Lenalidomide As Maintenance Therapy after R-CHOP Has No Protecting Effect for Central Nervous System Relapse in Frontline Treatment of Diffuse Large B-Cells Lymphoma. an Ancillary Studies of the Remarc Study

Background. Central Nervous System (CNS) relapse is a rare and usually fatal event in diffuse large B-cell lymphoma (DLBCL). Prophylaxis of CNS relapse is currently based on high-dose methotrexate (HD MTX), a strategy associated with significant toxicity in elderly pts. Lenalidomide (LEN), an immunomodulatory agent, has shown activity in DLBCL, but also in association with rituximab in R/R PCNSL. In this analysis we explore the potential role of LEN given as maintenance for 2 years after R-CHOP to decrease the risk of CNS relapse

Methods. The REMARC study is an international multicentre double-blind randomized phase III trial that assessed lenalidomide (LEN) maintenance therapy (n=323) versus placebo (PBO) (n=327) in 650 DLBCL patients aged 60-80, responding to R-CHOP (NCT01122472). CNS prophylaxis was permitted including either 4 intrathecal (IT) MTX; or 2 intraveinous (IV) HD MTX: MTX 1.5g/m² 3 weeks and 6 weeks after the D1 of the last cycle of R-CHOP. The primary objective was to analyze the CNS relapse risk comparing LEN and PBO arms. Secondary objectives were to determine the clinical and biological risk factors associated with the CNS relapse risk.

Results. Median age was 68 (58-80), 392 pts (63%) had elevated LDH, 575 pts (89%) presented disseminated disease and 374 pts (60%) an elevated aaIPI (2-3). 333 pts (51%) received CNS prophylaxis during the R-CHOP induction phase. In LEN arm, 168 (52%) pts received a CNS prophylaxis: IT alone in 164 pts (98%) and HD MTX in 4 patients (2%). In PBO arm, 165 (50%) pts received a CNS prophylaxis with 160 pts (97%) with IT alone, 2 pts (1%) with HD MTX and 3 pts (2%) with both IT MTX and HD MTX. We retrospectively classified patients into groups of low- (n= 30, 5%), intermediate- (n= 377, 61%), and high-risk (n=215, 35%) CNS-IPI. MYC, MYC/BCL2, MYC/BCL2/BCL6 rearrangements were present in 10, 5 and 1 pts respectively. MYC expression alone, double expressor (DEL) (MYC/BLC2) were present in 79 pts (40%) and 71 pts (34%), respectively. 180 pts (46%) were GC, and 213 pts (54%) non-GC according to Hans score. Using Nanostring, 202 pts (50%) were classified GC-like DLBCL, 144 (36%) ABC-like DLBCL and 57 (14%) unclassified.

With a median follow-up at 81.1 months, 274 (42%) patients presented a relapse. Among them, 22 pts presented a CNS relapse (3.4%). Median time for CNS relapse was 12.2 months (95%CI: 7-48) compared to 24 months (95%CI: 19-26) for relapse other than CNS. Median OS for the group with CNS relapse was 18 months (95%CI: 11-62) compared to 64 months (95%CI: 53-75) for the group with relapses other than CNS. Pts with a CNS relapse presented an aaIPI 2-3 and were classified as intermediate-risk (n=12, 54.5%) or high-risk CNS-IPI (n=10, 45.5%). Among them, 10/16 pts (63%) were ABC. MYC rearrangement, DHL and THL were not detected. DEL was found in 45% of data available patients. According to the maintenance arm, 16 CNS relapses (73%) occurred in LEN group and 6 (27%) in placebo group.

Risk factors significantly associated with CNS relapses compared to no relapse were ABC profile (63% vs 28%, p=0.037), IPI 3-5 (91% vs 67%, p=0.028), and elevated LDH (38% vs 15%, p=0.012). Patients with high CNS-IPI were significantly associated with a worse PFS and OS in comparison with low-CNS-IPI (respectively HR=2.09, p = 0.035 and HR=2.70, p=0.032). In multivariable analysis, high CNS-IPI, age≥70, ABC profile assessed by NanoString and partial response at the end of R-CHOP were associated with poor OS.

Conclusion

LEN in maintenance for 2 years was not associated with a lower rate of CNS relapse. High CNS-IPI was confirmed as a poor prognosis factor for CNS-specific PFS, PFS and OS. The best strategy to decrease the risk of CNS relapses in DLBCL remains to be defined.