Introduction
T-cell–redirected immunotherapy has altered the therapeutic landscape of hematologic malignancies. Chimeric antigen receptor (CAR) T-cell therapy, which redirects T cells through synthetic antigen receptors, has achieved durable remission in patients with otherwise refractory B-cell acute lymphoblastic leukemia, B-cell lymphoma, and multiple myeloma [ 1 – 4 ]. In parallel, T-cell–redirecting bispecific antibodies (BsAbs), particularly CD3 × CD20 antibodies in B-cell lymphoma and BCMA- or GPRC5D-directed antibodies in multiple myeloma, have translated the principle of T-cell redirection into readily available, off-the-shelf treatment platforms [ 5 , 6 ]. These advances have shifted immune effector cell therapy from a specialized experimental approach to a major component of modern hematologic oncology.
The clinical benefit of T-cell–redirected therapy is inseparable from a distinct spectrum of immune-mediated toxicities. Cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS) are the two best-characterized toxicities after CAR T-cell therapy and are also observed after BsAb therapy, although their incidence and severity patterns generally differ [ 7 , 8 ]. CRS reflects systemic inflammatory activation driven by immune effector cell expansion, cytokine amplification, myeloid cell activation, and endothelial injury. ICANS is a neuroinflammatory syndrome that may occur concurrently with CRS, but more often follows it, and is associated with endothelial activation, blood–brain barrier disruption, and cytokine or immune cell accumulation within the central nervous system [ 7 ]. Although most events are reversible with supportive care and timely immunosuppression, severe CRS or ICANS can require intensive care, contribute to non-relapse mortality, and compromise the safe delivery of otherwise highly effective therapies [ 9 , 10 ].
Therefore, management of CRS and ICANS has evolved from reactive treatment of established toxicity to early detection, risk-adapted monitoring, and personalized intervention. The ASTCT consensus grading system established a unified clinical language for CRS and ICANS, enabling consistent assessment across clinical trials and routine practice [ 8 ]. Recently, registry studies, biomarker-based models, and toxicity-focused clinical studies have emphasized that the risk and consequences of CRS and ICANS are influenced by disease burden, baseline inflammation, endothelial vulnerability, CAR construct or antibody platform, patient comorbidities, and the timing and intensity of the intervention [ 9 – 12 ]. In this review, we discuss practical approaches to early detection and personalized management of CRS and ICANS in patients receiving T-cell–redirected therapy, with particular attention to risk stratification, timely intervention, and balancing effective toxicity control with preservation of antitumor efficacy.
Biological basis of CRS and ICANS
Sequential immune activation after immune effector cell therapy
Although the biology of CRS and ICANS has been most extensively characterized in CAR T-cell therapy, the central mechanism is shared across T-cell–redirected platforms: immune effector cells engage with antigen-expressing tumor cells, become activated, and trigger a broad inflammatory cascade. After antigen recognition, CAR T cells undergo local activation and expansion, whereas BsAbs promote the repeated or continuous engagement of endogenous T cells and malignant cells. The magnitude and kinetics of this interaction vary by platform, target antigen, tumor burden, and treatment schedule; however, the downstream sequence commonly involves T-cell activation, cytokine release, myeloid cell recruitment, endothelial activation, and systemic inflammatory amplification [ 7 , 13 , 14 ].
Early cytokine signals such as interleukin (IL)−2, IL-7, and IL-15 support the proliferation and survival of immune effector cells. Subsequently, inflammatory mediators including IL-1, GM-CSF, interferon-γ, and tumor necrosis factor-α promote activation of monocytes and macrophages. Activated myeloid cells produce high levels of IL-6, IL-1, nitric oxide, and additional inflammatory cytokines, thereby amplifying the systemic inflammatory response. These findings highlight the central contribution of monocyte/macrophage activation in CRS pathophysiology [ 7 , 13 , 14 ]. In the broader context of hematologic oncology, macrophage polarization and macrophage-directed pathways are increasingly recognized as therapeutically relevant components of the tumor microenvironment [ 15 ]. This immune cell crosstalk explains why CRS is not simply a direct consequence of T-cell expansion but rather a multicellular process involving tumor cells, effector T cells, monocytes/macrophages, and the vascular endothelium.
Pathophysiology of CRS
CRS is a systemic manifestation of the inflammatory cascade triggered by T-cell redirection. Activated T cells and tumor-cell damage trigger cytokine release and danger-associated molecular signaling, further activating monocytes and macrophages [ 16 ]. Preclinical and translational studies have shown that myeloid cells are not merely bystanders but also central amplifiers of CRS, producing IL-1, IL-6, nitric oxide, and other inflammatory mediators after interacting with activated T cells and injured tumor cells [ 13 , 14 ]. Among these downstream mediators, IL-6 plays a central clinical role, as reflected by the efficacy of IL-6 receptor blockade with tocilizumab. However, IL-6 is only one component of a broader inflammatory network that also includes IL-1, GM-CSF, interferon-γ, tumor necrosis factor-α, monocyte chemoattractant protein-1, and other cytokines and chemokines [ 7 , 17 , 18 ]. This explains why IL-6 blockade may be highly effective in many cases of CRS but insufficient in severe or refractory cases, in which alternative inflammatory pathways remain active.
Endothelial activation is another key feature that links cytokine signaling to the clinical phenotype of CRS. In clinical cohorts of patients receiving CD19 CAR T-cell therapy, severe CRS has been associated with distinct cytokine kinetics and biomarkers of endothelial activation [ 17 ]. Inflammatory cytokines increase vascular permeability and promote capillary leakage, hypotension, hypoxia, coagulopathy, and, in severe cases, multiorgan dysfunction [ 7 , 16 , 17 ]. Thus, the clinical severity of CRS reflects not only the magnitude of immune activation, which is influenced by the tumor burden and baseline inflammation, but also host susceptibility, including endothelial vulnerability and limited organ reserve. These observations provide a biological rationale for early risk stratification and timely intervention before low-grade inflammation progresses to hemodynamic or respiratory compromise.
Pathophysiology of ICANS
ICANS is biologically related to CRS but is not simply a neurologic extension of systemic CRS. It frequently occurs after or during CRS; however, its timing, response to therapy, and clinical manifestations suggest distinct pathogenic features. Endothelial activation and blood–brain barrier disruption appear to be central processes. Systemic inflammation increases endothelial permeability, which facilitates the entry of cytokines and immune cells into the central nervous system. Within the central nervous system, local cytokine accumulation, microvascular injury, and neuroinflammation may lead to encephalopathy, aphasia, tremors, seizures, cerebral edema, and, rarely, fatal neurologic deterioration [ 7 , 19 ].
The distinct biology of ICANS also has therapeutic implications. Tocilizumab is effective for CRS, but has limited penetration into the central nervous system and does not directly suppress many of the inflammatory pathways implicated in ICANS. In some settings, early or prophylactic IL-6 receptor blockade may increase circulating IL-6 levels, raising concerns that it may not prevent, or could theoretically exacerbate, neurotoxicity in susceptible patients. Therefore, corticosteroids remain the mainstay of treatment for clinically significant ICANS, whereas IL-1 blockade and other anti-inflammatory strategies are being explored in refractory or high-risk cases. Understanding these mechanistic differences is essential for personalized management, because the optimal intervention for CRS may differ from that for ICANS [ 7 , 13 , 14 , 19 – 21 ].
Clinical manifestations and grading
Clinical presentation of CRS
CRS is the earliest clinically apparent toxicity associated with immune effector cell therapies. The initial manifestation is fever, which typically occurs within the first several days after CAR T-cell infusion or during the step-up dosing period of BsAbs [ 7 , 8 , 16 ]. Fever may be accompanied by constitutional symptoms such as chills, myalgia, fatigue, anorexia, and malaise. In mild cases, these symptoms may be self-limiting or respond promptly to antipyretics and supportive care. However, CRS can progress rapidly from isolated fever to hypotension, hypoxia, capillary leakage, coagulopathy, and multiorgan dysfunction, reflecting systemic endothelial activation and inflammatory amplification [ 7 , 16 , 17 ].
The timing and severity of CRS vary according to treatment platform, disease burden, target antigen, and product-specific pharmacology. CAR T-cell–associated CRS often follows in vivo T-cell expansion, whereas BsAb-associated CRS frequently occurs during initial or step-up dosing and may attenuate with subsequent doses. Although most CRS events are low grade, severe CRS remains clinically important because it may require vasopressor support, high-flow oxygen, intensive care, and prompt immunosuppressive therapy. Therefore, fever after immune effector cell therapy should not be regarded as a nonspecific symptom alone but as a potential early warning sign of evolving systemic inflammation.
Clinical presentation of ICANS
ICANS is characterized by a broad spectrum of neurologic symptoms ranging from subtle cognitive changes to life-threatening cerebral edema. Early manifestations may include inattention, impaired handwriting, difficulty finding words, expressive aphasia, confusion, tremors, and somnolence [ 7 , 8 , 19 ]. As the neurotoxicity progresses, patients may develop global encephalopathy, seizures, motor weakness, increased intracranial pressure, or coma. Aphasia and impaired attention are particularly characteristic and may precede a more obvious deterioration in arousal.
ICANS often occurs concurrently with or after CRS; however, the two syndromes are not always equally severe. Some patients develop clinically significant ICANS after CRS has improved, and neurologic toxicity may respond differently to anti-cytokine therapy than to systemic inflammation. Clinical and translational studies have associated severe neurotoxicity with high disease burden, robust CAR T-cell expansion, early inflammatory cytokine elevation, endothelial activation, and blood–brain barrier disruption [ 19 , 22 ]. Because early ICANS can be subtle and fluctuating, serial neurologic assessments are essential, particularly in patients receiving platforms associated with a higher neurotoxicity risk.
ASTCT grading system
The ASTCT consensus grading system provides a unified framework for defining and grading CRS and ICANS across clinical trials and routine practice [ 8 ]. The CRS grading is based on three core clinical elements: fever, hypotension, and hypoxia. Grade 1 CRS was defined as a fever without hypotension or hypoxia. Grade 2 CRS includes fever with hypotension that does not require vasopressors and/or hypoxia that requires low-flow oxygen. Grade 3 CRS is characterized by the need for a vasopressor, with or without vasopressin, and/or hypoxia requiring high-flow oxygen or similar noninvasive oxygen support. Grade 4 CRS includes life-threatening hypotension requiring multiple vasopressors and/or hypoxia requiring positive-pressure ventilation.
ICANS grading incorporates the immune effector cell–associated encephalopathy (ICE) score, level of consciousness, seizure activity, motor findings, and signs of elevated intracranial pressure or cerebral edema [ 8 ]. The ICE score is a practical bedside tool for detecting and monitoring encephalopathy in adults undergoing immune effector cell therapy. It assesses orientation, naming, command following, writing, and attention, with a maximum score of 10. A score of 10 indicates no impairment. Grade 1 ICANS generally corresponds to mild encephalopathy, with an ICE score of 7–9; grade 2 to an ICE score of 3–6 or more evident impairment requiring closer monitoring; grade 3 to an ICE score of 0–2, with severe impairment in arousal, seizure, or focal motor findings; and grade 4 to coma, life-threatening seizures, deep focal motor weakness, or evidence of cerebral edema. CRS and ICANS should be graded independently because their timing, clinical course, and optimal management may differ.
Clinical impact and risk prediction
Although advances in supportive care and standardized management algorithms have reduced the incidence of life-threatening CRS and ICANS, these toxicities remain clinically important complications of immune effector cell therapy. Severe events may require intensive care and prolonged hospitalization; however, their clinical significance extends beyond acute resource use. In a large CIBMTR registry analysis of patients with B-cell lymphoma receiving CD19 CAR T-cell therapy, grade ≥ 3 CRS and grade ≥ 3 ICANS were independently associated with inferior overall survival, with adjusted hazard ratios of 1.79 and 1.34, respectively [ 9 ]. Similarly, in a systematic review and meta-analysis of non-relapse mortality after CAR T-cell therapy, CRS and ICANS accounted for 4.7% and 5.2% of reported non-relapse deaths, respectively, although infections remained the predominant cause [ 10 ]. These findings support the clinical importance of preventing severe toxicity while underscoring that the association between CRS/ICANS and survival should be interpreted in the context of baseline disease burden, inflammatory status, and other adverse patient- or disease-related factors.
These observations have encouraged increasing efforts to shift from purely reactive management to proactive risk prediction. An ideal prediction model would identify high-risk patients before clinically significant toxicity develops, allowing closer monitoring, earlier intervention, and individualized supportive care. This approach is relevant as immune effector therapies expand beyond highly specialized inpatient settings and are increasingly administered to older patients or those with substantial comorbidities.
Multiple clinical variables have been recognized as risk factors for severe CRS or ICANS. Among disease-related factors, high tumor burden remains one of the most reproducible predictors, presumably because greater antigen availability can drive more extensive immune activation and cytokine release [ 16 – 18 ]. Baseline systemic inflammation, reflected by elevated inflammatory markers, such as C-reactive protein or ferritin, has also been associated with an increased risk of toxicity [ 11 , 17 , 18 ]. Treatment-related factors further modify the toxicity profiles, including CAR constructs, co-stimulatory domains, target antigens, lymphodepletion strategies, and therapeutic platform [ 7 , 9 , 23 ]. In particular, CD28-based CAR T-cell products generally show earlier expansion kinetics and higher rates of severe CRS and ICANS than 4-1BB-based constructs. In contrast, BsAbs are more commonly associated with lower-grade CRS during the initial step-up dosing and substantially lower rates of severe neurotoxicity [ 7 , 9 , 23 ].
Based on these observations, several prediction models have been proposed to estimate the risk of severe CRS or ICANS before overt toxicity develops. The Endothelial Activation and Stress Index (EASIX) combines lactate dehydrogenase, serum creatinine, and platelet count as surrogates of endothelial vulnerability, and its modified forms incorporating inflammatory markers such as C-reactive protein or ferritin have shown moderate ability to predict CAR T-cell–related toxicity [ 11 , 24 ]. However, the model performance has been inconsistent across cohorts. A recent systematic review and meta-analysis of prediction models for CAR T-cell therapy–related adverse events found that most models relied on overlapping clinical and laboratory variables, including platelet count, C-reactive protein, and IL-6. However, the overall discrimination was only modest, particularly for ICANS, and all included models were judged to have a high risk of bias [ 25 ]. These findings suggest that EASIX-based and related biomarker models are useful for risk awareness but are not yet sufficient as stand-alone decision tools.
Several disease- and platform-specific models have been developed. In the French DESCAR-T registry of patients with large B-cell lymphoma treated with axicabtagene ciloleucel (axi-cel) or tisagenlecleucel (tisa-cel), the CRS prognostic scoring system incorporated bulky disease, platelet count < 150 × 10 9 /L, C-reactive protein > 30 mg/L, and absent or ineffective bridging therapy, whereas the ICANS prognostic scoring system incorporated female sex, platelet count < 150 × 10 9 /L, absent or ineffective bridging therapy, and axi-cel use [ 12 ]. These scores outperformed the EASIX-based approaches in this setting, although a broader validation is needed. A Japanese nationwide registry analysis further proposed CRS-PRE, a pre-infusion model based on estimated glomerular filtration rate (eGFR), ferritin, C-reactive protein, and lactate dehydrogenase, which stratified patients into low-, intermediate-, and high-risk groups for grade ≥ 2 CRS with observed incidences of 2.8%, 26.0%, and 50.0%, respectively [ 26 ]. The inclusion of eGFR is notable because it supports the concept that host organ reserves, not only tumor burden or inflammation, may influence the clinical expression of CRS.
Imaging-based approaches may also contribute to risk assessment. The CAR-T cell-specific International Metabolic Prognostic Index (CAR-IMPI), which incorporates metabolic tumor volume, age, and stage, was primarily developed to refine survival prediction in patients with large B-cell lymphoma receiving CD19 CAR T-cell therapy [ 27 ]. However, patients with higher CAR-IMPI values also had higher rates of grade ≥ 3 CRS, high-grade ICANS, intensive care unit admission, and greater use of supportive therapies such as tocilizumab and corticosteroids [ 27 ]. Although CAR-IMPI should not be viewed as a dedicated CRS or ICANS prediction model, these findings suggest that metabolic tumor burden and host characteristics help identify patients who require closer toxicity monitoring.
Taken together, the current prediction models represent an important step from descriptive risk factors to practical risk stratification; however, their performance remains imperfect, particularly for ICANS. Until they are prospectively validated and integrated with dynamic biomarkers, these tools should complement, rather than replace, clinical judgment in early detection and personalized management.
Early detection and risk-adapted monitoring
Early detection of CRS begins with recognition that fever may represent the first sign of systemic immune activation, rather than a nonspecific post-treatment event. Therefore, temperature, blood pressure, oxygen requirements, fluid balance, and symptoms suggestive of capillary leakage should be closely monitored during the expected risk window. Laboratory trends, including complete blood counts, renal and hepatic function, coagulation parameters, C-reactive protein, ferritin, and lactate dehydrogenase, may provide additional context, particularly when interpreted together with baseline disease burden and inflammatory status. Although laboratory abnormalities alone should not define CRS severity, dynamic changes can support earlier recognition of evolving toxicity and help distinguish CRS from infection, disease progression, or other treatment-related complications [ 8 , 16 , 17 ].
Neurologic monitoring should be similarly proactive because early ICANS may be subtle and fluctuating. Serial ICE assessments, including handwriting evaluation, provide a simple, reproducible bedside method to detect early encephalopathy before a more obvious deterioration in arousal occurs [ 8 , 19 , 22 ]. However, the ICE score should be interpreted alongside clinical assessments of attention, language, level of consciousness, seizure activity, and focal neurologic findings. When patients are unable to complete the ICE test, simpler measures, such as the Alert, Voice, Pain, Unresponsive (AVPU) scale or the Glasgow Coma Scale, may help document the degree of impaired consciousness. Worsening mental status, new seizures, persistent aphasia, or focal deficits should prompt urgent evaluation for ICANS as well as alternative causes, including infection, metabolic disturbance, medication effects, stroke, or intracranial disease [ 8 , 19 ].
Several adjunctive tools are being explored to improve early ICANS recognition beyond conventional bedside assessments. In a prospective study using twice-daily single-channel electroencephalography after CAR T-cell infusion, an automated polymorphic delta activity score identified ICANS with good discrimination and preceded clinical symptoms by a median of 68 h [ 28 ]. A separate pilot study of a rapid neurocognitive testing tool, CART-NS, also suggested that brief serial cognitive testing detects subtle neurotoxicity before the deterioration of the ICE score, with early Stroop test changes showing promising predictive performance [ 29 ]. These approaches are not yet ready to replace ICE-based monitoring. However, they illustrate how objective digital or neurophysiologic tools may eventually support earlier detection in high-risk patients.
The intensity and setting of monitoring should be tailored to the treatment platform and baseline risk. Patients receiving CAR T-cell products associated with higher early expansion or neurotoxicity risk, such as CD28-costimulated products, and those with a high tumor burden, elevated inflammatory markers, thrombocytopenia, renal dysfunction, or high-risk prediction scores may require more intensive early observation [ 7 , 9 , 12 , 23 – 27 ]. In contrast, BsAb-associated CRS most often occurs during initial or step-up dosing, making monitoring around early doses and after dose escalation particularly important [ 30 ]. Risk-adapted monitoring does not imply that low-risk patients require minimal observation. Rather, it provides a framework for allocating inpatient monitoring, outpatient follow-up, laboratory assessment, neurologic surveillance, and escalation pathways according to the expected timing and severity of CRS and ICANS. This is particularly relevant for outpatient cellular therapy programs, in which standardized workflows are needed to detect CRS and ICANS symptoms outside routine clinic hours and ensure timely escalation when symptoms emerge [ 31 ]. Thus, risk-adapted monitoring should be understood not only as a bedside assessment strategy but also as a system-level process that links risk prediction to timely, individualized intervention.
Early intervention and personalized management
Grade-based management algorithms remain the backbone of CRS and ICANS treatments; however, standardized algorithms do not fully address all clinically important decisions. In practice, clinicians must determine not only whether a patient meets the criteria for tocilizumab or corticosteroid therapy but also how early to intervene, how much immunosuppression is necessary, and how rapidly treatment can be tapered or withheld. These decisions should be individualized according to baseline risk, treatment platform, product-specific toxicity profile, comorbidities, concurrent infection risk, and evolving patterns of CRS or ICANS [ 32 , 33 ].
The most informative prospective experience supporting earlier interventions came from the axi-cel safety management cohorts in ZUMA-1. Cohort 4 was an exploratory safety management cohort that incorporated levetiracetam prophylaxis and earlier corticosteroid and/or tocilizumab intervention, including treatment of grade 1 CRS if no improvement was observed after 3 days and corticosteroid use for grade ≥ 1 neurologic events [ 34 ]. This strategy reduced the incidence of severe CRS and neurologic events compared with pivotal cohorts 1 and 2 without an apparent adverse effect on CAR T-cell expansion or ongoing response [ 34 ]. In cohort 6, prophylactic dexamethasone 10 mg once daily from day 0 before axi-cel infusion through day 2, together with earlier toxicity intervention, was associated with no grade ≥ 3 CRS, low rates of severe neurologic events, and high response rates [ 35 ]. Longer follow-up of cohort 6 continued to support durable activity without a clear negative effect on axi-cel pharmacokinetics or efficacy outcomes [ 36 ]. These findings support the concept that early and measured corticosteroid exposure can mitigate severe toxicity, particularly for products associated with rapid expansion and higher inflammatory risk.
However, these data should not be interpreted as justifications for indiscriminate corticosteroid exposure. Several retrospective studies have suggested that higher cumulative corticosteroid doses, prolonged exposure, and early intensive use are associated with inferior outcomes; however, confounding by indication remains a major limitation, because patients receiving more steroids usually have more severe toxicity and a higher risk of disease [ 37 ]. More recent time-dependent analyses also support a cautious interpretation; corticosteroids used for toxicity management did not clearly compromise overall survival or relapse risk, but greater cumulative exposure was associated with a higher infection risk, late hematologic toxicity, and, in sensitivity analyses, shorter progression-free survival [ 38 ]. Thus, a practical steroid strategy should favor early use when clinically indicated, but at the lowest effective dose and shortest feasible duration, with frequent reassessment and rapid tapering once toxicity improves.
The role of anti-IL-6 therapy also requires personalization. Tocilizumab remains central to CRS treatment, particularly when fever is accompanied by hypotension or hypoxia. However, its role in isolated or evolving ICANS is limited, and repeated tocilizumab exposure without systemic CRS is unlikely to address the dominant pathophysiology of neurotoxicity. ZUMA-1 cohort 3 evaluated prophylactic tocilizumab and levetiracetam in patients receiving axi-cel; levetiracetam was started on day 0, whereas tocilizumab 8 mg/kg was administered 48 h after axi-cel infusion [ 39 ]. Although high-grade CRS was reduced, high-grade neurologic events were not, and the study did not support routine use of prophylactic tocilizumab to prevent CAR T-cell–related adverse events. Because tocilizumab suppresses IL-6–driven hepatic CRP production, CRP may decline rapidly after administration and should not be relied upon in isolation to assess ongoing CRS activity or to exclude concurrent infection. Therefore, in patients with prominent neurologic risk or early neurologic symptoms, a tocilizumab-sparing and steroid-forward approach may be reasonable unless concurrent CRS requires IL-6 blockade.
ICANS management should be individualized because early manifestations can be subtle, fluctuating, and dissociated from CRS severity. Corticosteroids remain the mainstay of clinically significant ICANS, with dexamethasone commonly used for low- to intermediate-grade events and high-dose methylprednisolone for severe or rapidly progressive neurotoxicity [ 32 , 33 ]. Seizure prophylaxis with levetiracetam is often used in patients receiving products associated with a higher neurotoxicity risk, such as axi-cel or brexucabtagene autoleucel, although the evidence base is not definitive. Personalized ICANS management also includes avoidance of sedating medications when possible; correction of metabolic abnormalities; evaluation for infection or disease involvement; and early involvement of neurology, electroencephalography, or neuroimaging when symptoms progress or the diagnosis is uncertain [ 32 , 33 ].
IL-1 blockade has emerged as an attractive strategy because IL-1 has been implicated in both CRS and ICANS biology. In a phase 2 study of patients with relapsed or refractory lymphoma receiving CD19 CAR T-cell therapy, prophylactic anakinra 100 mg every 12 h was started on day 2 after infusion or after two documented fevers, whichever occurred first, and was associated with low rates of severe ICANS without an obvious loss of antitumor activity [ 21 ]. A separate phase 1 study also supported the feasibility of prophylactic anakinra in mitigating ICANS in large B-cell lymphoma [ 40 ]. However, subsequent data have been less uniform. Single-cell analysis of breakthrough toxicities after once-daily anakinra prophylaxis suggested that IL-1 blockade alone is insufficient for some patients and identified IFN-γ-associated pathways as potential contributors to breakthrough CRS and neurotoxicity [ 41 ]. More recently, a real-world inverse probability of treatment weighting analysis of high-risk patients receiving prophylactic anakinra did not show a reduction in severe ICANS and raised concerns about increased immune-mediated toxicities and infectious complications [ 42 ]. Taken together, anakinra is promising for selected high-risk or refractory cases; however, routine prophylactic use should remain investigational until prospective studies better define the patients, dose, timing, and duration most likely to benefit.
Thus, personalized management should emphasize proportionate intervention rather than maximal immunosuppression. Earlier corticosteroid use, CRS-directed tocilizumab, steroid-centered ICANS treatment, and selective anakinra use should be guided by the patient’s baseline risk, treatment platform, and evolving toxicity profile. The goal is to prevent progression to organ-threatening CRS or ICANS while preserving antitumor efficacy and minimizing infection, cytopenia, and unnecessary treatment disruption [ 30 , 32 – 42 ].
Refractory CRS and ICANS
Refractory CRS and ICANS are uncommon, but clinically challenging because evidence-based therapeutic sequencing remains limited. In practice, refractory CRS may be considered when fever, hypotension, hypoxia, or organ dysfunction persists or progresses despite appropriate supportive care, tocilizumab, and corticosteroids. Refractory ICANS refers to neurologic deterioration or lack of improvement despite adequate corticosteroid-based treatment and exclusion of alternative causes. In both settings, escalation should not be delayed when toxicity is progressing; however, reassessment of mimicking or contributing conditions is essential, including infection, sepsis, metabolic disturbances, medication effects, disease progression, vascular events, and immune effector cell–associated hemophagocytic lymphohistiocytosis-like syndrome (IEC-HS) [ 43 , 44 ].
Anakinra has emerged as one of the most clinically developed salvage treatment options for patients with refractory CRS or ICANS. Its use is supported by the role of IL-1 in macrophage-driven inflammation and the pathobiology of neurotoxicity [ 13 , 14 ]. In a multicenter retrospective study of 43 patients treated for refractory CRS or ICANS after CAR T-cell therapy, most patients received anakinra for ICANS that persisted despite high-dose corticosteroids; a smaller subset received it for CRS refractory to tocilizumab [ 20 ]. Anakinra was feasible at higher intravenous doses with limited drug-related toxicity, and CAR T-cell efficacy appeared to be preserved, with an overall response rate of 77% after CAR T-cell therapy. In that study, higher-dose anakinra was associated with lower day-28 treatment-related mortality than the lower-dose treatment, supporting the concept that adequate dosing and early escalation may be important for severe refractory toxicity [ 20 ]. Nevertheless, the retrospective design, small cohort size, heterogeneous CAR T-cell products, and potential selection bias preclude definitive conclusions.
Siltuximab, a monoclonal antibody that directly binds to circulating IL-6, is another cytokine-directed option, particularly when CRS persists after tocilizumab treatment or when IL-6 receptor blockade may be insufficient. Unlike tocilizumab, which blocks the IL-6 receptor, siltuximab neutralizes IL-6 itself and may therefore reduce circulating free IL-6. In a multicenter retrospective analysis of 54 patients treated with siltuximab for CAR T-cell–related CRS or ICANS, the CRS grade improved in 75% of patients with CRS at the time of treatment, and the ICANS grade improved in 60% of patients with ICANS [ 45 ]. Among patients with CRS who were previously treated with tocilizumab, the response rate was 63%. These findings support siltuximab as a reasonable salvage treatment; however, prospective data are needed to clarify the optimal timing, patient selection, and its role relative to anakinra or corticosteroid escalation.
For patients with progressive toxicity despite treatment with corticosteroids, tocilizumab, anakinra, or siltuximab, treatment should be guided by the dominant inflammatory phenotype. Persistent shock and capillary leakage suggest uncontrolled systemic CRS; progressive encephalopathy, seizure, or coma suggests refractory ICANS; and delayed or biphasic inflammation with worsening cytopenia, hyperferritinemia, hypofibrinogenemia, coagulopathy, and transaminitis should raise concerns regarding IEC-HS [ 43 , 44 ]. This distinction is clinically important because IEC-HS may require management approaches that differ from those conventionally used for CRS or ICANS. The ASTCT consensus recommendations have provided diagnostic and grading criteria for IEC-HS; however, optimal treatment remains uncertain [ 44 ]. Case-based literature has described the use of ruxolitinib, emapalumab, and low-dose etoposide in progressive IEC-HS refractory to corticosteroids and anakinra, illustrating that later-line treatment may need to target the dominant pathway, such as JAK-STAT signaling, IFN-γ, or pathologic T-cell/macrophage activation [ 46 ].
Other investigational or last-resort strategies for refractory CRS or ICANS include JAK inhibition, tyrosine kinase inhibition, T-cell–depleting antibodies, cyclophosphamide, and other lymphotoxic approaches [ 43 , 46 ]. These interventions should be reserved for life-threatening cases and used with multidisciplinary input, because they may increase infection risk, worsen cytopenia, or compromise CAR T-cell persistence and antitumor efficacy. In this setting, the aim is not simply to suppress inflammation but to choose the least destructive intervention that can reverse organ-threatening toxicity. Early intensive care, neurologic consultation, infectious disease evaluation, and repeated assessments of inflammatory markers and organ function are essential.
Overall, refractory CRS and ICANS should be managed as dynamic syndromes rather than as fixed grades. Salvage therapy should be selected based on the timing of toxicity, prior exposure to tocilizumab or corticosteroids, dominant clinical phenotype, laboratory evidence of macrophage activation or IEC-HS, infection risk, organ dysfunction, and the need to preserve antitumor activity. Current data support anakinra and siltuximab as the most clinically developed salvage options. However, more intensive or targeted approaches should remain reserved for selected life-threatening cases or clinical trials.
Conclusions
CRS and ICANS remain central challenges in the safe delivery of T-cell–redirected therapies. Although standardized grading and management algorithms have improved clinical recognition and treatment, optimal care requires more than reactive management of established high-grade toxicities. Early identification of high-risk patients, structured monitoring during the expected toxicity window, and timely intervention before progression to organ-threatening inflammation or neurotoxicity are essential.
Risk prediction models, serial clinical assessments, laboratory trends, and neurologic monitoring can support individualized care; however, none can replace careful bedside judgment. Therefore, management should emphasize proportionate intervention: sufficient to control toxicity without compromising antitumor efficacy or increasing the risk of infection, cytopenia, and treatment disruption. For refractory CRS or ICANS, phenotype-directed escalation and multidisciplinary care remain critical; however, the optimal timing and sequencing of salvage therapies require further study.
As CAR T-cell therapies and BsAbs move into broader clinical use, CRS and ICANS management should evolve into practical, risk-adapted strategies that integrate early detection, personalized intervention, and preservation of therapeutic benefits.
Authors’ contributions
Jinchul Kim wrote the main manuscript text.
Funding
None.
Data availability
No datasets were generated or analysed during the current study.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
1. Maude SL, Laetsch TW, Buechner J, et al. Tisagenlecleucel in children and young adults with B-cell lymphoblastic leukemia. N Engl J Med. 2018;378:439–48. 10.1056/NEJMoa1709866.
2. Neelapu SS, Locke FL, Bartlett NL, et al. Axicabtagene ciloleucel CAR T-cell therapy in refractory large B-cell lymphoma. N Engl J Med. 2017;377:2531–44. 10.1056/NEJMoa1707447.
3. Wang M, Munoz J, Goy A, et al. KTE-X19 CAR T-cell therapy in relapsed or refractory mantle-cell lymphoma. N Engl J Med. 2020;382:1331–42. 10.1056/NEJMoa1914347.
4. Munshi NC, Anderson LD Jr, Shah N, et al. Idecabtagene vicleucel in relapsed and refractory multiple myeloma. N Engl J Med. 2021;384:705–16. 10.1056/NEJMoa2024850.
5. Dickinson MJ, Carlo-Stella C, Morschhauser F, et al. Glofitamab for relapsed or refractory diffuse large B-cell lymphoma. N Engl J Med. 2022;387:2220–31. 10.1056/NEJMoa2206913.
6. Moreau P, Garfall AL, van de Donk NWCJ, et al. Teclistamab in relapsed or refractory multiple myeloma. N Engl J Med. 2022;387:495–505. 10.1056/NEJMoa2203478.
7. Morris EC, Neelapu SS, Giavridis T, Sadelain M. Cytokine release syndrome and associated neurotoxicity in cancer immunotherapy. Nat Rev Immunol. 2022;22:85–96. 10.1038/s41577-021-00547-6.
8. Lee DW, Santomasso BD, Locke FL, et al. ASTCT consensus grading for cytokine release syndrome and neurologic toxicity associated with immune effector cells. Biol Blood Marrow Transplant. 2019;25:625–38. 10.1016/j.bbmt.2018.12.758.
9. Shouval R, Strouse C, Kim S, et al. Cytokine release syndrome and neurotoxicity following CD19 CAR-T in B-cell lymphoma. Transplant Cell Ther. 2025;31:419–33. 10.1016/j.jtct.2025.03.011.
10. Cordas dos Santos DM, Tix T, Shouval R, et al. A systematic review and meta-analysis of nonrelapse mortality after CAR T cell therapy. Nat Med. 2024;30:2667–78. 10.1038/s41591-024-03084-6.
11. Greenbaum U, Strati P, Saliba RM, et al. CRP and ferritin in addition to the EASIX score predict CAR-T-related toxicity. Blood Adv. 2021;5:2799–806. 10.1182/bloodadvances.2021004575.
12. Sesques P, Kirkwood AA, Kwon M, et al. Novel prognostic scoring systems for severe CRS and ICANS after anti-CD19 CAR T cells in large B-cell lymphoma. J Hematol Oncol. 2024;17:61. 10.1186/s13045-024-01579-w.
13. Giavridis T, van der Stegen SJC, Eyquem J, et al. CAR T cell-induced cytokine release syndrome is mediated by macrophages and abated by IL-1 blockade. Nat Med. 2018;24:731–8. 10.1038/s41591-018-0041-7.
14. Norelli M, Camisa B, Barbiera G, et al. Monocyte-derived IL-1 and IL-6 are differentially required for cytokine-release syndrome and neurotoxicity due to CAR T cells. Nat Med. 2018;24:739–48. 10.1038/s41591-018-0036-4.
15. Chen ZG, Yang H, Yang C, et al. Macrophage polarization in hematologic cancers: mechanisms and therapeutic strategies. Blood Res. 2026;61:8. 10.1007/s44313-025-00119-w.
16. Lee DW, Gardner R, Porter DL, et al. Current concepts in the diagnosis and management of cytokine release syndrome. Blood. 2014;124:188–95. 10.1182/blood-2014-05-552729.
17. Hay KA, Hanafi LA, Li D, et al. Kinetics and biomarkers of severe cytokine release syndrome after CD19 chimeric antigen receptor-modified T-cell therapy. Blood. 2017;130:2295–306. 10.1182/blood-2017-06-793141.
18. Teachey DT, Lacey SF, Shaw PA, et al. Identification of predictive biomarkers for cytokine release syndrome after chimeric antigen receptor T-cell therapy for acute lymphoblastic leukemia. Cancer Discov. 2016;6:664–79. 10.1158/2159-8290.CD-16-0040.
19. Gust J, Hay KA, Hanafi LA, et al. Endothelial activation and blood-brain barrier disruption in neurotoxicity after adoptive immunotherapy with CD19 CAR-T cells. Cancer Discov. 2017;7:1404–19. 10.1158/2159-8290.CD-17-0698.
20. Gazeau N, Liang EC, Wu QV, et al. Anakinra for refractory cytokine release syndrome or immune effector cell-associated neurotoxicity syndrome after chimeric antigen receptor T cell therapy. Transplant Cell Ther. 2023;29:430–7. 10.1016/j.jtct.2023.04.001.
21. Park JH, Nath K, Devlin SM, et al. CD19 CAR T-cell therapy and prophylactic anakinra in relapsed or refractory lymphoma: phase 2 trial interim results. Nat Med. 2023;29:1710–7. 10.1038/s41591-023-02404-6.
22. Santomasso BD, Park JH, Salloum D, et al. Clinical and biologic correlates of neurotoxicity associated with CAR T-cell therapy in patients with B-cell acute lymphoblastic leukemia. Cancer Discov. 2018;8:958–71. 10.1158/2159-8290.CD-17-1319.
23. Kim J, Cho J, Lee MH, Yoon SE, Kim WS, Kim SJ. CAR T cells vs bispecific antibody as third- or later-line large B-cell lymphoma therapy: a meta-analysis. Blood. 2024;144:629–38. 10.1182/blood.2023023419.
24. Pennisi M, Jain T, Santomasso BD, et al. Modified EASIX predicts severe cytokine release syndrome and neurotoxicity after chimeric antigen receptor T cells. Blood Adv. 2021;5:3397–406. 10.1182/bloodadvances.2020003885.
25. Ye L, Liao L, Wang L, et al. Prediction models for CAR-T cell therapy-related adverse events in hematologic malignancies: a systematic review and meta-analysis. BMC Cancer. 2026;26:624. 10.1186/s12885-026-15877-8.
26. Arai Y, Jo T, Sato T, et al. Decreased renal function predicts severe cytokine release syndrome after CAR-T-cell therapy for large B-cell lymphoma. Br J Haematol. 2026;208:2124–33. 10.1111/bjh.70488.
27. Winkelmann M, Raj SS, Jain MD, et al. Optimization and validation of the international metabolic prognostic index for CD19 CAR-T in large B-cell lymphoma. Blood Cancer J. 2025;15:144. 10.1038/s41408-025-01338-1.
28. Tienstra M, Storck TN, van der Galiën HT, et al. Early ICANS recognition with a novel electroencephalogram-based deltawave detection system in patients receiving CAR T-cell therapy. Blood Immunol Cell Ther. 2025;1:100006. 10.1016/j.bict.2025.100006.
29. Suresh A, Wishart HA, Arslan MN, et al. Novel neurocognitive testing tool for early neurotoxicity detection following anti-CD19 and anti-BCMA chimeric antigen receptor (CAR) T-cell therapy: a pilot study. Clin Lymphoma Myeloma Leuk. 2025;25:365–78. 10.1016/j.clml.2024.12.011.
30. Crombie JL, Graff T, Falchi L, et al. Consensus recommendations on the management of toxicity associated with CD3×CD20 bispecific antibody therapy. Blood. 2024;143:1565–75. 10.1182/blood.2023022432.
31. Piela A, Ryan M, Bauer S, et al. Development and implementation of a CRS and ICANS monitoring process for outpatient cellular therapy. Transplant Cell Ther. 2026;32(2 Suppl):S523.
32. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Management of Immunotherapy-Related Toxicities. Version 2.2026. National Comprehensive Cancer Network; 2026. https://www.nccn.org/professionals/physician_gls/pdf/immunotherapy.pdf . Accessed 3 Jul 2026.
33. Santomasso BD, Nastoupil LJ, Adkins S, et al. Management of immune-related adverse events in patients treated with chimeric antigen receptor T-cell therapy: ASCO guideline. J Clin Oncol. 2021;39:3978–92. 10.1200/JCO.21.01992.
34. Topp MS, van Meerten T, Houot R, et al. Earlier corticosteroid use for adverse event management in patients receiving axicabtagene ciloleucel for large B-cell lymphoma. Br J Haematol. 2021;195:388–98. 10.1111/bjh.17673.
35. Oluwole OO, Bouabdallah K, Munoz J, et al. Prophylactic corticosteroid use in patients receiving axicabtagene ciloleucel for large B-cell lymphoma. Br J Haematol. 2021;194:690–700. 10.1111/bjh.17527.
36. Oluwole OO, Forcade E, Muñoz J, et al. Long-term outcomes of patients with large B-cell lymphoma treated with axicabtagene ciloleucel and prophylactic corticosteroids. Bone Marrow Transplant. 2024;59:366–72. 10.1038/s41409-023-02169-z.
37. Strati P, Ahmed S, Furqan F, et al. Prognostic impact of corticosteroids on efficacy of chimeric antigen receptor T-cell therapy in large B-cell lymphoma. Blood. 2021;137:3272–6. 10.1182/blood.2020008865.
38. Luttwak E, Nemirovsky D, Devlin SM, et al. Patterns and safety of glucocorticosteroid use following CD19 CAR-T therapy for large B-cell lymphoma. HemaSphere. 2025;9:e70248. 10.1002/hem3.70248.
39. Locke FL, Neelapu SS, Bartlett NL, et al. Tocilizumab prophylaxis following axicabtagene ciloleucel in relapsed or refractory large B-cell lymphoma. Transplant Cell Ther. 2024;30:1065–79. 10.1016/j.jtct.2024.08.018.
40. Strati P, Jallouk A, Deng Q, et al. A phase 1 study of prophylactic anakinra to mitigate ICANS in patients with large B-cell lymphoma. Blood Adv. 2023;7:6785–9. 10.1182/bloodadvances.2023010653.
41. Frigault MJ, Yao N, Berger TR, et al. Single-cell dynamics of breakthrough toxicities after anakinra prophylaxis for axicabtagene ciloleucel in lymphoma. Blood Adv. 2025;9:2122–35. 10.1182/bloodadvances.2024015161.
42. Easton N, Andreoli M, van Besien H, et al. Real-world assessment of prophylactic anakinra on neurotoxicity and cytokine release syndrome after CD19 CAR T-cell therapy in relapsed/refractory B-cell lymphoma: an inverse probability of treatment weighting analysis. Transplant Cell Ther. 2026. 10.1016/j.jtct.2026.06.008.
43. Jain MD, Smith M, Shah NN. How I treat refractory CRS and ICANS after CAR T-cell therapy. Blood. 2023;141:2430–42. 10.1182/blood.2022017414.
44. Hines MR, Knight TE, McNerney KO, et al. Immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome. Transplant Cell Ther. 2023;29:438.e1–438.e16. 10.1016/j.jtct.2023.03.006.
45. Bajwa A, Zhao Q, Geer M, et al. Siltuximab for chimeric antigen receptor T-cell therapy-related CRS and ICANS: a multicenter retrospective analysis. Blood Adv. 2025;9:170–5. 10.1182/bloodadvances.2024013688.
46. Scala JJ, Eckrich MJ, Lipak K, et al. Treatment strategies for progressive immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome: case series. Haematologica. 2024;109:3439–45. 10.3324/haematol.2023.284784.
Associated Data
Data Availability Statement
No datasets were generated or analysed during the current study.