IntroductionRationale

Chimeric antigen receptor (CAR)-T cell products have emerged as a paradigm-shifting therapeutic modality, demonstrating remarkable efficacy in the treatment of diseases, including relapsed and refractory B-cell neoplasms.1

To date, the Food and Drug Administration (FDA) has approved six CAR-T cell products:2 KYMRIAH (tisagenlecleucel,3 CD19 CAR-T cells), YESCARTA (axicabtagene ciloleucel,4 CD19 CAR-T cells), TECARTUS (brexucabtagene autoleucel,5 CD19 CAR-T cells), Breyanzi (lisocabtagene maraleucel,6 CD19 CAR-T cells), ABECMA (idecabtagene vicleucel,7 BCMA CAR-T cells) and CARVYKTI (ciltacabtagene autoleucel,8 BCMA CAR-T cells).

Emergence of neuropsychiatric symptoms

The realm of CAR-T cell therapy, while offering unprecedented therapeutic potential, is characterized by the emergence of neuropsychiatric symptoms that necessitate vigilant monitoring and prompt intervention. Neuropsychiatric symptoms can manifest at varying degrees within days to weeks following T-cell engaging therapies.9 It is imperative to underscore that most acute neuropsychiatric symptoms ensuing from CAR-T cell infusion are deemed reversible and typically ameliorate with judicious management strategies.10 Nevertheless, severe cases have been documented, necessitating intensive care unit monitoring and mechanical ventilation for respiratory support, alongside intracranial pressure management.

Rapidly evolving malignant cerebral edema

In the direst circumstances, fatalities have been reported, with the primary cause of death attributed to the precipitous development of malignant cerebral edema.11 This phenomenon was first documented in the context of the ROCKET trial, involving JCAR015, wherein five patients tragically succumbed to fatal cerebral edema, ultimately leading to the termination of the trial.12 Understanding the spectrum of neuropsychiatric manifestations, from reversible symptoms to the dire specter of malignant cerebral edema, is of paramount importance.

Therefore, clinicians and researchers alike are compelled to delve deeper into the etiological underpinnings and management modalities for these neuropsychiatric phenomena, thereby ensuring the safety and optimization of CAR-T cell therapy for patients with hematological malignancies.

Hypotheses on the origin of neuropsychiatric symptoms following CAR-T cell infusion

The emergence of neuropsychiatric symptoms subsequent to CAR-T cell infusion is a multifaceted and intricate phenomenon. It is important to acknowledge that the precise mechanisms underlying these symptoms may exhibit variations among individual patients. Ongoing research in this field holds promise for refining our comprehension of neuropsychiatric symptoms within the context of CAR-T cell therapy, potentially leading to enhanced strategies for both prevention and management. Numerous hypotheses have been advanced to elucidate the genesis of neuropsychiatric symptoms that ensue following CAR-T cell infusion, and current investigations are steadily illuminating this complex phenomenon. Key factors and hypotheses encompass the following:

Cytokine release syndrome: One prominent hypothesis suggests that the neuropsychiatric symptoms observed after CAR-T cell infusion might be linked to cytokine release syndrome (CRS), a common side effect of CAR-T cell therapy. During CRS, the activation and proliferation of CAR-T cells can trigger the release of a cascade of cytokines, such as interleukin (IL)-6 and IL-1, along with other inflammatory mediators. These immunological molecules may contribute to neuroinflammation, leading to neurological symptoms. However, it is important to note that immune effector cell-associated neurotoxicity syndrome (ICANS) can also occur independently of CRS.

Additional cytokines contributing to ICANS: TNF-α and IFN-γ are two additional proinflammatory cytokines implicated in the neuroinflammatory process that can contribute to the onset and severity of ICANS. TNF-α is known to disrupt the blood-brain barrier (BBB), thereby facilitating the entry of immune cells and other inflammatory mediators into the central nervous system (CNS), exacerbating neurotoxicity. IFN-γ, another key cytokine, is involved in activating microglia and enhancing the inflammatory response in the brain, leading to further neuronal damage. IL-2, often elevated in CAR-T cell therapies, contributes to the activation and expansion of CAR-T cells, which can indirectly increase cytokine levels, enhancing neuroinflammation and ICANS.13–15

Endothelial activation and BBB disruption: Additionally, a hypothesis suggests that CAR-T cells or cytokines released during CRS have the potential to induce brain vascular endothelial activation and disrupt the integrity of the BBB. This protective barrier separates the bloodstream from the brain, and any compromise in its function could allow immune cells and inflammatory molecules to infiltrate the brain, thereby leading to neurotoxicity.

Activation of microglia: Within the CNS, microglia are immune cells capable of activation in response to inflammation. Activated microglia may contribute to neuroinflammation and inflict damage on neurons, potentially playing a role in the genesis of neuropsychiatric symptoms.16

Patient-specific factors: Individual patient-related factors, including overall health status, previous treatments, and the distinct characteristics of their cancer, can influence the susceptibility and severity of neuropsychiatric symptoms. Additionally, genetic factors may predispose certain individuals to developing these symptoms. Variations in genes related to cytokine production and regulation, such as those involved in the IL-6 and IL-1 pathways, can affect the extent of inflammatory responses. Genetic polymorphisms in these pathways may lead to higher cytokine levels, contributing to increased neuroinflammation and subsequent neurotoxicity.17 Furthermore, genetic predispositions affecting BBB integrity have been implicated. Genes involved in endothelial function and BBB maintenance, such as those encoding tight junction proteins and endothelial growth factors, vary among individuals, influencing their susceptibility to BBB disruption and the infiltration of inflammatory cells into the CNS.16

CAR-T cell attributes: The design and inherent characteristics of CAR-T cells themselves can affect the likelihood of neuropsychiatric symptoms. For instance, specific CAR constructs or types of CAR-T cell therapies exhibit a heightened propensity to induce neurotoxicity. Certain CAR designs, including the selection of costimulatory domains and the affinity of the CAR for its target antigen, can influence the activation and expansion of CAR-T cells, thereby affecting the severity of inflammatory responses and the risk of neurotoxicity.

Costimulatory domains in CAR-T cells: The two most commonly employed co-stimulatory domains in CAR-T therapies are CD28 and 4-1BB. Each domain plays a pivotal role in modulating CAR-T cell expansion, persistence, and cytokine release, all of which are central to both the therapeutic efficacy and the risk of neurotoxicity.

CD28-based CAR-T cells: CD28 is known to drive rapid CAR-T cell expansion and higher cytokine release, which can result in an enhanced inflammatory response. Studies have shown that patients receiving CD28-based CAR-T cells may experience a higher incidence and earlier onset of ICANS. This is likely due to the strong immune activation and cytokine storm triggered by CD28, which leads to endothelial activation and BBB disruption, exacerbating neurotoxicity. The heightened inflammatory environment associated with CD28-based CAR-T cells has been linked to more severe and acute cases of ICANS.14 15

4-1BB-based CAR-T cells: In contrast, CAR-T cells with the 4-1BB costimulatory domain exhibit a slower and more sustained activation, resulting in a gradual and prolonged immune response. This slower proliferation typically leads to a lower incidence of severe ICANS, as it causes less intense cytokine release and systemic inflammation. Consequently, the risk of endothelial damage and BBB disruption is reduced in patients receiving 4-1BB-based CAR-T cells, lowering the likelihood of severe neurotoxicity. While 4-1BB-based CAR-T cells present a lower acute risk, they may be associated with more chronic, although milder, forms of ICANS.13

Inflammatory signaling: Anomalies within signaling pathways and immune system responses triggered by CAR-T cell therapy may also contribute to the emergence of neuropsychiatric symptoms.18

Direct CAR-T cell infiltration: There is evidence that CAR-T cells can cross the BBB and directly infiltrate brain tissue, causing localized inflammation and neuronal damage. This mechanism may be more prominent with certain CAR-T constructs.1 19 Research has shown that CAR-T cells, which are designed to target specific antigens expressed on tumor cells, may also recognize and attack healthy brain cells that express the same antigens. The exact mechanisms are not fully understood, and ongoing research aims to elucidate these processes.20 21

Delayed neurotoxicity: Delayed neurotoxicity refers to the onset of neurological symptoms that appear well after the initial administration of CAR-T cells, often outside the acute treatment window. Unlike immediate neurotoxicity, which is typically related to the acute inflammatory response, delayed neurotoxicity may involve more complex mechanisms, including persistent inflammation, autoimmune responses, and direct neuronal damage.22 The pathophysiological mechanisms underlying delayed neurotoxicity are not fully understood, but several hypotheses have been proposed. Prolonged activation of the immune system may lead to sustained inflammation within the CNS.23 Persistent BBB disruption can allow immune cells and cytokines to infiltrate the CNS, causing ongoing damage.16 Additionally, CAR-T cells or antibodies generated against them may cross-react with neuronal antigens, leading to an autoimmune attack on the nervous system.22

Parkinsonian-like syndromes following BCMA-targeted CAR-T cells

In one patient the onset of parkinsonian-like symptoms following BCMA-targeted CAR-T cell therapy showed a close temporal correlation with elevated peripheral circulating CAR-T cell counts. Notably, the symptoms resolved entirely following the depletion of CAR-T cells through a single administration of cyclophosphamide. These observations imply a potential causal relationship between the presence of circulating CAR-T cells and the manifestation of parkinsonism symptoms.20 24

The enigma surrounding neuropsychiatric symptoms following CAR-T cell infusion is progressively unraveling through scientific inquiry. A comprehensive grasp of these mechanisms holds promise for refining therapeutic strategies and ultimately improving patient outcomes.

Primary outcomesNeuropsychological symptoms following CAR-T cell infusion

We conducted a thematic analysis to categorize symptoms reported after CAR-T cell infusion.

Cognitive dysfunction symptoms such as aphasia, attention deficits, confusion, and memory impairment are primarily linked to cortical regions of the brain, including the frontal lobe (responsible for executive functions and attention), the temporal lobe (involved in memory and language), and the parietal lobe (important for processing sensory information and spatial awareness). Subcortical structures like the hippocampus (crucial for memory formation) and the thalamus (which relays sensory and motor signals to the cerebral cortex) may also be affected. Damage or dysfunction in these areas can lead to the cognitive impairments observed following CAR-T cell infusion.16 17

Movement disorders and motor dysfunction can arise from cerebellar or spinal involvement and neuroinflammatory processes affecting motor pathways. Ataxia and tremors are often linked to disruptions in the cerebellum, responsible for coordinating movement and balance. Fine motor impairment, such as difficulty with small muscle movements required for handwriting, can result from dysfunction in the corticospinal tract and basal ganglia, crucial for precise motor control and coordination. This impairment is often an early detectable symptom of ICANS. Akathisia and myoclonus can occur due to disruptions in specific motor pathways affecting motor control regions in the brain and spinal cord. Neuroinflammatory damage to the basal ganglia, which regulates voluntary motor movements, procedural learning, and routine behaviors, can lead to involuntary movements and motor restlessness. Additionally, damage to the cortical motor areas, including the primary motor cortex and supplementary motor area, as well as the involvement of the spinal cord and motor neurons, can contribute to these movement disorders. Inflammation and damage in these regions can lead to akathisia, characterized by inner restlessness and the need to move, and myoclonus, characterized by sudden, involuntary muscle jerks.17 22

Sensory dysfunction, such as auditory and visual hallucinations, sensory loss, and paresthesia, can be attributed to the effects of neurotoxicity on sensory pathways. Auditory hallucinations and sensory loss can be linked to dysfunction in the temporal lobe, which processes auditory information. Visual hallucinations are typically associated with disruptions in the occipital lobe, responsible for visual processing. Paresthesia and other sensory disturbances are often related to the somatosensory cortex in the parietal lobe, which interprets tactile information, as well as the possible involvement of peripheral nerves and spinal pathways that convey sensory information to the brain. Damage or inflammation in these areas can lead to the sensory impairments observed following CAR-T cell infusion.16

Behavioral and emotional symptoms, including aggression, agitation, anxiety, and depression, are thought to be linked to the brain’s limbic system and other regions involved in emotion regulation. Specifically, the amygdala, which is crucial for processing emotions such as fear and anxiety, and the hippocampus, which is involved in memory and emotional responses, are often affected when such symptoms occur. The prefrontal cortex, which plays a key role in regulating behavior, decision-making, and moderating social behavior, may also be affected. Additionally, disruptions in the hypothalamus, which regulates stress responses, and the anterior cingulate cortex, which is associated with mood regulation, can contribute to these behavioral and emotional symptoms. The psychological stress of the treatment process and the underlying illness may further exacerbate these symptoms.17 26

Decreased levels of consciousness and alertness, catatonia, drowsiness, and insomnia following CAR-T cell infusion are linked to disruptions in brain regions responsible for maintaining consciousness and regulating sleep-wake cycles. Specifically, the reticular activating system in the brainstem, which plays a critical role in maintaining wakefulness and alertness, can be affected. The hypothalamus, which regulates circadian rhythms and sleep–wake cycles, may also be disrupted. Additionally, the thalamus, which acts as a relay center for sensory and motor signals and contributes to consciousness, can be affected.16

Pain and discomfort due to headaches, migraines, nerve compression, neuralgia, and general tiredness are common and non-specific symptoms resulting from both direct neurotoxic effects and secondary inflammatory responses. These symptoms are often exacerbated by cytokine release and other systemic inflammatory processes associated with CAR-T cell therapy. Headaches and migraines are typically linked to inflammation and irritation in the meninges and blood vessels within the brain. The brainstem is involved in the initiation of migraines. The dysfunction in the brainstem and its interaction with the trigeminal nerve may play a critical role in triggering migraine attacks. Nerve compression and neuralgia following CAR-T cell therapy can result from inflammation affecting peripheral nerves or nerve roots, often involving the dorsal root ganglia or spinal nerves.26

Severe and fatal symptoms, such as acute obtundation, disseminated intravascular coagulation with multifocal brainstem hemorrhage, and cortical laminar necrosis, represent the most extreme manifestations of neurotoxicity. Acute obtundation involves a severe reduction in alertness, which may progress to coma and brain death. This condition is often due to extensive neuroinflammation and cerebral edema, affecting widespread areas of the brain, including the cerebral cortex, thalamus, and brainstem, which are crucial for maintaining consciousness and alertness. Disseminated intravascular coagulation with multifocal brainstem hemorrhage involves widespread clotting and bleeding within the brainstem. The brainstem, which includes structures such as the midbrain, pons, and medulla oblongata, is critical for regulating vital functions like heart rate, breathing, and consciousness. Hemorrhage in this area leads to significant and often irreversible brain damage. Cortical laminar necrosis is characterized by the death of neurons in the cortex of the brain. This condition results from prolonged and severe neuroinflammatory processes, particularly affecting the cortical layers responsible for higher cognitive functions, sensory processing, and voluntary movement. These severe outcomes underscore the critical importance of early detection and intervention in managing neurotoxicity associated with CAR-T cell therapy.16 17 22

Online supplemental table 2 shows a detailed categorization of neuropsychological symptoms observed following CAR-T cell infusion, grouped thematically to highlight the specific brain regions and pathways affected. These groupings provide insight into the underlying mechanisms and potential areas of damage or dysfunction associated with each type of symptom.

Different neurotoxicity profiles in FDA-approved CAR-T cell products

The analysis of neurotoxic symptoms across different CAR-T cell therapies revealed significant variations in the frequency of adverse events. The incidence of encephalopathy, confusion, aphasia, seizures, headache, tremors, and agitation were compared among all six FDA-approved CAR-T cell products: KYMRIAH, YESCARTA, TECARTUS, BREYANZI, ABECMA, and CARVYKTI.

It is important to note, however, that the type and severity of neurotoxic symptoms can also be influenced by other factors, including the underlying hematological malignancies being treated. For example, patients with diffuse large B-cell lymphoma treated with YESCARTA show a higher prevalence of severe neurotoxic symptoms compared with those with multiple myeloma treated with BCMA-targeted therapies.17 27 This difference may be partly attributed to the varying disease burden and the immune environment specific to each malignancy.

Online supplemental table 3 shows the different neurotoxicity profiles in FDA-approved CAR-T cell products, including the incidence of various neurotoxic symptoms and the specific studies investigating these effects. These findings highlight the varying degrees of neurotoxicity risk associated with different CAR-T cell therapies, underscoring the importance of tailored management strategies to mitigate these adverse effects and improve patient outcomes.

Secondary outcomes

Online supplemental table 4 shows data for the methodology used in previous studies for the assessment of neuropsychological symptoms within 30 days of CAR-T cell therapy. The standard daily clinical evaluation for ICANS includes the following components:

· Conducting a physical examination and reviewing vital signs: It is advisable to perform more frequent clinical assessments and bedside evaluations for patients at a high risk of neurotoxicity, such as those with fever, signs of CRS, preexisting neurological deficits, or any observed changes in mental status.26 28

· Regular neurological examinations: These examinations should be conducted with particular attention to identifying subtle deficits in attention and monitoring changes in alertness, language function, and handwriting as these are typically the earliest indicators of ICANS.22 26 28

Grading of neuropsychological symptoms during CAR-T cell therapy using the ICE-score28 and CARTOX-10 grading system29

Over the years, the grading and assessment of ICANS have evolved with the development of various scoring systems. Two of the most commonly used tools are the ICE (Immune Effector Cell-Associated Encephalopathy) score and the CARTOX-10 (CAR T-cell Therapy-Associated TOXicity) grading system. While both are essential for monitoring neurotoxicity during CAR-T cell therapy, they differ in their scope and depth of assessment.

ICE score: The ICE score is focused on evaluating basic cognitive functions, including orientation (time and place), command-following, and basic arithmetic. This system’s simplicity makes it ideal for frequent assessments, particularly in intensive care settings. However, the ICE score may miss subtle or later-onset neurotoxic symptoms, as it is primarily focused on gross cognitive impairments such as disorientation or failure to follow commands.

CARTOX-10 grading system: This tool provides a more comprehensive assessment of neurotoxicity and includes additional cognitive and motor evaluations such as writing ability and seizures. One distinctive feature of CARTOX-10 is its inclusion of a question about the president or prime minister of the patient’s country of residence, testing both temporal and spatial awareness. The CARTOX-10 is better suited for capturing a broader range of neurotoxic symptoms, including more nuanced neurological deficits that may develop during therapy.

The choice between the ICE score and CARTOX-10 for grading ICANS has implications for how neurotoxicity is reported across clinical trials. Simpler grading systems, like ICE, may under-report subtle or delayed-onset neurotoxic symptoms, making it difficult to compare neurotoxicity incidence across studies that use more detailed tools like CARTOX-10. As newer trials adopt more comprehensive grading systems, higher ICANS reporting rates are observed, reflecting improved detection rather than an actual increase in incidence.13 14 17 26 28

Online supplemental table 5 shows the Encephalopathy Assessment Tools for Neurotoxicity Grading, specifically the ICE30 and CARTOX-10 scales.

Treatment of neurological and psychiatric symptoms following CAR-T cell infusion

Management of neurological and psychiatric symptoms following CAR-T cell infusion typically involves a multifaceted approach. Each treatment modality is tailored to the patient’s specific clinical presentation and response, ensuring optimal outcomes while minimizing adverse effects. The primary treatment options include supportive care, steroids, antiseizure medication, anti-IL-6 receptor monoclonal antibodies, and IL-1 receptor antagonists. Patients receiving early intervention with corticosteroids and anti-IL-6 therapies showed better neurological outcomes compared with those receiving supportive care alone.31

IL-1 receptor antagonists as a treatment for ICANS

Anakinra, an IL-1 receptor antagonist, has emerged as a potential therapeutic option in the management of ICANS. ICANS is driven by a complex inflammatory response involving several cytokines, including IL-1, and targeting this pathway with anakinra has shown efficacy in reducing neuroinflammation. In clinical use, anakinra has been administered to patients with severe or corticosteroid-refractory ICANS, resulting in rapid symptom resolution in many cases.

Research has highlighted that blocking IL-1 with anakinra may mitigate ICANS without impairing CAR-T cell efficacy. Small studies and retrospective analyses have supported its potential, with high-dose anakinra proving particularly effective in resolving both ICANS and CRS in some patients. Given these positive findings, anakinra’s role in ICANS management is likely to expand, especially in high-risk patients. However, more prospective trials are needed to further validate its use and optimize treatment protocols.32–36

Online supplemental table 6 shows data for the treatment of neurological and psychiatric symptoms following CAR-T cell infusion.

Prophylactic strategies against neurotoxicity in CAR-T cell therapy

Recent clinical trials and guidelines have underscored the importance of prophylactic strategies to reduce the risk of neurotoxicity in patients undergoing CAR-T cell therapy. These preventative approaches aim to either avert the onset of neurotoxic symptoms or diminish their severity. The American Society for Transplantation and Cellular Therapy (ASTCT) guidelines highlight the significance of early intervention and the implementation of prophylactic measures to manage neurotoxicity effectively.28 These guidelines advocate for the use of steroids, antiseizure medications, and anti-inflammatory agents as integral components of a comprehensive strategy to prevent neurotoxicity. These measures have been supported by recent clinical trials and guidelines, demonstrating efficacy in lowering the incidence and severity of neurotoxic symptoms, thus enhancing the safety and effectiveness of CAR-T cell treatments.

In addition, recent studies have provided evidence supporting the use of reduced-intensity lymphodepletion prior to CAR-T cell therapy as a means to lower the incidence of ICANS. One study compared 2-day and 3-day lymphodepletion regimens using fludarabine and cyclophosphamide in patients receiving CD19-directed CAR-T cell therapy. The results revealed no significant differences in response rates or safety outcomes, including the incidence of ICANS, between the two regimens. This suggests that a shorter lymphodepletion regimen may maintain therapeutic efficacy while potentially reducing toxicity and resource utilization.37 Moreover, another study investigated alternative agents for lymphodepletion, such as bendamustine, either alone or in combination with fludarabine. The combination was associated with longer CAR-T cell persistence and elevated levels of circulating cytokines, which are critical for antitumor activity.38 39 Importantly, recent analyses have demonstrated that bendamustine-based lymphodepletion regimes may also affect ICANS incidence. Studies reported a reduced incidence of ICANS in patients receiving bendamustine-containing lymphodepletion compared to those treated with the standard fludarabine/cyclophosphamide regimen.40–43 These findings underscore the potential of alternative lymphodepletion strategies, not only to preserve anti-tumor efficacy and CAR-T cell persistence, but also to mitigate neurotoxicity. Further randomized controlled trials (RCTs) are needed to determine the ideal lymphodepletion regimen that effectively balances therapeutic efficacy with reduced toxicity.

Online supplemental table 7 shows data for prophylactic strategies against neurotoxicity following CAR-T cell infusion.

Correlation between biomarkers and severity of neuropsychiatric symptoms

Studies indicate that elevated levels of certain cytokines, such as IL-6 and IL-1, are significantly correlated with the severity of neuropsychiatric symptoms. Specifically, patients with higher levels of IL-6 and IL-1 post-infusion were more likely to experience severe neurotoxicity, including cognitive impairment and seizures. Additionally, increased CAR-T cell counts were associated with a higher incidence of neuroinflammatory symptoms, such as headaches and agitation. Statistical analysis revealed that patients with elevated cytokine levels (eg, IL-6 and IL-1) had a higher mean severity score for neuropsychiatric symptoms (p<0.01). Furthermore, a strong correlation was observed between the peak CAR-T cell counts and the severity of symptoms.16 17 22 26 44 These findings underscore the importance of monitoring cytokine levels and CAR-T cell counts as potential predictors for severe neuropsychiatric outcomes. Early identification of patients at risk can facilitate timely intervention and improve management strategies.

Impact of tumor burden on neuropsychiatric symptoms

Recent studies have demonstrated that tumor burden can significantly influence the severity and type of neuropsychiatric symptoms experienced by patients undergoing CAR-T cell therapy. Patients with a higher tumor burden are more likely to experience severe neurotoxic manifestations, potentially due to the increased levels of cytokine release and inflammatory responses triggered by a larger number of CAR-T cells targeting a higher quantity of malignant cells. Higher tumor burdens are associated with an increased risk of severe CRS, which in turn can exacerbate neurotoxic symptoms. The larger the tumor burden, the greater the activation and proliferation of CAR-T cells, leading to a more intense cytokine release16 44 Studies have shown that patients with a high tumor burden are more prone to BBB disruption and neuroinflammation. This is likely due to the higher levels of inflammatory cytokines, such as IL-6 and IL-1, which are produced in response to the large-scale destruction of tumor cells17 22 The severity of neuropsychiatric symptoms, including cognitive impairment, seizures, and motor dysfunction, has been found to correlate with tumor burden. Higher tumor burdens necessitate more aggressive CAR-T cell activity, which can result in increased neurotoxicity.26 45

Understanding the relationship between tumor burden and neuropsychiatric symptoms can help in risk stratification and early intervention. Patients with a high tumor burden may benefit from pre-emptive measures, such as prophylactic steroids or other anti-inflammatory treatments, to mitigate the risk of severe neurotoxicity.22

Novel CAR-T cell targets

Recent advances in CAR-T cell therapy have expanded beyond traditional targets like CD19 and BCMA, with new antigens such as CD22, CD33, and EGFRvIII offering promising results in both hematological and solid tumors.30 46 47 CD22-directed CAR-T cells have shown efficacy in B-cell malignancies, especially for patients who relapse after CD19 therapies. Neurotoxicity, including encephalopathy and seizures, has been observed, but it tends to be less severe compared with CD19 therapies.48 49 CD33-targeted CAR-T cells, used in acute myeloid leukemia, also carry neurotoxic risks, although these vary across studies, likely due to CD33’s expression on healthy hematopoietic cells.50 51 Ongoing research aims to refine the safety profile of CAR-T therapies targeting novel antigens, with larger trials needed to fully understand and mitigate neurotoxicity risks.

Discussion

To the best of our knowledge, this is the first review to provide a systematic, comprehensive, and detailed overview of current research on neuropsychological symptoms emerging subsequent to CAR-T cell infusion and the methods for their detection in the fields of hemato-oncology and psycho-oncology.

The understanding of CAR-T cell-associated neurotoxicity has significantly evolved since the introduction of CAR-T therapies. Initially, the neurotoxic effects were underestimated, with early clinical trials primarily focusing on the efficacy and immediate side effects of the therapy. As CAR-T cell treatments became more widespread, it became evident that neurotoxicity, particularly ICANS, was a significant adverse effect that required thorough investigation and management. In the early stages, neurotoxic effects were often conflated with CRS. The overlapping symptoms made it challenging to distinguish between the two conditions. Early clinical trials, such as those involving CD19-targeted CAR-T cells like tisagenlecleucel and axicabtagene ciloleucel, documented incidences of neurotoxicity but did not fully elucidate the mechanisms behind it.1 19

The CARTOX and ASTCT ICANS grading systems represent significant advancements in the clinical management of neurotoxicity. These systems provided a structured approach to evaluating and grading the severity of neurotoxic symptoms, which include a range of cognitive and motor deficits.28 The introduction of these grading systems allowed for better clinical monitoring and standardization across different studies and treatment centers.

It is noteworthy that neuropsychiatric symptoms following CAR-T cell administration may manifest independently of CRS52 ; however, severe neurotoxicity almost invariably occurs in a temporal correlation with CRS.17 22 23 Recognizing the early stages of these symptoms is imperative for attending physicians, as their severity can escalate to life-threatening complications if not promptly managed.

Recent research has provided deeper insights into the mechanisms underlying CAR-T cell-associated neurotoxicity. Studies have shown that endothelial activation and BBB disruption play crucial roles in the pathophysiology of ICANS. Studies have also demonstrated that the release of cytokines, such as IL-6 and IL-1, and other inflammatory mediators during CRS can lead to neuroinflammation, contributing to the development of neurotoxic symptoms.16 17 This understanding has been instrumental in developing targeted interventions.

The evolution in the understanding of neurotoxicity has directly influenced clinical practice. The use of corticosteroids like dexamethasone and methylprednisolone has become a cornerstone in managing ICANS, especially in severe cases. Additionally, anti-IL-6 receptor monoclonal antibodies, such as tocilizumab and siltuximab, are employed, particularly when CRS is present concurrently with neurotoxicity.17

Tocilizumab, an IL-6 receptor antagonist, has become a cornerstone in managing CRS following CAR-T cell therapy, but its efficacy in treating or preventing ICANS remains inconsistent. The primary reason for this disparity lies in the differing pathophysiologies of CRS and ICANS. While CRS is driven by systemic inflammation and elevated IL-6 levels, ICANS involves endothelial activation and BBB disruption, which are not solely IL-6 mediated.15 18

Clinical studies have shown that while tocilizumab is highly effective in reducing CRS-related inflammation, it may have limited benefit in addressing ICANS. Some reports even suggest that blocking IL-6 peripherally with tocilizumab could increase circulating IL-6 in the cerebrospinal fluid, potentially exacerbating neurotoxicity in the CNS. This phenomenon is due to the disruption of the BBB, which plays a crucial role in ICANS development.53 54 As a result, current guidelines, including those from the ASTCT, recommend corticosteroids as the primary treatment for ICANS, reserving tocilizumab mainly for cases where ICANS co-occurs with CRS.55,18 Continued research, including RCTs, is essential to clarify its exact role in neurotoxicity management.

The inclusion of anakinra represents an important advancement in the therapeutic options available for treating ICANS. Given that ICANS is characterized by a complex inflammatory cascade involving various cytokines, including IL-1, targeting this specific pathway with anakinra offers a promising strategy. Anakinra has shown efficacy in reducing neuroinflammation, particularly in patients who exhibit elevated levels of IL-1 alongside severe neurotoxicity.33 35 As the role of IL-1 in ICANS becomes clearer, anakinra’s inclusion in treatment protocols may increase, especially in high-risk patients. Continued research into this cytokine blockade and its application in CAR-T-related neurotoxicity is necessary to solidify anakinra’s place in the management of ICANS. Recent studies emphasize the importance of early detection and intervention to prevent the progression of neurotoxic symptoms. Intensive monitoring and supportive care, including the management of intracranial pressure and the prevention of seizures, are critical components of the current treatment protocols.22 56

Research has illuminated that post CAR-T cell infusion, neuropsychological symptoms encompass a wide spectrum of manifestations, including aphasia, attention deficits, impaired consciousness, disorientation, confusion, cognitive impairment, memory loss, writing difficulties, fatigue, headache, agitation, tremor, seizures, and psychomotor retardation. Significantly, the severity and duration of these symptoms can vary considerably among individuals, emphasizing the necessity for tailored approaches to management.

As CAR-T cell therapies evolve, the exploration of novel antigen targets such as CD22, CD33, and EGFRvIII offers exciting new therapeutic possibilities. Emerging research indicates that neurotoxicity is not limited to CD19- and BCMA-directed CAR-T therapies.49 50 The risk of neurotoxicity may be inherent to CAR-T cell therapies in general, rather than specific to a single target. Continued investigation is necessary to better understand the underlying mechanisms of these neurotoxic effects and to develop effective management strategies to improve patient outcomes and safety across all CAR-T platforms.

Strengths and limitations of current studies

Our review of existing studies highlights several strengths and limitations in the context of neurotoxicity management in CAR-T cell therapy. One of the primary strengths is the systematic approach to neurotoxicity grading facilitated by standardized systems such as CARTOX and ASTCT ICANS. These grading systems provide a consistent framework for evaluating neurotoxic symptoms across different studies, enhancing the comparability of results.28 Additionally, targeted interventions like corticosteroids and anti-IL-6 therapy have demonstrated efficacy in clinical practice, supporting their use in managing neurotoxic effects associated with CAR-T cell therapy.17 However, there are notable limitations. The variability in study designs and the heterogeneity of patient populations can lead to inconsistent findings. Different methodologies and patient characteristics make it challenging to draw definitive conclusions across studies. Furthermore, the absence of large-scale RCTs directly addressing neurotoxicity management in CAR-T therapy limits the generalizability of treatment strategies. This review underscores the need for future studies to identify patients at the highest risk for long-term or fatal neurological complications following CAR-T cell infusion. By recognizing these strengths and limitations, future research can be better directed toward overcoming these challenges, ultimately improving the safety and efficacy of CAR-T cell therapies.

Future research should aim to address these limitations by conducting large-scale, RCTs to validate current findings and explore additional biomarkers that may contribute to neuropsychiatric manifestations. In addition to optimizing and refining clinical strategies for managing side effects, there is an urgent imperative to develop preclinical assays capable of accurately predicting adverse events linked to engineered T cells. This imperative extends to gaining a comprehensive understanding of the underlying pathophysiological processes. However, the non-clinical safety testing of engineered T cells, crucial for clinical trial and marketing approval, faces challenges due to the complex nature of these therapies and limitations in current animal models. Recent years have witnessed notable efforts to create hematopoietic stem and progenitor cell (HSPC)-humanized mouse models,57 58 proving valuable in predicting CRS and neurotoxicity observed in patients post-CAR-T cell infusions. Despite their utility, humanized mice exhibit limitations in developing a comprehensive human immune system, and their implementation is hindered by high costs, complexity, and ethical considerations. Moreover, regulatory acceptance of the 3R principles (reduce, refine, replace) emphasizes the importance of minimizing animal use in preclinical development.

The evolving field acknowledges the need for tailored toxicity programs and innovative designs in clinical trials, incorporating risk mitigation strategies. T2EVOLVE, an Innovative Medicine Initiative consortium with the objective of expediting the clinical translation of T cell immunotherapies in Europe through a collaborative effort between the public and private sectors, stands as a significant endeavor. This European consortium is dedicated, among others, to refining current preclinical models specifically for CAR-T cell therapies, addressing the intricate challenges associated with toxicity prediction.59 60 Moreover, the project strives to foster the emergence of ex vivo human models, notably organoids and organotypical systems, which hold great promise. These models, mirroring human biology, offer a compelling avenue for predicting some of the toxicities induced by engineered T cells in a personalized manner. Regrettably, despite the potential inherent in these models, certain hurdles persist, particularly in translating quantitative effects observed in vitro to in vivo settings. Furthermore, there is an ongoing challenge in incorporating all pertinent cells and components responsible for adverse events, such as myeloid cells and endothelial cells, into these models. Addressing these challenges remains crucial for advancing the reliability and applicability of preclinical models in predicting the complexities associated with CAR-T cell therapy toxicities. Additionally, as suggested by the T2EVOLVE consortium, clinical trials that include innovative designs should consider phase I studies characterized by a de-escalating/ escalating approach, split doses, and tumor burden-adjusted doses.