Journal of Clinical Oncology, Vol 21, Issue 12 (June), 2003: 2407-2414

Conventional Chemotherapy
The primary chemotherapy regimen in PCNSL patients should include intravenous HD-MTX (MTX ? 1 g/m2), which is the most effective drug against these malignancies. HD-MTX produces a response rate of 52% to 88% as a single agent and 70% to 94% when associated with other drugs; these chemotherapeutic approaches followed by WBRT are associated with a 2-year overall survival of 58% to 72% and 43% to 73%, respectively. The efficacy of this drug depends on the duration of exposure and drug concentration,which are determined by the administration schedule and pharmacokinetics. Because MTX clearance from plasma is triphasic, an initial rapid administration to overcome the distribution phase, followed by a more prolonged infusion, seems to be the most rational schedule for this drug. However, this strategy has not been used in most published trials. The optimal duration of HD-MTX infusion is still unknown; in most trials using doses of 1 to 5 g/m2, MTX has been administered in a 4-hour infusion, whereas 24-hour infusions have been used for higher doses. In a study using an MTX dose of 100 mg/kg, a 3-hour infusion has been associated with a significantly higher response rate and CSF levels compared with a 6-hour infusion. The optimal dose of MTX has not been defined. CSF MTX concentration is strictly related to the dose administered (see Leptomeningeal Lymphoma). The best timing of MTX administration remains undefined, but no significant difference in efficacy or toxicity was observed when MTX at 3.5 g/m2 was administered every 3 weeks versus every 10 days.

Any regimen without HD-MTX is associated with outcomes no better than with radiotherapy alone. At least partially because of their poor blood-brain barrier (BBB) penetration, the most effective drugs against NHL, doxorubicin and cyclophosphamide, are associated with unsatisfactory results. Therefore, the CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) regimen is not an effective treatment against PCNSL.

Corticosteroids alone may produce a rapid and substantial tumor regression in up to 40% of PCNSL patients. Thus, the concurrent use of corticosteroids and investigational agents should be avoided in phase II trials because it may not be clear which drug caused tumor regression. Moreover, because many patients with brain masses are treated with corticosteroids before definitive therapy, the baseline cranial magnetic resonance imaging scan should be obtained immediately before the initiation of the experimental treatment.

Several drugs have been added to HD-MTX to improve outcome. These drugs were selected based on their capacity to penetrate the BBB and on their demonstrated efficacy against systemic NHL. However, none of these drugs had been previously evaluated as effective single agents in patients with relapsed or refractory PCNSL. Preliminary results from small pilot studies in relapsed patients are now available with topotecan, rituximab, temozolomide, and the procarbazine, lomustine, and vincristine regimen. A recently reported survival improvement resulting from the addition of high-dose cytarabine immediately after HD-MTX deserves to be prospectively confirmed. Although there is no proven benefit of additional drugs, it is likely that an MTX-based polychemotherapy regimen will emerge as the standard combination for PCNSL. The identification of new active drugs and combinations in phase I/II trials in relapsed or refractory PCNSL should receive high priority.

BBB Disruption (BBBD)
Increasing drug delivery to the lymphoma-infiltrated brain and intracerebral lymphoma could significantly enhance survival. Investigators at the Oregon Health & Science University (Portland, OR) have focused on delivery of agents across the BBB by intra-arterial infusion of hypertonic mannitol, resulting in reversible BBBD. This procedure has been performed at Oregon Health & Science University and at the collaborating institutions of the multicenter BBBD consortium, and high rates of good and excellent degrees of BBBD, acceptable complication rates, and high response and survival rates have been obtained. The estimated 5-year survival in patients treated with MTX-based chemotherapy in conjunction with BBBD is 42%.46 Moreover, 86% of patients in complete response after 1 year from BBBD have demonstrated no cognitive loss over time. Given its good efficacy and safety profiles, the role of BBBD as part of first-line treatment deserves to be investigated in future trials.

BBBD may also be an effective strategy in PCNSL patients who have experienced relapse after initial treatment with HD-MTX. Carboplatin-based chemotherapy in conjunction with BBBD produced a 36% response rate and a median survival after relapse of 6.8 months (range, 1 to 91 months); 16% of patients survived beyond 3 years from salvage therapy without cognitive loss in the absence of prior radiotherapy.47 Finally, the technique of BBBD may prove most useful in the delivery of agents unlikely to traverse an intact BBB, such as unconjugated or radiolabeled monoclonal antibodies, which deserves to be assessed in future trials.

High-Dose Chemotherapy With Autologous Peripheral-Blood Stem-Cell Transplantation (APBSCT)
High-dose chemotherapy supported by APBSCT has been used as one strategy to dose-intensify chemotherapy given to patients with newly diagnosed or relapsed PCNSL. Theoretically, this strategy can be used to replace WBRT in an effort to avoid treatment-related neurotoxicity. In patients with newly diagnosed PCNSL, there have been two small APBSCT phase II trials. In one study, 28 patients received five cycles of MTX 3.5 g/m2 and two cycles of cytarabine 3 g/m2 daily for 2 days, followed by carmustine, etoposide, cytarabine, and melphalan consolidation chemotherapy in those patients with chemosensitive disease. Fourteen patients completed the planned therapy, and five remained in remission at a median of 26 months after transplantation. Significant treatment-related toxicity was rare; however, only 50% of patients had chemosensitive disease, and a significant proportion relapsed after transplantation. In another ongoing study, a combination of MTX, thiotepa, and cytarabine is being used as the induction regimen followed by high-dose chemotherapy with carmustine and thiotepa and hyperfractionated radiotherapy.49 Nineteen of 24 patients enrolled to date have achieved a complete remission, and there have not been any unexpected acute toxicities. In a study on 22 patients with recurrent or refractory primary CNS or intraocular lymphoma, induction cytarabine and etoposide followed by high-dose chemotherapy with thiotepa, busulfan, and cyclophosphamide produced a complete remission rate of 72%, with a 3-year overall survival of 64%.50 However, there was a significant incidence of neurotoxicity as well as significant treatment-related morbidity/mortality in patients over the age of 60 years.

The preliminary results from these trials using high-dose chemotherapy with APBSCT clearly indicate that this strategy is feasible in patients with PCNSL. It is possible that the patients treated at relapse who previously received WBRT will have a higher risk of neurotoxicity. As with conventional therapy, cytostatic drugs for induction and conditioning chemotherapy have been selected on the basis of their safety, efficacy against systemic lymphomas, and ability to cross the BBB. The lack of cross-resistance with MTX has been an advantage when this strategy has been used as salvage therapy.50 The role of high-dose chemotherapy and APBSCT in PCNSL remains to be defined considering that the worldwide experience is still limited, and further studies will need to be performed to identify the optimal induction and high-dose chemotherapy regimens.

Radiotherapy alone is rarely curative in PCNSL patients because response is usually short-lived, with a median survival of 12 to 14 months. Consolidation after chemotherapy may represent the best role for radiotherapy, although the optimal field and doses have not been identified. Because PCNSL is often multifocal, the target for radiotherapy is the whole brain, whereas the added value of the tumor-bed boost is questionable.  The inclusion of the posterior two thirds of the eyes into the radiation field is advisable.  The radiation dose should be decided on the basis of response to primary chemotherapy, and, until definitive conclusions from well-designed trials are available, radiotherapy parameters should follow the widely accepted principles used for other aggressive NHLs. Doses of  40 Gy or 36 to 40 Gy may be advisable in patients with or without residual disease, respectively, after primary chemotherapy.

Combined chemoradiotherapy is associated with severe neurologic impairment in 40% of patients and a neurotoxicity-related mortality of 30% especially in patients older than 60 years of age. In fact, a direct relationship between age and risk of neurotoxicity has been reported, and female sex, MTX dose more than 3 g/m2, intrathecal chemotherapy, and higher tumor radiation dose have also been proposed as risk factors for this complication.  Avoiding radiotherapy in patients older than 60 years of age in complete remission after primary chemotherapy has been proposed as a strategy to minimize neurotoxicity (see Chemotherapy as Exclusive Treatment).

New strategies to improve the tolerance and efficacy of radiotherapy should be investigated in future trials. An important issue will be to define the risk of neurotoxicity in younger patients. In a recently published study, HD-MTX–based chemotherapy, followed by WBRT (45 Gy) and postradiation cytarabine, has been associated with severe neurotoxicity in 15% of patients; this complication was seen as frequently in patients younger than 60 years as in those who were 60 years or older. Interestingly, in the same study,  the use of hyperfractionated WBRT (1.2 Gy/fraction twice daily; total dose, 36 Gy) did not seem to reduce the risk of neurotoxicity. Substantial dose reduction or WBRT withdrawal in patients younger than 60 years should be critically discussed considering that a detrimental survival effect has been reported with a WBRT dose reduction from 45 Gy to 30.6 Gy in these patients in a nonrandomized trial.  A major question in older patients will be to define whether reduced radiation doses and restricted treatment fields may reduce the incidence of neurotoxicity without compromising efficacy.

As described above, the use of chemotherapy alone is of particular importance in PCNSL patients over the age of 60 years who achieve a complete remission after HD-MTX–based chemotherapy. In small series, this strategy has produced response rates in excess of 90%, and patients who relapsed were effectively treated with additional salvage chemotherapy or radiotherapy.  In published prospective trials, HD-MTX alone produced a 52% to 100% response rate and a 2-year survival rate of 61% to 63%,  whereas HD-MTX–based polychemotherapy regimens resulted in a 65% to 100% response rate and 2-year survival rate of 65% to 78%. In a comparison of older patients treated with or without WBRT after HD-MTX–based chemotherapy, chemotherapy alone markedly reduced the risk of neurotoxicity, and although there was a higher relapse rate in patients treated without WBRT, there was no difference in survival (median, 32 months) between these two subgroups. In a retrospective analysis of 378 patients, it was observed that WBRT did not improve survival in patients achieving complete remission after HD-MTX.

These data seem to indicate that it is feasible to treat PCNSL using chemotherapy alone. Given the extremely high risk of treatment-related neurotoxicity, chemotherapy alone should be considered in patients over the age of 60 years. Future studies, in larger series, should validate the chemotherapy-alone strategy, as well as other strategies to dose-intensify chemotherapy and eliminate the need for WBRT (BBBD, APBSCT, and prolonged-maintenance MTX).