CNS prophylaxis in DLBCL: first do no harm

In this issue of Blood, Wilson et al¹ present a retrospective analysis of the timing of intravenous high-dose methotrexate (HD-MTX) during first-line treatment of diffuse large B-cell lymphoma (DLBCL). They found that intravenous HD-MTX given at the end of treatment (EOT) was equally effective for central nervous system (CNS) prophylaxis and less toxic than MTX given earlier (ie, between cycles of anthracycline-based immunochemotherapy).

Secondary progression of DLBCL into the CNS after initial treatment with standard anthracycline-based immunochemotherapy is a devastating event. Several groups have identified risk factors for secondary CNS progression and have developed the CNS International Prognostic Index (CNS-IPI)² that identifies high-, intermediate-, and low-risk groups. Because the high-risk group has an ∼10% risk of progression in the CNS at 2 years, strategies to prevent CNS progression have been explored.³ The blood-brain barrier prevents many treatments from entering the CNS space, and MTX has long been used as a method for treating and preventing CNS progression. It has been speculated that CNS progression occurs early in the course of treatment, and thus, should be intercalated into standard immunochemotherapy.³ However, guidance has largely come from expert opinion developed from subgroup analyses of prospective and retrospective studies. Rigorous evaluations focused on the utility of CNS prophylaxis are lacking.

Thus, Wilson et al performed a wide-ranging retrospective analysis of the timing of CNS prophylaxis with intravenous HD-MTX occurring during or after immunochemotherapy. Their study includes 1384 patients with DLBCL or high-grade B-cell lymphoma not otherwise specified who were diagnosed between 2007 and 2020 at 47 centers in Europe, Australia, and North America. It should be noted that roughly half the patients (46.1%) also received intrathecal MTX, with more patients in the group that received MTX after immunochemotherapy than in the group that received MTX during immunochemotherapy. Ultimately, there was no significant difference in CNS progression rates when HD-MTX was given during or after immunochemotherapy; however, patients had more chemotherapy delays and increased toxicity when MTX was given during immunochemotherapy. Ultimately, the authors thoughtfully conclude that CNS prophylaxis based on HD-MTX is best performed at the end of primary immunochemotherapy (EOT), a clinically meaningful finding which will limit toxicity to our patients.

Although this study was not meant to evaluate the utility of CNS prophylaxis in general, Wilson et al did note a rate of CNS relapse in their patients similar to that of historical controls who had received limited CNS prophylaxis.² Other recently reported retrospective studies that were focused on other aspects of MTX-based CNS prophylaxis in DLBCL (intrathecal vs intravenous HD-MTX,⁴ outcomes in older patients,⁵ and single-institution analyses^6,7) have also failed to show improvements compared with historical CNS relapse rates. These retrospective studies will likely be the highest level of data we will have. A prospective randomized study that evaluates the utility of MTX-based prophylaxis for CNS progression will need a very large numbers of participants to detect a small difference in CNS progression rates, and it may detract from other potential studies of emerging biologics-based targets.

The utility of CNS prophylaxis in DLBCL is increasingly questioned. The Wilson et al study, in the context of other recently presented studies, suggests that MTX, given intrathecally or intravenously or during or after immunochemotherapy, is not adequately preventing CNS progression. However, there is evidence that suggests that in the groups with highest risk, such as those with extranodal disease involving testes, kidneys, or adrenal glands, MTX may provide a level of protection.⁸ Thus, it is my current practice to limit MTX prophylaxis to these situations. On the basis of the study by Wilson et al, intravenous HD-MTX (if used) should be given at the EOT for at least 2 cycles and likely reserved for fit patients. In addition, the priority for treatment should be on anthracycline-based immunochemotherapy and not on CNS prophylaxis. If CNS prophylaxis delays or otherwise interferes with curative-intent immunochemotherapy, it should be abandoned.

What does the future hold? The lymphoma field is hopeful that therapies directed at the biology of DLBCL will replace MTX-based strategies for CNS prophylaxis. Small-molecule inhibitors have a chance of reaching the CNS through the blood-brain barrier. Some that are already known to do so (ibrutinib⁹ and lenalidomide¹⁰) have already been tested with rituximab, plus cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP). It is possible that these and other agents may be effective in preventing secondary CNS progression, particularly if therapy is chosen on the basis of predicted activity identified with new classification systems that are focused on DLBCL biology.⁹ It is also possible that better, more effective therapies against systemic DLBCL will prevent secondary CNS progression. Finally, improvements in the detection of CNS involvement through DNA testing of cerebrospinal fluid, or better risk assessment with circulating tumor DNA could potentially define a group of patients with extremely high risk who warrant additional treatment.

Given the uncertainty around CNS prophylaxis, upcoming studies of first-line treatment for DLBCL should not mandate CNS prophylaxis, particularly if it is paired with novel agents known to cross the blood-brain barrier. CNS-specific progressions should be reported so that researchers can identify new ways of preventing devastating CNS relapses. With the currently available data, intravenous HD-MTX should be limited to the EOT and to patients with the highest risk.