A focus on pharmacotherapy for primary central nervous system lymphoma

Abstract

The prognosis of primary central nervous system lymphoma (PCNSL) is still poor compared to other diffuse large B-cell lymphomas. High-dose methotrexate-based chemotherapy plays a central role in the management of PCNSL; however, the optimal regimen has not been established thus far. Consolidating whole-brain radiotherapy is currently not routinely being performed at most centers due to lack of survival advantage in a randomized Phase III trial and increased risk of cognitive impairment. Promising results with the combination of high-dose methotrexate-based chemotherapy with the CD20 antibody rituximab and consolidating the primary chemotherapy by other non-cross-resistant conventional chemotherapy or high-dose chemotherapy followed by autologous stem cell transplantation have been reported in non-comparative studies. Results of well-designed randomized trials must be awaited to better define the optimal management of PCNSL.

Keywords:

• blood-brain barrier
• chemotherapy
• high-dose methotrexate
• primary CNS lymphoma
• rituximab

The ability to penetrate the CNS by crossing the intact blood–brain and blood–CSF barrier is considered a condition sine qua non for systemic therapy to be effective in primary central nervous system lymphoma (PCNSL). Thus, the cyclophosphamide, vincristine, doxorubicine, prednisolone regimen, the current standard chemotherapy for aggressive lymphomas, is not an adequate treatment of PCNSL. Unfortunately, only a limited number of drugs are able to enter the CNS. The CSF concentrations (regarded a surrogate marker for CNS penetration) are >100% of the corresponding plasma concentrations for high-dose cytarabine (HDAraC), temozolomide and thiotepa, approximately 30% for ifosfamide and <5%, but still cytotoxic, for high-dose methotrexate (HDMTX). These drugs are currently most commonly used in patients with PCNSL.

The paucity of randomized trials to support therapeutic decisions is the main problem in the management of PCNSL. The current knowledge on PCNSL treatment is derived from one randomized Phase III [1] and one randomized Phase II study [2] and a number of relatively small non-comparative Phase II studies and retrospective analyses. Based on these findings, HDMTX is the most important drug in the treatment of PCNSL. Unfortunately, its optimal dose and schedule is not known. It seems likely that not only MTX dose intensity (optimally >3 g/m² in maximum 2-weeks intervals), but also the infusion rate (optimally 3–4 h) contribute to effectivity.

HDMTX monotherapy is considered less effective than HDMTX combined with other CNS penetrating drugs. This belief is based on the indirect comparison of trials using HDMTX alone to those using HDMTX-based polychemotherapy and on only one randomized Phase II trial. In the German NOA-03 trial with 37 patients, six biweekly courses of HDMTX (8 g/m²) resulted in a complete remission (CR) rate of 29.7% and a median relapse-free survival of CR patients of 13.7 months [3]. In contrast, in non-comparative trials with 52–102 patients using HDMTX-based polychemotherapy CR rates were 33–61% and median progression-free survival (PFS) of all patients ranged between 10 and 21 months [4–8], however, mostly with WBRT given directly after chemotherapy. In the G-PCNSL-SG1 trial, the largest trial in PCNSL published thus far, patients were treated with HDMTX monotherapy (4 g/m² every 2 weeks) within the first study period (N = 409) and a combination of HDMTX (same dose) and ifosfamide (1.5 g/m2 over 3 days) thereafter (N = 128), as introduced per amendment. The CR rate increased by the addition of ifosfamide from 32 to 42% and primary progression rate was reduced from 26 to 15% [1]. The only published randomized Phase II study with 79 patients directly compared HDMTX 3.5 g/m² on day 1 every 3 weeks alone or combined with HDAraC 2g/m² twice daily on day 2–3, both followed by WBRT. The CR rate (primary end point) was significantly higher with 46 versus 18%, p = 0.006, and 3-year PFS significantly longer with 38 versus 21%, p = 0.01, with the combination [2]. The authors concluded that based on these results, HDMTX combined with HDAraC should be considered standard chemotherapy for PCNSL. In the opinion of the author of this article, the findings of this study should be interpreted with caution due not only to the relatively small number of patients but also to the unusually poor treatment results in the monotherapy arm probably due to suboptimal HDMTX schedule.

As the vast majority of PCNSL are diffuse large B-cell lymphoma (DLBCL), the inclusion of rituximab in the treatment protocols analogous to systemic DLBCL appears reasonable. However, as a large protein, rituximab poorly penetrates the CNS, with less than 1% of systemic concentration reaching the leptomeningeal compartment [9] as an indirect measure for parenchymatous concentration. Thus, its maximal concentration/effectivity in the CNS might be assumed in the early treatment phase at blood–brain barrier breakdown within the tumor. Improved effectivity of chemotherapy by addition of rituximab was suggested by small non-comparative studies [10]. The benefit of rituximab could at least in part be explained by the systemic effects on sub-clinical (occult) systemic lymphoma manifestation. In a retrospective analysis of 81 patients, the addition of rituximab to HDMTX was demonstrated to improve CR rates as well as PFS and overall survival (OS) [11]. It is of note, however, that the results in the HDMTX monotherapy group of this study with a CR rate of 36%, median PFS of 4.5 months and median OS of 26.7 months were poorer than reported in the literature.

Intra-CSF therapy appears a logical replenishment of systemic treatment of PCNSL – a tumor known for its tendency to seed to the leptomeninges. Moreover, cytotoxic MTX concentrations can be better sustained within the CSF by repeated application via an Ommaya reservoir, as compared to short-lived cytotoxic CSF concentration for 24–72 h only after intravenous HDMTX [12]. Thus far, there is no evidence that intra-CSF therapy adds benefit when given in conjunction with HDMTX-based systemic chemotherapy. The experience of a German group only suggests the role of an intensive intraventricular chemotherapy in PCNSL: With a regimen including HDMTX (5 g/m²), HDAraC (3 g/m²), vincristine, alkylating agents and dexamethasone combined with intensive intraventricular chemotherapy via an Ommaya reservoir, CR rate of 61%, median PFS of 21 months and median OS of 54 months were reached in 65 patients [4]. These results could not be reproduced in a second trial using the same regimen without intraventricular treatment [13]. The key point is probably the use of an intraventricular access rather than an intrathecal and a high treatment intensity, since no benefit from intra-CSF chemotherapy has been demonstrated in retrospective analyses including less-intensive intrathecal (via lumbar puncture) regimens.

It would be interesting to evaluate the combined systemic/intraventricular therapy in a direct comparison to other intensified protocols.

The only published randomized Phase III study, the G-PCNSL-SG1 study, defined the role of consolidating WBRT with 45 Gy after HDMTX-based primary chemotherapy in newly diagnosed PCNSL, showing no survival advantage for WBRT compared to chemotherapy alone. The results were confirmed in a long-term analysis including an as-treated evaluation [14]. Although there were some problems with the execution of this study, all detailed in the first publication [1], the question about the role of WBRT could clearly be answered, and as a consequence, WBRT should not routinely be given as a part of initial therapy in PCNSL. The increased risk of delayed neurotoxicity after combined chemo-/radiotherapy compared to chemotherapy alone [15] is another important argument against the use of WBRT in the primary management of PCNSL. Thus, the concept of consolidating of the effect of HDMTX-based primary chemotherapy by other non-cross-resistant chemotherapy has gained interest and was evaluated in more recent studies. In the CALGB 50202 study, 44 patients were initially treated with four courses of rituximab/HDMTX/temozolomide, which in CR patients was followed by etoposide/HDAraC. The results were promising with a CR rate to rituximab/HDMTX/temozolomide of 66%, a 2-year rate of PFS of 57% and median time to progression of 4.0 years [16]. In younger patients, the consolidation therapy can be intensified up to high-dose chemotherapy followed by autologous stem cell transplantation. In a German study with 80 patients <65 years, the induction chemotherapy with rituximab/HDMTX was followed by rituximab/HDAraC/thiotepa, stem cell harvest and high-dose chemotherapy followed by autologous stem cell tramsplantation with carmustine/thiotepa followed by ASCT. The 3-year PFS of the first 79 patients was 63% and 3-year OS 78% [17]. In an American study, 32 patients ≤67 years received five courses of rituximab/HDMTX/procarbacine/vincristine, which was in responders consolidated by HD-therapy with thiotepa/cyclophosphamide/busulfan followed by ASCT. The CR rate to rituximab/HDMTX/procarbacine/vincristine was 66% and the 3-year PFS estimate 79%, however, with a relatively high toxic death rate of 10% [18]. These regimens should not be considered standard until confirmation by well-designed randomized trials.

It is very welcome that new drugs start to find their way into the treatment of PCNSL, with relapsed and refractory disease being predestined due to lack of treatments standards and poor prognosis of these patients. In a Phase I/II German trial, the mammalian target of rapamycin inhibitor temsirolimus was evaluated in 37 elderly patients (median age 69 years) with refractory or relapsed PCNSL. A relatively high response rate of 61% was seen, however, with considerable toxicity, mostly pulmonary infections [19]. The immunomodulating drugs lenalidomide and pomalidomide are other drugs worth being tested in clinical studies [20]. Since PCNSL tumors were shown to exhibit multiple genetic properties shared with ABC-type DLBCL, it appears attractive to address targets and pathways known to be involved in the pathogenesis and survival of this kind of lymphoma such as NF-kB, the B-cell receptor, the janus kinase (JAK)/Signal transducer and activator of transcription (STAT) pathway, interferon regulatory factor 4, B-Cell Lymphoma 6 Protein as well as PIM kinases [20]. Also, the tumor microenvironment might therapeutically be exploited by addressing chemokine pathways, macrophages and T-cell responses. Studies with the Bruton tyrosine kinase inhibitor ibrutinib, the dual PI3K-mammalian target of rapamycin inhibitor PQR309 and the checkpoint inhibitor nivolumab are currently in preparation.

In summary, considerable advantage has been achieved in the treatment of PCNSL by introducing HDMTX in the 1990s, HDMTX-based polychemotherapy since 2000 and abandonment of consolidation WBRT from primary treatment since 2010. However, a better definition of the optimal treatment is still urgently needed. Ongoing randomized trials address many important questions such as the best combination partners for HDMTX, the role of rituximab and the value of HD-ASCT as consolidation after HDMTX-based chemotherapy. Unfortunately, the trials are not always adequately sized or include inadequate comparator arms. Thus, the results will have to be interpreted with caution. The other goal should be a more precise definition of risk factors and, as a consequence, a risk-tailored therapy with intensified treatment, when possible, in poor-risk patients and less intensive treatment in good-risk patients to spare toxicity, particularly delayed neurotoxicity.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.