Is there a future for maintenance temozolomide chemotherapy in PCNSL? Free

Primary central nervous system lymphoma (PCNSL) is a rare, aggressive extra-nodal non-Hodgkin lymphoma arising in the central nervous system, accounting for only 3% of intracranial tumors and an incidence of 0.4–0.44/100.000/year. Because of its rarity only very few randomized phase III studies have been completed and published,^{1, 2} and many questions regarding optimal treatment of these patients remain. Despite a currently accepted standard induction treatment with high-dose methotrexate (HD-MTX)-based polychemotherapy, many different regimens are still in use, though almost all indeed HD-MTX based.³ Response rates to induction treatment, with complete responses (CR) rates of 30%–60%, leave room for improvement and consolidation treatment with (low dose) whole brain radiotherapy (WBRT), high-dose chemotherapy and autologous stemcell transplantation (ASCT) or possibly non-myeloablative chemotherapy appears essential to improve chances of long-term survival. Although survival has much improved in recent years, relapses continue to occur even after consolidation, albeit perhaps more so after radiotherapy than after high-dose chemotherapy and ASCT.^{4, 5} Thus the optimal treatment has yet to be defined.

In the current issue of Neuro-Oncology, Mishima et al. describe their randomized phase III study for patients with newly diagnosed PCNSL, treated with HD-MTX followed by WBRT with/without temozolomide (TMZ) concomitant with WBRT and as maintenance for 2 years.⁶ They included 134 patients with a median age of 62 years of whom 122 could be randomized, after induction treatment with at least 1 of the planned 3 cycles of HD-MTX (3.5 g/m² every 2 weeks). Patients were subsequently consolidated with 30 Gy WBRT, with 10 Gy integrated boost in case of residual or recurrent tumor, and randomized for the addition of daily concurrent TMZ followed by maintenance TMZ in 5/28 day cycles during 2 years (arm B). The study was terminated for futility at the second interim analysis because of a HR of 2.18; 95% confidence interval (CI): 0.95–4.98; one-sided P = .97, and a predicted probability of showing superiority of the experimental arm of 1.3%. In part this was likely the result of very good outcomes in arm A with 87% (95% CI 75%–93%) overall survival (OS) at 2 years, considerably more than the expected 70% in the power calculations. A complicating factor in the study is that the randomization was stratified for presence or absence of residual tumor after HD-MTX induction based on local evaluation, but that the arms turned out not to have been balanced for CR/complete response, unconfirmed (CRu) rate on central review with 26.9% in arm A and 18.4% in arm B. Since, however, OS at 2 years was not better in patients in arm A with CR/CRu (77%, 95% CI 44–92) than in patients in arm A with partial response/stable disease/progressive disease (89%, 95% CI 74–96), it seems unlikely that this unbalance was the cause of the lack of effect of TMZ in the study. However, it does underscore the limited prognostic value of response categories in PCNSL, be it the result of interobserver variability or an intrinsic feature of PCNSL resulting in residual disease that is insufficiently visible on MRI imaging.⁷

Of the TMZ-treated patients 30% experienced gr 3–4 lymphopenia during radiotherapy and 44% during maintenance treatment. These high rates of lymphopenia may have played a role in the lack of effect of TMZ, as suggested by the authors, though this seems unlikely to be the only cause since the 31% lymphopenia observed in glioblastoma is not so dissimilar to the rate found in the current study.⁶ It seems likely that the added value of TMZ is at best limited, at least in patients already treated with chemo- and radiotherapy. Does this then mean that there is no role for maintenance treatment with TMZ in PCNSL? The concept of maintenance treatment with an agent that is well tolerated even by older patients is very attractive especially for patients unsuited for intensive chemotherapy or (reduced dose) radiotherapy. Induction treatment alone leads to a high rate of relapses⁸ and an effective and well-tolerated option of reducing these is dearly needed, especially since the high median age of patients with PCNSL results in a large proportion of patients with PCNSL being too old or frail for current consolidation strategies. Although, after the study by Mishima et al., TMZ may not be the most attractive agent to investigate further, it is by no means ruled out that maintenance treatment instead of consolidation after induction chemotherapy could be an effective way of reducing the relapse rate in more frail patients with PCNSL.^{9, 10} Several randomized studies are currently active to generate more clear answers to this question (NCT02313389, NCT02498951, and NCT03495960).

The results of the randomized trial in this issue, although disappointing, underscore the importance of well designed prospective phase III trials, after a positive signal in phase II trials.

Declaration

Editorial. The text is the sole product of the authors and no third party had input or gave support to its writing.

References