Primary CNS lymphoma (PCNSL) is a rare disease representing less than 3% of all non-Hodgkin lymphoma. It refers to lymphomas that arise in the brain parenchyma, spinal cord, eyes, cranial nerves and meninges. Most of these (90%) are diffuse large B-cell lymphomas (DLBCL) but other B-cell and T-cell lymphomas are rarely encountered in the CNS. Interestingly, many epidemiological studies show an increased incidence of PCNSL over recent years. PCNSL can occur in both HIV negative and positive patients. HIV-associated PCNSL is virtually always associated with EBV and biologically is a different disease from PCNSL (B-cell type) in immunocompetent patients (HIV negative), which is virtually never EBV positive. PCNSL is typically confined to the CNS and disease spreads outside the CNS in less than 5% of cases. The following study is for PCNSL (aggressive B cell type) in HIV negative and immunocompetent patients. Treatment of PCNSL Compared to patients with systemic DLBCL, the outcome for patients with PCNSL is a great deal worse. Treatment of PCNSL differs significantly from systemic DLBCL because many chemotherapy agents do not adequately penetrate the blood-brain barrier. Historically, radiotherapy has been a mainstay of treatment because it is effective and side stepped the limitations of chemotherapy in penetrating the CNS. However, when used alone, responses are typically short-lived and virtually all patients relapse. Due to the poor outcome of radiotherapy and significant cognitive side effects, investigators have been developing chemotherapy regimens to be given with radiotherapy and more recently without radiotherapy, respectively. The development of the chemotherapy-based treatments relied heavily on strategies used for solid brain tumors and were not optimized for the treatment of DLBCL. High dose methotrexate (HD-MTX), an agent with good CNS penetration, has been the centerpiece of PCNSL treatment for years; however, when used alone it produces a progression-free survival of only 7 months. When HD-MTX was followed by whole brain radiotherapy, it produced an impressive 82-88% complete remission rate and median progression free-survivals of 32-40 months, but no evidence of cure. Unfortunately, such combined modality treatment is associated with severe long-term neurotoxicity. For this reason, there has been much interest in developing regimens that obviate or defer the need for radiation until relapse. Most promising in this regard are combinations of high dose methotrexate with systemic agents that cross the blood brain barrier, such as cytarabine and ifosfamide, particularly in patients under 60 years of age. The Bonn group and others have adopted such an approach and have reported promising results with chemotherapy and deferred radiation in younger patients. In their trial, 9% died of treatment related toxicity. Overall, the median time to treatment failure was 21 months with a median follow-up of 26 months; for those over 60 years of age, the TTF was 15 months and for those under 60 years old, 60% were progression-free at 4 years. While these are promising results for PCNSL, they nonetheless show that current strategies have a poor outcome with an unacceptably high treatment related mortality, and low rates of cure. Thus, there is an important need to develop more effective and less toxic strategies for PCNSL. While one approach is the addition of rituximab, there is a need to develop chemotherapy platforms that are based on the most effective strategies in systemic DLBCL and to incorporate new ideas based on the molecular biology of PCNSL. Studies that have incorporated rituximab, such as the CALGB 50202 study, are still based on HD methotrexate and cytarabine. In a preliminary report, the CALGB reported median PFS of 2.3 years, indicating that few patients are cured even with the addition of rituximab to these agents (Rubenstein et al ASH Abstract 763; 2010). Genetics and Molecular Biology of PCNSL PCNSL has a high load of somatic mutations. Approximately, 30-40% of cases have mutations of BCL6. In addition, aberrant hypermutation in the following proto-oncogenes and tumor suppressor genes are found: PIM1 (50%), c-MYC (60%), TTF (70%), PAX5 (60%), and Fas (CD95) (20%). Interestingly, PCNSL almost never leaves the CNS and though many theories have been hypothesized, the reason for this neuro-tropism is not well understood. Theories that may explain this phenomenon include the following: 1) the brain is an immune sanctuary and the malignant B-cells are eliminated in the periphery by a specific antitumor immune response; 2) the malignant B-cells are dependent on a chemokine that is present in high concentrations within the brain such as BCA-1 - B-cell attracting chemokine encoded by CXCL13; 3) the malignant B-cells are dependent on specific adhesion signals; 4) the malignant B-cells are dependent on an antigen that is only found in the CNS. Accessing tissue to better understand the biology of PCNSL has been very challenging due to the rarity of the disease and the technical challenges of obtaining tissue from the CNS. In addition, interpretation and analysis of CNS tissue has been impeded by the architecture of the CNS - characteristically there is variable cell density, neovascularization and many infiltrating immune cells. However, four molecular profiling studies have been performed and they all revealed different messages about the biology of PCNSL. In the first of these, Rubenstein et al. looked at 23 cases of PCNSL and contrasted PCNSL with systemic DLBCL. They showed that the oncogenes PIM1 and MYC where highly expressed in PCNSL relative to systemic DLBCL and there was also evidence of high expression of IL4 genes. They also demonstrated high expression of STAT 6 - which is important in IL4 signaling - and high levels of this were associated with poor survival, albeit in a small series. In the second of these studies, Tun et al., evaluated 13 cases of PCNSL. They described a characteristic signature that was composed of extracellular matrix and adhesion-related genes. A third study, by Montesinos-Rongen et al. looked at 21 cases and demonstrated a signature of a late germinal center B-cell and finally Booman et al looked at 9 cases and described a signature that was enriched in genes that were involved in apoptosis and the immune response12.
