CNS DLBCL: Causes, Symptoms, Treatment, and Prognosis

Central nervous system diffuse large B-cell lymphoma (CNS DLBCL) is an aggressive cancer of immune cells that arises in or spreads to the brain, spinal cord, eyes, or the fluid surrounding them. When the disease starts in the CNS itself, it is called primary CNS lymphoma (PCNSL), and the overwhelming majority of these cases are DLBCL. Treatment has improved significantly since the widespread adoption of high-dose methotrexate-based chemotherapy, but the disease remains harder to treat than DLBCL elsewhere in the body, largely because the blood-brain barrier limits which drugs can reach the tumor.

What Causes CNS Lymphoma

Most cases of primary CNS DLBCL occur in people with apparently normal immune systems, and a clear external cause is rarely identified. What researchers have found is a striking pattern of genetic mutations inside the tumor cells themselves. Two mutations in particular, in genes called MYD88 and CD79B, show up far more often in CNS lymphoma than in DLBCL arising in lymph nodes or other organs. One leading explanation is that the brain is an immunologically sheltered site with fewer of the signals that normally stimulate B cells. The tumor cells may compensate by switching on these mutations, which activate growth-promoting signaling pathways on a permanent basis.

A smaller but clinically distinct group of cases arises in people whose immune systems are compromised. Organ transplant recipients on immunosuppressive drugs, people living with HIV, and patients on long-term immune-suppressing therapy for autoimmune conditions can all develop CNS lymphoma. These immunosuppression-related cases are usually driven by Epstein-Barr virus (EBV) and behave differently from their immunocompetent counterparts, with generally worse outcomes and distinct biological features.¹ Even in non-HIV patients, EBV-positive CNS lymphoma can appear in the context of aging-related immune decline, chronic inflammation, or medication side effects.²

How Symptoms Typically Present

Because CNS lymphoma can form almost anywhere in the brain or spinal cord, symptoms vary widely depending on where the tumor sits. The most common complaints are cognitive changes: difficulty concentrating, confusion, memory problems, and personality shifts that family members often notice before the patient does. These changes can be subtle at first and are frequently mistaken for depression, dementia, or normal aging, which can delay diagnosis by weeks or months.

Focal neurological deficits are also common. A tumor pressing on motor pathways can cause weakness on one side of the body; one near language areas can impair speech. Headaches, nausea, and visual disturbances suggest rising pressure inside the skull. When the disease involves the eyes, a subtype known as primary vitreoretinal lymphoma, patients may notice blurred vision, floaters, or painless vision loss. This ocular form carries a high risk of eventually spreading into the brain itself.³ Seizures occur in a minority of patients but are another presenting symptom to watch for.

Getting the Diagnosis Right

Diagnosing CNS DLBCL usually starts with brain MRI, which typically shows one or more enhancing masses, most often in the deep white matter or near the ventricles. But imaging alone cannot distinguish lymphoma from other brain tumors or infections. The gold standard is a stereotactic biopsy, where a neurosurgeon guides a needle to the lesion and retrieves a tissue sample for pathology.

One widely recognized pitfall involves corticosteroids. Because brain tumors cause swelling, doctors frequently prescribe steroids such as dexamethasone before a biopsy is scheduled. In CNS lymphoma specifically, steroids can shrink or even temporarily erase the tumor on imaging, because they are directly toxic to lymphoma cells. The problem is that this “vanishing” can leave too few intact tumor cells in the biopsy specimen for a pathologist to make a diagnosis. A meta-analysis found that preoperative steroids roughly tripled the risk of a nondiagnostic stereotactic biopsy.⁴ A multicenter review of nearly 800 patients confirmed similar odds, with steroids raising the chance of an inconclusive surgery more than threefold.⁵ The practical takeaway: if CNS lymphoma is suspected, clinicians try to biopsy before starting steroids whenever the patient’s neurological condition allows it.

Liquid Biopsy From Spinal Fluid

An emerging alternative or complement to tissue biopsy involves analyzing cell-free tumor DNA (ctDNA) in the cerebrospinal fluid. In a study of patients with confirmed CNS lymphoma, ctDNA was detectable in the spinal fluid of every patient whose disease was confined to the brain, whereas plasma ctDNA was found in only a minority and at lower levels.⁶ Even more promising, spinal fluid ctDNA picked up residual disease that standard flow cytometry missed, and in two cases it flagged an impending relapse months before it became visible on imaging. This technology is not yet routine, but it could eventually reduce the need for repeat biopsies and catch recurrences earlier.

First-Line Treatment

High-dose methotrexate is the backbone of initial therapy for CNS DLBCL. Unlike most chemotherapy drugs, methotrexate at sufficiently high doses can cross the blood-brain barrier and reach meaningful concentrations in brain tissue. It is the single drug most responsible for improved survival in this disease over the past few decades.⁷

In practice, methotrexate is rarely given alone. Most regimens combine it with other agents, and the addition of rituximab, an antibody that targets a protein on B cells, has become standard at many centers.⁸ Induction typically involves four to eight cycles administered in a hospital setting, because high-dose methotrexate requires aggressive hydration, urine alkalinization, and careful monitoring of kidney function. Side effects include mouth sores, low blood counts, fatigue, and occasional kidney injury, all of which are manageable with close supportive care.

Consolidation After Remission

Once induction chemotherapy achieves a response, the next decision is how to consolidate that response and lower the risk of the disease coming back. Two main strategies have been studied head-to-head: whole-brain radiation therapy (WBRT) and high-dose chemotherapy followed by autologous stem-cell transplant (ASCT).

A landmark randomized trial compared the two approaches in patients aged 60 and younger. At eight years of follow-up, patients who received ASCT had an event-free survival rate of about 67%, compared with 39% for those who received WBRT, a statistically significant difference driven by far fewer relapses in the transplant group.⁹ Overall survival was similar between the two arms, largely because patients who relapsed after WBRT could still be salvaged with other treatments. But the neurocognitive data told a stark story: balance deteriorated in about half of WBRT patients versus about 10% of ASCT patients, and neurocognition worsened in roughly two-thirds of the radiation group compared with about 13% in the transplant group.

A separate trial within a larger international study found both consolidation strategies met a minimum efficacy bar, but again showed no clear survival advantage for radiation, while a systematic meta-analysis of the available evidence confirmed similar overall outcomes but noted a higher degree of neurocognitive decline with WBRT.¹⁰¹¹ For younger, fit patients, the trend in practice has moved decisively toward ASCT as the preferred consolidation approach. WBRT at full dose in the first-line setting is increasingly seen as something to avoid when alternatives exist, given the long-term cognitive harm.

Neurocognitive Harm From Treatment

The brain damage caused by whole-brain radiation deserves its own attention, because it is one of the most feared consequences of CNS lymphoma treatment and a major reason the field has worked to find alternatives. Survivors treated with WBRT score significantly lower on tests of attention, executive function, and motor skills compared with survivors who received chemotherapy alone or chemotherapy followed by transplant.¹² Brain imaging in these patients shows extensive white-matter abnormalities, with the volume of those abnormalities correlating with worse cognitive and quality-of-life scores. The damage tends to progress over years, meaning a patient who seems fine at one year may deteriorate noticeably by year three or five. This is the central tension in CNS lymphoma treatment: radiation works against the tumor, but the organ it treats is the one you can least afford to damage.

When the Disease Comes Back

Relapse remains common, and no single salvage approach has become the established standard. For patients who are fit enough, retreatment with high-dose methotrexate can still produce responses if the initial remission lasted long enough. Otherwise, alternative chemotherapy combinations are used, and WBRT is an option for patients who have not previously received radiation.¹³ One retrospective study found that patients who went on to receive a stem-cell transplant after salvage chemotherapy had a median progression-free survival of about 20 months, compared with roughly 7 months for those who received chemotherapy alone.¹⁴

BTK Inhibitors

The discovery that MYD88 and CD79B mutations are nearly universal in CNS DLBCL has opened the door to targeted drugs that exploit this biology. Bruton’s tyrosine kinase (BTK) inhibitors block a key step in the signaling pathway that those mutations hijack. Ibrutinib, the first-generation BTK inhibitor, has produced response rates between roughly 50% and 89% in patients with relapsed or refractory CNS lymphoma, though responses tend to be short-lived, with median progression-free survival around five months.¹⁵ Newer BTK inhibitors such as zanubrutinib have shown measurable drug levels in spinal fluid, and when corrected for protein binding, a substantial fraction of the free drug does cross into the CNS.¹⁶ Ibrutinib has also been confirmed to reach the cerebrospinal fluid after oral dosing, crossing the blood-brain barrier at levels considered therapeutically relevant.¹⁷ The question now is how best to combine these agents with chemotherapy or other drugs to extend the duration of response.

CAR T-Cell Therapy

Chimeric antigen receptor (CAR) T-cell therapy, in which a patient’s own immune cells are engineered to attack lymphoma, has begun to be tested in CNS disease after dramatic success in systemic DLBCL. A study on behalf of two European consortia reported outcomes for patients with active CNS disease at the time of CAR T infusion. Two-year overall survival was 37%, and two-year progression-free survival was 28%, with a relapse rate of about 59% at two years.¹⁸ These numbers may not sound impressive in isolation, but they represent a heavily pretreated population with few remaining options. The treatment carried real risks: most patients developed cytokine release syndrome, about 42% experienced immune-related neurotoxicity, and two patients died of neurotoxicity. Patients with poor functional status before treatment were at considerably higher risk for severe neurological side effects.

Immune Checkpoint Combinations

A phase II trial (SPECTRUM) tested the anti-PD-1 antibody tislelizumab combined with pemetrexed in patients with relapsed or refractory primary CNS DLBCL. The overall response rate was about 71%, including a complete response rate of 50%, and median overall survival reached nearly 28 months.¹⁹ These results are early-phase and need confirmation in larger studies, but they add to a growing body of evidence that the tumor microenvironment in CNS lymphoma may be susceptible to immune checkpoint blockade.

Predicting Outcomes

Doctors use prognostic scoring systems to estimate how a given patient is likely to fare. The most validated model for primary CNS lymphoma is the Memorial Sloan Kettering Cancer Center (MSKCC) score, which sorts patients into three risk classes based on age and functional status. In one validation cohort, median overall survival ranged from about 68 months in the best group to roughly 19 months in the worst.²⁰ A competing model developed by the International Extranodal Lymphoma Study Group (IELSG) uses additional variables such as protein levels in the spinal fluid and involvement of deep brain structures, but multiple independent studies have found it less reliably discriminates between risk groups than the MSKCC score.²¹ A newer model called the 3F score has shown the best statistical precision in a recent multicenter comparison, outperforming both older systems.²²

At the molecular level, the MYD88 L265P mutation, which is the most common version of the MYD88 change in CNS lymphoma, has been associated with better survival in at least one single-center study.²³ This may seem counterintuitive, since the mutation drives tumor growth, but it also makes cells more dependent on the signaling pathway it activates, potentially making them more sensitive to drugs that target that pathway.

When Systemic DLBCL Spreads to the Brain

Primary CNS lymphoma is not the only way DLBCL reaches the brain. Patients with DLBCL that starts in lymph nodes or other organs can experience secondary CNS relapse, which happens in a small but devastating fraction of cases. A widely used clinical risk tool, the CNS International Prognostic Index (CNS-IPI), identifies high-risk patients by counting factors such as age over 60, advanced stage, elevated lactate dehydrogenase, poor functional status, multiple sites of disease outside lymph nodes, and kidney or adrenal gland involvement. Patients with four or more of these risk factors face a roughly 10% chance of CNS relapse within two years, compared with well under 1% for low-risk patients.²⁴

Preventing CNS relapse has proven frustratingly difficult. Historically, doctors injected methotrexate or cytarabine directly into the spinal fluid (intrathecal prophylaxis) for high-risk patients. But most retrospective analyses conducted in the modern treatment era have failed to show that intrathecal prophylaxis actually reduces CNS relapse rates. High-dose intravenous methotrexate has been proposed as a better strategy, since most relapses occur in the brain tissue itself rather than the spinal fluid, but several large retrospective studies have also failed to demonstrate a clear benefit.²⁴ The field is still searching for reliable ways to prevent secondary CNS disease, and better biomarkers to identify truly high-risk patients would help target prevention efforts more precisely.²⁵

For patients with high-risk systemic DLBCL who have negative spinal fluid results, routine MRI screening of the brain may not add much. One recent study concluded that MRI can reasonably be skipped when spinal fluid analyses are negative, avoiding the cost and anxiety of imaging in patients who are unlikely to have occult brain disease.²⁶

The Tumor’s Immune Neighborhood

CNS lymphoma does not exist in isolation inside the brain. It creates a local immune environment filled with infiltrating immune cells, and the interplay between these cells and the tumor is drawing increasing research attention because it may predict how patients respond to treatment and point to new therapeutic targets.

The immune cells around CNS DLBCL tumors are predominantly a type of macrophage (CD163-positive, often called M2 macrophages) and T cells that express exhaustion markers like PD-1. The tumor cells themselves express PD-L1 in a minority of cases, roughly 18% in one study, often driven by a genetic rearrangement involving the PD-L1 gene.²⁷ Interestingly, PD-L1 expression on the surrounding immune cells, rather than on tumor cells, has been linked to better outcomes, suggesting the immune system is actively trying to control the cancer when these markers are present.²⁸ Beyond PD-1/PD-L1, other immune checkpoint pairs such as TIM-3/galectin-9 have been found to be highly expressed in CNS lymphoma, offering additional potential targets for immunotherapy.²⁹

Getting Drugs Past the Blood-Brain Barrier

The fundamental challenge in treating any brain cancer is the blood-brain barrier, a tightly sealed layer of cells lining brain blood vessels that blocks most drugs from entering the brain. High-dose methotrexate works partly because flooding the bloodstream with extremely high concentrations forces enough drug across the barrier to be effective. But this brute-force approach limits which drugs can be used and at what doses.

Several experimental strategies are trying to solve this problem more elegantly. One approach uses tiny exosomes, naturally occurring cell-derived vesicles, coated with antibodies against CD19, a protein found on lymphoma cells. In laboratory models, these exosomes loaded with methotrexate were taken up by the cells lining brain blood vessels and transported across the barrier, delivering the drug directly to tumor tissue.³⁰ A similar exosome system carrying doxorubicin, a drug that cannot normally reach the brain, showed evidence of crossing the barrier through endothelial cell uptake in preclinical experiments.³¹ Another group developed a nanoparticle carrying both methotrexate and lenalidomide, an immunomodulatory drug, that bypassed the barrier by routing through lymphatic channels near the injection site in the neck.³² All of these technologies remain preclinical, but they represent a shift in thinking from “give more drug and hope some gets through” toward engineered delivery systems that could eventually make a wider range of therapies available for brain tumors.