Japanese researchers have identified cerebrospinal fluid (CSF) proteins that accurately predict which patients are at high risk for developing immune effector cell-associated neurotoxicity syndrome (ICANS) during Chimeric Antigen Receptor (CAR)-T cell therapy for B-cell non-Hodgkin lymphoma.
“If validated in larger studies, these CSF biomarkers could help identify high-risk patients before treatment, allowing for early intervention with preventive strategies such as anti-inflammatory or complement system-targeting drugs,” said senior author Yuya Kunisaki, from Kyushu University in Fukuoka.
Although CAR-T cell therapy has shown high response rates in B-cell malignancies, it can be associated with serious toxicities like cytokine release syndrome and ICANS, with ICANS reported in up to 64% of clinical trial patients.
“Finding reliable biomarkers for ICANS is crucial because this condition can range from mild symptoms to severe, life-threatening complications such as seizures, brain hemorrhages, and loss of consciousness,” Kunisaki told Inside Precision Medicine. “Identifying high-risk patients before treatment allows doctors to take early interventions, potentially reducing the severity or even preventing ICANS altogether. Having a predictive biomarker test could help manage these risks and ensure safer treatment.”
In their search for such a biomarker, Kunisaki and team analyzed residual CSF samples collected prior to CAR-T therapy from 29 patients with B-cell non-Hodgkin lymphoma who were treated with Tisagenlecleucel (n=10), Axicabtagene ciloleucel (n=12), or Lisocabtagen maraleucel (n=7). Of these, 11 (38%) developed ICANS.
In all, 1,350 CSF proteins were identified using data-independent acquisition mass spectrometry. The 100 proteins that showed the greatest difference in expression between people with and without ICANS were then classified into five clusters. The researchers found that these corresponded to biological pathways that could be involved in the development of ICANS including hydrolytic enzyme functions acting on lysosomes or glycosyl bonds (Cluster 1), complement activation or humoral immunity (Cluster 2), low-density lipoprotein particle clearance (Cluster 3), and extracellular matrix proteins (Cluster 4).
Kunisaki and colleagues next screened 46 proteins as candidates for discrimination between ICANS and no ICANS. Of these, six were increased in the ICANS-positive group, while the remaining 40 were decreased. They then created a composite ratio index for pairs of increased and decreased proteins and found that the ratio of two proteins—C1RL and FUCA2—discriminated between the ICANS-positive and -negative groups with an area under the curve (AUC) value of 0.95, indicating 95% accuracy.
The researchers validated their findings in a cohort of 10 patients (seven ICANS-positive and three ICANS-negative). In this group the AUC was 1.0 for the C1RL/FUCA2 ratio factor, and was also 1.0 for a second protein pair—F12/FUCA2.
“We were very surprised by the exceptionally high predictive accuracy of these biomarkers,” Kunisaki remarked. “In previous studies, blood-based biomarkers for ICANS, such as serum neurofilament light chain or plasma fibrinogen, have generally shown an AUC of around 0.7, indicating only moderate predictive power. In contrast, our CSF biomarkers demonstrated an AUC of 1.0, significantly surpassing the predictive performance of blood-based markers. This suggests that CSF provides a much clearer and more direct window into the neuroinflammatory processes associated with ICANS.”
However, CSF collection before CAR-T cell therapy is not standard practice. “Given this limitation, we are actively exploring the possibility of detecting similar biomarkers in blood samples, which would make the test far more accessible and practical for clinical use,” said Kunisaki. “If we can identify comparable markers in blood serum, it could significantly enhance early risk assessment for ICANS.”
The researchers are now planning a larger, prospective clinical study to assess the robustness of the findings across different patient populations and ensure that the CSF biomarkers can be effectively applied in real-world clinical settings. They also aim to expand their research to include patients with other types of blood cancers beyond B-cell non-Hodgkin’s lymphoma.
Kunisaki believes that it should be fairly straightforward to integrate the ICANS biomarkers into routine clinical practice, particularly if they can develop a blood-based test.
“Proteomics analysis is becoming more accessible, and with proper standardization, this predictive test could be incorporated into routine pre-treatment evaluations for CAR-T cell therapy,” he said.
The team is also investigating whether the biomarkers could be measured using a simple immunochromatographic test that could be used in a point-of-care setting. “By implementing these biomarkers in a simple and efficient manner, we aim to provide a tool that enhances patient safety and supports a more personalized approach to CAR-T cell therapy,” said Kunisaki.
The study findings are published in Leukemia.
