Risk of CNS Relapse with Diffuse Large B-Cell Lymphoma (DLBCL) in the Rituximab Era: Results from the UK NCRI R-CHOP 14 v 21 Trial

Background: Central Nervous System (CNS) relapse of Diffuse Large B-cell Lymphoma (DLBCL) is associated with a poor prognosis. Prior to the advent of rituximab the incidence of CNS relapse without CNS prophylaxis was around 5% in most unselected series. The risk in the rituximab era appears to be decreased in some but not all reported series (2.2-6.4%). CNS prophylaxis may be administered to patients deemed to be at high-risk but indications are not standardised. Methods: The randomised phase III UK R-CHOP-14 vs 21 trial assessed rituximab, cyclophosphamide, doxorubicin, vincristine, prednisolone given 2 weekly versus 3 weekly in 1080 DLBCL patients. We previously reported that R-CHOP-14 was not superior to R-CHOP-21 for OS, response rate, PFS or safety. Administration of CNS prophylaxis was at the discretion of investigators but recommended for patients with involvement of bone marrow, peripheral blood, nasal/paranasal sinuses, orbit and testis (12.5mg intrathecal methotrexate (IT MTX) for the first 3 cycles of treatment or according to local guidelines).We assessed the incidence of CNS relapse in the R-CHOP 14 vs 21 cohort and performed a retrospective analysis of risk factors for CNS relapse. The Chi-Squared Test was used to compare demographics between the groups: CNS relapse (n=16) vs no relapse (n=815) vs systemic (without CNS) relapse (n=249). Demographics assessed were sex, age (<60 vs ≥60 yrs) stage (I-IV), stage (IV vs I/II/III), performance status (PS; 0-1 vs 2), presence or absence of bulky disease, B symptoms, raised LDH, IPI group (low-high), high-risk IPI group (high vs non-high), cell-of-origin (COO; GCB vs non- GCB), presence or absence of BCL-2, BCL-6, C-MYC or “double-hit” rearrangement, MIB1 <80 vs ≥80%, <90 vs ≥90% and treatment arm (R-CHOP 14 vs 21). Details of CNS prophylaxis were retrospectively collected from participating sites using case report forms. Results: With a median follow-up of 5.7 years the overall incidence of CNS relapse was 1.48% (16/1,080), with isolated CNS relapse in 8 (0.74%) patients. 16.4% (157/953) of patients received CNS prophylaxis: IT MTX=94% (n=147), high-dose IV MTX=1.3% (n=2), unknown=4.5% (n=7). CNS prophylaxis was administered to patients with baseline involvement of bone marrow, bone, breast, kidney +/- adrenal gland involvement in 51%, 53%, 50% and 30% of cases respectively. All patients with sinus involvement and 75% of patients with testicular involvement received CNS prophylaxis. If CNS prophylaxis was given the incidence of CNS relapse was 3.8% (6/157) and 1.3% (10/796) if it was not. The demographics of patients with CNS relapse (n=16) are shown in Table 1. Risk factors for CNS relapse vs no relapse by univariate analysis were PS 2 (p=0.004), B symptoms (p=0.03), >1 site of extranodal disease (p=0.0056), stage IV disease (p=0.0063), IPI group (p=0.005) and high-risk IPI group (p=0.0015). Raised LDH was not significant (p=0.12). No histopathological feature including COO (p=0.43), C-MYC (p=0.45) or “double-hit” rearrangement (p=0.63) was associated with an increased risk of CNS relapse. Stage IV disease and B symptoms did not differentiate between the risk of developing a CNS vs systemic relapse. No difference in incidence was seen by treatment arm. No risk factor was significant on multivariate analysis. Conclusion: The incidence of CNS relapse in our total study population is very low (1.48%). Risk factors for CNS relapse versus no relapse after R-CHOP were PS 2, >1 site of extranodal disease, high-risk IPI group, B symptoms and stage IV disease by univariate analysis. None of these risk factors were significant on multivariate analysis however and we concluded that this may have been due to the low numbers of patients (n=16) in the CNS relapse group. The incidence of CNS relapse in selected patients identified to be at increased risk of CNS relapse was 3.8% which may suggest some benefit of this therapy.

Key: *Isolated CNS Relapse