Version May 2024
INTRODUCTION — Primary central nervous system lymphoma (PCNSL) is a subtype of non-Hodgkin lymphoma (NHL) restricted to the brain, spinal cord, cerebrospinal fluid (CSF), and/or eyes. PCNSL is a rare but well-described central nervous system (CNS) malignancy with distinct treatment and prognostic implications compared with other brain tumors.
Historically, PCNSL was treated primarily with whole brain radiation therapy (WBRT), similar to gliomas and other primary brain tumors. Treatment has evolved over the years, however, and initial treatment now consists of polychemotherapy with high-dose methotrexate-based regimens, which are distinct from those used to treat systemic NHL [1-4].
This topic will review the treatment and prognosis of PCNSL. Related topics on clinical features and diagnosis of PCNSL as well as other forms of lymphoma involving the nervous system are presented separately.
●(See "HIV-related lymphomas: Primary central nervous system lymphoma".)
●(See "Secondary central nervous system lymphoma: Clinical features and diagnosis".)
●(See "Secondary central nervous system lymphoma: Treatment and prognosis".)
PRETREATMENT EVALUATION
Pathology review — All patients should have histopathologic confirmation of PCNSL by brain biopsy, cerebrospinal fluid (CSF) cytology, or vitrectomy. The diagnosis should not be made solely on the basis of imaging or glucocorticoid responsiveness. (See "Primary central nervous system lymphoma: Clinical features, diagnosis, and extent of disease evaluation", section on 'Diagnosis'.)
Approximately 90 percent of PCNSLs are CD20-positive diffuse large B cell lymphomas (DLBCLs). The remaining cases consist of other histologies including T cell lymphomas, low-grade lymphomas, and Burkitt lymphoma confined to the central nervous system (CNS). Provided the diagnosis of PCNSL is confirmed (as opposed to systemic lymphoma with secondary CNS involvement), all high-grade histologies are generally treated with the same approach, except that rituximab is omitted for CD20-negative and T cell lymphomas. (See "Primary central nervous system lymphoma: Clinical features, diagnosis, and extent of disease evaluation", section on 'Pathology and pathogenesis'.)
Extent of disease and staging — PCNSL can involve brain parenchyma, eyes, CSF, and/or spinal cord. Patients should undergo CNS and ocular staging as well as body imaging to exclude lymphoma outside the nervous system including in the testes, for males [5]. This evaluation is reviewed separately. (See "Primary central nervous system lymphoma: Clinical features, diagnosis, and extent of disease evaluation", section on 'Extent of disease evaluation'.)
PCNSL is stage IE (extranodal or involvement of one organ outside the lymph nodes) per the Ann Arbor staging system for non-Hodgkin lymphoma (NHL) (table 1). (See "Pretreatment evaluation and staging of non-Hodgkin lymphomas", section on 'Staging'.)
Laboratories and kidney function — Baseline pretreatment laboratory studies should include:
●Complete blood count with differential
●Serum chemistries, including electrolytes, kidney and liver function tests, calcium, phosphorus, and lactate dehydrogenase (LDH; for prognostic purposes) (see 'Prognostic assessment' below)
●Hepatitis B surface antigen and core antibody (because chemoimmunotherapy increases the risk of hepatitis B reactivation), hepatitis C testing, and a human immunodeficiency virus (HIV) test
Kidney function is a key pretreatment consideration in patients with PCNSL because of the primary role of high-dose methotrexate, which is eliminated by the kidneys. We typically calculate the estimated glomerular filtration rate (eGFR) based on serum creatinine using the 2021 chronic kidney disease epidemiology (CKD-EPI) creatinine equation (calculator 1). When available, cystatin C can be used to confirm the accuracy of creatinine-based estimations using the calculator available through the National Kidney Foundation (algorithm 1). A 24-hour urine collection is an alternative confirmatory method. (See "Assessment of kidney function", section on 'Estimation of GFR'.)
Prognostic assessment — Age and performance status are the strongest independent predictors of overall survival (OS) in patients with PCNSL [6-8]. Two prognostic scoring systems developed specifically for PCNSL are available [6,7].
●IELSG model – The International Extranodal Lymphoma Study Group (IELSG) model is based on five independent predictors of poor prognosis: age >60 years, Eastern Cooperative Oncology Group performance status >1 (table 2), elevated serum LDH, high CSF protein, and involvement of deep regions of the brain (periventricular regions, basal ganglia, brainstem, and/or cerebellum) [7]. Two-year OS rates based on these factors are estimated as follows:
•Low risk (zero to one factor) – 80 percent
•Intermediate risk (two to three factors) – 48 percent
•High risk (four to five factors) – 15 percent
●MSKCC model – The Memorial Sloan Kettering Cancer Center (MSKCC) model divides patients into three groups based on age and Karnofsky Performance Status (KPS) score (table 2) [6]. Median OS estimates by class are as follows:
•Class 1 (<50 years old) – 5.2 to 8.5 years
•Class 2 (≥50 years old with a KPS score ≥70) – 2.1 to 3.2 years
•Class 3 (≥50 years old with a KPS score <70) – 0.9 to 1.1 years
ACUTE SYMPTOM MANAGEMENT — Patients with PCNSL may present with acute neurologic manifestations of the disease, including altered mental status, weakness, headache, seizures, and deterioration of performance status. After diagnostic biopsy has been obtained, an initial course of glucocorticoids, when indicated, often results in significant improvement in symptoms and performance status prior to initiating definitive chemotherapy.
A typical regimen of glucocorticoids for moderate to severe symptoms due to central nervous system (CNS) disease is dexamethasone 8 to 16 mg daily, oral or intravenous (IV), in one or two divided doses [1,2]. Lower doses can be used for milder symptoms. Maximal symptomatic effects are typically seen within two to three days. Once lymphoma treatment is initiated, glucocorticoids should be tapered gradually to the lowest effective dose to minimize steroid toxicities. Most patients with responding disease are able to be weaned off steroids completely by the second or third cycle of induction chemotherapy. (See "Management of vasogenic edema in patients with primary and metastatic brain tumors".)
Pneumocystis pneumonia prophylaxis is particularly important in patients receiving glucocorticoids and concomitant chemotherapy. Due to overlapping effects on folate metabolism, we avoid using trimethoprim-sulfamethoxazole in patients receiving high-dose methotrexate. Alternative prophylactic regimens are reviewed separately. (See "Treatment and prevention of Pneumocystis pneumonia in patients without HIV", section on 'Prophylaxis'.)
INDUCTION CHEMOTHERAPY — Upfront treatment for PCNSL has evolved over the years and now generally includes two phases: induction to achieve remission and consolidation to prevent disease recurrence. Treatment should be initiated as soon as the diagnosis is confirmed; early treatment helps to reverse neurologic deficits and has been associated with improved survival [9].
The mainstay of treatment for PCNSL is high-dose methotrexate, which is combined with rituximab and, in patients with adequate fitness, at least one additional chemotherapeutic agent. Modifications to this standard of care are appropriate in patients with inadequate kidney function or functional status, some older adults, and patients with eye involvement.
Patients with adequate fitness and kidney function
Approach and rationale — For patients with adequate fitness and kidney function (eg, estimated glomerular filtration rate [eGFR] ≥30 mL/minute per 1.73 m2), we recommend a high-dose methotrexate-based combination regimen rather than high-dose methotrexate alone or primary radiation therapy. Patients with an eGFR <30 mL/minute per 1.73 m2 are at higher risk for toxicity even with appropriate dose reductions, and such patients should be managed individually in close collaboration with renal consultants.
The evidence base for induction therapy in PCNSL includes multiple single-arm trials and few randomized trials. In sum, studies indicate that a higher number of high-dose methotrexate cycles (eg, at least four to six doses with two more after maximum response) and the addition of chemotherapy to high-dose methotrexate lead to increased response rates [9-15]. In the singular phase 2 trial that included a randomized comparison between a methotrexate-based combination chemotherapy induction regimen (methotrexate plus cytarabine) and methotrexate alone, objective response (OR) rates were higher with combination therapy (69 versus 40 percent) and overall survival (OS) was nonsignificantly improved (hazard ratio [HR] 0.65, 95% CI 0.38-1.13), despite a higher number of toxic deaths in the combination arm [13].
A dose of methotrexate greater than 1 g/m2 is considered high dose, requiring continuous hydration and renal support. Methotrexate has been used in varying doses in clinical trials over the years. For the purposes of PCNSL treatment, any dose ≥3 g/m2 delivered intravenously (IV) achieves adequate cerebrospinal fluid (CSF) concentrations [16]. Specific doses ≥3 g/m2 vary based on the combination regimen, as reviewed below and in the table (table 3). Safe administration of high-dose methotrexate, including appropriate dose reductions for patients with altered kidney function, is reviewed in detail separately. (See "Therapeutic use and toxicity of high-dose methotrexate", section on 'Kidney function'.)
Rituximab has also been studied in combination regimens with mixed results, and its exact role remains to be determined. While we generally give rituximab as part of induction therapy in all patients with CD20-positive disease, others await definitive evidence of improved survival. We have a low threshold to omit rituximab in patients with increased risk for immunosuppression and its complications.
Prospective data addressing the value of rituximab directly are accumulating. In the only randomized, phase 3 trial to compare high-dose methotrexate-based chemotherapy with and without rituximab for newly diagnosed PCNSL, response rates, event-free survival, and OS were similar between groups with a median follow-up of 33 months, but mature survival data are not yet available [17]. In the randomized phase 2 International Extranodal Lymphoma Study Group-32 (IELSG32) trial with a median follow-up of 7.3 years, the group assigned to methotrexate, cytarabine, and rituximab had similar progression-free survival (PFS) but improved OS (37 versus 21 percent at seven years; HR 0.64, 95% CI 0.41-0.99) compared with the methotrexate plus cytarabine group [18].
Specific regimens — Few trials have compared different chemotherapy combination regimens, and institution and clinician preference and familiarity may guide selection. In cooperative group trials, the complete response (CR) and objective response (OR) rates for most regimens are similar.
Absent robust comparative data, we therefore suggest using one of the following regimens, which have been studied in single-arm or randomized phase 2 studies in patients with PCNSL (table 3):
●MTR – MTR as an induction regimen consists of high-dose methotrexate at 8 g/m2, temozolomide, and rituximab for four 28-day cycles [19].
MTR induction was initially studied in a multicenter phase 2 trial (Alliance 50202) in 44 patients with newly diagnosed PCNSL (median age 61 years, range 12 to 76 years); patients who achieved a CR with induction received consolidation with nonmyeloablative etoposide and cytarabine (EA) [9]. The CR and OR rates with MTR induction were 66 and 77 percent, respectively. Two-year PFS in the entire cohort was 57 percent, and the estimated four-year OS was 65 percent.
In a subsequent randomized phase 2 trial (Alliance 51101) in which patients received MTR followed by one cycle of high-dose cytarabine (MTR-A) for induction and were then randomly assigned to consolidation with either EA or high-dose chemotherapy with autologous stem cell transplantation (HDC-ASCT), the CR rate with MTR-A was 50 percent [19].
●R-MPV – R-MPV as an induction regimen consists of rituximab, high-dose methotrexate at 3.5 g/m2, procarbazine, and vincristine for five to seven 14-day cycles [11].
R-MPV induction followed by reduced-dose whole brain radiation therapy (WBRT) and cytarabine was initially studied in a single-center Memorial Sloan Kettering Cancer Center (MSKCC) trial in 30 patients with newly diagnosed PCNSL [11]. The CR rate after up to seven cycles of R-MPV was 78 percent, and the two-year PFS for the entire regimen (including WBRT and cytarabine) was 57 percent [11]. In two subsequent multicenter trials, OR rates following R-MPV ranged from 79 to 83 percent, and two-year PFS with R-MPV plus one cycle of high-dose cytarabine (R-MPV-A) was 54 percent [20,21].
Two other regimens are also evidence based but associated with increased risk for hematologic toxicity. Therefore, these regimens should only be considered in younger, fit adults.
●MATRix – The MATRix induction regimen consists of four 21-day cycles of high-dose methotrexate at 3.5 g/m2, cytarabine (dosed as 2 g/m2 every 12 hours for four doses), thiotepa, and rituximab [12,22].
In the IELSG32 randomized phase 2 trial of three different methotrexate-based regimens in 227 patients (18 to 70 years old) with newly diagnosed PCNSL, the CR and OR rates with MATRix were 49 and 87 percent, respectively [12]. With a median follow-up of 7.3 years, 7-year OS in the MATRix arm was 56 percent [18].
●R-MBVP – The R-MBVP induction regimen consists of rituximab, high-dose methotrexate at 3 g/m2, carmustine, teniposide (no longer available) or etoposide, and prednisolone for two 28-day cycles [23]. In most cases, this is followed by cytarabine 2 g/m2 every 12 hours for four doses (R-MBVP-A).
The R-MBVP-A regimen with variable consolidation strategies has been studied in cooperative group trials in France (PRECIS study) and the Netherlands/Australia/New Zealand (HOVON105 study), primarily in patients ≤60 years of age [17,23]. The CR and OR rates with this regimen ranged from 43 to 45 and 70 to 82 percent, respectively.
Patients with low performance status or organ dysfunction — Patients with PCNSL may have low performance status at the time of diagnosis for a variety of reasons, some but not all of which require modifications in the induction approach. Importantly, patients whose performance status is low primarily based on neurologic dysfunction from bulky central nervous system (CNS) disease should not be excluded from traditional induction regimens, as performance status can improve markedly with successful treatment of the lymphoma.
As discussed above, kidney dysfunction is not an absolute contraindication to methotrexate-based induction therapy, but safe administration requires close monitoring and nursing support, appropriate dose reductions, and contingency plans for kidney failure and prolonged high plasma methotrexate levels.
Patients with very low ejection fraction are poor candidates for high-dose methotrexate due to risk of pulmonary edema and volume overload. In addition, pleural effusions and ascites are "third spaces" where methotrexate can accumulate, causing delayed elimination. Most protocols require at least 2.5 to 3.5 liters/m2 of IV fluid hydration per day, starting 4 to 12 hours prior to methotrexate and continuing for two to four days afterward. We advise obtaining a baseline echocardiogram in patients with a history of heart failure. For those with reduced ejection fraction, a 24-hour fluid challenge is advised before committing to methotrexate-based therapy. (See "Therapeutic use and toxicity of high-dose methotrexate", section on 'Pretreatment assessment'.)
Treatment options are limited for patients who are not eligible for high-dose methotrexate induction therapy. For those without an absolute contraindication to methotrexate but who have multiple comorbidities or low functional status, single-agent methotrexate with or without rituximab may be an option, with a plan to increase treatment intensity if functional status improves with treatment. Temozolomide with or without rituximab can be considered in patients ineligible for high-dose methotrexate or cytarabine [24,25]. Radiation therapy is an option, particularly in patients with poor functional status for whom the goals of care are primarily palliative. (See 'Radiation therapy' below.)
Outside of clinical trials, we do not generally try single-agent novel therapies such as ibrutinib, based on a lack of single-agent data in this setting.
Older adults — Clinical trials and retrospective studies have demonstrated that patients ≥65 years of age are able to tolerate treatment with high-dose methotrexate, and it is kidney function and functional status rather than age alone that should guide treatment decisions [26-28]. In fact, in one retrospective study, patients ≥80 years of age were able to safely receive high-dose methotrexate-based therapy [29].
Most fit older adults are good candidates for a methotrexate-based combination regimen that includes an oral alkylating agent (eg, MTR or R-MPV). Similar regimens (without rituximab) were studied in a French intergroup phase 2 trial, in which 98 patients ≥60 years of age (median 72 years) were randomly assigned to receive either MPV-A or methotrexate plus temozolomide [30]. Both regimens were associated with moderate toxicities and no statistically significant differences in PFS or OS.
Cytarabine and other aggressive consolidative options are often not suitable for older patients due to increased risk of morbidity and mortality from treatment. In an individual patient data meta-analysis of 783 older adults (≥60 years) with PCNSL, OS was similar for patients treated with high-dose methotrexate and oral alkylating agents when compared with other aggressive combination therapies (HR 1.39, 95% CI 0.9-2.15) [26].
Patients with eye involvement — Eye involvement in PCNSL can occur in conjunction with brain and CSF involvement or can be isolated, as in primary vitreoretinal lymphoma (PVRL).
For patients with eye involvement in conjunction with brain lesions, we use the same approach to induction and consolidation therapy as for other patients with newly diagnosed PCNSL. Based on retrospective studies with multivariable analysis, eye involvement does not appear to be associated with worse outcomes [8,31]. (See 'Induction chemotherapy' above and 'Response assessment' below.)
For patients with PVRL, no standard of care has been established, and care patterns are variable [32]. Local therapies to treat eye disease include intravitreal methotrexate or rituximab, administered by ophthalmologists with experience treating PVRL, and ocular external-beam radiation therapy; high-dose methotrexate administered systemically also has good eye penetration [10,33,34]. Patients have a high risk of both ocular and CNS relapse after initial therapy, yet it is unclear whether any specific therapy is associated with risk reduction, including high-dose methotrexate-based chemotherapy, hematopoietic cell transplantation, or WBRT.
In the absence of better data, treatment decisions should be individualized and reviewed by a multidisciplinary team that includes neuro-oncology, ophthalmology, and radiation oncology. Our usual practice is to establish ocular disease control with local therapies and offer simultaneous or sequential high-dose methotrexate therapy to younger, fit adults who are expected to be at low risk for toxicity, in the hope of extending PFS. We do not typically offer consolidation with high-dose chemotherapy or WBRT.
Others have proposed a more intensive approach in young, fit patients that includes high-dose methotrexate-based combination chemotherapy supplemented by intravitreal therapy and followed by ocular radiation or HDC-ASCT [32].
Limited role of intrathecal therapy — Intrathecal methotrexate has been eliminated from most induction chemotherapy regimens for PCNSL, even for patients with evidence of CSF dissemination. This is largely based on research showing that IV methotrexate, at least at doses >3 g/m2, consistently achieves micromolar concentrations in the CSF [35-39]. In addition, placement of an Ommaya reservoir for repeated doses of intrathecal drug confers risk of infection and other complications [40,41].
RESPONSE ASSESSMENT — Patients with PCNSL should be examined regularly during induction therapy to assess clinical response to treatment. Patients with responsive disease typically show neurologic improvement within a cycle or two of methotrexate, sometimes within days of the first dose.
Brain magnetic resonance imaging (MRI) with contrast should be performed at regular intervals (eg, every two to four doses of methotrexate) during induction in order to guide therapy. Formal response criteria are available from the International PCNSL Collaborative Group (IPCG) [5]. Eye examination, spine MRI, and cerebrospinal fluid (CSF) cytology should be followed in patients with abnormalities at baseline.
Patients with responsive disease — We aim for a complete response (CR) before proceeding with consolidation therapy, up to a usual maximum of eight doses of high-dose methotrexate. For patients who achieve a CR within the first four doses, we typically provide one to two additional cycles (ie, two to four doses of methotrexate) based on renal and overall tolerance and then proceed with further consolidation, if eligible (see 'Consolidation' below). Others cap the number of induction cycles and proceed directly to consolidation in patients with at least a partial response (PR).
Experts vary in the approach to patients with a PR to induction methotrexate. For patients with clinical and radiographic improvement yet residual enhancing disease on MRI despite up to eight doses of high-dose methotrexate, we most often give a cycle of high-dose cytarabine (eg, 2 g/m2 every 12 hours for four doses). Patients with stable disease or better after cytarabine are considered for further consolidation. An alternative for transplant-eligible patients with a PR is to proceed directly to high-dose chemotherapy. (See 'Consolidation' below.)
Patients with primary refractory disease — Primary refractory disease includes patients with progression despite two or more doses of methotrexate at adequate doses to penetrate the central nervous system (CNS; ie, ≥1 g/m2 accounting for dose reductions, ideally ≥3.5 g/m2). We also consider patients with a best response of stable disease to be refractory, given the usual sensitivity of PCNSL to high-dose methotrexate.
The approach to primary refractory disease is not standardized. Patients are encouraged to consider clinical trials for novel therapies. Absent clinical trial options, we often use high-dose cytarabine by itself or in combination with etoposide as a second-line therapy in patients who fail high-dose methotrexate [42]. Responding patients are then considered for consolidation.
Patients who fail second-line therapy with cytarabine have a poor prognosis. Some patients can be stabilized with focal radiation therapy as a bridge to further salvage therapies. Our experience with single-agent ibrutinib in this setting is that responses are short lived. (See 'Relapsed/refractory disease' below.)
CONSOLIDATION — High-dose chemotherapy with autologous stem cell transplantation (HDC-ASCT), nonmyeloablative chemotherapy alone, and reduced-dose whole brain radiation therapy (WBRT) are the three options for consolidation in patients who have responded to induction. All three are supported by prospective and in some cases randomized trials; however, few comparative trials have been performed, toxicity profiles vary considerably, and many older adults are not good candidates for any of the three.
Our approach — Our approach to selection of consolidation is as follows:
●For young, fit patients who achieve a complete response (CR) or a good partial response (PR) with induction therapy, we offer thiotepa-based HDC-ASCT. We typically refer patients for bone marrow transplantation evaluation early in induction therapy in order to minimize delays between induction and consolidation. If a delay of longer than three to four weeks is anticipated, we may give a cycle of high-dose cytarabine as a bridge. (See 'Hematopoietic cell transplantation' below.)
●For patients with responding disease who are not eligible for HDC-ASCT on the basis of advanced age or medical comorbidities, we consider nonmyeloablative chemotherapy (eg, high-dose cytarabine or etoposide plus cytarabine [EA]). Hematologic toxicity is high even with these nonmyeloablative regimens, however, and risks and benefits must be weighed in each patient individually. We rarely offer cytarabine or EA in patients older than 75 years of age. (See 'Nonmyeloablative chemotherapy' below.)
●Reduced-dosed WBRT is reserved for rare patients who have responsive disease and preserved functional status but who are not candidates for high-dose chemotherapy on the basis of organ dysfunction or patient preference. The risk of delayed neurotoxicity rises with advancing age, and we limit use of consolidative WBRT to patients younger than 65 years. (See 'Whole brain radiation therapy' below.)
Hematopoietic cell transplantation — HDC-ASCT is a feasible and effective consolidation strategy in select young, fit patients with chemotherapy-responsive PCNSL.
The optimal conditioning regimen has not been established, but the available data support use of thiotepa-based regimens such as TBC (thiotepa, busulfan, and cyclophosphamide) or TT-BCNU (thiotepa and carmustine) rather than conditioning regimens more commonly used in patients with systemic diffuse large B cell lymphoma (DLBCL), such as BEAM (carmustine, etoposide, cytarabine, and melphalan) or CBV (cyclophosphamide, etoposide, and carmustine) [43-45]. A retrospective review of data from the Center for International Blood and Marrow Transplant Research registry comparing TBC, TT-BCNU, and BEAM conditioning regimens concluded that thiotepa-based regimens were associated with improved survival rates [46]. Compared with TT-BCNU, TBC was associated with improved survival but higher rates of toxicity and nonrelapse mortality.
Compared with WBRT, HDC-ASCT is similarly effective and is associated with lower risk of neurocognitive decline. Although early treatment-associated mortality is higher with HDC-ASCT, late treatment-associated mortality appears to be higher with WBRT due to neurocognitive and vascular complications. Two randomized trials have shown similar results:
●IELSG32 trial – In the two-stage phase 2 International Extranodal Lymphoma Study Group-32 (IELSG32) trial, 118 patients with responsive or stable disease after induction therapy (out of 227 patients enrolled overall, age 18 to 70 years) were randomly assigned to consolidation with either HDC-ASCT or WBRT [22]. With median follow-up of 30 months, two-year progression-free survival (PFS) was similar in both groups (69 and 88 percent, respectively; hazard ratio [HR] 1.50, 95% CI 0.83-2.71). There were two toxic deaths due to infection, both in the HDC-ASCT arm. With long-term follow-up (median 88 months), outcomes remained comparable, including seven-year PFS (50 and 55 percent for HDC-ASCT and WBRT, respectively) and seven-year overall survival (OS; 57 and 63 percent, respectively) [18].
Neurocognitive function and quality of life were measured prospectively in both groups at varying times of follow-up and compared with baseline testing. With both short-term and long-term follow-up, objective cognitive decline was observed in patients who had received WBRT (primarily attention and executive function), whereas improvement was observed in the HDC-ASCT group [18,22].
●PRECIS trial – In the randomized phase 2 PRECIS trial, 140 patients (≤60 years of age) with newly diagnosed PCNSL were enrolled and randomly assigned to HDC-ASCT or WBRT (40 Gy) consolidation after high-dose methotrexate and cytarabine-based induction chemotherapy [23]. With a median follow-up of approximately 30 months, two-year event-free survival (EFS) was nonsignificantly greater in the HDC-ASCT arm (70 percent; 95% CI 59-82) compared with the WBRT arm (58 percent; 95% CI 47-71), and two-year OS was similar (66 and 75 percent for HDC-ASCT and WBRT, respectively). There were more early deaths due to treatment toxicity in the HDC-ASCT arm (five versus one).
With long-term follow-up in the randomized population (median 8.2 years), OS remained similar between the arms (eight-year OS 54 versus 58 percent for HDC-ASCT and WBRT, respectively) [47]. A disproportionate number of patients in the HDC-ASCT arm (13 versus 5 in the WBRT arm) had progressive disease during induction and were not eligible to receive the assigned consolidation. Among the remaining 44 patients who completed HDC-ASCT per protocol, eight-year EFS was 67 percent. This compared favorably with patients who completed WBRT consolidation per protocol (eight-year EFS 39 percent).
Neurocognitive function was stable or improved in the HDC-ASCT arm at 30 months but had worsened in the WBRT arm (primarily attention and executive function) [23]. With longer-term follow up, higher rates of deterioration were observed in the WBRT arm for both neurocognition (64 versus 13 percent) and balance (52 versus 10 percent), and there were four late deaths related to WBRT toxicity [47].
Further studies are needed to define the long-term risk of neurocognitive decline after HDC-ASCT as well as after reduced-dose WBRT. The induction regimens and doses of WBRT were different in the completed trials, making comparison difficult. In at least one longitudinal study that followed patients with PCNSL for up to five years after consolidation with either HDC-ASCT or reduced-dose WBRT, comprehensive neurocognitive evaluations showed improvement in the first three years followed by a decline with time, regardless of the consolidative modality used [48].
Nonmyeloablative chemotherapy — Nonmyeloablative chemotherapy (eg, high-dose cytarabine alone or EA) is a slightly less intensive alternative to HDC-ASCT in patients with responding disease who are not candidates for transplantation due to age or comorbidities. We generally consider this option only in patients ≤75 years of age who are otherwise fit.
Supportive evidence for consolidation with EA includes results of two Alliance cooperative group trials in the United States. In the first trial (CALGB 50202), 44 patients (median age 61 years, range 12 to 76 years) with newly diagnosed PCNSL were treated with MTR induction (high-dose methotrexate at 8 g/m2, rituximab, and temozolomide); patients who achieved a CR (n = 29) then received consolidative EA (etoposide 5 mg/kg every 12 hours for eight doses and cytarabine 2 g/m2 every 12 hours for eight doses) [9]. The two-year PFS was 57 percent and median OS was not reached. There was one treatment-related death.
This study was followed by a randomized phase 2 trial (CALGB 51101) of consolidative EA versus HDC-ASCT in 113 patients with newly diagnosed PCNSL (≤75 years of age); all patients received induction MTR-A (MTR followed by one cycle of cytarabine) [19]. In preliminary results presented at the 2021 American Society of Clinical Oncology annual meeting, two-year PFS was superior in the ASCT arm, although there were a statistically higher number of patients with primary refractory disease on the EA arm during induction, prior to consolidation. When comparing two-year PFS from the start of consolidation, there was trend towards improvement with ASCT (HR 0.58, 95% CI 0.25-1.36). The estimates of three-year survival were similar between groups (83 and 72 percent for HDC-ASCT and EA, respectively).
Importantly, hematologic toxicity with EA is substantial, and patients who are not felt to be candidates for HDC-ASCT consolidation due to age or comorbidities are also at high risk for toxicity with EA. We use cytarabine alone in such patients at a usual dose of 2 g/m2 every 12 hours for four doses or 3 g/m2 every 24 hours for two doses.
Whole brain radiation therapy — WBRT as a consolidation therapy for newly diagnosed PCNSL is controversial, and the risk of delayed neurotoxicity limits its acceptance as a standard approach. However, ongoing studies are examining whether reduced-dose WBRT (eg, 23.4 Gy) may offer benefits in terms of disease control while reducing the risk of neurotoxicity, and existing data support this approach in selected patients who are not candidates for chemotherapy-based consolidation [20,21,23,49].
The feasibility of reduced-dose WBRT after induction chemotherapy was initially demonstrated in a single-arm study in which patients in CR after R-MPV induction received consolidation with reduced-dose WBRT (23.4 Gy) and one cycle of cytarabine; those achieving less than a CR received standard-dose WBRT (45 Gy) and cytarabine [20]. In the 31 patients who received reduced-dose WBRT, two-year PFS and three-year OS were 77 and 87 percent, respectively, and cognitive assessment scores were relatively stable over time. A subsequent randomized phase 2 trial of R-MPV-A with or without reduced-dose WBRT (23.4 Gy) was completed by the NRG cooperative group, and preliminary results demonstrated improved two-year PFS on the low-dose WBRT arm as well as early evidence of preserved neurocognition [21]. The data for long-term neurotoxicity based on comprehensive neuropsychiatric testing and OS are yet to mature. Additional outcomes data on reduced-dose WBRT as consolidation in the PRECIS trial are reviewed above. (See 'Hematopoietic cell transplantation' above.)
The largest study of consolidative WBRT used a relatively high dose (45 Gy delivered in 30 daily fractions of 1.5 Gy) and was designed to test whether WBRT could be omitted as consolidation after high-dose methotrexate therapy without compromising OS [50]. The trial enrolled 551 patients with newly diagnosed PCNSL at 75 centers in Germany. Patients were randomly assigned to receive methotrexate-based induction therapy followed by WBRT or methotrexate-based induction therapy followed by response-based consolidative chemotherapy (ie, patients with less than a CR after induction received high-dose cytarabine, and patients in CR received no further therapy). With a median follow-up of approximately 50 months in the intent-to-treat population, median PFS was longer in the WBRT arm (HR 0.79, 95% CI 0.63-0.99), but median OS was similar between groups (HR 1.01, 95% CI 0.79-1.30). Results were very similar in an analysis restricted to the 318 patients who received per-protocol therapy. The risk of clinically defined neurotoxicity among 80 evaluable patients was twofold higher in the WBRT group (49 versus 26 percent).
Other studies also confirmed that the combination of high-dose methotrexate and WBRT can lead to progressive and irreversible neurotoxicity [51,52]. This risk is particularly high in patients over 60 years of age. It typically manifests as a decline in psychomotor speed, attention, executive function, and memory. White matter abnormalities and atrophy have been noted on MRI, but their correlation with neurocognitive status has been inconsistent [51,53].
Patients who are not candidates for consolidation — An effective and safe consolidation strategy for older adults and those with low performance status or high risk for neurotoxicity has not been established, and treatment decisions are individualized. It is not known whether lower-intensity interventions offer any advantage over observation, and some CRs are durable for many months without any further treatment.
Outside of a clinical trial, a watch-and-wait strategy may be appropriate for many of these patients, particularly those who are in CR at the completion of induction. For patients who have responded to and tolerated high-dose methotrexate, some experts suggest continuing high-dose methotrexate monthly for up to one year [54]. Oral options, such as temozolomide, procarbazine, or ibrutinib monotherapy, have also been proposed, but our experience anecdotally has not been positive.
Maintenance therapy with temozolomide, procarbazine, or ibrutinib has been studied in several single-arm trials and has been shown to be feasible and tolerable in select patients, including adults ≥65 years of age [55-58]. However, adequately powered randomized trials have not yet been performed [59]. In a phase 3 trial conducted by the Japan Clinical Oncology Group, 122 patients were randomly assigned to receive high-dose methotrexate induction chemotherapy followed by either WBRT or WBRT with concurrent plus adjuvant temozolomide [60]. At the time of a planned interim analysis, two-year OS was nonsignificantly worse in the temozolomide maintenance arm (71 versus 87 percent; HR 2.18, 95% CI 0.95-4.98), and the trial was terminated early for futility.
There are ongoing studies using novel agents such as immunomodulatory drugs and checkpoint inhibitors as maintenance.
MONITORING — Regular clinical follow-up and brain imaging are required in all patients with PCNSL in order to detect relapse as early as possible. This is especially important in PCNSL, as patients with relapsed disease can have progressive neurologic deficits, which can be challenging to reverse.
We follow the International PCNSL Collaborative Group (IPCG) guidelines, which recommend a brain MRI with contrast at least every three months for the first two years after completion of therapy, then every six months for three years, and then yearly for five years, for a total of 10 years [5] or indefinitely [54], depending on patient and clinician preference. Late relapses have been seen beyond 10 years but are rare [61]. In a study of 256 patients with relapsed/refractory PCNSL, relapse was asymptomatic in 25 percent of patients and was identified on serial surveillance imaging, highlighting the importance of surveillance imaging [62].
Long-term survivors of PCNSL may develop worsening neurocognitive function years after completing treatment. The IPCG recommends an assessment of neurocognitive function and quality of life in survivors [51]. We monitor cognition and quality of life by history and targeted examination at each follow-up visit. Formal neuropsychological testing is useful in patients reporting a decline in function to guide speech and cognitive therapy and allow for objective comparisons over time.
RELAPSED/REFRACTORY DISEASE — Approximately 25 percent of patients with PCNSL do not respond to initial high-dose methotrexate-based therapy, and 50 percent experience a relapse. The prognosis of relapsed/refractory PCNSL is poor, and there is no standard-of-care guidance or consensus due to the lack of randomized studies.
Treatment selection is individualized based on a variety of factors, including patient age, performance status, neurologic status, comorbidities, site of relapse, prior treatment, and duration of prior response. Participation in clinical trials is encouraged.
Prior response to methotrexate, prolonged remission — For patients who previously responded to methotrexate and have had a prolonged period of remission (eg, 12 to 24 months or more), we suggest a rechallenge with high-dose methotrexate, provided they remain eligible for such therapy. We typically give high-dose methotrexate plus rituximab, but other methotrexate-based regimens are also reasonable.
Based on retrospective studies of patients with PCNSL at first relapse, rechallenge with high-dose methotrexate is associated with a response rate as high as 80 to 90 percent and median overall survival (OS) of 40 to 60 months [63-65].
Fit patients with responsive disease who have not previously undergone high-dose chemotherapy with autologous stem cell transplantation (HDC-ASCT) can be considered for this consolidative approach, based on the experience in patients with newly diagnosed PCNSL [44,45] and in patients with methotrexate-refractory disease [42,66]. (See 'Candidates for high-dose chemotherapy' below.)
Methotrexate-refractory disease
Candidates for high-dose chemotherapy — Select fit patients with methotrexate-refractory disease who have not previously undergone thiotepa-based HDC-ASCT may be eligible for this approach at first relapse [42,66].
In a prospective phase 2 trial conducted by a French cooperative group, 43 patients with relapsed/refractory PCNSL (median age 52 years, range 23 to 65 years) were treated with induction therapy with etoposide and cytarabine (EA) followed by conditioning with thiotepa, busulfan, and cyclophosphamide (TBC) and ASCT [42]. In the overall cohort with a median follow-up of three years, median progression-free survival (PFS) and OS were 11.6 and 18.3 months, respectively, and there were three treatment-related deaths during induction therapy. Among 27 patients who completed ASCT, median PFS and OS were 41 and 59 months, respectively.
A separate trial by a German cooperative group enrolled 39 patients ≤65 years of age (median 57 years) with relapsed/refractory PCNSL who had failed high-dose methotrexate [66]. Induction consisted of two courses of rituximab, high-dose cytarabine, and thiotepa with collection of stem cells in between. Thirty-two patients completed ASCT, and there were four treatment-related deaths. With a median follow-up of 45 months in the entire cohort, median PFS was 12.4 months, and median OS was not yet reached (two-year OS 56.4 percent).
All other patients — For all other patients with relapsed/refractory PCNSL, treatment options are limited, and durable responses are difficult to achieve. Available options include single-agent or combination treatment with a variety of systemic therapies, radiation therapy, and clinical trials of novel agents.
Systemic therapies
●Chemotherapy – Prospective trials of chemotherapy agents like pemetrexed, topotecan, rituximab, and temozolomide have yielded reasonable response rates but poor PFS, ranging from two to six months [24,67-71].
●Bruton tyrosine kinase (BTK) inhibitors – Ibrutinib and tirabrutinib are two BTK inhibitors that have been studied in phase 1 and 2 trials as monotherapy in relapsed/refractory PCNSL with dramatic responses but poor durability.
•In a phase 1 dose escalation study (doses of 560 and 840 mg) in patients with relapsed/refractory PCNSL or secondary central nervous system (CNS) lymphoma, the objective response (OR) rate was 75 percent but median PFS was only 4.6 months [72]. In a multicenter French cooperative group phase 2 study, 52 patients with PCNSL (including primary vitreoretinal lymphoma [PVRL]) were treated with ibrutinib at a dose of 560 mg daily [73]. The OR rate at two months was 52 percent and the median PFS was 4.8 months in the entire cohort; excluding patients with PVRL, median PFS was only two months.
•Tirabrutinib is a second-generation, selective, irreversible BTK inhibitor that has been studied in a phase 1/2 trial in Japan with an OR rate of 63.4 percent and median PFS of 2.9 months [74]. Tirabrutinib received approval for treatment of relapsed/refractory PCNSL in March 2020 in Japan [75]. A phase 2 trial is ongoing in the United States.
Ibrutinib has been studied in combination with chemotherapy in phase 1 studies. In a phase 1 study of ibrutinib in combination with temozolomide, etoposide, liposomal doxorubicin, dexamethasone, rituximab (TEDDI-R), and intrathecal cytarabine in patients with newly diagnosed (n = 5) and relapsed/refractory (n = 13) PCNSL, the OR rate was 93 percent (mostly complete responses [CRs]) and median PFS was 15.5 months in 14 evaluable patients [76]. In another phase 1 study of ibrutinib with rituximab and high-dose methotrexate in relapsed/refractory PCNSL or secondary CNS lymphoma, the OR rate was 80 percent, with a median PFS of 9.2 months [63].
Of note, there is a risk of aspergillosis with ibrutinib, although the overall incidence in most studies has been <5 percent. An effort must be made to lower glucocorticoid doses while using ibrutinib or other BTK inhibitors. It is unclear if there is any added benefit of using antifungal agents concurrently because the incidence is low and there is a concern for drug-drug interaction.
●Immunomodulatory drugs – Lenalidomide has been studied in combination with rituximab in phase 1 and 2 studies in relapsed/refractory PCNSL, yielding an OR rate of 64 to 67 percent and median PFS of six to eight months [77,78]. Lenalidomide maintenance has also been noted to improve PFS when used in the salvage setting [78].
A third-generation agent, pomalidomide, was studied in combination with dexamethasone in a phase 1 dose escalation study in 29 patients, with an OR rate of 48 percent and median PFS of 5.3 months [79].
Ibrutinib and lenalidomide are being investigated in combination in the relapsed/refractory setting, as well as with induction chemotherapy in the upfront setting. Given that the durability of the agents is poor as monotherapy and combination therapy clinical trials exclude prior therapy with similar agents, we typically do not suggest using these targeted therapies outside of trials, unless no other options are available.
●CAR-T cell therapy – Chimeric antigen receptor T (CAR-T) cell therapies targeting CD19 are of interest in PCNSL and are being studied in clinical trials; however, further studies are needed to define their role.
A phase 1/2 study of tisagenlecleucel in 12 patients with recurrent PCNSL demonstrated safety, with grade 3 immune effector cell-associated neurotoxicity syndrome (ICANS) in only one patient and no treatment-related deaths [80]. The OR rate was 58 percent, although all patients had received prior bridging therapy. There is an ongoing study of axicabtagene ciloleucel in patients with relapsed/refractory PCNSL (NCT04608487).
Radiation therapy — Like lymphoma in other areas of the body, PCNSL lesions tend to be radiosensitive in the short term, and thus radiation can be an effective palliative or bridging strategy in certain patients. PCNSL is a multifocal disease at a microscopic level, and coverage of the entire brain is necessary when radiation is used as a definitive or standalone treatment.
When no appropriate systemic options are available or feasible, as in the setting of rapidly declining neurologic function, whole brain radiation therapy (WBRT) is a reasonable salvage option, with the caveat that it could lead to further worsening of neurocognitive deficits. In retrospective studies, WBRT has been associated with a radiographic response in 74 to 79 percent of patients with relapsed/refractory PCNSL and median OS of 10 to 16 months [81,82].
Focal radiation (delivered conventionally or stereotactically) can be considered as a temporizing measure in select patients with a single symptomatic PCNSL lesion, provided there is a plan to follow with additional systemic therapy for microscopic disease.
PROGNOSIS — PCNSL responds well to chemotherapy, and select patients can achieve long-term (>10 year) durability and survival [83]. Tumor regression is seen in up to 70 to 80 percent of patients with induction high-dose methotrexate-based therapy, but complete response (CR) rates are lower at approximately 50 percent. In addition to the approximately 25 percent of patients with primary refractory disease, recurrences after initial response are also quite common and mostly occur within the central nervous system (CNS). Age and performance status are independently associated with prognosis. (See 'Prognostic assessment' above.)
In population studies, the median overall survival (OS) of patients with PCNSL remains poor (less than two years), particularly in patients with advanced age. Based on results from two national databases, the Surveillance, Epidemiology, and End Results (SEER) and Central Brain Tumor Registry of the United States (CBTRUS), the median OS doubled from 12 months in the 1970s to 26 months in the 2010s [84]. However, the median OS in older patients remains at seven months despite optimization of treatments.
Prognosis of relapsed/refractory disease is worse, with a median OS of two months without treatment and seven months with treatment [8,84].
SUMMARY AND RECOMMENDATIONS
●Pretreatment evaluation – Patients with newly diagnosed primary central nervous system lymphoma (PCNSL) require an extent of disease evaluation as well as pretreatment laboratories and functional assessment to determine fitness for high-dose methotrexate-based chemotherapy. (See 'Pretreatment evaluation' above.)
Age and performance status (table 2) are the strongest independent predictors of overall survival (OS) in patients with PCNSL. Two prognostic scoring systems developed specifically for PCNSL are available to assist in counseling and treatment planning. (See 'Prognostic assessment' above and 'Prognosis' above.)
●Induction chemotherapy – For patients with adequate fitness and kidney function, we recommend a high-dose methotrexate-based combination regimen rather than high-dose methotrexate alone or primary radiation therapy (Grade 1B). We suggest including rituximab in induction regimens (Grade 2C), although evidence is mixed. (See 'Approach and rationale' above.)
Few trials have compared different chemotherapy combination regimens (table 3); objective response (OR) rates for most regimens are similar. For most patients, we suggest either MTR (methotrexate at 8 g/m2, temozolomide, and rituximab) or R-MPV (rituximab, methotrexate at 3.5 g/m2, procarbazine, and vincristine) (Grade 2C). The MATRix regimen is also reasonable to consider in younger, fit adults. (See 'Specific regimens' above.)
Treatment options are limited for patients who are not eligible for high-dose methotrexate induction therapy due to low performance status or organ dysfunction. Importantly, patients ≥65 years of age are able to tolerate treatment with high-dose methotrexate, and it is kidney function and functional status rather than age alone that should guide treatment decisions. (See 'Patients with low performance status or organ dysfunction' above and 'Older adults' above.)
●Response assessment – Brain MRI with contrast should be performed at regular intervals (eg, every two to four doses of methotrexate) during induction. We aim for a complete response (CR) before proceeding with consolidation therapy, up to a usual maximum of eight doses of high-dose methotrexate. (See 'Response assessment' above.)
Experts vary in the approach to patients with a partial response (PR) to induction methotrexate. For patients with clinical and radiographic improvement yet residual enhancing disease on MRI despite up to eight doses of high-dose methotrexate, we most often give a cycle of high-dose cytarabine and, for patients with stable disease or better after cytarabine, proceed with further consolidation. An alternative for transplant-eligible patients with a PR is to proceed directly to high-dose chemotherapy. (See 'Patients with responsive disease' above.)
Patients with progression during induction or a best response of stable disease are considered to have primary refractory disease and have a poor prognosis. (See 'Patients with primary refractory disease' above.)
●Consolidation in patients with responsive disease – The three options for consolidation in patients who have responded to induction are high-dose chemotherapy with autologous stem cell transplantation (HDC-ASCT), nonmyeloablative chemotherapy alone, and reduced-dose whole brain radiation therapy (WBRT). Few comparative trials have been performed, toxicity profiles vary considerably, and many older adults are not good candidates for any of the three. We engage patients in shared decision-making and encourage participation in clinical trials. (See 'Our approach' above.)
•For most young, fit patients, we offer thiotepa-based HDC-ASCT. Patients should be referred for bone marrow transplantation evaluation early in induction therapy in order to minimize delays between induction and consolidation. (See 'Hematopoietic cell transplantation' above.)
•For most patients up to 75 years of age who are not eligible for or interested in HDC-ASCT, we suggest consolidation with nonmyeloablative chemotherapy (cytarabine with or without etoposide) rather than WBRT or no consolidation (Grade 2C). (See 'Nonmyeloablative chemotherapy' above and 'Whole brain radiation therapy' above.)
•A watch-and-wait strategy is appropriate for many patients who are not candidates for consolidation, particularly those who are in CR at the completion of induction. The value of continuing high-dose methotrexate or temozolomide monthly as maintenance therapy is uncertain. (See 'Patients who are not candidates for consolidation' above.)
●Relapsed/refractory disease – Treatment of relapsed/refractory PCNSL is individualized, and prognosis is poor. Treatment selection is individualized based on a variety of factors, including patient age, performance status, neurologic status, comorbidities, site of relapse, prior treatment, and duration of prior response.
•Rechallenge with high-dose methotrexate is appropriate in patients who previously responded to therapy and have had a prolonged period of remission (eg, 12 to 24 months or more), and highly select patients may be candidates for thiotepa-based HDC-ASCT as salvage therapy. (See 'Prior response to methotrexate, prolonged remission' above and 'Candidates for high-dose chemotherapy' above.)
•For all other patients, options include systemic therapies such as ibrutinib or lenalidomide, radiation therapy, and participation in clinical trials with agents such as chimeric antigen receptor T (CAR-T) cells. (See 'All other patients' above.)
ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges the seminal contributions of Fred H Hochberg, MD [1-4,10,28,34,36-39,64,71,82], who contributed to earlier versions of this topic review.
آیا می خواهید مدیلیب را به صفحه اصلی خود اضافه کنید؟
