For patients with primary central nervous system lymphoma (PCNSL), the treatment journey often becomes increasingly difficult after relapse. Once conventional therapies fail, many patients are left with few effective options. As a rare subtype of diffuse large B-cell lymphoma (DLBCL), PCNSL is associated with a high relapse rate and poor prognosis—patients with relapsed or refractory (R/R) PCNSL have a median overall survival (OS) of only 6.8 months.
Although chimeric antigen receptor T-cell (CAR-T) therapy has shown remarkable results in systemic lymphomas, its use in PCNSL has been limited due to concerns about immune effector cell–associated neurotoxicity syndrome (ICANS). While small studies and meta-analyses have suggested potential benefits, the lack of large real-world data, long-term follow-up, and comparative analyses have made it difficult to establish the true value of CAR-T in PCNSL.
To address this, the French LOC (Lymphomes Oculo-Cérébraux) network conducted the largest real-world study to date on CAR-T therapy in PCNSL, with the aim of better defining its efficacy and safety through a larger cohort and extended follow-up. The results were recently published in the American Journal of Hematology.
1. Study Focus: Relapsed and Refractory PCNSL Patients
The researchers selected two groups from the LOC network database.
The CAR-T group included patients who had either relapsed after autologous stem cell transplantation (ASCT) or were ineligible for ASCT, had received at least two prior lines of therapy, and underwent leukapheresis for CAR-T preparation. A total of 27 patients met these criteria—25 received CAR-T infusion, while 2 patients died of disease progression before infusion.
The control group included 247 patients aged 18–80 years with DLBCL-type PCNSL who had received high-dose methotrexate (HD-MTX)–based first-line therapy, experienced CNS relapse (excluding isolated vitreoretinal relapse), had no systemic involvement, and were ineligible for or had relapsed after ASCT.
Among the 25 infused patients, 16 received tisagenlecleucel (tisa-cel) and 9 received axicabtagene ciloleucel (axi-cel). All but one patient underwent bridging therapy to control tumor burden before CAR-T infusion. Response was assessed using IPCG criteria (CR, PR, SD, PD), and safety evaluation focused on CRS and ICANS per ASTCT 2019 guidelines.
- Patient Characteristics: Elderly, Heavily Pretreated, and Difficult-to-Treat Population
Patients in the CAR-T cohort represented a challenging group:
Median age: 68 years (range 34–76).
Performance status: 40% had Karnofsky Performance Status (KPS) <70, indicating limited daily functioning.
Prior treatments: Median of 3 prior lines (range 2–6).
52% had undergone ASCT.
78% had received ibrutinib.
56% had received lenalidomide.
63% were refractory to first-line therapy or relapsed within 6 months.
Disease sites: 81% had parenchymal brain involvement, and 30% had cerebrospinal fluid (CSF) involvement, including 4 with isolated CSF relapse—indicating complex CNS disease.
- How Effective Was CAR-T Therapy?
1) Response Rates: CR Rate Reached 64% with Consistent Efficacy Across Sites
At 1 month post-infusion, the overall response rate (ORR) was 76%—32% achieved complete response (CR) and 44% partial response (PR). With longer follow-up, the best response rate improved further:
64% (16/25) achieved CR.
16% (4/25) achieved PR.
20% (5/25) were refractory to CAR-T.
Remarkably, CAR-T therapy demonstrated efficacy across all CNS compartments:
Brain parenchyma: CR 60%
CSF involvement: CR 87%
Ocular lesions: CR 67%
This finding dispelled previous concerns that certain CNS regions might respond poorly to CAR-T therapy.
Importantly, disease status prior to infusion significantly influenced outcomes:
Patients who were in CR or PR before infusion achieved a subsequent CR rate of 86%, compared to 36% among those with SD or PD (p=0.02). This underscores the importance of tumor control before CAR-T therapy.
Additionally, two cases exhibited a “flare effect”—a transient increase in MRI lesions after infusion that later resolved spontaneously, emphasizing the need to differentiate it from disease progression.
2) Long-Term Survival: Median OS Exceeded 21 Months
With a median follow-up of 20.8 months from leukapheresis (19.4 months post-infusion):
Median progression-free survival (PFS): 8.4 months.
6-month and 12-month PFS rates: 63% and 43%, respectively.
The PFS curve plateaued after 12 months, suggesting durable remission in some patients.
Overall survival (OS):
Median OS: 21.2 months.
6-, 12-, and 18-month OS rates: 77%, 58%, and 51%, respectively.
These outcomes were far superior to the control group’s median OS of 4.7 months (p<0.001).
For patients achieving CR, relapse risk was significantly reduced:
12-month relapse-free survival (RFS): 87% for CR patients vs. 0% for non-CR (p<0.001).
Patients in CR/PR before infusion had 12-month RFS of 79%, compared to 24% for SD/PD (p=0.02).
3) Comparison with Conventional Treatments: CAR-T Shows Clear Superiority
After inverse probability weighting (IPW) adjustment, CAR-T still showed significant advantages:
PFS hazard ratio (HR): 0.39 (95% CI: 0.30–0.48, p<0.001) — a 61% reduction in progression risk.
OS HR: 0.40 (95% CI: 0.31–0.51, p<0.001) — a 60% reduction in mortality risk.
Even when compared with patients treated with novel agents like ibrutinib, CAR-T remained superior, confirming its therapeutic value in PCNSL.
- Safety: Manageable Neurotoxicity and Acceptable Tolerability
CRS: Occurred in 92% (23/25), including 2 cases of grade 3–4.
ICANS: Occurred in 68% (17/25).
89% (8/9) with axi-cel vs. 56% (9/16) with tisa-cel (p=0.2).
5 cases (20%) were grade ≥3.
Median onset times were 1 day for CRS and 5 days for ICANS; median resolution times were 4 days and 6 days, respectively.
Three patients developed grade 4 neurotoxicity lasting over 3 months:
One patient died from neurotoxicity.
One fully recovered at 6 months.
One was still recovering at 9 months.
Management included tocilizumab in 72% and corticosteroids in 56%.
Prolonged grade ≥3 cytopenia (>28 days) occurred in 36%, and 36% required ICU admission within one month post-infusion.
- Summary
This large real-world study from the French LOC network represents the largest cohort to date of CAR-T therapy in PCNSL. It demonstrated that CAR-T therapy is not only effective but also durable and manageable in safety, with a median follow-up of 20.8 months confirming both high response rates and long-term survival benefits compared with standard therapies.
The results appear even comparable to outcomes in systemic DLBCL (12-month PFS ≈ 40%), suggesting that CAR-T therapy should not be withheld in PCNSL due to neurotoxicity concerns.
However, further studies with longer follow-up and larger cohorts are needed to determine whether CAR-T can achieve true cure in PCNSL and to identify prognostic factors that guide patient selection. Moreover, there is room to enhance efficacy through combination strategies—for example, integrating CAR-T therapy with PD-1 inhibitors, immunomodulatory drugs, or BTK inhibitors, or developing next-generation CAR-T constructs tailored for CNS diseases.
