Cognitive dysfunction associated with cellular immunotherapy for cancer

Abstract

Immunotherapies engage the individual’s immune system to fight both hematologic and solid malignancies and have led to durable clinical responses for many patients. A subset of immunotherapies, including T cell engagers (TCEs) and chimeric antigen receptor T-cell therapies (CAR-T), use T cell mediated mechanisms to induce cancer cell death. These therapies can be very effective against the cancer while also causing significant systemic and neurologic toxicity, including cognitive dysfunction. Here, we provide a review on the existing and emerging literature describing the acute, subacute, and potential long-term effects of CAR-T cells and other T-cell-engaging therapies on cognition. While the acute effects on cognition, including immune effector cell-associated neurotoxicity syndrome are becoming better characterized, there is mixed data on the long-term subjective and objective cognitive outcomes after T-cell-based immunotherapies. Further research is needed, especially as these promising and effective therapies become more commonly used for both hematologic and solid malignancies.

Key Points.

• Immunotherapies including CAR-T and TCE can cause acute cognitive dysfunction including ICANS and as part of tumor inflammation associated neurotoxicity (TIAN; a unique on-target and on-tumor neurotoxicity encountered after CAR-T cell therapies targeting CNS malignancies).

• Less is known about subacute and chronic cognitive effects from immunotherapies, as currently available clinical data is mixed. It remains controversial if and to what extent patients may experience long-term subjective and objective cognitive dysfunction.

Immunotherapies have revolutionized oncology, improving outcomes for many patients with both hematologic and solid tumors. Immunotherapies use a variety of distinct mechanisms to engage the individual’s immune system in fighting cancer. Types of immunotherapies include personalized vaccines, TCEs, checkpoint inhibitors, and cellular therapies, such as chimeric antigen receptor T-cell therapies (CAR-T).¹

While cell-based immunotherapies are effective for many patients in treating cancer, they are associated with a high frequency of systemic toxicities affecting multiple organ systems, including the central nervous system (CNS).² Neurological symptoms termed immune effector cell-associated neurotoxicity syndrome (ICANS) classically develop after cytokine release syndrome (CRS) and evolve within hours to a few days; with supportive therapy they are usually fully reversible within days. Other neurologic symptoms may have a delayed onset by weeks and can present as movement and cognitive disorders, cranial nerve palsies, and peripheral neuropathies, which are termed non-ICANS neurologic toxicities.^3,4

As applications of cellular therapy have expanded and outcomes have improved, studies of the subacute and chronic effects of immunotherapies on cognitive functions—so critically relevant to quality of life for those treated with CAR-T^5,6—have gradually proliferated. Early reviews of cognitive outcomes in these patients demonstrated that the literature was limited by heterogeneity among disease groups, treatments, and assessment methods, and had little information about cognitive outcomes beyond the acute period.⁷

Here, we define acute cognitive changes as occurring within 2 weeks of therapy, subacute changes as occurring between 2 weeks and 6 months of therapy, and chronic changes as occurring or persisting after 6 months of therapy.

Brief Overview of T-cell–Based Therapies

Both TCEs and CAR therapies utilize molecules engineered to target specific cancer antigens and subsequently activate the T cell-mediated immune system to induce cytotoxicity. Many such therapies are in use or in development.

At their simplest design, TCEs bind a cancer associated antigen on 1 arm and a T cell in the other arm (often via the T-cell receptor CD3); when the TCE simultaneously binds the antigen and the T cell, the T cell is activated leading to the desired T-cell-mediated cytotoxicity.^8,9 Multiple TCEs have been approved, including for treatment of acute lymphoblastic leukemia, myeloma, melanoma, and small cell lung cancer.^10–14 Several other TCEs are currently in development, including for primary brain tumors such as glioblastoma.^15,16 Additionally, there is ongoing research for their use in other diseases beyond cancers.^17,18

CAR therapies utilize genetically engineered T cells with synthetic receptors designed to target specific antigens.¹⁹ These engineered T cells proliferate in the treated patient and ultimately lead to T-cell-mediated cytotoxicity of the cancer cells. Prior to treatment with CARs, patients typically undergo lymphodepleting chemotherapy with fludarabine and cyclophosphamide to enhance effectiveness of tumor-cell kill by promoting survival of the CARs through depletion of native T cells.²⁰ The first CAR-T therapies were developed to target CD19 for treatment of B-cell acute lymphoblastic leukemia and B-cell lymphomas, with further treatments targeting B-cell maturation antigen (BCMA) for treatment of myeloma.^21–26 There is ongoing work investigating CAR-T therapies for solid tumors, including glioblastoma,^27,28 along with autoimmune diseases affecting the nervous system, such as multiple sclerosis, systemic lupus erythematous, and myasthenia gravis.^29,30

Cognitive Assessment of Patients Undergoing T-cell–Based Therapies

Cognition is a fundamental but complex set of neurologic functions that allow an individual to sense, perceive, interpret, plan, and act in their environment.³¹ Both objective and self-reported (eg, subjective) cognitive function measurements are relevant to outcome, and may not correlate.^32,33 Strategies to assess objective cognitive performance include screening tests like Montreal Cognitive Assessment (MOCA) or Mini-Mental State Examination,³⁴ structured neurobehavioral exams such as the immune effector cell-associated encephalopathy (ICE) score,³⁵ and formal neuropsychological evaluation. Subjective cognitive surveys, such as the Patient Reported Outcome Measurement Information System Cognitive Function Survey (PROMIS-Cog) and the Functional Assessment of Cancer Therapy-Cognition (FACT-Cog)^34,36,37 provide complementary information. Guidelines for cognitive assessment in the clinical setting (eg, from the National Comprehensive Cancer Network [NCCN]^38,39 and the International Partnership for Action Against Cancer [IPAAC])⁴⁰ and in research studies (from the international Cognition and Cancer Task Force [ICCTF])⁴¹ emphasize the importance of combining both objective and subjective methods of assessment. Ultimately, the goal of cognitive assessment for patients with cancer undergoing complex regimes such as immunotherapy is to balance sensitivity to the domains of interest with efficiency of administration.^33,41

Acute Cognitive Dysfunction After T-cell–Based Therapies

Clinical Presentation

Patients undergoing treatment with CAR therapies and TCEs are at risk for acute cognitive and neurologic dysfunction within days after treatment. Patients may experience a strong systemic inflammatory response characterized as CRS, characterized by fevers, elevated inflammatory markers, and cytopenias.⁴² Acute to subacute neurotoxicity is also common, with patients often developing a distinct pattern of encephalopathy with confusion, disorientation, aphasia, and impaired fine motor function which is termed ICANS.³ The main cognitive domains affected in ICANS are executive functioning, attention, and language.² Patients are also at risk of cerebral edema and seizures, with most severe cases having fatal outcomes.^2,43

To monitor for ICANS, patients receiving CAR-T and TCE therapies will undergo regular neurologic assessment via the immune effector cell associated encephalopathy (ICE) score, a simple bedside scale from 1 to 10 that includes tests for attention, language (expressive, receptive, and writing) and orientation.³⁵ ICANS classically develops within $$5-10 days of treatment with CAR or TCEs.^10,44 Of note, TCE treatments are repeated over time, and development of ICANS often leads to a dose interruption, dose reduction, or discontinuation of the treatment.¹⁰ Further research is needed to determine if development of ICANS with the first cycle of TCE therapy is associated with increased risk in subsequent cycles; conversely, it is not yet known if absence of ICANS in the first cycle is associated with lower risk of ICANS in future cycles.

Workup for patients with ICANS usually demonstrates elevated systemic inflammatory markers⁴⁵ while cerebrospinal fluid sampling can be bland or show mild inflammatory pleocytosis or elevated protein.⁴⁵ Electroencephalograms (EEGs) can be helpful for patient management, often showing electrical slowing, generalized period discharges (GPDs) or seizures^44,46,47; a recent study also found frontal intermittent rhythmic delta activity (FRIDA) on EEG to be a highly sensitive as a marker for ICANS.⁴⁸ Magnetic resonance imaging (MRI) is often obtained to exclude other causes of neurologic deterioration. Brain MRI in patients affected by ICANS is commonly normal but in some cases varying may abnormalities have been reported, including T2 hyperintensities, hemorrhage or microhemorrhage, diffusion restriction, and contrast enhancement in multiple areas of the brain including the corpus callosum, thalami, basal ganglia, and brainstem.⁴⁹ Transcranial doppler ultrasound can show increased velocities.⁵⁰ With supportive treatment, as further discussed below, symptoms often resolve in days.

CAR-T and TCEs targeting tumors in the brain including lymphoma, glioblastoma and diffuse midline glioma have been associated with neurologic symptoms termed tumor inflammation associated neurotoxicity (TIAN).⁵¹ TIAN is considered an on-target and on-tumor neurotoxicity syndrome characterized by local inflammation at the site of the tumor and classically presents with worsening of ­previous neurologic deficits. Depending on the location of the tumor and severity of the inflammatory response, these focal neurologic deficits may impact cognition; for example, a patient experiencing TIAN with a left temporal tumor may have worsened aphasia.^3,52 However, in severe cases, TIAN may cause brain herniation of obstructive hydrocephalus with impaired cognition and consciousness.⁵¹

Epidemiology

ICANS is a common toxicity of CAR-T therapies and somewhat less common, but still prevalent, for many TCEs. It occurs in patients treated with T-cell–based immunotherapies, independent of the specific target of the therapy. Rates of ICANS following CAR-T cell therapy were initially reported in the range of 20-80%, with about 5%-10% of patients having severe or grade 3-4 effects.^{2,24,46,50,53–55} For patients with TCEs, studies suggest that approximately 20% of patients will develop ICANS, with smaller percentages experiencing severe ICANS, though this appears to be more specific to the type of TCE.⁴

TIAN was noted in 100% of patients receiving CAR-T targeted on GD2 for treatment of diffuse midline gliomas, including TIAN grades 3-4⁵² but may be less common in patients treated for CNS Lymphoma with a reported frequency of ∼20%.⁵⁶ Ongoing safety data and rates of TIAN are still being collected for patients with glioblastoma receiving epidermal growth factor receptor (EGFR) targeted CAR-T.^27,28

Risk Factors

Because of how commonly ICANS develops in patients receiving CAR-T and TCE therapies, there has been significant effort to identify patients who are at risk for ICANS, especially grade 3-5 ICANS. ICANS appears to be more common in those with more aggressive and extensive disease at initiation of the therapy.^45,57 Molecular risk factors include baseline inflammation which has also been associated with higher rates of ICANS. Specifically, higher levels of lactate dehydrogenase (LDH), C-reactive protein (CRP), procalcitonin, ferritin, and IL-6 and lower levels of platelets are associated with high rates of serious ICANS.^44,57,58 Additionally, those who have CRS are more likely to develop ICANS.^57,59 CNS disease does not seem to be an independent risk factor for ICANS for patients with B-cell malignancies⁶⁰ nor does bridging therapy with chemotherapy or systemic radiation.^57,61 Untreated brain metastases in small cell lung cancer may be associated with higher rates of acute neurotoxicity for patients treated with tarlatamab, a TCE against delta-like ligand 3 (DLL-3).⁶²

Subacute Cognitive Dysfunction After T-cell–Based Therapies

Clinical Presentation

Generally, the acute cognitive dysfunction from ICANS resolves within a few days with supportive care and treatment (see below). Subacute cognitive dysfunction, developing between a few weeks to the first few months has also been observed, and specifically in the context of CAR-T cell therapy targeting BCMA for treatment of refractory multiple myeloma. This treatment has been linked to a unique neurocognitive and movement disorder syndrome termed Movement and Neurocognitive Treatment (MNT) emergent adverse events. Symptoms of MNT classically develop within 2-4 weeks, though can also present later⁶³ and are characterized by features of Parkinsonism (rigidity, tremor, bradykinesia), and cognitive dysfunction (confusion, recent memory dysfunction, cognitive slowing, and ­inattention).^46,63–66 In some of the reported cases, MRI has shown T2 hyperintensities in the basal ganglia, while [¹⁸F] fluorodeoxyglucose-positron emission tomography (FDG-PET) has revealed hypometabolism in the basal ganglia.^46,66 Studies have identified variable CSF findings, with 1 study showing lymphocytic pleocytosis⁶⁵ and another showing bland CSF.⁶⁶ Patients have been found to have elevated circulating peripheral CARs in high levels as well.⁶⁶ In some cases, these symptoms have not resolved even after many months; for example, 1 patient in the original Ciltacabtagene Autoleucel trial who developed MNT was noted to have ongoing grade 2 symptoms at the time of the data cut in 2022, which was 2 years after the final participant joined the study.⁶⁷ Another patient was noted to have persistent movement symptoms at day 198 after treatment.⁶⁵ Of note, more extensive cognitive testing has not been reported for patients with MNT.

Epidemiology

In the original study (CARTITUDE-1) investigating BCMA-targeted CAR-T ciltacabtagene autoleucel (cilta-cel), MNT was reported in 5% of patients⁶³ (5 out of 97). After more than 2 years of follow up, 1 further case was identified 914 days after infusion.⁶⁷ It has also been reported in 2 cases treated with idecabtagene vicleucel via FDA adverse event monitoring.⁶⁸ A retrospective study of 76 patients who received BCMA targeting CAR-T identified 2 who developed MNT.⁴⁶ In the CARTITUDE-4 study, a phase 3 trial, 208 patients received cilta-cel, and only 1 patient developed grade 2 MNT.⁶⁹

Risk Factors

Because of the low numbers of patients treated with CAR-T therapies targeting solid tumors like glioblastoma and the low numbers of patients who have developed MNT, risk factors for these toxicities have not been clearly identified. In the original CARTITUDE-1 study, the 5 patients who developed MNT were noted to have at least 2 of the following risk factors: high tumor burden, CRS grade 2 or higher, ICANS of any grade, or high levels of CAR-T expansion and persistence.⁶³ Emerging data suggest that a rapid increase in absolute lymphocyte counts (ALC) post CAR-T therapy and persistent high levels of circulating CARs may be important risk factors as well.⁶⁴

Long-term or Delayed Cognitive Dysfunction After T-cell–Based Therapies

Clinical Presentation

There is much less known about the chronic effects of immunotherapies on cognition. Overall, studies have not identified significant numbers of patients who experience new or worsening cognitive dysfunction long-term, though studies are limited by small sample sizes, high level of attrition with loss to follow up, and multiple confounding variables including disease progression and other oncologic treatments⁷⁰ (Table 1). High rates of attrition are problematic for longitudinal outcome studies because attrition may not be random. It is possible that patients with greater symptom burden, including cognitive dysfunction are more likely to drop out over time thus confounding cognitive outcome results.⁷¹ When cognition is evaluated with neuropsychological tests, a handful of studies have shown that cognitive performance appears to be stable 6-12 months after CAR-T compared to baseline,^53,72–74 on the group level.

Table 1.

Selection of studies examining cognition on a sub-acute to chronic time course after CAR-T therapy.

Studies with subjective cognitive outcome measures
Reference	Number of subjects and how subjects were recruited	Diseases	Immunotherapy product	Time period studied	Cognitive tests used	Major Findings
Ruark et al.8	40 subjects from a phase I/II trial, contacted to fill out a survey (response rate of 76.9%)	ALL (27.5%) NHL (35%), CLL (37.5%); 15% had CNS involvement	CD-19 targeted CAR-T (product not specified)	1-5 years after CAR-T	Subjective cognitive dysfunction with concentrating, word finding, with memory, or with solving problems (yes or no)	15/40 patients reported cognitive difficulties 1-5 years after CAR-T
Wang et al.0	60 subjects, who received CAR-T at a single medical center. Response rate not reported	ALL (5%), DLBCL (95%)	Axicabtagene ciloleucel (86.7%), Tisagenlecleucel (13.3%)	0-12 months after CAR-T	Subjective difficulty remembering on scale from 1 to 10, with >7 representing severe difficulty	19 patients had received CAR-T >3 months prior and on average reported minimal difficulty remembering (mean score of 1.1)
Barata et al.7	115 patients prospectively recruited getting CAR-T; 86 did 90-day assessment (11 died, 15 withdrew, 6 refused the survey); 70 did 360-day assessment (13 died, 2 withdrew, 4 refused survey)	NHL	Axicabtagene ciloleucel (85%), 13% tiasgenlecleucel (13%), brexucabrtagene autoleucel (2%)	Baseline, 90- and 360 days post CAR-T	Subjective cognition via Everyday cognition (ECog) scale (includes memory, language, visuospatial abilities, planning, organization, and divided attention)	No change in global cognition or any subdomain at day 90. From 90 to 360, global cognition, memory, language, organization, and divided attention worsened, though absolute changes were small. Worse global cognition with higher ICANS grade
Sidana et al.9	34 patients prospectively recruited who underwent CAR-T. 34 completed assessment at week 2. 24 completed assessments at 3 months. 15 completed assessments at 6 months (∼40% had death or change in therapy/progression; ∼20-30% did not complete surveys by choice)	Lymphoma (68%), myeloma (24%), leukemia (9%)	Not specified	Baseline, 2 weeks, 1 month, 2 months, 3 months, 4 months, 5 months, 6 months after CAR-T	Subjective cognition with NeuroQoLv2 questionnaire	No change in self-reported cognitive function over 6 months
Delforge et al.	126 subjects prospectively recruited who underwent CAR-T as part of a clinical trial. 114 were included. 70-90% of patients completed assessments until month 6. 60-70% completed assessments through month 12. 10% completed assessments at 18 months.	Multiple myeloma	Idecabtagene vicleucel	Baseline, monthly for 1-6 months, then every 3 months until 24 months	Subjective cognition as part of larger surveys regarding quality of life	Cognitive functioning improved by month 2 and was stably improved from months 2 to 9 on average. 80% of patients had no change or improvement in cognition over the time points.
Akinola et al.5	40 patients recruited after CAR-T and 15 caregivers	Hematologic malignancy	Any CAR-T product	2 weeks to 6 months after CAR-T	Subjective cognition using semi-structured interviews	Qualitative analysis showed patients and caregivers noted cognitive dysfunction, especially forgetfulness
Perthus et al.6	59 patients prospectively recruited before undergoing CAR-T. 53 completed 3-month survey. 35 completed 6-month survey. 23 completed 12-month survey. Patients did not complete if disease progression or death or if not followed up. Of note, only 70-80% of these patients completed the cognition survey.	Lymphoma	Axicabtagene ciloleucel (62.7%), tiasgenlecleucel (30.5%), brexucabrtagene autoleucel (6.8%)	Baseline, 3, 6, and 12 months after CAR-T	Subjective cognition (FACT-Cog Scale)	No population level changes in subjective cognition. 13-33% of patients reported significant cognitive decline between 3- and 12 months.

Studies with objective cognitive outcome measures
Reference	Number of subjects and how subjects were recruited	Diseases	Immunotherapy product	Time period studied	Cognitive tests used	Major Findings
Hoogland et al.3	106 prospectively recruited patients receiving CAR-T; 84 completed 30-day assessment; 58 completed 90-day assessment; 54 completed 360-day assessment (combination of patients dying, withdrawing from study, and refusing survey)	NHL	Axicabtagene ciloleucel (87%), tiasgenlecleucel (13%)	Baseline, 30-, 90-, and 360 days post CAR-T	Full neuro-psychological battery	Global cognition decreased slightly from baseline to 90 days, then improved from 90 to 360 days. Patients with grade >2 neurotoxicity had worse cognition. Responders to therapy had better cognition at day 360
Shalabi et al.4	52 children and young adults; 37 completed neuro-psychological testing but only 18 completed both baseline and day 21-28 testing (reasons not specified)	Relapsed B-cell malignancies	CD19/CD28 CAR-T	Baseline and 21-28 days after CAR-T	Formal neuro-psychological battery (if age >4)	No difference in executive functioning, attention, working memory, or attention at 1 month when compared to baseline (on average). 88% of individuals had stable to increased score; 10% had declining score
Studies with both subjective and objective cognitive outcome measures
Reference	Number of subjects and how subjects were recruited	Diseases	Immunotherapy product	Time period studied	Cognitive tests used	Major Findings
Maillet et al.3	52 consecutive adult patients treated with CAR-T at a single center. 27 patients underwent repeat testing (1 lost to follow; 18 died; 6 progressed and were excluded)	Relapsed B-cell lymphoma	Axicabtagene ciloleucel (63%), Tisagenlecleucel (27%)	Baseline prior to CAR-T and then 6-12 months after CAR-T	Full neuro-psychological battery and subjective memory difficulty	3 patients (11%) reported memory difficulty after 6 months (decreased from 30% pre-CAR-T) No worsening in formal neuro-psychological testing per patient between baseline and 6-12 months post-CAR-T. No worsening of cognition for 12 patients who had acute neurotoxicity
Ursu et al.4	52 prospectively recruited patients getting CAR-T. Follow up at 2 years for 19 (excluded 20 who died and 13 with progression)	Relapsed B-cell lymphoma patients	CD19 CAR-T	Baseline and then 2 years after (if no progression)	Formal neuro-psychological battery and subjective cognition	No difference at 2 years post CAR-T compared to baseline cognition (both subjective and objective)
Barata et al.2	106 subjects completed subjective cognition at baseline; 70 completed it at 6 months. 39 underwent baseline neuro-psychological testing; 26 did at 6 months.	NHL (56.6%), CLL (2.8%), HL (0.9%), ALL (0.9%), MM (25.5%), solid tumor (13.2%)	Axicabtagene ciloleucel (10.4%), tisagenlecleucel (24.5%), lisocabtagene maraleucel (12.3%), tisagenlecleucel (24.5%), brexucabrtagene autoleucel (5.7%), BCMA CAR (24.5%), other (22.6%)	Baseline and 6 months after CAR-T	Formal neuro-psychological battery and subjective cognition (FACT-Cog)	No changes in subjective cognition, in overall cognitive performance on objective testing, or by individual domains. At 6 months, 20% of patients reported declined subjective cognition.

Abbreviations: ALL, acute lymphoblastic leukemia; BCMA, B-cell maturation antigen; CLL, chronic lymphocytic leukemia; CNS, central nervous system; DLBCL, diffuse large B-cell lymphoma; MM, multiple myeloma; NHL, non-Hodgkin’s lymphoma.

Studies of subjective cognitive symptoms have shown mixed results. Qualitative analysis of semi-structured interviews have found that patients report subjective issues with cognition and especially memory dysfunction 4-6 months after receiving CARs.^5,75 Qualitative analysis of semi-structured interviews of caregivers also found subjective cognitive dysfunction to be a common theme with 47% of caregivers in 1 study noting cognitive dysfunction in their loved ones.⁷⁵ Patients can report subjective changes in cognition after CAR-T therapies long-term, with some patients describing a decrease in subjective cognition over the first year from CAR-T.⁷⁶ A subset of patients (25%-38%) note worsening of their global cognition that persists 1-5 years after treatment.^77,78 Other groups have found that subjective cognition is stable over the first 6 months after CAR-T⁷⁹ or even improved.⁶ Patients with higher grades of ICANS in the acute period are more likely to self-report cognitive issues months later.^77,80 It is important to note that even in studies where no changes in cognitive performance were demonstrated at a group level, there have been a range of subjective cognitive outcomes, with 33% of patients reporting improvement, 47% reporting stable, and 20% reporting worsening of their cognitive function on subjective measures in 1 study.^53,72–74

In the pediatric population, less is known about long-term cognitive effects of cell-based immunotherapy, with 1 study showing no change in cognition 1 month after CAR-T infusion.⁵⁴ Additional, it is important to note that most patients studied for the long-term cognitive effects of immunotherapies have been patients with B-cell malignancies treated with CAR-T, with much less work done on patients with other malignancies and those treated with TCEs. There is need for further research on the objective and subjective long-term cognitive outcomes following CAR-T and TCEs, including in the pediatric, adult, and geriatric populations.

Mechanisms of Cognitive Dysfunction

Immune System and Cytokines

The precise mechanisms leading to altered cognition in patients treated with cell-based therapies remain unknown and poorly understood. Cognitive changes associated with ICANS are likely multifactorial in etiology, with a robust and supraphysiological immune response affecting the CNS considered as a major factor. Mouse models of CAR-T have shown cognitive impairment after treatment, affecting attention, short-term memory, and spatial memory.⁸¹ Mice were found to have elevated inflammatory markers including in the cerebrospinal fluid and heightened microglial reactivity, which in turn has been shown to affect neuroglial function, myelin integrity and impairs neuronal function and plasticity.⁸¹ Other studies suggest a role of endothelial dysfunction and blood-brain barrier disruption, with abnormally high levels of inflammatory cytokines and chemokines entering the CNS across a disrupted blood-brain barrier.^19,44,45

CNS Tumor–Specific Mechanisms

TIAN is generally felt to be due to localized inflammation and edema causing focal neurocognitive and other neurologic symptoms.^51,56 Two types of TIAN have been proposed.⁵¹ Type I has been postulated to originate from localized inflammation and peri-tumoral edema leading to increased intracranial pressure, obstructive hydrocephalus, and potential brain herniation at its most severe form. Type II has been thought to occur from local neuronal dysfunction secondary to an inflammatory response and ultimately resulting in disruption of local circuits. While somnolence and more severe impairment of consciousness can be seen in the setting of type I TIAN as a consequence of elevated intracranial pressure, type II TIAN could drive worsening of language impairment, attention, and executive function depending on the location of the tumor and associated worsening edema. Of note, both types of TIAN can present simultaneously.

BCMA-Specific Mechanisms

BCMA is not exclusively expressed in hematopoietic cells relevant to the treatment of multiple myeloma. It is also expressed in neurons and astroglia, though there is conflicting data on the level of expression in human brain, especially in the adult brain. A key area of expression is in astroglia and neurons of the basal ganglia, suggesting that MNT as a delayed neurological complication from BCMA-targeted therapy may represent an on-target and off-tumor effect of these cellular immunotherapies.^3,66,82,83

Other Contributing Factors for Cognitive Impairment

Of note, patients undergoing cellular immunotherapies often have had many other prior lines of treatments for their cancer, including cytotoxic chemotherapy, stem cell transplantation and lymphodepleting chemotherapy as a conditioning regimen prior to CAR-T cells. In studies of patients undergoing cellular immunotherapies, up to 20% of patients demonstrate cognitive impairment on performance-based measures at the time of immunotherapy and ∼25% report subjective cognitive dysfunction,⁷² likely reflecting the combined effects of other neurotoxic factors, including the neurocognitive effects of the disease and prior treatments.^70,84 In this context, the use of fludarabine, known to be associated with subacute and delayed myelin damage and leukoencephalopathy, is an important consideration in the pathophysiology of delayed and chronic cognitive impairment.⁸⁵ Additionally, patients are at risk for infection, ischemic strokes, cerebral hemorrhage, CNS disease, disease progression, seizure, or pre-existing or undiagnosed neurodegenerative disorders. These other conditions can be contributing factors to alterations in cognition, and they could have a confounding or additive effect on top of the cellular therapies themselves.

Management Options

Acute/Subacute Cognitive Dysfunction Including ICANS

The backbone of treatment for patients who develop acute to subacute cognitive dysfunction after immunotherapy is supportive care and close monitoring for neurologic deterioration, potentially requiring management in the intensive care unit setting. We would like to refer here to other recent publications which have extensively reviewed the treatment for ICANS, including refractory ICANS.^3,42,86 In brief, moderate to severe cases are usually successfully treated with corticosteroids and other immunomodulatory agents, with refractory cases sometimes treated with systemic and intrathecal chemotherapy and ongoing research on adapter and kill switches to inactivate CARs when needed.^87–92

MNT

There are currently no established treatment guidelines for the management of MNT’s and treatment is usually individualized and guided by the treating physician. Typically, similar approaches as for ICANS are employed, with corticosteroids and anakinra as first line intervention⁴⁶ followed by chemotherapy to kill CAR-T cells including intrathecal methotrexate, cytarabine⁶⁵ and cyclophosphamide.^3,64 Empiric treatment with levodopa/carbidopa has unfortunately not been effective in most reported cases to treatment of delayed onset Parkinsonism.⁶⁶

TIAN

Because the mechanism for cognitive impairment from immunotherapies for CNS tumors is felt to be on-target/on-tumor effect, management involves aggressive supportive care and immune modulation as needed.^51,52 With optimal supportive care, TIAN is often fully reversible, similar to the reversibility of ICANS.^3,56

Multidisciplinary Treatment of Cognitive Dysfunction

As detailed above, the extent of chronic cognitive dysfunction experienced by patients with cancer that can be attributed to CAR T-cell therapies is an ongoing area of investigation. It is likely that multiple factors contribute to these symptoms, as patients are typically extensively treated with a wide range of regimens than can contribute to cancer-related cognitive impairment (CRCI).⁷² Management of CRCI requires a multidisciplinary team, including cognitive rehabilitation services (including occupational and speech therapy), neuropsychology, and oncology.⁹³ Identification and treatment of other causes of cognitive impairment is essential. With increasing knowledge about the pattern and frequency of potential long-term cognitive alterations post CAR-T cells, there is hope that novel interventions can be developed.

Prevention of Cognitive Dysfunction

Prevention of cognitive dysfunction has historically been based on preventing ICANS. After the original CARTITUDE-1 study identified ∼5% of patients with MNT, the clinical trial team implemented changes to try to prevent and identify MNT early, including reducing tumor burden prior to CAR-T and aggressively treating CRS and ICANS. Lower rates of MNT were observed after these strategies were implemented,^63,69 suggesting that such efforts are beneficial in preventing and managing the toxicity.

Regarding other methods used to try to prevent cognitive dysfunction, prophylactic anakinra in a phase 2 study was safe, with suggestion of lower rates of high-grade CRS and ICANS compared to historical data.⁹⁴ Similar findings were seen with prophylactic use of corticosteroids.⁹⁵ There are multiple studies investigating the use of immune modulatory therapy prophylactically to prevent development of ICANS.³

Outlook

CAR-T cells and TCEs are highly effective immunotherapies with the potential to treat a wide range of oncologic conditions, but they carry a risk of cognitive toxicity. Management of acute and subacute neurocognitive syndromes has improved substantially in recent years. However, the extent and frequency of long-term cognitive effects associated with cellular immunotherapies remain poorly defined, underscoring the need for further research.

Emerging evidence suggests that integrating longitudinal subjective and objective cognitive assessments, alongside other patient-centered outcome measures, in larger patient cohorts will be critical to determining whether cell-based therapies lead to persistent cognitive consequences in individual patients. Combining these approaches with conventional and advanced neuroimaging techniques may further elucidate the mechanisms and functional impact of toxicity on neural network systems.

Identifying the precise risk factors associated with cognitive changes will be essential to improving outcomes and developing strategies to mitigate or prevent neurotoxicity. As current cellular immunotherapies continue to be refined and newer, potentially safer CAR-T designs emerge, sustained attention to cognitive outcomes will remain vital to optimizing both clinical efficacy and patients’ quality of life.

Conflict of Interest Statement

A.G.H. and M.W.P. report no conflict of interest. J.D. serves as a consultant and advisory board member for Novartis and Johnson & Johnson. J.D. is a consultant for Ono Pharmaceuticals and has received royalties from Wolters Kluwer.

Supplement Sponsorship

This article appears as part of the supplement “Cancer Therapy–Related Cognitive Impairment,” supported by an educational grant from Novocure, Inc.