Primary central nervous system lymphoma

What every physician needs to know:

Primary central nervous system lymphoma (PCNSL) is a rare variant of non-Hodgkin lymphoma (NHL) representing 1 to 2% of all cases of NHL. PCNSL is defined as lymphoma which presents in the brain, spinal cord, leptomeninges, or within the orbits, and which is restricted entirely to the central nervous system (CNS). Among primary brain tumors, primary CNS lymphoma presents unique prognostic and treatment considerations, and the therapeutic approaches used in PCNSL are distinct from systemic NHL.

Over the past 3 decades, the incidence of PCNSL has increased in parallel with a rise in the number of cases of systemic NHL. PCNSL constitutes 3.1% of all primary brain tumors in the United States of America. At present, this rise cannot be completely explained by the HIV epidemic and does not appear to be a consequence of advances in diagnostic techniques. Among immunocompetent patients with PCNSL, there is a slight male predominance (3:2 ratio of males to females), while in the AIDS population, 90% are men.

Greater than 90% of PCNSL cases in the immunocompetent patient population exhibit the pathologic characteristics of diffuse large B-cell lymphoma (DLBCL); the remainder includes other subtypes of B-cell lymphoma such as lymphoblastic B-cell lymphoma, Burkitt and Burkitt-like lymphoma, as well as T-cell lymphomas. Both B-and T-cell variants are histologically identical to their systemic counterparts, however PCNSL, as an extranodal lymphoma of stage 1E, has a worse prognosis than other localized extranodal lymphomas.

Continue Reading

Are you sure your patient has primary central nervous system lymphoma? What should you expect to find?

While PCNSL can affect all ages, the mean age of affected patients is 55 years for immunocompetent patients. For immunocompromised PCNSL patients, the mean age of diagnosis is 30 years. After thorough evaluation, the absence of systemic involvement confirms the primary nature of this disease. If lymphoma is found concomitantly outside the CNS, the diagnosis is not PCNSL, but rather stage IV NHL with CNS involvement.


PCNSL may disrupt virtually any aspect of normal CNS function, depending on the location(s) of the tumor(s). Greater than 50% of patients present with focal neurologic deficits such as hemiparesis or aphasia, 35% exhibit alteration in cognition, behavior, or personality, 11% present with seizure, and 32% present with headache, nausea, vomiting and papilledema as a manifestation of increased intracranial pressure.

PCNSL most commonly involves the cerebral hemispheres, followed by the basal ganglia, corpus callosum, and cerebellum. PCNSL less commonly arises in deep brain structures such as the brainstem and spinal cord. PCNSL tumors are typically in contact with cerebrospinal fluid spaces. In 20 to 40% of immunocompetent patients, multiple lesions are observed at initial diagnosis, while nearly 100% of AIDS patients present with multifocal lesions.

Approximately 10 to 20% of patients present with intraocular lymphoma (IOL) manifested by ocular floaters and blurry vision, and clumped cellular infiltrates are detected on slit lamp examination. It is often difficult to distinguish IOL from vitreitis/uveitis. Fewer than 10% of patients exhibit spinal dissemination marked by cauda equina syndrome, back pain, or radiculopathy. It is useful to obtain information relating to truncal anesthesias, bladder problems, or non-specific pain, in order to approximately evaluate these symptoms.

Beware of other conditions that can mimic primary central nervous system lymphoma:

The differential diagnosis includes glioma, CNS metastasis of other malignancies, vasculitis, sarcoid, toxoplasmosis or other infection, and multiple sclerosis.

Which individuals are most at risk for developing primary central nervous system lymphoma:

Risk factors for the development of PCNSL are similar to those associated with other types of NHL, namely chronic inflammation.

Profound immunodeficiency associated with AIDS, the Wiskott-Aldrich syndrome, immunoglobulin-A deficiency syndrome, and severe combined immunodeficiency (SCID), or even more subtle immunodeficiency states such as those associated with rheumatoid arthritis and systemic lupus erythematosus, clearly represent predisposing factors to the development of PCNSL.

In AIDS, PCNSL tends to be a manifestation of advanced disease and is associated with a CD4 count of less than 50 cells/mm3.

What laboratory studies should you order to help make the diagnosis and how should you interpret the results?

Diagnosis and clinical evaluation

Gadolinium-enhanced magnetic resonance imaging (MRI) is the standard imaging modality used to facilitate diagnostic work-up of PCNSL. Usually PCNSL is hypointense on T-1-weighted images, and hypo to isointense on T2-imaging, with intense homogeneous contrast enhancement. The majority of radiographic lesions are supratentorial and exhibit periventricular localization.

There is a significant tendency for PCNSL tumors at relapse to progress within the leptomeninges. This manifestation of biologic invasiveness is a significant factor in terms of therapeutic management, as PCNSL is virtually always refractory to surgical resection or focal radiotherapy. Due to the high intrinsic metabolic activity of the brain, PET imaging is rarely used to facilitate diagnosis of PCNSL.

In patients who present with only neurologic signs and symptoms (and a pathologic diagnosis of NHL is established by detection of tumor within the CNS), the diagnosis is PCNSL in 90% of cases. Nevertheless, a thorough staging work-up is necessary to definitively exclude systemic disease and to define the extent of CNS involvement. This process requires careful evaluation of distinct compartments within the CNS axis. An MRI with contrast is needed to evaluate brain and spine; an ophthalmologic evaluation including slit-lamp exam is mandated to evaluate the vitreous.

Lumbar puncture is appropriate unless contraindicated because of increased intracranial pressure, not only to facilitate diagnosis, but also to assess for evidence of leptomeningeal involvement. MRI of the total spine with gadolinium is necessary in patients with suspicion for leptomeningeal dissemination.

In addition, staging bone marrow biopsy as well as chest, abdomen and pelvic computed tomography (CT) scan are standard. Because approximately 30% of testicular lymphomas metastasize to the brain, a testicular examination is indicated at presentation of CNS lymphomas in male patients.

HIV antibody testing is indicated in all patients with suspicion for CNS lymphoma, as HIV-associated lymphoma has different implications in terms of therapy and prognosis. Serum lactate dehydrogenase (LDH) measurement informs prognosis. Ten percent of patients who present with a tentative diagnosis of PCNSL will ultimately be found to have occult systemic NHL on rigorous systemic staging evaluation, which includes bone marrow biopsy. In addition, there is recent preliminary evidence that the identification of clonally rearranged immunoglobulin-heavy chain (IgH) genes, identified by polymerase chain reaction, may be useful as biomarkers for the identification of subclinical disease in bone marrow.

Establishing a diagnosis of PCNSL may be a significant challenge in the patient who presents with neurologic symptoms and suspicious findings on radiographic imaging. The differential diagnosis may be broad, given the non-specific findings on neuroimaging and often includes glioma, metastatic malignancy, vasculitis, sarcoid, infection, or multiple sclerosis. There is often significant diagnostic latency in the presentation of PCNSL – in particular in immunocompetent patients in whom the disease may present with a smoldering course.

A critical problem in the initial diagnostic work-up of patients with PCNSL is the common practice of administering glucocorticoids as an early intervention to attenuate or reverse progressive neurologic decline at first presentation, before diagnosis. Because glucocorticoids usually induce tumor regression, likely by induction of apoptosis, administration of steroids such as dexamethasone increases the risk for a nondiagnostic specimen at initial work-up. The rate of complete response to single agent glucocorticoids may be as high as 20%. The ideal approach is to withhold corticosteroids prior to biopsy, unless there is active clinical decompensation.

Stereotactic biopsy is the preferred method for tissue diagnosis in those suspected of having PCNSL. There is up to a 4% risk of intracranial hemorrhage associated with stereotactic biopsy. In addition, stereotactic biopsy is associated with 9% rate of diagnostic failure.

In subsets of patients who present with signs and symptoms of intraocular lymphoma, pathologic diagnosis of lymphoma may be obtained by vitreal biopsy after detection of a suspicious intraocular infiltrate by slit lamp fluorescein angiography or ocular ultrasound. Another means of establishing tissue diagnosis is by evaluation of the cerebrospinal fluid: 10 to 20% of patients with PCNSL present with overt leptomeningeal lymphoma. In these patients diagnostic information may be achieved via lumbar puncture with cytologic analysis and/or flow cytometric analysis to identify kappa or lambda light chain restriction in lymphoma cells.

Pathologic diagnosis of either leptomeningeal or brain parenchymal lymphoma are essential in order to establish the diagnosis of PCSNL, and have identical prognostic and therapeutic implications. Intraocular lymphoma (IOL) may exist as an independent entity, with an approximate 80% risk of eventual CNS dissemination; because the intraocular compartment is a chemotherapeutic sanctuary site, the diagnosis of IOL raises an independent set of therapeutic concerns.

Recent evidence suggests that diagnostic brain biopsy may be obviated in selected cases in the setting of suspected AIDS-associated PCNSL. The association between Epstein-Barr virus (EBV) infection and AIDS-related PCNSL is well-established and the detection of EBV DNA in CSF (cerebrospinal fluid) by polymerase chain reaction has been shown to be a useful biomarker with specificity in excess of 95%. Because single photon emission computed tomography (SPECT) is also able to differentiate between cerebral lymphoma and non-neoplastic lesions in patients with AIDS, such as toxoplasmosis, concordant positive SPECT, and EBV DNA polymerase chain reaction, results appear to obviate the need for brain biopsy, and may facilitate immediate clinical intervention in the treatment of AIDS-associated PCNSL.

Positron emission tomography has also been shown to facilitate distinction between CNS lymphoma and cerebral toxoplasmosis in AIDS patients with contrast-enhancing lesions. Unfortunately there is, at present, no established CSF marker to facilitate the non-invasive diagnosis of PCNSL arising in immunocompetent patients and definitive diagnosis is usually dependent upon histopathologic analysis.

What imaging studies (if any) will be helpful in making or excluding the diagnosis of primary central nervous system lymphoma?

MRI of the brain plus gadolinium. An MRI of the spine is useful if there are relevant symptoms. Staging CT of the chest, abdomen and pelvis are critical to rule out systemic lymphoma.

If you decide the patient has primary central nervous system lymphoma, what therapies should you initiate immediately?

Once the diagnosis has been definitively established, glucocorticoids (dexamethasone) may be initiated, while organizing definitive chemotherapy.

More definitive therapies?

First line therapy is high-dose methotrexate-based chemotherapy (typically 3 to 8gm/m2) given every 2 weeks. Leucovorin rescue is essential to mitigate methotrexate toxicity. The addition of rituximab is likely to be beneficial, as is concurrent administration of a CNS penetrant alkylating agent such as procarbazine or temozolomide.

The role of consolidation (whole brain radiotherapy versus high-dose chemotherapy) is under investigation.

Untreated, most patients with PCNSL succumb to this aggressive neoplasm within 3 months of diagnosis. Treatments for this disease are in evolution and the optimal therapy for PCNSL has yet to be defined. However, a number of therapeutic principles have emerged with the accumulation of formal clinical experience and with the perspective of multicenter phase II clinical trials in PCNSL.


Glucocorticoids are a cornerstone of treatment in PCNSL given their direct lymphocytotoxic effects, as well as their ability to reduce tumor-associated edema. There is evidence that the oncolytic effect of glucocorticoids may be more marked in PCNSL than in systemic lymphomas, with single-agent response rates approaching 70% in lymphomas which involve the brain. As above, a common problem is that initial complete responses yielded by glucocorticoids may increase the risk for nondiagnostic biopsies, resulting in significant delays both in diagnosis, as well as in the initiation of definitive therapy.

Radiation therapy

PCNSL is a radiation-sensitive disease, that is, response rates to whole brain irradiation exceed 90% and rates of complete response are approximately 60%. Whole brain irradiation therapy has historically been a cornerstone of the therapeutic management of PCNSL and survival, compared to no therapy; median survival with radiation alone is between 12 and 18 months. Approaches using focal irradiation are not used because of the multifocal highly infiltrative nature of this tumor. There also is currently no data to support the use of a radiation boost to the tumor bed.

Total radiation dose may also be a significant factor. In a study of patients who received identical chemotherapy programs followed by whole-brain irradiation, reduction in the dose of radiation from 4,500cGy to 3,060cGy was associated with a significantly increased risk of relapse and lower overall survival. For this reason, the standard dose of whole-brain irradiation is generally still considered to be between 4,000cGy and 4,500cGy.

There is however, recent preliminary evidence to support the use of lower dose whole-brain irradiation (2,340cGy) in patients who achieve a complete remission to induced-based upon combination methotrexate, rituximab plus procarbazine. There is no evidence that craniospinal axis irradiation provides meaningful therapeutic benefit, unless there is evidence of spinal dissemination or meningeal drop metastases. It is important to note that spinal cord irradiation may ultimately compromise therapeutic options, because of radiation-induced bone marrow suppression.


Conventional therapy for systemic large B-cell lymphoma includes combination chemotherapy, usually with cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP). Multicenter studies of CHOP (or of CHO with dexamethasone substituting for prednisone) followed by cranial radiation, demonstrate that these agents are ineffective for PCNSL, and do not improve survival over whole brain irradiation alone. The biological basis for the differential sensitivity may relate to the blood-brain barrier; the majority of compounds in the CHOP regimen penetrate poorly into the CNS due to the specific molecular characteristics of the CNS vasculature, including the unique expression of tight junctions. Another possible contributing factor is the molecular distinctions between PCNSL and systemic lymphomas which potentially are responsible for differential chemotherapeutic sensitivity.

Intravenous administration of methotrexate greater than 1gm/m2yields cytotoxic levels in the CSF (greater than 1μM) in patients with brain tumors. Canellos and colleagues in the 1970’s were among the first to describe dramatic anti-lymphoma effects in recurrent primary and secondary CNS lymphomas with the use of high dose methotrexate monotherapy. Bokstein et al provided evidence that the rate of complete response to high-dose intravenous methotrexate for metastatic NHL within the brain was two-fold higher than the rate of response for systemic lymphoma in the same patients. As above, the use of high-dose methotrexate has emerged in multivariate analyses as the most important treatment-related prognostic variable in PCNSL.

Combined modality therapy

Combined-modality therapy for PCNSL was developed in the 1980’s by investigators at Sloan-Kettering Cancer Center. This approach involved methotrexate, administered both by intrathecal injection through an Ommaya reservoir as well as high-dose intravenous injection at doses equal to or greater than 1gm/m2body surface area, in combination with the alkylating agent procarbazine plus the vinca alkaloid, vincristine. Repeat administration of combination chemotherapy preceded the initiation of whole brain irradiation. High-dose intravenous cytarabine was administered at the completion of irradiation. This resulted in prolongation of survival with median overall survival of 42 months and 5 year survival of 22.3 months, significantly better than the 3 to 5% rate of 5 year survival, which results with the use of whole brain irradiation without pretreatment chemotherapy.

With the awareness that long-term survival can be expected in a significant fraction of PCNSL patients, there has been an accumulation of interest and data regarding the problem of treatment-related neurotoxicity in PCNSL, this has become a central issue in the consideration of treatment planning for these patients. The use of whole brain irradiation has been associated with an at least 80% rate of delayed neurotoxicity characterized by a syndrome of progressive dementia, gait ataxia, and urinary incontinence, presenting with an average latency of 13 months in patients age 60 or greater.

Neurotoxicity associated with the combination of whole-brain irradiation, plus high-dose methotrexate is associated with diffuse white matter disease and cortical-subcortical atrophy. Autopsy data have demonstrated white matter damage with gliosis, small vessel thickening, and demyelination. Because the use of chemotherapy alone in PCNSL patients has not been associated with a high rate of profound neurotoxicity, there is great interest in the implementation of more effective chemotherapy regimens, which eliminate the use of whole-brain irradiation.

The regimen developed at Memorial Sloan-Kettering which involves high-dose methotrexate, procarbazine, and vincristine, followed by whole brain irradiation was recently studied in a multicenter evaluation involving the RTOG (Radiation Therapy Oncology Group) in 102 immunocompetent patients with PCNSL. The rate of complete response to preirradiation chemotherapy (methotrexate, procarbazine, and vincristine) was 58%; the median overall survival was 36.9 months. Fifteen percent of patients (N=12) exhibited delayed neurologic toxicity: eight of these died, thus confirming the potential for severe neurotoxicity with combined modality therapy. No prospective neuropsychological testing was performed and the 15% incidence of neurotoxicity reported was clearly a minimum estimate.

Intrathecal methotrexate

The role of intrathecal chemotherapy during the induction chemotherapy for PCNSL is also a significant consideration under debate. A recent study of high-dose intravenous methotrexate which evaluated matched CSF and serum pharmacokinetic data demonstrated that cytotoxic CSF and serum methotrexate concentrations were maintained much longer after intravenous administration, compared to intrathecal dosing of methotrexate.

Outcomes in 16 patients who received high-dose intravenous methotrexate at 8gm/m2over 4 hours for solid tumor neoplastic meningitis malignancies were retrospectively compared to a reference group of patients treated only with intrathecal methotrexate. The median survival in the group treated with high-dose methotrexate was significantly longer than the group treated with intrathecal methotrexate 13.8 months versus 2.3 months (p=0.003). Two retrospective studies suggest that PCNSL patients do not benefit from additional intrathecal chemotherapy through an Ommaya reservoir if they receive high-dose intravenous methotrexate. There is, at present, no evidence to support the routine administration of intrathecal methotrexate in newly diagnosed PCNSL patients who receive intensive high-dose methotrexate therapy exceeding a dose of at least 1gm/m2body surface area.

Moreover, complications of intrathecal chemotherapy include neurotoxicity, chemical meningitis, and infection of the Ommaya reservoir. Ommaya reservoir infection occurs commonly in cancer patients treated with cytotoxic chemotherapy, affecting 19% of patients in one recent trial. This serious complication can lead to considerable morbidity, including disruption of treatment schedules. At present, many neuro-oncologists therefore restrict the use of intrathecal chemotherapy to patients with positive CSF cytology.

High dose methotrexate

In parallel with the development of combined modality therapy for PCNSL, clinical investigators at Massachusetts General Hospital performed evaluations of intensive intravenous methotrexate monotherapy with deferral of radiotherapy until progression. Methotrexate at 8gm/m2was given every 2 weeks until complete response, followed by four more cycles every 2 weeks, then monthly methotrexate treatments through 1 year of therapy. The initial responses in 31 patients were promising, with an overall rate of myelosuppression and mucositis. Quality-of-life assessment in 11 long-term survivors was similar to normative groups and there was no evidence of delayed encephalopathy related to methotrexate exposure.

This pilot experience led to a recent multicenter study of intensive high-dose methotrexate (8gm/m2) administered every 2 weeks through the New Approaches to Brain Tumor Therapy Consortium (NABTT). The rate of complete response to high-dose methotrexate in the multicenter study was 52%; no delayed neurotoxicity was detected in this multicenter study. Median progression-free survival was 12.8 months.

A similar trial of high-dose methotrexate monotherapy, performed in Germany, demonstrated weaker activity of high-dose methotrexate monotherapy in the induction treatment of PCNSL. Only 30% of 37 patients achieved a complete remission and 38% had progressive disease while on treatment. After 4 years of follow-up, the rate of leukoencephalopathy in patients who survived more than 12 months was 58% for the group that received salvage therapy with whole-brain irradiation, and only 10% for the patients who did not receive whole-brain irradiation.

Because there is evidence that the blood-brain barrier may contribute to resistance to chemotherapeutic agents and to therapeutic failure in brain tumors, there has been considerable interest in developing methods to disrupt the tight junctions and specializations of CNS endothelia which normally restrict entry of macromolecules into the brain. One approach has been to use osmotic agents to disrupt the blood-brain barrier, followed by intra-arterial or intravenous chemotherapy.

Seventy-four PCNSL patients were treated with osmotic blood-brain barrier disruption, followed by intra-arterial methotrexate, intravenous cyclophosphamide, and either oral procarbazine or intravenous or intra-arterial etoposide. The estimated 5-year survival in this group was 42% over a 15 year period. The rate of complete response was 65%, and the estimated median survival was 40.7 months. While each disruption procedure requires general anesthesia and is associated with significant acute toxicities including stroke and seizure, comprehensive neuropsychological testing performed in 36 patients who achieved a durable complete remission (greater than 1 yr) revealed no evidence of cognitive loss in those patients who ultimately did not receive whole brain irradiation.

Autologous stem cell transplant

For a subset of patients with high-risk or relapsed systemic NHL, this strategy is being evaluated in PCNSL, both to maximize the potential benefit of chemotherapy and to avoid the toxicity of whole brain irradiation in patients with newly-diagnosed PCNSL. Abrey et al were among the first to conduct a trial involving autologous stem cell transplant (ASCT) in this disease. These investigators used high-dose methotrexate and cytarabine as induction therapy, followed by BEAM (carmustine, etoposide, cytarabine and melphalan) as the preparative regimen without whole-brain irradiation. Fourteen of 28 patients enrolled, exhibited sufficient response to induction to ultimately warrant ASCT. Median progression-free survival for transplanted patients was only 9.3 months; eight of 14 patients who received ASCT developed progressive disease, at a median of 2.3 months after transplantation.

While these results demonstrate that ASCT is feasible for patients with PCNSL, the unacceptable rate of early relapse in the study using the BEAM combination raises the question of whether other preparative regimens may exhibit greater disease control in PCNSL. Soussain et al. used a conditioning regimen consisting of busulfan, thiotepa, and cyclophosphamide in 22 patients with recurrent (10 patients) or refractory (12 patients) PCNSL, and/or ocular lymphoma.

Transplantation was associated with a 53% rate of 3 year event-free survival and 20% rate of relapse in this heavily pretreated population. Patients over 60 years suffered a 71% rate of treatment related mortality. An unexpected aspect of this study was the efficacy of combination high-dose cytarabine plus etoposide as salvage therapy before ASCT. Combination high-dose cytarabine plus etoposide resulted in 12 responses (eight complete) among 14 treated patients. A recent successor study performed by this French group confirms their preliminary data regarding the efficacy of intensive chemotherapy plus autologous stem cell transplant in patients with recurrent disease.

Several principles have thus emerged with respect to therapeutic options for patients with PCNSL. An accumulation of evidence demonstrates that long-term survival and cure is now possible in PCNSL. While approximately 20% of patients exhibit refractory disease in response to cytotoxic therapy, a significant fraction of these patients (between 20 and 30%) exhibit durable responses and long-term survival with simple high-dose methotrexate-based chemotherapy without receipt of whole brain irradiation. In addition, there is increasing evidence that augmentation of high-dose methotrexate with other chemotherapeutic and/or biologic agents may result in durable responses for a greater fraction of patients with acceptable toxicity.


Management of relapsed or refractory disease is an area of active clinical study. A number of agents are under investigation including the anti-CD20 monoclonal antibody rituximab. Rituximab is the first monoclonal antibody to receive FDA (Food and Drug Administration) approval in the treatment of cancer. CD20 is a cell-surface protein that is expressed exclusively on mature B cells and is not expressed by neurons or by glia.

While rituximab has activity as a single agent in the treatment of relapsed large cell lymphoma, there is an accumulation of data demonstrating synergistic interaction with chemotherapy. Randomized data generated by the Groupe d’Etude des Lymphomas de l’Adulte (GELA) comparing CHOP chemotherapy plus rituximab, versus CHOP alone in 399 elderly patients with systemic, non-CNS diffuse large-B-cell lymphoma, revealed that the rate of complete response was significantly higher in the group that received CHOP plus rituximab than in the group that received CHOP alone (76 percent vs.% versus 63 percent, p=0.005). The addition of rituximab to standard CHOP chemotherapy also significantly reduced the risk of treatment failure and death.

This data is likely relevant to PCNSL as large B-cell lymphoma is the most common lymphoma histology to present in the brain. For this reason, a number of studies are exploring use of rituximab in this disease, both as a single agent and in combination with chemotherapy. One of the limitations of the systemic administration of rituximab, however, is that only approximately 0.1% of this monoclonal antibody appears to penetrate the intact blood-brain-barrier to enter the cerebrospinal fluid.

While anecdotal response in the CNS have been described after the systemic administration of rituximab, retrospective analysis of the GELA data demonstrated that systemic administration of rituximab did not affect the rate of CNS relapse in this population of patients. For these reasons, there is, at present, interest in the evaluation of the intrathecal administration of rituximab, both in the treatment of PCNSL as well as in the prophylaxis against CNS dissemination of non-Hodgkin’s lymphoma. A recent Phase I study of intraventricular administration of rituximab in patients with refractory PCNSL provided preliminary evidence for the safety as well as therapeutic efficacy against recurrent CNS lymphoma involving the leptomeninges, brain, and intraocular compartments.

Other investigational therapy

A number of other chemotherapeutic agents which have been shown to have activity in the treatment of PCNSL are under clinical investigation. One example is the alkylating agent temozolomide, a congener of dacarbazine (DTIC), initially approved for the treatment of malignant glioma. Temozolomide is administered orally and exhibits excellent CNS penetrance, as well as a favorable side effect profile. Its use is associated with improvements in health-related quality-of-life in brain tumor patients.

There is increasing evidence that temozolomide has activity in PCNSL, both at diagnosis and at relapse, with response rates on the order of 26%. The administration of temozolomide in combination with rituximab is also under investigation. At the University of California, San Francisco (UCSF) the combination of high-dose methotrexate plus temozolomide and intravenous rituximab is currently being investigated as an intensive induction regimen, followed by high-dose cytarabine plus etoposide as consolidation. This UCSF regimen is also being evaluated through the national consortium CALGB (Cancer and Leukemia Group B).

Another chemotherapeutic agent being studied in PCNSL is the topoisomerase I inhibitor topotecan, a derivative of camptothecin. Fischer et al treated 16 immunocompetent patients with refractory or relapsed PCNSL: six of these patients exhibited responses, four of which were complete.

Leptomeningeal dissemination of NHL represents a common pathway of relapse and dissemination, both in PCNSL as well as systemic disease. Standard treatment of lymphomatous meningitis consists of radiation to sites of radiographically visible disease and the intrathecal administration of chemotherapy with either cytarabine or methotrexate. A randomized evaluation of intrathecal administration of the long-acting form of cytarabine, Depocyt, in the treatment of lymphomatous meningitis, demonstrated its favorable response rate (71%) compared to free cytarabine (15%). Arachnoiditis associated with the intrathecal administration of depocyt appears to be common and may be ameliorated by the concomitant administration of dexamethasone.

Intraventricular administration of chemotherapy drugs such as cytarabine or methotrexate using an Ommaya reservoir device is sometimes favored in neoplastic meningitis because of evidence suggesting more reliable efficacy, compared to intrathecal administration of drugs by lumbar puncture. Ommaya reservoir placement requires a neurosurgical procedure and is associated with an approximate 10% risk of complications such as infection. Therefore, the decision to place such a device needs to be balanced with respect to probability of benefit and expectation for long-term survival.

A significant and avoidable complication may occur when there is tumor-associated obstruction in the ventricular system; delayed clearance of chemotherapy drugs from the cerebral ventricles can result in severe neurotoxicity. For this reason a radionuclide CSF flow study may be indicated to evaluate the patency of the ventricular outflow system before administration of cytotoxic agents using an Ommaya reservoir.

AIDS related PCNSL

AIDS-associated PCNSL has, until recently, been associated with a dismal prognosis and a median survival of only 3 months. The standard approach in treating AIDS-associated PCNSL has been to use whole brain irradiation, largely because of concern that patients with this condition have significant comorbidities which contraindicate the use of chemotherapy. A pilot study demonstrated encouraging results with the use of high-dose intravenous methotrexate at 3gm/m2every 14 days in AIDS-related PCNSL: seven out of 15 patients exhibited complete responses, with a median overall survival of 290 days. The treatment was well-tolerated and the median Karnofsky performance score improved from 50 to 80. Similar results have not been achieved using whole brain irradiation monotherapy in this population. There is also significant evidence that highly active antiretroviral therapy (HAART) has resulted in a decreased incidence of AIDS-related PCNSL and the use of HAART is associated with a survival benefit in this disease.

Primary intraocular lymphoma

Primary intraocular lymphoma is characterized by invasion of malignant lymphoid cells in the retina, vitreous, and/or optic nerve head and is closely related to PCNSL. In 80% of cases, the disease is bilateral. Intraocular lymphoma complicates up to 20% of cases of primary CNS lymphoma and 60 to 80% of patients who present with primary intraocular lymphoma will develop primary CNS lymphoma within 2 years. Intraocular lymphoma is a distinct entity from systemic lymphomas that metastasize via the circulation to the uvea and/or develop de novo in the external eye (conjunctiva, lacrimal gland, or orbit).

Treatment options for intraocular lymphoma are somewhat limited because of the blood-ocular barrier resulting from tight junctions between vascular endothelial cells and pigmented epithelial cells of the anterior uvea and retina. However, while both intravenous methotrexate and intravenous cytarabine are able to penetrate the blood-ocular barrier and have activity in intraocular lymphoma, responses with these agents are variable.

Radiotherapy alone has long been used to treat patients with isolated ocular lymphoma. Radiation-associated ocular toxicities include retinopathy, visual loss, and cataracts. Additionally, radiotherapy to the eye cannot be repeated should tumor relapse in the irradiated eye. For this reason, intravitreal administration of methotrexate is being evaluated on an experimental basis for isolated and recurrent ocular disease. Intravitreal methotrexate yields therapeutic concentrations in the vitreous humor for up to 5 days, significantly longer than after intravenous administration. Intravitreal injection of rituximab, an anti-CD20 monoclonal antibody is also being investigated in this disease.

One approach to treatment of the ocular involvement of PCNSL is combined modality therapy with systemic high dose methotrexate-based chemotherapy and with radiation to the ocular globes. Recurrence of intraocular lymphoma alone may be treated with intravitreal methotrexate as a single agent at some centers. There is evidence that intrathecal rituximab also has activity in recurrent ocular lymphoma.

What other therapies are helpful for reducing complications?

Patients with CNS lymphoma are at high risk for infections typically seen in immune-suppressed patients such as Pneumocystis pneumonia, as well as herpes zoster. Prophlylactic measures are indicated.

What should you tell the patient and the family about prognosis?

In the past 5 to 10 years, clinical investigators have recognized that unlike the vast majority of patients with malignant gliomas, long term lymphoma free survival is possible in PCNSL, with 5 year rates of overall survival approaching 30 to 40%.

In general, at least 70% of patients exhibit major clinical and radiographic responses to methotrexate-based induction chemotherapy. Conversely, multiple clinical series demonstrate however, that it has also become clear that approximately 20% of patients exhibit primary refractory disease to chemotherapeutic interventions within the first 6 months of therapy and succumb to CNS lymphomas. Approximately 50-60% of patients will obtain a complete remission at 1 year and there is increasing evidence to suggest that the majority of these may experience long-term remission and potentially be cured.

Late toxicity of radiation including dementia is a significant cause of late mortality in patients receiving whole brain radiotherapy at 45Gy.

Two major prognostic schemes have recently been proposed for newly-diagnosed patients with PCNSL. In 2003, Ferreri, on behalf of the International Extranodal Lymphoma Study Group (IELSG) proposed a disease-specific prognostic score based upon five variables: age, ECOG performance status (PS), lactate dehydrogenase (LDH) serum level, CSF protein concentration, and involvement of deep regions of the brain.

Age greater than 60 years, ECOG performance status greater than one, elevated serum LDH level, elevated CSF protein concentration and involvement of deep regions of the brain (periventricular regions, basal ganglia, brainstem, and/or cerebellum) were significantly and independently associated with a worse survival. Each of these five factors, when considered as discrete variables added together, constitute the prognostic score according to this system.

Each variable was assigned a value of either 0, if favorable, or 1, if unfavorable. The cumulative prognostic score was then tested in 105 assessable PCNSL patients for which complete data for all five variables were available. The 2 year overall survival (OS) was 80%, 48%, and 15% (p=0.00001) for patients with zero to one, two to three, and four to five unfavorable features, respectively. The prognostic value of this score was validated by analysis of PCNSL patients who were treated with high-dose methotrexate based chemotherapy. Neither the histologic subtype of NHL (large cell versus non-large cell), nor the presence of leptomeningeal or ocular disease, had an impact on survival in PCNSL.

The only treatment-related variable which has been shown to reproducibly impact survival is the use of high-dose systemic methotrexate containing regimens. Use of intrathecal chemotherapy, high-dose cytarabine, and the use of anthracyclines did not impact survival in multivariate analysis. Memorial-Sloan-Kettering Cancer Center has proposed a simpler prognostic model based upon clinical variables of age and Karnofsky performance status which was developed on the basis of a retrospective analysis of 282 patients treated on RTOG studies.

"What if" scenarios.

A major pitfall to avoid is methotrexate-based renal failure. This can be mitigated by dose-adjustment of methotrexate by repeat evaluation of the creatinine clearance with each cycle of methotrexate.


At least 95% of PCNSL are B-cell neoplasms which express the B-cell marker CD20. A central puzzle in the pathogenesis of PCNSL is what accounts for the presentation of lymphoma within the brain, normally devoid of resident lymphocytes. In addition, while the lymphoma spreads aggressively within the CNS, these tumors rarely metastasize elsewhere in the body.

At least two hypotheses may explain the phenomena. One is that the lymphoma represents the malignant transformation of an inflammatory process which developed in situ within the CNS. A second is that malignant B-lymphocytes acquire selective adhesion receptors for ligands expressed by the brain, for example the cerebral vascular endothelium.

A recent gene expression profile analysis of brain lymphomas of PCNSL with nodal large B-cell NHL, identified several hundred genes with significant differential expression between PCNSL and its systemic counterpart. Interleukin-4 (IL-4), a B-cell growth factor, was also demonstrated to be ectopically expressed both by the tumor vasculature as well as by the lymphoma cells in PCNSL. The study proposed that IL-4 elaborated by the tumor microenvironment may facilitate lymphoma survival signaling and stimulate a favorable tumor microenvironment, perhaps facilitating lymphoma survival and CNS tropism.

In support of this hypothesis was the observation that tumors with high expression of activated STAT-6, a transcriptional mediator of IL-4 signaling, exhibit significantly shorter survival among patients treated with high-dose methotrexate containing regimens, compared to tumors with lower expression of activated STAT-6.

The vast majority of PCNSL in immunocompetent individuals are large B-cell neoplasms which express the transcription factor BCL-6 as well as the oncogene BCL-2. Molecular genetic studies have revealed that the p14 ARF gene and p16INK4a gene are frequently inactivated by either homozygous deletion or by hypermethylation in approximately 56% of cases of PCNSL. In addition to BCL-6 protein expression, genetic evidence supports the notion that primary CNS lymphomas originate from germinal centers, given a high frequency of mutations in the immunoglobulin variable gene region.

The notion that PCNSL are related to germinal center large B-cell lymphomas is, in a sense paradoxical, given that the prognosis for patients diagnosed with PCNSL is significantly worse than the prognosis of patients with most cases of aggressive NHL. This concept is in disagreement with data that germinal center large cell lymphomas are associated with a favorable prognosis relative to the activated B-cell type. Recent gene expression profile data of PCNSL tumors obtained from immunocompetent patients provide evidence for germinal center ontogeny, but also demonstrate that PCNSL expresses features ofgerminal center, as well as activated B-cell subtypes of lymphoma, and thus represents a unique histogenic time slot of B-cell tumorigenesis.

In AIDS, the majority of PCNSL are EBV-driven neoplasms of either Burkitt’s or large-cell histology. In the latter, PCNSL arises from Epstein Barr virus (EBV) infection of B-lymphocytes which proliferate unchecked forming neoplasms in the immunoprivileged environment of the CNS. PCNSL in this setting is observed mainly in patients afflicted with HIV or congenital immunodeficiency syndromes, as well as patients who are status-post organ transplantation.

What other clinical manifestations may help me to diagnose primary central nervous system lymphoma?


What other additional laboratory studies may be ordered?


What’s the evidence?

Skarin, AT, Zuckerman, KS, Pitman, SW. “High-dose methotrexate with folinic acid in the treatment of advanced non-Hodgkin lymphoma including CNS involvement”. Blood.. vol. 50. 1977. pp. 1039-47. [This study describes six patients with CNS involvement by lymphoma treated as part of a study in patients with relapsed disease.]

[This is a paper describing the risk of relapse and overall survival in patients receiving 4,500cGy, versus 3,060cGy of radiotherapy following chemotherapy.]

[This paper describes outcomes in patients treated with osmotic blood-brain barrier disruption with intra-arterial chemotherapy.]

[In this pilot study, patients with HIV associated PCNSL were treated with single agent high dose methotrexate.]

Soussain, C, Suzan, F, Hoang-Xuan, K. “Results of intensive chemotherapy followed by hematopoietic stem-cell rescue in 22 patients with refractory or recurrent primary CNS lymphoma or intraocular lymphoma”. J Clin Oncol.. vol. 19. 2001. pp. 742-9. [This is a phase 2 study of patients with relapsed CNS lymphoma undergoing autologous stem cell transplantation with high dose thiotepa, busulfan, and Cytoxan conditioning.

DeAngelis, LM, Seiferheld, W, Schold, SC, Fisher, B, Schultz, CJ. “Combination chemotherapy and radiotherapy for primary central nervous system lymphoma: Radiation Therapy Oncology Group Study 93-10”. J Clin Oncol.. vol. 20. 2002. pp. 4643-8. [This paper describes the results of a phase 2 study of 102 patients treated with methotrexate chemotherapy, followed by whole brain radiotherapy.]

Abrey, LE, Moskowitz, CH, Mason, WP. “Intensive methotrexate and cytarabine followed by high-dose chemotherapy with autologous stem-cell rescue in patients with newly diagnosed primary CNS lymphoma: an intent-to-treat analysis”. J Clin Oncol.. vol. 21. 2003. pp. 4151-6. [This study reports the outcome of 28 newly diagnosed patients treated with methotrexate and cytarabine induction, followed by autologous transplantation.]

Batchelor, T, Carson, K, O’Neill, A. “Treatment of primary CNS lymphoma with methotrexate and deferred radiotherapy: a report of NABTT 96-07”. J Clin Oncol.. vol. 21. 2003. pp. 1044-9. [In this paper, 25 patients with CNS lymphoma were treated with high dose (8gm/m²) of single agent methotrexate.]

Abrey, LE, Batchelor, TT, Ferreri, AJ. “Report of an international workshop to standardize baseline evaluation and response criteria for primary CNS lymphoma”. J Clin Oncol.. vol. 23. 2005. pp. 5034-43. [This paper describes consensus recommendations for the initial staging and assessment of response to therapy in primary CNS lymphoma.]

Rubenstein, JL, Fridlyand, J, Shen, A. “Gene expression and angiotropism in primary CNS lymphoma”. Blood. vol. 107. 2006. pp. 3716-23. [This study compares the results of gene expression profiling in primary CNS lymphoma and nodal B-cell lymphomas.]

Ferreri, AJ, Reni, M, Foppoli, M. “High-dose cytarabine plus high-dose methotrexate versus high-dose methotrexate alone in patients with primary CNS lymphoma: a randomised phase 2 trial”. Lancet.. vol. 374. 2009. pp. 1512-20. [In this study, patients with primary CNS lymphoma were randomized to high dose methotrexate, with or without high dose cytarabine.]

Thiel, E, Korfel, A, Martus, P. “High-dose methotrexate with or without whole brain radiotherapy for primary CNS lymphoma (G-PCNSL-SG-1): a phase 3, randomised, non-inferiority trial”. Lancet Oncol.. vol. 11. 2010. pp. 1036-47. [This paper describes the results of a large randomised study evaluating the role of radiotherapy after high dose methotrexate.]

Villano, JL, Koshy, M, Shaikh, H, Dolecek, TA, McCarthy, BJ. “Age, gender, and racial differences in incidence and survival in primary CNS lymphoma”. Br J Cancer.. vol. 105. 2011. pp. 1414-8. [This paper analyzes the epidemiology and outcome of patients with primary CNS lymphoma.]