5 % of pts achieved low, intermediate and high MRD respectively; 64.2% of pts had CD4≤200/mm 3 while 35.8% had CD4>200/mm 3 . At landmark analysis of EOT, multivariate Cox model identified IgHV unmutated (HR 2.46), MRD intermediate (HR 3.3) or high (HR 19.57) (Figure 1), CD4 level (HR 1.22) as factors associated with PFS, and only MRD (HR 1.1) and CD4 counts (HR 1.2) were found to be associated with improved OS (Figure 2). In EOT intermediate MRD pts (n=34), those having EOT CD4≤200/mm 3 had better 5-year PFS than those having CD4>200/mm 3 (53.5% vs 7.7%, p=0.001) (Figure 3). This difference was not significant for low MRD pts (65% vs 58.9%, p=0.7). In landmark analysis at 21 months (=one year after EOT assessment), intermediate MRD pts (n=44) having CD4≤200/mm 3 had significantly better 5-year PFS (85.7% vs 52.8%, p=0.0138). When performing successive landmark analyses on the whole population, CD4>200/mm 3 was associated with worse 5-year PFS at EOT (34.1% vs 59.6%, p=0.016), 6 months (49.5% vs 58.8%, p=0.017) and 12 months (46.7% vs 68.3%, p=0.008) (Figure 4). In total, 20.4% of patients developed opportunistic infections, not correlated to CD4/CD8/NK counts. Conclusion: complementarily with MRD, CD4 count is an independent predictive factor of PFS and OS after FCR. As expected, MRD increase during FU is independently associated with worse PFS and OS, but CD4 counts at EOT and the first year post-FCR refine prognosis of the "MRD detectable" risk group (independently of CLL associated risk factors such as del17p/IgHV status). This is important in the interpretation and design of clinical trials dedicated to prolong PFS/OS using consolidation regimens based on MRD detection>10 -4 in CLL. On the other hand, CD4>200/mm 3 was not predictive of infectious events. Disclosures Recher: Celgene, Sunesis, Amgen, Novartis: Research Funding; Novartis, Celgene, Jazz, Sunesis, Amgen: Consultancy. Ysebaert: Janssen: Consultancy, Research Funding, Speakers Bureau.