Immunotherapy-related neurotoxicity in the central nervous system of children with cancer

Abstract

Significant gaps remain in our understanding of immunotherapy-related neurotoxicity in pediatric patients, largely because much of our knowledge comes from studies in adults. Accurately identifying the adverse effects of immunotherapy in children is also challenging, owing to variations in terminology and grading systems. Moreover, the manifestation of immunotherapy-related neurotoxicity differs greatly across different diseases, various modalities, dosages, and delivery methods. Combining immunotherapy with other treatments might improve outcomes but introduces new complexities and potential for increased toxicities. Additionally, pediatric patients with intracranial malignancy have unique responses to immunotherapies and distinct neurotoxicity compared to those with extracranial malignancy. Consequently, we must enhance our understanding of the pathophysiology, prevalence, severity, and management of immunotherapy’s neurotoxic effects in this vulnerable group. This review consolidates the current knowledge of immunotherapy-related neurotoxicity in pediatric oncology, highlighting various types of neurotoxicity including cytokine release syndrome (CRS), immune effector cell-associated neurotoxicity syndrome (ICANS), and tumor inflammation-associated neurotoxicity (TIAN), among others. Furthermore, we examine the unique features of neurotoxicity associated with adoptive cellular therapy (ACT), antibody-based therapies, immune checkpoint inhibitors (ICIs), oncolytic viruses (OV), and cancer vaccines.

Graphical Abstract

Graphical Abstract.

Key Points.

• Immunotherapy in children with cancer presents unique challenges in recognizing, grading, and managing neurotoxicity, varying by agent, dose, and delivery route.

• Immunotherapy offers opportunities for children with brain tumors but involves specific risks.

Immunotherapy has emerged as a formidable treatment for pediatric cancer, with chimeric antigen receptor (CAR) T-cell therapy in particular showing effectiveness across a broad spectrum of hematologic malignancies.^1,2 Immune checkpoint inhibitors (ICIs) have been investigated in numerous clinical trials enrolling pediatric patients, but these agents’ success in these patients has occasionally diverged from that seen in adult populations.^3,4 Given the differences in the efficacy of immunotherapy between adults and pediatric populations, as well as between hematologic and solid malignancies, the respective toxicity profiles are likely to differ as well.

The aim of immunotherapy is to activate the host immune system or the adoptive cellular therapy (ACT) to target neoplastic cells. However, this heightened immune response, owing to the immune system’s cross-reactivity with the nervous system, can lead to neurotoxicity.^5,6 Additionally, neurotoxicity can be caused by systemic inflammation syndromes such as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS).⁷Furthermore, children with intracranial diseases treated with immunotherapy can also have tumor inflammation-associated neurotoxicity (TIAN), resulting from edema and inflammation within the intracranial tumors.⁸These neurotoxicities can swiftly become life-threatening, necessitating immediate recognition and intervention. Variations in target antigens, dosing, and delivery methods can result in significantly different neurotoxicity profiles. Therefore, comprehending the immunotherapy-related neurotoxicity landscape in pediatric cancer patients, which diverges from its counterpart in adults, is of paramount importance. Previous reviews have provided detailed descriptions of immune-related neurotoxicity in pediatric cancer patients.^9,10 With the recent surge in immunotherapy trials for pediatric solid tumors, particularly brain tumors, this review presents the most up-to-date information and offers a comprehensive overview of immunotherapy-associated neurotoxicity including children with and without intracranial disease. Specifically, we focus mainly on the diverse neurotoxicities involving the central nervous system (CNS) caused by different immunotherapies including ACT, antibody-based therapies, ICIs, oncolytic viruses (OV), and cancer vaccines.

Assessment and Grading Systems of Immunotherapy-Related Neurotoxicity in Children

Assessment

Immunotherapy can lead to a wide range of neurological symptoms, including headaches, encephalopathy, and seizures,¹¹ and it may cause delayed or chronic neurotoxicity that results in substantial neurocognitive deficits. The symptom patterns in pediatric patients may be distinct from those in adults. Owing to their reliance on language-based measures, tools designed to assess adult cognitive function are often inadequate in pediatric populations.⁷ In children, symptoms can manifest as subtle changes in behavior, and their assessment frequently relies on caregivers’ observations. Moreover, certain common symptoms, such as headaches, may be challenging to discern in young children. The diagnostic approach can also differ, given that children often require sedation for extended imaging studies and procedural interventions. Furthermore, children with cancer who are candidates for immunotherapy often receive heavy chemotherapy beforehand and have implantable devices like Ommayas, raising their risk for CNS infections.¹² Therefore, it is crucial to rule out CNS infections during the assessment of neurotoxicity.^13,14 Early identification and intervention are critical, as unchecked neurocognitive deficits can significantly affect future functional abilities.^15,16

Pediatric patients receiving immunotherapy benefit from caregivers working closely with a multidisciplinary team that includes neuro-oncologists, neurologists, neurosurgeons, pediatric intensivists, and other related specialists like endocrinologists and infectious disease physicians. The neurological exam is a crucial diagnostic tool, as subtle symptomatic changes often appear earlier than radiographic findings, which can lag behind, especially in TIAN.⁸ Currently, there is no single symptom, laboratory finding, or imaging modality that can diagnose immunotherapy-related neurotoxicity, as this is a complex condition that requires consideration of multiple factors. There are efforts to incorporate key parameters including maximum daily temperature, C-reactive protein, interleukin (IL)-6, and procalcitonin to develop predictive models for ICANS in adults.¹⁷ Future studies are needed to create similar predictive models specifically for children.

Long-term follow-up and ongoing assessment for neurologic functions in pediatric cancer patients are urgently needed, requiring strong cooperation with caregivers. Specific cellular injuries and neuroinflammation in the CNS have been linked to cognitive impairments that may progress to dementia.^18,19 In adults receiving B-cell Maturation Antigen (BCMA)-CAR T therapy, a hypokinetic movement disorder with features of parkinsonism has been observed in some patients weeks to months after infusion.²⁰ Given that children’s brains are still developing, the impact of acute and chronic neuroinflammation following immunotherapy on future neurologic functions remains poorly understood. Several cerebrospinal fluid (CSF) biomarkers, including total tau, chitotriosidase, and glial fibrillary acidic protein, have proven useful in identifying survivors at risk for abnormal neurodevelopment after chemotherapy.²¹ Since immunotherapy is likely to induce central neuroinflammation, particularly in brain tumor patients, it is essential to prioritize long-term neurologic evaluations in the design of novel immunotherapy trials. Moreover, efforts should focus on identifying and evaluating appropriate biomarkers to early identify patients at higher risk for abnormal neurodevelopment.

Grading Systems

There are various grading tools available for assessing immune-related neurotoxicity in both clinical trials and clinical practice; however, these grading systems are not consistent across different settings. The Common Terminology Criteria for Adverse Events (CTCAE) is a widely used tool for assessing adverse events, including neurotoxicity. It categorizes toxicities into several grades: in general, grade 1 indicates asymptomatic and mild disease; grade 2 signifies moderate disease that limits instrumental activities of daily living; grade 3 requires hospitalization; grade 4 denotes life-threatening situations necessitating urgent intervention; and grade 5 represents death.²² The CTCAE is user-friendly but sometimes falls short of accurately reflecting the true extent of immunotherapy-related neurotoxicity.^23–25 Thus, various organizations have devised specialized grading systems to characterize specific toxicities more accurately. For instance, the American Society for Transplantation and Cellular Therapy (ASTCT) introduced a grading system for ICANS, whose grade is determined by the most severe event not attributable to other causes (Table 1).⁷ Owing to differences in developmental stages, pediatric patients who are older than 12 years are assessed using the immune effector cell-associated encephalopathy score, whereas those who are 12 years old or younger or who are developmentally delayed are assessed using the Cornell Assessment for Pediatric Delirium (CAPD) score together with level of consciousness, seizure occurrence, motor findings or elevated intracranial pressure (ICP), and/or cerebral edema.^26,27 Notably, a change in the CAPD score from baseline is one of the earliest indicators of ICANS.²⁸ In the JULIET trial, the ASTCT’s scale of ICANS offered a more accurate assessment of tisagenlecleucel, a CAR T cell therapy, than the CTCAE did.²⁵ However, because most experience with cellular therapy, including its toxicity, is based on the systemic administration of intravenous cellular therapy for extracranial tumors, new grading systems for the toxicity of CNS tumors are needed. Mahdi et al. recently created a grading system to describe tumor inflammation-associated neurotoxicity (TIAN), a toxicity syndrome that is distinct from CRS and ICANS, in patients who receive immunotherapy for CNS tumors⁸ (Table 2). TIAN has 2 subcategories; type 1 TIAN primarily presents with acute neurological symptoms and signs due to mechanical factors, including elevated ICP, whereas type 2 TIAN reflects local neural dysfunction.^8,29 Of note, the concept of TIAN is still in development and requires further validation. Most data on TIAN comes from ACT, but it can also apply to other immunotherapies, including ICIs, OV, and vaccines. In the context of brain tumors, it can often be hard to distinguish between tumor progression, pseudo-progression, and various types of neurotoxicity in the acute phase.^29–31 We often need to wait to safely diagnose the underlying condition or a possible tumor progression. Further investigations into additional biomarkers are needed to better guide our treatment approaches.

Table 1.

Grading and Management of Neurotoxicity in Patients Receiving Cellular Therapy^*

Neurotoxicity domain	Grade 1	Grade 2	Grade 3	Grade 4
ICE score (age ≥ 12 y)	7–9	3–6	0–2	0 (unarousable/unable to perform ICE)
- Or -
CAPD score (age < 12 y)	1–8	1–8	≥9	Unable to perform CAPD
Depressed level of consciousness	Awakens spontaneously	Awakens to voice	Awakens only to tactile stimulus	Unarousable or requires vigorous/repetitive tactile stimulus; stupor or coma
Seizure	NA	NA	Any clinical seizure (focal or generalized) that resolves rapidly; or non-convulsive seizures on EEG that resolve with intervention	Life-threatening prolonged seizure (> 5 min); or repetitive clinical or electrical seizures without return to baseline in between
Motor findings	NA	NA	NA	Deep focal motor weakness such as hemiparesis or paraparesis
Raised ICP/cerebral edema	NA	NA	Focal/local edema on neuroimaging	Diffuse cerebral edema on neuroimaging; decerebrate or decorticate posturing; or cranial nerve VI palsy; or papilledema; or Cushing’s triad

^*Adopted from ASTCT consensus.⁷ Used with permission.

Abbreviations: ICE, immune effector cell encephalopathy; CAPD, Cornell Assessment for Pediatric Delirium; NA, not available; ICP, intracranial pressure.

Table 2.

Grading Scale for Tumor Inflammation-Associated Neurotoxicity

Grade	Definition
1	Headache with fever OR mild worsening from the baseline neurological function with only supportive symptomatic management required.
2	Moderate neurological changes from baseline that substantially affect function
3	Severe neurological clinical signs and symptoms that may affect critical cardiorespiratory functions. OR clinical signs and symptoms of increased intracranial pressure (>20 mmHg) that are responsive to intervention
4	Life-threatening, clinically significant elevated ICP (>20 mmHg) refractory to CSF drainage with no improvement in clinical symptoms in response to CSF drainage, possibly warranting urgent escalation of neurosurgical intervention (such as with emergent EVD or VPS placement) OR concerning clinical signs and symptoms of impending/early herniation OR severe medullary dysfunction requiring endotracheal intubation for airway protection and/or mechanical ventilation

^*Adapted from Mahdi J, Dietrich J, Straathof K, et al. Tumor inflammation-associated neurotoxicity. Nat Med. 2023;29(4):803-810. doi:10.1038/s41591-023-02276-w⁸ Used with permission.

Abbreviations: ICP, intracranial pressure; CSF, cerebrospinal fluid; EVD, extraventricular drain; VPS, ventriculoperitoneal shunt.

In summary, immunotherapy-associated neurotoxicity has a wide range of symptoms that are influenced by multiple factors, including the immunotherapy type, delivery method, dosing regimen, and patient characteristics, including age and developmental stage. Unraveling the mechanisms underlying these neurological toxicities and establishing standardized grading systems tailored to each type of immunotherapy can ensure early recognition and prompt intervention, effectively mitigating potential neurotoxic effects.

Adoptive Cellular Therapy

ACT harnesses the power of immune cells, such as T cells or natural killer cells, to target and combat tumor cells. There are 3 main types of ACT: tumor-infiltrating lymphocytes, lymphocytes with genetically engineered T-cell receptors, and CAR-transduced cells.³² To date, most clinical data and notable successes in adoptive cellular therapy stem from the use of CAR T cells, particularly in patients with hematologic diseases. Toxicities related to the use of CAR T cell therapy in pediatric patients vary based on disease type, CAR T cell type, dosage, and delivery methods. Here, we summarize the pediatric trials using CAR T cells for various diseases, targets, dosages, and the associated rates of CRS and neurotoxicity in Table 3. Most of these trials enroll patients with hematologic diseases and involve systemic administration. Only a few include patients with active intracranial disease.^35,37,48,50 The CAR T-cell dosage is typically 1–10 × 10⁶/kg or 1–100 × 10⁶/m². In most trials, patients undergo lymphodepletion. CD19 is the most common target in patients with B-cell acute lymphoblastic leukemia, whereas disialoganglioside (GD2) and HER2 are popular targets in patients with solid tumors including brain tumors. As seen in Table 3, the rate of any-grade CRS is around 80%, and that of grade 3 or higher CRS is around 20%.^{2,35,37,33,34,36,38,39,41,52} The larger trials in this table have also demonstrated an overall neurotoxicity rate of around 40%, and grade 3 or higher neurotoxicity rates of 0-28%.^{2,35,37,33,34,36,38,39,41,52}

Table 3.

Neurotoxicity in Clinical Trials of Adoptive Cell Therapy in Pediatric Patients

Study	NCT number	Phase	Patient number	Age, y	Target disease	Intra-cranial disease included	Lympho-depletion	CAR target	Co-stimulator	CAR T dosage, × 106/kg	MTD reached a	RP2D, × 106/kg	CRS %, all grades (≥grade 3)	Neuro-toxicity %, all grades (≥grade 3)
Systemic administration
Levine JE, et al.33	2435849, 2228096	2	137	3–21	B-ALL	No	Yes	CD19	CD28	<50 kg: 0.2-5.0 >50 kg: 0.1-2.5	NA	NA	79 (22)	36 (10)
Lee DW, et al.34	01593696	1	21	5–27	B-ALL	No	Yes	CD19	CD28	DL1: 1 DL2: 3	Yes	1	76 (29)	29 (5)
Gardner RA, et al.35	02028455	1/2	45	1.3–25.3	B-ALL	Yes	Yes	CD19	D28, 4-1BB	DL1: 0.5 DL2: 1 DL3: 5 DL4: 10	Yes	1	93 (23)	49 (21)
Shah NN, et al.36	02315612	1	58	4.4–30.6	B-ALL	No	Yes	CD22	CD28	DL1: 0.3 DL2: 1 DL3: 3	No	3	86 (10b)	33 (2)
Wang T, et al.37	ChiCTR 2000032211	2	225	<20	B-ALL	Yes	Yes	CD19/ CD22	CD28, 4-1BB	CD19 CAR T: 5.6 (4.1-7.6)c CD22 CAR T: 2.8 (2.1-4.0)c	NA	NA	88 (29)	21 (5)
Benjamin R, et al.38	02808442, 02746952	1	21	0.8–62	B-ALL	No	Yes	CD19	4-1BB (UCAR19)	1.1-2.3	No	2.3	91 (14)	38 (0)
Fry TJ, et al.39	02315612	1	21	7–30	B-ALL	No	Yes	CD22	4-1BB	DL1: 0.3 DL2: 1 DL3: 3	Yes	1	76 (0)	38 (0)
Shalabi H, et al.40	03448393	1	20	5.4–34.6	B-ALL	No	Yes	CD19/ CD22	CD28, 4-1BB	DL1: 0.3 DL2: 1 DL3: 3	No	3	50 (15)	5 (5)
Curran KJ, et al.41	01860937	1	25	1–22.5	B-ALL	No	Yes	CD19	CD28	DL1: 1 DL2: 3 DL1: 10 DL2: 30	Yes	NA	80 (16)	72 (28)
Heczey A, et al.42	03294954	1	12	2–12	NB	No	Yes	GD2 CAR NKT	CD28	DL1: 3 DL2: 10 DL3: 30 DL4: 100d	No	100d	8 (0)	8 (8)
Del Bufalo F, et al.43	03373097	1/2	27	2.7–18.6	NB	No	Yes	GD2	CD28, 4-1BB	DL1: 3 DL2: 6 DL3: 10	No	10	74 (4)	0 (0)
Straathof K, et al.44	02761915	1	12	4–15	NB	No	Yes	GD2	CD28	DL1: 10 DL2: 100 DL3: 1000d	No	1000d	50 (8)	17 (0)
Heczey A, et al.45	01822652	1	11	4.1–23.6	NB	No	Cohort 1: No 2: Yes 3: Yes	GD2	CD28, OX40 with PD1	C1: 10~100 C2: 100~150 C3: 150d	No	NA	9 (0)	90e (18)
Kaczanowska S,46	02107963	1	13	8–28	NB, osteosarcoma	No	Yes	GD2	CD28, OX40	DL1: 0.1 DL2: 1 DL3: 3 DL4: 10	No	10	31 (0)	77 (0)
Ahmed N, et al.47	00902044	1/2	19	7.7–29.6	Sarcoma	No	Yes	HER2	CD28	0.01, 0.03, 0.1, 1, 3, 10, 30, 100dd	No	100d	0	32 (0)
Ahmed N, et al.48	01109095	1	17	10–69	GBM	Yes	No	HER2	CD28	1, 3, 10, 30, 100d	No	100d	0	18 (0)
Hedge M, et al.49	00902044	1	13	4–55	Sarcoma	No	Yes	HER2	CD28	Cohort A and B: 100d (T cells) Cohort C: 100d (CAR+ T cells)		100d (T cells)	79 (14)	0(0)
Lin FY, et al.50	04099797	1	11	4–18	Brain tumors	Yes	Yes	GD2	41BB	DL1: 10 DL2: 30d (CAR+ T cells)	No	30d (CAR+ T cells)	55(9)	64(0)

Study	NCT number	Phase	Patient number	Age range, y	Target disease	Lympho-depletion		CAR target	Co-stimulator	Delivery route	Dosage		CRS %, all grades (≥grade 3)	Neuro-toxicity %, all grades (≥grade 3)
Locoregional administration
Monje M et al.29,51	04196413	1f	11	4–30	H3K27M-mutated midline gliomas	Yes g		GD2	4-1BB	i.v., i.c.v	i.v. DL1: 1 × 106/kg DL2: 3 × 106/kg i.c.v. DL1: 10-30 × 106 with multiple infusions for some patients		i.v.:100 (36) i.c.v: 67 (11)h	i.v.:100 (55)i i.c.v:100 (44)i
Vitanza NA, et al.30	03500991	1f	3	15–26	CNS tumors	No		HER2	4-1BB	Intra-tumoral, i.c.v.	DL1: 1 × 107 DL2: 2.5 × 107 with multiple infusions		0	100 (66)
Vitanza NA, et al.31	04185038	1f	5	5–22	DIPG, HGG	No		B7-H3	4-1BB	Intra-tumoral, i.c.v.	DL1: 1 × 107 with multiple infusions		75 (0)j	100 (50)j

Abbreviations: CAR, chimeric antigen receptor; MTD, maximum tolerated dose; RP2D, recommended phase 2 dose; B-ALL, B-cell acute lymphoblastic leukemia; NA, not available; AML, acute myeloid leukemia; NB, neuroblastoma; GBM, glioblastoma multiforme; i.v., intravenous; i.c.v., intracerebroventricular; CNS, central nervous system; DIPG, diffuse intrinsic pontine glioma; HGG, high-grade glioma.

^aPhase 1 trials only.

^bTwenty-four percent patients had grade 3 or higher CRS per American Society for Transplantation and Cellular Therapy CRS consensus grading retrospectively.

^cValues are median (interquartile range).

^dDosages are × 10⁶/m².

^eMostly peripheral neurotoxicity.

^fOngoing.

^gLymphodepletion was not given before some i.c.v. infusions.

^hData were available for 9 patients.

ⁱBoth TIAN and ICANS were included.

^jData were available for 4 patients.

Given the growing interest in extending ACT to solid tumors, research into its specific adverse events is increasing. Among these adverse events, neurotoxicity, including ICANS, is one of the most significant concerns, given its risk of death.²⁶ ICANS commonly develop just after CRS resolution, but it can also develop concurrently with CRS within the first 5 days after CAR T-cell infusion.^2,53,54 Delayed-onset ICANS (occurring > 3 weeks after infusion) have been described in roughly 10% of patients.⁵⁵ It usually presents as toxic encephalopathy and can start with dysphasia, aphasia, impaired fine motor control, and confusion; it can progress to seizure, motor weakness, cerebral edema, and coma.^55–57 In most cases, ICANS is reversible and transient.⁵⁸ The pathophysiology of ICANS is still poorly understood. On the basis of data from clinical trials and animal models, researchers have proposed that ICANS starts with the increased secretion of pro-inflammatory cytokines from CAR T cells and the activation of bystander immune cells.⁴⁹ ICANS is likely the result of cumulative effects of inflammation, endothelial activation, systemic capillary leak, and subsequent blood-brain barrier dysfunction.⁵⁹ Importantly, Myeloid cell-driven inflammation is believed to play a key role in the development of ICANS.⁵⁹Clinical evidence shows that an increase in myeloid cells and myeloid cell-derived cytokines, such as IL-1 and IL-6, is correlated with the development of severe ICANS.^56,60Additionally, granulocyte-macrophage colony-stimulating factor (GM-CSF)-mediated stimulation of monocytes after CAR T cell therapy has been linked to neuroinflammation in mice,⁶¹ and GM-CSF was the cytokine most strongly associated with ICANS development in the ZUMA-1 clinical trial.⁵³ Moreover, severe ICANS has been associated with elevated levels of MCP-1, IP-10, IL-6, and IL-8, indicating activation of macrophages and microglia.^57,59A recent study identified the TGFβ-activated kinase-1 (TAK1)-NF-κB-p38 MAPK pathway as key to microglia activation, laying the foundation for potential using TAK1 inhibition in ICANS prevention.⁶² High-risk patients must be identified and monitored for symptoms. A higher baseline inflammatory state (with elevated c-reactive protein, ferritin, and d-dimer), a higher tumor burden, and more intense conditioning regimens are related to higher rates of ICANS.^58,60,63,64 Occasionally, CAR T cells can elicit a hyperinflammatory state that eventually causes immune system dysfunction, resulting in damage to end organs, termed “secondary hemophagocytic lymphohistiocytosis (HLH)/macrophage activation syndrome.”^{36,55,65–67} Sharing similarities with CRS, CAR T cell-related HLH may fall within the same spectrum of systemic hyperinflammatory syndromes. However, it tends to be more aggressive and is usually refractory to treatment compared to CRS.⁵⁵ CAR T cell-related HLH is rare, occurring in approximately 1% of patients who receive CAR T-cell therapy, but it carries a high mortality rate if not treated promptly.^68,69 CAR T cell-related HLH can manifest as CNS disturbances.⁵⁵ Direct evidence for managing CAR T cell-related HLH, particularly for neurotoxicity, is scarce. However, because of the high mortality rate, treatment with intrathecal cytarabine with or without hydrocortisone should be considered.⁵⁵

In addition to causing ICANS, ACT for intracranial tumors can cause TIAN. TIAN is defined as the neurological symptoms related to CAR T cell-mediated inflammation in sites of CNS disease. As indications for immunotherapy expand in brain cancer, cases of TIAN are likely to increase. The diagnosis of TIAN can be challenging. Although it is sometimes associated with T2/FLAIR changes on MRI, radiographic changes may be delayed or even absent. Type 1 TIAN is related to acutely elevated ICP, which is most likely due to inflammation-induced edema and subsequent communicating or obstructive hydrocephalus. This emergent type of TIAN can be life-threatening if not managed immediately and appropriately, and treating institutions should maintain adequate neuro-intensive care and appropriate clinical management algorithms. Type 2 TIAN is often milder, more transient, and related to the inflammation of the tumor-affected brain or spinal cord, resulting in local neural dysfunction. However, neural dysfunction in areas such as the lower brainstem or in medullary tumors is critical, as it can compromise cardiopulmonary function and other vital physiological processes. Its management is usually conservative and depends on the affected neurological functions. Because TIAN is caused by the in-site inflammation of the brain tumor, its onset largely depends on the delivery route of immunotherapy delivery. Type 2 TIAN can happen within 48 hours after intracerebroventricular administration, and most cases of TIAN in patients receiving intravenously administered CAR T cells occur concurrently with CRS 6–7 days after administration.²⁹

In patients who are suspected of having TIAN, measuring ICP is critical, and CSF drainage can usually reverse neurological deficits if accomplished in a timely manner. In a recent trial of GD2-targeting CAR T-cell therapy for diffuse intrinsic pontine glioma (DIPG), all patients underwent Ommaya reservoir placement for ICP monitoring and CSF drainage.²⁹ It is important to identify high-risk patients in this population. For example, the trial of GD2-targeting CAR T-cell therapy for DIPG excluded patients with bulky disease in thalamic or cerebellar regions to reduce the risk of TIAN.²⁹For patients with mild symptoms and no increase in ICP, treatment should focus on addressing specific neurologic deficits and decreasing the effects of pro-inflammatory cytokines.²⁹ Corticosteroids should be initiated promptly for life-threatening swelling, including obstructive hydrocephalus. Although distinguishing between ICANS and TIAN often poses challenges, particularly in patients with intracranial tumors, clinicians must promptly recognize the neurotoxicity and the underlying cause, so that appropriate treatment can quickly be applied.

There is a growing interest in exploring the locoregional delivery of cellular therapy to brain tumors to enhance efficacy. Currently, the only available data regarding the use of locoregional CAR T-cell therapy in pediatric patients with brain tumors are preliminary data from several phase 1 studies (Table 3). Notably, some patients in these studies who underwent locoregional delivery did not undergo lymphodepletion.^29–31 The trials’ limited sample sizes make drawing definitive conclusions about toxicity challenging; however, they have generally shown lower rates of systemic hyperinflammation, with fever being the most common symptom observed when compared to the systemic administration of CAR T cells.^29–31 Different targets of CAR T cell therapy may result in different levels of neurotoxicity. For instance, in the trial using both intravenous and intrathecal GD2 CAR T cell therapy, most patients experienced significant neurotoxicity that required intensive monitoring and management in the intensive care unit.²⁹ In the trial using CAR T cells targeting HER2, B7-H3, and EGFR, the patients were reported to have less intense neurotoxicity, allowing many of them to receive multiple doses of intracranial CAR T therapies in the outpatient setting.^30,31 However, given the limited data available and the complexity of different diseases treated and varying dosages, it is challenging to conclude that one target is definitely more toxic than another.^29–31 Even with the same target, such as GD2, differences in product construct, dosage, administration route, and lymphodepletion regimen can result in varying neurotoxicity patterns.^29,50 Complicating matters further, because all patients enrolled in these studies had brain tumors, distinguishing among neurotoxicities stemming from ICANS, TIAN, or disease progression is exceedingly difficult.^29–31 Further research derived from larger clinical trials is urgently needed to standardize evaluation and reporting practices and to develop more consistent management recommendations.

Immune Checkpoint Inhibitors

ICIs are a novel class of agents that have shown great antitumor effects against several human malignancies. ICIs block negative regulation of T-cell activation and further enhance the antitumor immune response. Owing to their special antitumor mechanism, ICIs generate a wide range of immune-related adverse eventes (irAEs). Neurological toxicities are of special interest because of their potential severity.²⁴ The mechanism of ICI-related neurotoxicity remains unclear due to the limited number of cases. Postmortem autopsy findings from patients with ICI-related encephalitis often reveal a T cell-dominated inflammatory process with perivascular lymphocytic infiltration, together with reactive astrogliosis, and microglial activation.⁷⁰

In general, ICI-related irAEs typically occur 6-12 weeks after ICI initiation and respond well to corticosteroid treatment.²⁴ Common ICIs include anti-programmed death-1 (PD-1) antibodies, antibodies against their ligand PD-L1, and cytotoxic T lymphocyte antigen-4 (CTLA-4) antibodies. Reports indicate that anti-PD-1/PD-L1 therapy is associated with higher rates of encephalitis and neuromuscular junction dysfunction, such as myasthenia gravis, but with lower rates of Guillain-Barré syndrome and meningitis compared to anti-CTLA-4 therapy.⁷¹

Neurologic irAEs related directly to ICIs are relatively uncommon but a wide range of neurological symptoms has been reported in these trials as seen in Table 4. In one trial, 6%–12% of patients receiving an ICI had mild, nonspecific neurologic symptoms (headache, dizziness, and sensory impairment), and fewer than 1% had grade 3 or higher neurologic irAEs.¹³ The 2 most severe forms of CNS toxicity are encephalitis and meningitis. (ICIs can also exacerbate demyelination, especially in patients with a history of multiple sclerosis, but the relationship and mechanisms are unclear.^13,84) Encephalitis, which occurs in 0.1%–0.2% of patients receiving ICIs,¹³ often presents as seizure, confusion, ataxia, abnormal behavior, or alterations of consciousness. Meningitis usually emerges 1–7 weeks after ICI initiation and almost always first manifests as fever. Most CSF analyses show lymphocytic meningitis and elevated proteins, and most cases respond to steroids with complete recovery from meningitis.¹³

Table 4.

Neurotoxicity From Other Immunotherapies in Children With Cancer

Drug	Study	Phase	Target	Patient number	Age range, y	Target disease	Intra-cranial disease included	Delivery route	Dosage	RP2D	Grade 1 or 2	Grade 3-5
Ipilimumab	Merchant MS, et al.72	1	CTLA4	33	2–21	Solid tumors	No	i.v.	1, 3, 5, or 10 mg/m2 every 3 weeks	NA	None listed	Headache 3% Vision changes 3%
Nivolumab	Davis KL,et al.73	1/2	PD1	75	1–30	Solid tumors and lymphoma	No	i.v.	3 mg/kg every 2 weeks	3 mg/kg every 2 weeks	Headache 15%	CNS dysfunction 1.3%
Pembrolizumab	Geoerger B, et al.74	1/2	PD1	154	0.5–17	Solid tumors and Lymphoma	Yes	i.v.	2 mg/kg every 3 weeks	2mg/kg every 3 weeks	Ataxia 0.6% Headache 20.8% Hemiparesis 1.9% Neuralgia 0.6 Seizure 4.5% Somnolence 2.6% Tremor 0.6% Vision changes 3.2%	Ataxia 0.6% Balance disorder 0.6% Depressed LOC 0.6% Headache 1.3% Hemiparesis 0.6% Ischemic stroke 0.6% Neuralgia 0.6% Opisthotonos 0.6% Seizure 1.3% Somnolence 0.6% Spinal cord compression 1.2%, Tremor 1.2% Vision changes 0.6%
Atezolizumab	Geoerger, B. et al.75	1/2	PDL1	87	10–17	Solid tumors and lymphoma	No	i.v.	15 mg/kg every 3 weeks	15mg/kg every 3 weeks	Agitation 1.1% Headache 19.5%, Neuralgia 1.1%, Paresthesia 2.3%,	Agitation 2.2% Cranial nerve VI Palsy 1.1% Headache 1.1% Hydrocephalus 1.1% Migraine 1.1% Neuralgia 2.2% Papilledema 1.1% Paresthesia 1.1%
Blinatumomab	Locatelli F. et al.76	3	CD19	54	1–17	B-cell ALL	No	i.v.	5 μg/m2/d for 4 weeks, continuous intravenous infusion	N/A	Agitation 7.4% Anxiety 3.7% Depression 3.7% Dizziness 1.9% Encephalopathy 1.9% Headache 35.2% Irritability 1.9% Nervous system disorder 1.9% Neurological symptom 3.7% Seizure 1.9% Tremor 9.3%	Nervous system disorder 1.9% Neuralgia 1.9% Seizure 1.9%
Blinatumomab	Brown PA. et al.77	3	CD19	102	5–16	B-cell ALL	No	i.v.	2 continuous 28-day infusion cycles of blinatumomab, 15 µg/m2 per day	N/A	Encephalopathy 11% Seizure 4%	Encephalopathy 4% Seizure 1%
Blinatumomab	Locatelli, F. et al.78	N/A	CD19	110	0.4–17	B-cell ALL	No	i.v.	15 µg/m2 per day	N/A	Headache 22.7% Pain in extremity 12.7%	Agitation 0.91% Decreased LOC 0.91% Headache 1.8% Seizure 0.91% Trigeminal nerve disorder 0.91%
Blinatumomab	von Stackelberg, A. et al.79	1/2	CD19	70	1–17	B-cell ALL	No	i.v.	5 µg/m2/d for the first 7 days, followed by 15 µg/m2/d thereafter		Ataxia 1% Atonic seizure 1% Convulsion 3% CSF leak 1% Decreased LOC 1% Dizziness 4% Dysgeusia 1% Encephalopathy, 1% Headache 30% Hypesthesia 1% Mental disorder 1% Nystagmus 1% Pain in extremity 11% Paresthesia 3% Somnolence 1% Syncope 1% Tremor 6%	Confusional state 1% Neuralgia 1% Somnolence 3%
Dinutuximab	Mody, R. et al.80	2	GD2	53	0.2–13.3	Neuroblastoma	No	i.v.	17.5 mg/m2 per day IV on days 2–5	N/A	Not given	Peripheral neuropathy 2%
Dinutuximab	Hara, J.et al.81	1/2	GD2	25	2–14	Neuroblastoma	No	i.v.	17.5 mg/m2	17.5 mg/m2	Not given	Pain 4%
Dinutuximab	Yu AL. et al.82	3	GD2	137	0–31	Neuroblastoma	No	i.v.	25 mg/m2/d for 4 days	N/A	Not given	CNS cortical symptoms 3.65% Neuropathic pain 52% Seizure 0.73%
Dinutuximab	Handgretinger R83	1	GD2	9	2–10	Neuroblastoma	No	i.v.	30,40 and 50 mg/m2/day for 5 days	50 mg/m2/day	Not given	Optic atrophy 11.1 % Pain 84%, Pupil atonia 11.1%,

Abbreviation, RP2D, Recommended Phase 2 Dose; CTLA4, Cytotoxic T-Lymphocyte-Associated Protein 4; i.v., Intravenous; N/A, Not Applicable; CNS, Central Nervous System; PD1, Programmed Cell Death Protein 1; B-cell ALL, B-cell Acute Lymphoblastic Leukemia; CSF, Cerebrospinal Fluid; LOC, Level of Consciousness.

ICIs can also cause neurotoxicity in the peripheral nervous system. The spectrum of peripheral neurotoxicities includes mild to moderate peripheral neuropathies, Guillain-Barré syndrome, necrotizing myositis, and myasthenic syndromes.¹³ Infection and tumor invasion should be excluded as confounding diagnoses. Complete neurological examination with necessary investigational studies, including CSF analyses and MRI, are recommended based on symptoms. The management of this type of irAE usually includes supportive care, and corticosteroids should be considered for significant (grade 3 or 4 irAEs). The diagnostic process should not delay the use of corticosteroids, and the response to corticosteroids can have diagnostic value.¹³

Neuroendocrine toxicity, particularly hypophysitis, is a unique and challenging irAE of ICIs in pediatric patients due to its impact on neurodevelopment. Hypophysitis and subsequent panhypopituitarism are reported to be less common in children (around 3%) compared to adults.⁸⁵ Adult data shows that different ICIs present distinct patterns.⁸⁶CTLA-4 inhibitors often cause both isolated ACTH deficiency (IAD) and hypophysitis within months of treatment, while anti-PD-1/PD-L1 antibodies primarily cause IAD, which can develop several months to over a year after treatment.⁸⁶ Pediatric patients may be asymptomatic or have nonspecific symptoms, making early and comprehensive hormonal evaluation by an endocrinologist essential.⁸⁵

Antibody-Based Therapies

Blinatumomab

Blinatumomab is a biphasic fusion protein with 2 single-chain variable fragments arranged in tandem. One of the fragments targets the CD3 antigens of T cells, and the other targets CD19 surface antigens on B-cell leukemia cells. As a bispecific T-cell engager, blinatunomab can induce B-cell apoptosis. The mechanism behind neurotoxicity from blinatumomab is still under active investigation. Klinger et al. proposed a 2-step process for its development, involving a B-cell-independent redistribution of peripheral T cells, together with B-cell-dependent T cell activation and cytokine release.⁸⁷

The prevalence of blinatumomab-induced neurotoxicity is approximately 30% in the pediatric population.⁷⁹ The most common presentation of neurotoxicity includes headache and tremor, but it can also present as dizziness, somnolence, disorientation, seizure, and encephalopathy (Table 4). Symptoms usually occur during the first few days after receiving the drug or after dose escalation.^79,88,89 Grade 3 or 4 neurotoxicity includes seizure, somnolence, and encephalopathy and it usually resolves after blinatumomab discontinuation with adequate supportive care.^79,88,89 In the phase 2 TOWER trial, 25 of the 267 adults (9.4%) receiving blinatumomab had grade 3 or higher neurotoxicity.⁹⁰ In contrast, von Stackelberg et al. found that 24% of pediatric patients receiving the drug had neurologic/psychiatric events, whereas only 4% had grade 3 or higher neurologic events, including somnolence and neuralgia.⁷⁹ The significance of neurotoxicity for patients with acute lymphoblastic leukemia and active CNS disease who receive blinatumomab is unclear since active CNS disease was an exclusion criterion in clinical trials.^{76,77,79,88,91,92} The pathophysiology of blinatumomab-induced neurotoxicity is unclear; Klinger et al. have proposed that the adhesion of T cells to endothelial cells might be associated with neurological events.^76,77,87

For pediatric patients, the usual blinatumomab dose is 5–15 µg/m²/day.⁷⁹ In a study of patients with relapsed, refractory non-Hodgkin lymphoma, the blinatumomab dose was 0.5–90 µg/m²/day; neurological events were reported to be dose-limiting and 60 µg/m²/day established as the maximum tolerated dose.⁸⁹ Because blinatumomab-related neurotoxicity often resolves after the interruption, de-escalation, or discontinuation of the drug, the main treatment for neurotoxicity is supportive care and, if the neurotoxicity is severe, systemic steroids. The concurrent use of intrathecal chemotherapy with dexamethasone might help prevent blinatumomab-related neurotoxicity; however, neurotoxicity has also been attributed to the concurrent use of blinatumomab and intrathecal chemotherapy, so the efficacy and safety profile of the combination requires further investigation.^93,94

Dinutuximab

Dinutuximab is a chimeric anti-GD2 monoclonal antibody. Strong surface expression of GD2 is found on pediatric tumor cells including; neuroblastoma, diffuse midline glioma, melanoma, and osteosarcoma, whereas normal pediatric and adult CNS, peripheral nerves, and skin melanocytes have only weak expression of GD2.⁹⁵ Dinutuximab was the first anti-GD2 antibody approved by US Food and Drug Administration (FDA) for the treatment of high-risk neuroblastoma, in 2015, and has mainly been used to treat that disease since.⁹⁶

While GD2 expression is weak in normal tissue, there is still some expression in the nerve system; therefore, neurotoxicity is one of the common side effects of antibodies targeting GD2, regardless of the antibody’s origin (mouse, chimeric, or humanized).⁹⁷ IL-2, a T-cell enhancing cytokine, has been reported to result in neurological side effects, including coma, convulsions, and paralysis,⁹⁸ and the combination of dinutuximab and IL-2 has been reported to increase the risk of severe neurotoxicity.^97,99

In a combined analysis of the HR-NBL1/SIOPEN study and LTI/SIOPEN study (International Society of Pediatric Oncology Europe Neuroblastoma Group), where dinutuximab was added as maintenance therapy either along or with IL-2, 4% of the patients receiving dinutuximab experienced grade 3 or 4 neurotoxicity, and 2.45% of the patients had severe neurotoxicity. However, the neurotoxicity was transient in most patients, as only 0.45% of the patients experienced persistent and severe neurological deficits. The most common severe neurotoxicities were paresis, hypotonia, neurogenic bladder, seizures, ataxia, and cranial nerve palsy, and the most common single symptoms were seizure and cranial nerve disturbance.⁹⁷In another trial, 3 patients who received dinutuximab with concomitant IL-2 and granulocyte-macrophage colony-stimulating factor or chemotherapy developed transverse myelitis, but their symptoms resolved after treatment with steroids, intravenous immunoglobulin (IVIG), and/or plasmapheresis.¹⁰⁰

When neurological symptoms occur in patients receiving dinutuximab (Seen in Table 4), immunosuppression with steroids and IVIG should be strongly considered, and imaging studies and CSF analyses should be performed to identify other possible causes, including infection and disease progression.

Others

Oncolytic Viruses

Oncolytic viruses (OVs) have emerged as one of the promising novel anticancer therapies. The US FDA and the European Medicines Agency have approved talimogene laherparepvec, a genetically modified herpes simplex virus type 1 (HSV-1), for the treatment of melanoma. More recently, several viruses have been tested in children with brain tumors including intratumoral HSV-1, adenovirus, and poliovirus/rhinovirus recombinant (PVSRIPO), and IV reovirus. Because OVs activate the immunosuppressive tumor microenvironment, researchers have investigated their use for solid tumors, and one of the significant challenges is neurotoxicity. OV-induced neurotoxicity is primarily caused by the localized inflammatory reaction, especially in cases where the virus is infused intratumorally. It can present as headache, pyramidal tract syndrome (impaired voluntary movement and reflex abnormalities due to damage to the corticospinal tracts), dysphasia, cognitive disturbance, confusion, paresthesia, gait disturbance, dizziness, seizure, and facial nerve disorder.^101–103 Most symptoms can be treated with supportive care. In the phase 2 trial of oncolytic polio/rhinovirus recombinant (PVSRIPO), 1 patient died from grade 4 cerebral edema and grade 5 seizure, which may have been attributable to PVSRIPO.¹⁰³The role of prophylactic antiepileptics in patients receiving OVs is unclear.¹⁰³Of note, response assessment to distinguish a true tumor progression from pseudo-progression after intratumoral immunovirotherapy is challenging because tumor enlargement may be associated with necrotic tumor cell death and inflammation caused by an influx of immune cells or even delayed immune responses.¹⁰²

Cancer Vaccines

Cancer vaccination works much like vaccinations against microbial infections. It involves exposing the immune system to tumor-associated antigens and/or tumor-specific antigens in various forms, including dendritic cells, peptides, nucleic acids, and viral vectors, to activate the immune system to target and specifically kill tumor cells.¹⁰⁴ Most cancer vaccines are administered subcutaneously or intramuscularly, but their intravenous administration is a potential method to promote a stronger T cell-mediated immune response.^104,105 The first cancer vaccine, sipuleucel-T, was approved by the US FDA in 2010 for the treatment of prostate cancer. In the phase 3 trial of the vaccine, more than 40 million activated autologous peripheral-blood mononuclear cells were administered IV over approximately 60 minutes, following premedication with acetaminophen and an antihistamine. Most adverse events occurred within 1 hour after infusion, and the symptoms were generally nonspecific and included chills, fever, fatigue, nausea, and headache; without more severe neurotoxicity.¹⁰⁶ Similarly, pediatric phase 1 or 2 vaccine trials of the vaccine have not reported significant neurotoxicity.¹⁰⁴

To date, most cancer vaccines have only modest efficacy, as they tend to elicit weak immunogenic responses, but they are generally well-tolerated. Ongoing clinical trials are investigating combinations of cancer vaccines with other immunotherapies. The efficacy and safety profiles of these combinations have not been established. Most trials of these combinations have been conducted in adults, with only a few enrolling pediatric patients. Notably, an ongoing phase 1 trial (NCT0423040) is investigating the combination of a granulocyte-macrophage colony-stimulating factor cell-based vaccine (GVAX) with nivolumab and ipilimumab in pediatric patients with neuroblastoma.

Combinational Immunotherapy

Researchers are investigating how to combine immunotherapy with other treatment modalities, including radiotherapy and chemotherapy, to achieve maximal efficacy. They are also investigating how to combine different types of immunotherapies. Most data about combinational immunotherapy come from studies of combinations of ICIs. For example, nivolumab and ipilimumab have been used together in the treatment of advanced melanoma. In the adult trial CheckMate 067 trial, the combination of nivolumab and ipilimumab significantly increased the 5-year survival rate to 52% but also raised the incidence of grade 3 or 4 treatment-related toxicities to 59% compared to 23% for nivolumab group and 28% for ipilimumab.¹⁰⁷Neurotoxicity was relatively uncommon, with only a few patients reporting headaches, mostly of mild intensity. Notably, in this trial, patients with active brain metastases were excluded.¹⁰⁷ The combination of nivolumab and ipilimumab was tested in children with extra-cranial solid tumors. It was reported that this combination was generally well-tolerated but did not show the same level of response as seen in adult melanoma patients. The toxicity was dose-related: with nivolumab (3 mg/kg) plus ipilimumab (1 mg/kg), 2 out of 41 patients experienced dose-limiting toxicities, whereas, with nivolumab (1 mg/kg) plus ipilimumab (3 mg/kg), 3 out of 8 patients had dose-limiting toxicities. Notably, none of these toxicities were neurological, with only low-grade headaches and paresthesia reported as neurological symptoms.¹⁰⁸ In the CheckMate 908 trial for pediatric patients with CNS malignancies, nivolumab (3 mg/kg) plus ipilimumab (1 mg/kg) showed a manageable safety profile but did not improve survival. Treatment-related toxicity was one of the common reasons for discontinuation. Neurological treatment-related adverse events of any grade occurred in 16.5% of patients receiving nivolumab alone and 13.6% of those receiving the combination. The most frequent neurological TRAE was headache, reported in 7.1% of patients with nivolumab alone and 6.2% with the combination, mostly grade 1 or 2.¹⁰⁹

Combination immunotherapy can also involve the same immunotherapy that is administered in different ways. For instance, one phase 1 trial tested the concurrent intrathecal and intravenous administration of nivolumab in melanoma patients with leptomeningeal disease. This trial demonstrated that when the same immunotherapy is used, different routes of administration can enhance efficacy but also lead to varying adverse effects.¹¹⁰

Other methods of combination immunotherapy have been explored, but available data are limited, particularly for the pediatric population. There are ongoing investigations into combining immunotherapy with other modalities such as chemotherapy, radiation therapy, and targeted therapy. However, due to the complexity of isolating specific immune-related neurotoxicity and because it is not within the scope of this review, we will not discuss these combinations here.

Acute Management of Immunotherapy-Related Neurotoxicity

Immunotherapy-related neurotoxicity is distinct from chemotherapy- or radiotherapy-related neurotoxicity. Whereas the latter results from direct cellular damage, immunotherapy-induced neurotoxicity stems from an overactivated immune response. Although the mechanisms of immune activation vary depending on the specific agents and administration routes and a clear understanding of the underlying pathophysiology is lacking, the primary management for such neurotoxicity generally involves early intervention with immunosuppression (with corticosteroids), targeted immunomodulation (with agents such as tocilizumab and anakinra), and intensive supportive care. For children with brain tumors who are at high risk for acute increases in intracranial pressure, preemptive implantation of an Ommaya reservoir before immunotherapy should be considered to facilitate close monitoring and decompression.^8,29–31 Because immunotherapy-related neurotoxicity can progress rapidly and be fatal, patients with neurological symptoms should be evaluated promptly by a pediatric neurologist and, if warranted, immediately transferred to intensive care.

As outlined above, different immunotherapy approaches carry different levels of neurotoxicity risk. Although management centers on immunosuppression, there is a lack of uniform guidelines regarding the recommended intensity and duration of corticosteroids and other immunosuppressive agents. Additionally, it is important to recognize that most treatment protocols are primarily designed with adults in mind.

The Society for Immunotherapy of Cancer (SITC) has specific guidelines for managing irAEs in patients receiving ICIs. They recommend starting with prednisone (or its equivalent) at a dosage of 0.5–1 mg/kg/day. For severe or life-threatening neurotoxicity, the SITC suggests a more aggressive approach involving intravenous pulse-dose methylprednisolone at 1000 mg/day for 3–5 days and the consideration of the use of plasmapheresis and IVIG.¹¹¹ These recommendations are generally consistent with guidelines from the American Society of Clinical Oncology and the European Society for Medical Oncology.^112,113 Guidelines specifically for pediatric patients are not yet available.

For ICANS, corticosteroids are the mainstay of treatment.^55,114 For pediatric patients specifically, Ragoonanan et al. recommended using corticosteroids escalating as needed until the condition is reduced to grade 1 or lower. American Society of Clinical Oncology recommended that Dexamethasone is preferred for grade 2 and 3 ICANS, particularly in cases without concurrent CRS, while high-dose methylprednisolone is recommended for the treatment of grade 4 ICANS.¹¹⁵ Some CAR T-cell products are designed with different killing switch systems to enable dose-dependent control of cell products.¹¹⁶There are 3 types of killing switches: caspase-9 suicide gene system (iCas9), ganciclovir-activated, and antibody-dependent cell-mediated cytotoxicity switches. When activated by specific molecules, they initiate programmed cell death, leading to irreversible termination of the immune response.¹¹⁷For refractory grade 4 ICANS lasting more than 24 hours, additional therapies on top of corticosteroids, as well as the activation of killing switches, if available, should be strongly considered.²⁶ These guidelines are primarily derived from the findings of trials of CAR T-cell therapy in patients with extracranial diseases, as detailed in Table 3. However, data on the management of immunotherapy-related neurotoxicity in patients with intracranial disease are sparse. In such patients with TIANs, the inflammation associated with the intracranial disease can escalate rapidly. Owing to the sensitive nature of the location, acute cerebral edema and elevated ICP must be managed, and systemic corticosteroids should be considered. Anakinra, an IL-1 receptor inhibitor, may benefit patients with TIAN.^29,50,51 For patients with CNS tumors in high-risk areas, Mahdi et al. suggested placing an Ommaya reservoir or similar device before initiating immunotherapy, which would enable both monitoring and intervention for elevations in ICP.⁸

Importantly, while steroids can be effective in treating immunotherapy-related neurotoxicity, they must be used cautiously, as they are known to cause immunosuppression and promote lymphocyte apoptosis which may eventually compromise the efficacy of the immunotherapy.^118,119 The optimal dose, timing, and duration of steroids that effectively manage neurotoxicity without significantly impacting treatment efficacy are still under active investigation.¹¹⁸ For ICIs, several studies have indicated that administering steroids for irAEs is not associated with a decrease in overall survival compared to non-steroidal treatments.^120–122 However, early use of steroids has been linked to poorer survival outcomes.^120,123 In the context of ACT, particularly CAR T cell therapy, the impact of corticosteroids on long-term patient outcomes remains controversial.¹²⁰ Some studies suggest that steroids do not affect survival^63,124 even at the higher dose,¹²⁵ while others indicate that steroid administration may negatively impact survival outcomes.^126,127 Research focused on the pediatric population is even more limited, and there are still unanswered questions regarding the long-term persistence of CAR T cells after steroid treatment in children.¹²⁸

As previously mentioned, corticosteroids are a fundamental component of the medical management of immunotherapy-related neurotoxicity. However, a significant proportion of cases are refractory to steroids, and corticosteroids can diminish the effects of immunotherapy.¹²⁸ Consequently, agents with more targeted properties are being developed for use in these situations. Next-generation inhibitors like anakinra have shown some clinical benefits in both adults¹¹⁶ and children¹²⁹ with refractory ICANS, even in the prophylactic setting¹³⁰; however, further validation is necessary.^29,116 The IL-6 receptor inhibitor tocilizumab has been shown to effectively reduce CRS, but its efficacy against ICANS is debatable owing to its limited penetration of the blood-brain barrier.^55–57 Siltuximab, an IL-6 inhibitor, has demonstrated better penetration through the blood-brain barrier and has been shown to be effective for refractory ICANS in adults.¹³¹ In addition, broader-spectrum cytokine inhibitors such as the JAK1/2 inhibitor ruxolitinib and the JAK1 inhibitor itacitinib are being investigated in preclinical models and may one day become viable treatment options.¹³²

Early recognition, intervention with corticosteroids and other immunomodulatory agents, along with intensive supportive care and potential decompression, are key to management. Questions remain regarding identifying high-risk patients, prophylactic strategy, optimal corticosteroid dosing and timing, and when and whether to resume immunotherapy after neurotoxicity.

Conclusion

Immunotherapy is emerging as a promising treatment with significant potential in pediatric malignancies. However, its Immunotherapy-related neurological toxicities, which can be fatal, may limit the application of immunotherapy in pediatric patients. Understanding the differences in the pathophysiology, clinical presentation, and management of these toxicities is crucial, especially for patients with intracranial diseases. Most data and recommendations regarding immunotherapy-related neurotoxicity are derived from trials in adults, and pediatric patients may exhibit distinct patterns of neurological toxicities. Although the pattern of immunotherapy-related neurotoxicity has some core characteristics, it varies considerably across different agents, delivery routes, and dosages. Owing to anatomical factors, the use of immunotherapy for intracranial malignancies poses unique risks, necessitating distinct grading systems and monitoring and management approaches for which current data are limited. The management of immunotherapy-related neurotoxicity hinges on early recognition and intervention and relies on corticosteroids supplemented with other promising but yet-to-be-validated cytokine inhibitors. Combination therapies aimed at enhancing efficacy may introduce unique, potentially more severe neurotoxicities, and this should be taken into consideration when designing future clinical trials.

Funding

J.H. reports grants from ChadTough Defeat DIPG Foundation (Co-funded by Cal’s Angels and the Violet Foundation for Pediatric Brain Cancer) during the conduct of this work. GKF is supported by the U.S. FDA (R01FD006368), CureSearch for Children’s Cancer, V Foundation for Cancer Research, Cannonball Kids’ cancer Foundation, Trial Blazers for Kids, Rally Foundation for Childhood Cancer Research, Andrew McDonough B + Foundation, National Pediatric Cancer Research Foundation, and Pediatric Cancer Research Foundation.

Conflict of Interest statement

Authors declare no conflicts of interest.

Authorship statement

Conceptualization: J.H., J.C., Z.S., J.S, and A.A. Writing of the manuscript drafts: All. Revising manuscripts: All. Preparation of tables and figures: J.H., J.C., A.M., S.X., H.M., G.F. Approving final content of manuscript: all.