Comprehensive Minimal Residual Disease (MRD) Analysis of the Fondazione Italiana Linfomi (FIL) MCL0208 Clinical Trial for Younger Patients with Mantle Cell Lymphoma: A Kinetic Model Ensures a More Refined Risk Stratification

Abstract

Abstract Background and Aims. Minimal residual disease (MRD) detection by PCR-based methods is a relevant outcome predictor in MCL, however it is not clear which might represent the most effective methodology (nested vs real-time quantitative PCR, RQ-PCR), the most informative tissue source (bone marrow, BM, vs peripheral blood, PB), the best timing of analysis (midterm vs post-therapy) and the added value of performing multiple MRD determinations. To address these issues a systematic MRD detection program was performed in the Fondazione Italiana Linfomi (FIL) MCL0208 trial (NCT02354313), a prospective, randomized phase III trial comparing lenalidomide maintenance vs observation after an intensive citarabine containing chemo-immunotherapy (R-HDS) program followed by ASCT in 300 frontline MCL patients <66 years [Cortelazzo EHA2015]. Patients and methods. MRD was assessed with ASO primers on either IGH or BCL-1/IGH rearrangements by both nested and RQ-PCR in a Euro-MRD certified lab, both in PB and BM samples at the following time points (TP): diagnosis, after 3 R-CHOP-21 and R-high-dose cyclophosphamide (R-HD-CTX), after R-high-dose Ara-C (R-HDAC), after ASCT and every six months thereafter. Landmark analysis starting 12 months after consent using Cox models was performed based on MRD negativity at each TP. To evaluate the effect of MRD on PFS and OS, we considered also the whole follow-up (FU) period, including all available MRD evaluations as time-varying covariates, both in a dichotomous (pos vs. neg) and cumulative manner (0, 1, 2 or more consecutive MRD-negative results). Finally, the discrimination ability of MRD vs clinical evaluation after ASCT was assessed in the randomized population in terms of C index. All effects were estimated adjusting for MIPI score. Results. A total of 1476 BM and 1482 PB samples were collected, for a sampling compliance rate of 93%. 250 patients (83%) had a molecular marker and showed higher median baseline tumor infiltration by flow cytometry than no marker patients (BM 8.70% vs 0.35). 231/250 (92%) patients presented at least one FU sample and were thus studied for MRD by nested PCR, while 163/231 (71%) were studied also by RQ-PCR, according to the EuroMRD guidelines. Rates of MRD negativity in BM and PB by nested-PCR, as well as by RQ-PCR, were 29%, 46%, 36% and 49% after R-HD-CTX, 53%, 78%, 73% and 87% after R-HDAC, and 54%, 79%, 81% and 89% after ASCT, respectively. MRD positivity at every TP (either by nested or RQ-PCR, either in BM or PB) showed a two-fold higher risk of relapse or death during the six months following the sampling, independently of MIPI. Remarkably, similar two-fold HRs were recorded in terms of OS, too (Table 1A). In detail, RQ-PCR showed a higher risk increase than nested-PCR, as well as BM than PB. In the landmark analysis we found that the risk of relapse gradually increased, the more MRD negativity occurs later during therapy; actually, compared to patients with MRD response after R-HD-CTX, the HR was 1.24 for MRD responders after R-HDAC, 1.51 after ASCT and 2.04 for patients never achieving MRD response by RQ-PCR in BM (Table 1B). Therefore, 3y-PFS for patients MRD positive vs negative in BM by RQ-PCR was 53% vs 66% (HR=1.57, p=0.033) after R-HD-CTX, 47% vs 64% (HR=1.47, p=0.241) after R-HDAC and 25% vs 66% (HR=2.47, p=0.037) after ASCT. Overall, the PFS discrimination ability of MRD negativity after ASCT was better than the clinical response in terms of C-index (0.67 vs 0.62), according to Cox models including MIPI and randomization arm. Most importantly, the PFS risk seemed to follow a downward trend, according to the accumulation of MRD negative results, independently of the single TP. Actually, the presence of 2 or 3 consecutive MRD negative results conferred a significantly reduced risk of relapse, refining the risk stratification of a single MRD negativity (Table 1C). E.g., focusing on RQ-PCR in BM, the HR for relapse was 0.60 for a single negativity, 0.40 for 2 consecutive negative results and 0.27 for 3 or more. Conclusions. 1) MRD results are predictive both for PFS and OS in MCL; 2) RQ-PCR is the most reliable MRD technique and data derived from BM samples provide the best risk stratification; 3) MRD analysis performed at the TP post R-HD-CTX and post ASCT well describes patients' relapse risk, independently from MIPI, however: 4) a kinetic model, based on the combination of 2 or more MRD TP, provides a powerful risk stratification tool, suitable for MRD-guided treatment. Disclosures Vitolo: Celgene: Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; Sandoz: Speakers Bureau; Gilead: Speakers Bureau; Takeda: Speakers Bureau; Janssen: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Roche: Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau.