A review on toxicities of T-cell engager-based therapies in solid tumors Authors: Etienne Bastien, Clara Vacheret, Raphael Sanchez, Francesca Lucibello, Nicolas Girard, Emanuela Romano Published: 2026 ABSTRACT T-cell engager (TCE)-based therapies represent an emerging and rapidly expanding class of immunotherapies in oncology. Initially developed for hematologic malignancies, TCEs are now entering the field of solid tumors. To date, two agents have received regulatory approval in this setting: Tebentafusp for metastatic uveal melanoma and Tarlatamab for small-cell lung cancer (SCLC). By physically linking a tumor-associated antigen (TAA) to a T-cell activating domain, most commonly cluster of differentiation (CD) 3, TCEs induce an immunological synapse that promotes targeted T-cell mediated cytotoxicity. This narrative review summarizes the most clinically relevant toxicities associated with TCEs in solid tumors, drawing on data from clinical trials and real-world experience. Tebentafusp exemplifies the safety profile of this therapeutic class, which is dominated by cytokine release syndrome (CRS) and cutaneous adverse events (AEs). These toxicities are typically early in onset, predictable, and reversible, occurring predominantly during the initial step-up dosing phase and decreasing substantially with subsequent administrations, consistent with immune adaptation. Tarlatamab, a Delta-like ligand 3 (DLL3) × CD3 bispecific TCE recently approved for SCLC, displays a toxicity profile similarly driven by immune activation, with CRS as the most frequent AE, generally of low grade and confined to early treatment cycles. In contrast, neurotoxicity including immune effector cell-associated neurotoxicity syndrome (ICANS) occurs more frequently, though events are usually low grade, delayed relative to CRS. Other TCEs currently in clinical development target a wide range of TAAs and display a broadly similar toxicity profile, with CRS emerging as a class effect related to CD3-mediated T-cell activation. The incidence and severity of CRS are influenced by dosing strategies, with step-up dosing mitigating early toxicity. Cutaneous, immune-related, and hematologic AEs are generally infrequent and rarely severe. Continued optimization of antigen selection, molecular engineering, and pharmacokinetic parameters will be critical to improving the safety profile of TCEs gathering data from clinical trials and real-world experience. Introduction T-cell engager (TCE) therapies represent an emerging and rapidly expanding class of immunotherapies for the treatment of solid tumors. While they are already widely used in hematologic malignancies, where the monoclonal nature of these diseases facilitates the identification of suitable targets, TCEs are now entering the field of solid oncology. The EMA and FDA have recently approved two agents in this setting: Tebentafusp for uveal melanoma and Tarlatamab for small-cell lung cancer (SCLC).1,2 Structurally, TCEs combine a tumor-associated antigen (TAA)-binding domain with a T-cell engaging domain, thereby promoting the formation of an immunological synapse between tumor cells and T lymphocytes and inducing targeted cytotoxic activity. 3 Although the safety profile of TCEs is well characterized in hematology-oncology, data on their toxicities in solid tumors remain limited. 4 As these agents are expected to play an increasingly important role in cancer therapy, we review here the current knowledge on TCE-associated most-relevant toxicities in solid tumors. Tebentafusp Tebentafusp is a soluble T-cell receptor (TCR)-based bispecific protein targeting the TAA glycoprotein-100 (gp100) and is the first TCE approved for a solid tumor. It represents the current standard first-line therapy for metastatic uveal melanoma in human leukocyte antigen (HLA)-A02:01-positive patients. Its safety profile reflects its dual mechanism of action, which combines a high-affinity soluble TCR specific for gp100 presented by this HLA allele with a cluster of differentiation 3 (CD3) T-cell engaging domain that recruits and activates a broad polyclonal T-cell repertoire. The resulting rapid and systemic T-cell activation accounts for the predominance of acute treatment-related adverse events (TRAEs), particularly cytokine-mediated toxicities such as cytokine release syndrome (CRS), whereas on-target engagement of gp100 on cutaneous melanocytes likely underlies the frequent dermatologic manifestations observed during treatment. In the pivotal phase III trial, CRS occurred in the vast majority of Tebentafusp-treated patients (89%), predominantly grade 1–2 (12% and 76%, respectively), with only 1% experiencing grade 3 events. CRS typically developed within hours of the first three infusions. Clinical manifestations reflected systemic immune activation, including fever, chills, hypotension, or hypoxia, and were generally transient, resolving within hours with antipyretics, intravenous fluids, glucocorticoids, or a combination of these interventions (Table 2).5,6 No treatment-related deaths were reported.7,8 Cutaneous toxicities constitute the other major family of adverse events (AEs), including rash (83%), pruritus (70%), and vitiligo (25%). Their mechanism combines systemic inflammatory activation induced by polyclonal T-cell recruitment and an on-target-off-tumor effect related to the gp100 expression by melanocytes. 9 Rash and pruritus, frequent and appearing within hours following the first injections, are mostly mild to moderate in severity and rarely dose-limiting. Long-term vitiligo is often observed in Tebentafusp-treated patients.9,10 Histopathologic findings include vesicular interface dermatitis, CD8+ T-cell infiltration, epidermal spongiosis, and subepidermal blistering in severe inflammation. Interestingly, keratin downregulation has been observed, involving pathways similar to those in atopic dermatitis, including IL-13, IL-17, IL-4, and IL-17A. Interestingly, rash occurrence was not predictive of clinical outcome. 9 Altered keratinization may explain the skin exfoliation observed in approximately 25% of patients. The pooled safety analysis of the IMCgp100-01, IMCgp100-102, and IMCgp100-202 trials, recently updated in a combined dataset of 410 patients, provides a refined characterization of Tebentafusp-associated toxicity.11–14 The main AEs, CRS and cutaneous reactions, occurred almost exclusively during the three infusions set-up dosing phase (20, 30, and 68 µg). These events were largely predictable, rapidly reversible, and generally manageable with antipyretics and intravenous hydration, with corticosteroids or anti-IL-6 therapy required only rarely. Close early monitoring is recommended, particularly for prompt identification of fever or hypotension. A 16-h observation period was advised for the initial infusions, consistent with the temporal window in which hypotensive episodes occurred. 1 Severe events remained uncommon (2%), and no neurotoxicity, including immune effector cell-associated syndrome (ICANS), was observed. The incidence and severity of treatment-related toxicities decreased markedly after the initial doses, suggesting a phenomenon of immunologic adaptation. Transient transaminase elevations were also frequent but predominantly low-grade and resolved spontaneously. Across trials, no treatment-related deaths were reported. Real-world data from the EURACAN network further corroborate the favorable safety profile of Tebentafusp observed in clinical trials. In routine clinical practice across multiple European centers, most grade 3 TRAES mainly consisted of cutaneous rash, while other less frequent grade 3 events included transient oxygen desaturation and pain flare following palliative radiotherapy; a single case of grade 4 rash was reported. Importantly, all toxicities were transient and decreased over time. The incidence of grade 3–4 AEs was comparable between patients aged <70 and ⩾70 years, supporting the use of Tebentafusp in older patients without increased safety concerns. No new or unexpected safety signals were identified. 15 Tarlatamab Delta-like ligand 3 (DLL3) emerged as a therapeutically relevant target in SCLC following translational studies demonstrating high expression in high-grade pulmonary neuroendocrine tumors and accessibility for antibody-based targeting. 16 DLL3 is an atypical Notch ligand that is predominantly intracellular under physiological conditions but becomes aberrantly expressed at the tumor cell surface in SCLC, enabling immune redirection strategies. 16 After the clinical failure of the DLL3-targeted antibody-drug conjugate (ADC) Rovalpituzumab tesirine, DLL3 targeting shifted toward TCE formats designed to improve tolerability.17–19 Tarlatamab is a DLL3 × CD3 bispecific TCE approved for patients with extensive-stage SCLC progressing on or after platinum-based chemotherapy and represents, the first approved TCE in thoracic oncology. 20 Its safety profile reflects its mechanism of action, combining DLL3-directed tumor engagement with CD3-mediated activation of a polyclonal T-cell repertoire, resulting in immune activation-driven toxicities rather than classical cytotoxic effects. 20 Findings from the phase I DeLLphi-300 study informed the step-up dosing strategy used in subsequent Tarlatamab trials. 21 In the pivotal phase II DeLLphi-301 trial, CRS was the most frequent TRAE, occurring in 53% of patients, predominantly grade 1–2, with grade ⩾3 CRS in approximately 1%–2%. 20 CRS occurred mainly during the initial step-up dosing phase in cycle 1 and typically manifested within hours of infusion as fever, chills, hypotension, or hypoxia, resolving with supportive care, corticosteroids, and anti-IL-6 therapy when required. 20 Early dose-escalation cohorts identified a relationship between initial dose and CRS severity, leading to adoption of step-up dosing (1 mg followed by 10 mg QW2); with the standardized regimen used in DeLLphi-301, no consistent dose-toxicity relationship was observed. 20 Neurotoxicity, including ICANS, represents a distinctive feature of Tarlatamab among TCEs in solid tumors. In phase II DeLLphi-301, neurologic symptoms (confusion, impaired attention, somnolence, and motor disturbances) were reported in a substantial proportion of patients, while ICANS occurred in approximately 10% and was predominantly low grade. 22 Observational data suggest higher incidence and grade for patients with untreated brain metastases and it is lower than observed with chimeric antigen receptors (CAR)-T cell therapies (26%).23,24 The mechanisms underlying ICANS include excessive peripheral immune activation, endothelial activation leading to blood-brain barrier disruption, and inflammation of the central nervous system. 25 ICANS typically occurred later than CRS and could arise independently of cytokine release, with most events resolving following treatment interruption and corticosteroid administration (Table 2); a similar neurotoxicity profile without new safety signals was observed in the randomized phase III DeLLphi-304 trial.26–28 Cutaneous TRAEs were infrequent (<15%) and did not constitute a defining toxicity category for Tarlatamab, in contrast to earlier DLL3-targeted ADCs, where skin toxicity was largely attributable to the cytotoxic payload. 20 Immune-related AEs typically associated with immune checkpoint inhibitors (ICIs) were uncommon, whereas infections occurred in approximately 40% of patients, with grade ⩾3 infections in approximately 10%–15%, reflecting the vulnerability of a heavily pretreated SCLC population. 20 Hematologic toxicities were generally limited. Anemia occurred in 30% of patients (grade ⩾3: 5%–6%), while neutropenia and thrombocytopenia occurred in approximately 10%–15% and 30%, respectively, and were rarely severe; febrile neutropenia was uncommon (<1%). 20 Other common TEAEs included decreased appetite (36%), dysgeusia (32%), constipation (29%), asthenia (25%), and fatigue (24%). 22 The toxicity profile initially characterized in DeLLphi-301 was confirmed in the randomized phase III DeLLphi-304 study, which demonstrated an OS benefit compared with chemotherapy and did not reveal new safety signals. 26 In pooled regulatory safety analyses supporting approval, fatal adverse reactions occurred in approximately 2%–3% of patients and consisted predominantly of infectious or respiratory complications, without evidence for immune-mediated pneumonitis or an on-target off-tumor pulmonary mechanism. 2 Beyond Tarlatamab, other DLL3-directed TCEs such as Obrixtamig (BI 764532) are in early-phase clinical development, supporting DLL3 as a viable immunotherapy target. However, emerging data indicate that toxicity profiles, including CRS and neurotoxicity, are molecule- and regimen-specific and should not be extrapolated across DLL3-directed platforms. 29 Other TCEs Beyond Tebentafusp and Tarlatamab, multiple TCEs are currently under investigation. 30 Although the safety data remain limited, available evidence allows the identification of recurring toxicities, which, in addition to being target specific, reflect the intensity of T-cell activation that also depends on TCE design and dosing strategies. We identified seven additional TCEs that target distinct TAAs, with CD3 as the predominant T-cell engaging domain (Tables 1 and 2; Supplemental Table 1).31–35 Table 1.Characteristics of trials investigating T-cell engager as monotherapy.TrialClinicalTrial labelPublication datePhaseTumor typeDrugFormatTAA binding domain and T-cell engaging domain n DevelopmentIMCgp100-01NCT012112622020IMelanomaTebentafuspSoluble TCRgp100-HLA-A*02:01 and CD384Marketed for UMIMCgp100-102NCT025703082022IIUMTebentafuspSoluble TCRgp100-HLA-A*02:01 and CD3127Marketed for UMIMCgp100-202NCT030703922021IIIUMTebentafuspSoluble TCRgp100-HLA-A*02:01 and CD3245Marketed for UMDeLLphi-300NCT033199402023ISCLCTarlatamabHalf-life extended 1 + 1DLL3 and CD3107Marketed for SCLCDeLLphi-301NCT050600162023IISCLCTarlatamabHalf-life extended 1 + 1DLL3 and CD3220Marketed for SCLCDeLLpro-300NCT047027372025INeuroendocrine prostate cancerTarlatamabHalf-life extended 1 + 1DLL3 and CD340Marketed for SCLC/NCT039369592024IPan-tumorReozalimabIgG likePDL-1 and PD-1103Phase I terminated/NCT045513522024IMelanomaTYRP1-TCB2 + 1TYRP1 and CD320Phase I terminatedMT110-101NCT006355962018IPan-tumorSolitomab1 + 1EpCAM and CD365Discontinued/NCT037928412024IProstate cancerAcapatamabHalf-life extended 1 + 1PSMA and CD3133DiscontinuedMI-CP216NCT012842312016IGIMEDI-5651 + 1CEA and CD339Discontinued/NCT017234752020IProstate cancerPasotuxizumab1 + 1PSMA and CD316DiscontinuedERTUSONCT015694122016IPan-tumorErtumaxomabIgG likeHER2 and CD314DiscontinuedFormats of T-cell engagers are illustrated in Figure 1.CD3, cluster of differentiation 3; CEA, carcinoembryonic antigen; DLL3, delta-like ligand 3; EpCAM, epithelial cell adhesion molecule; gp100, glycoprotein 100; HER2, epidermal growth factor receptor 2; HLA, human leukocyte antigen; IgG, immunoglobulin G; PD-1, programmed death 1; PDL-1, programmed death-ligand 1; PSMA, prostate-specific membrane antigen; SCLC, small-cell lung cancer; TAA, tumor-associated antigen; TCR, T-cell receptor; TYRP1, tyrosinase-related protein 1; UM, uveal melanoma. Characteristics of trials investigating T-cell engager as monotherapy. Formats of T-cell engagers are illustrated in Figure 1. CD3, cluster of differentiation 3; CEA, carcinoembryonic antigen; DLL3, delta-like ligand 3; EpCAM, epithelial cell adhesion molecule; gp100, glycoprotein 100; HER2, epidermal growth factor receptor 2; HLA, human leukocyte antigen; IgG, immunoglobulin G; PD-1, programmed death 1; PDL-1, programmed death-ligand 1; PSMA, prostate-specific membrane antigen; SCLC, small-cell lung cancer; TAA, tumor-associated antigen; TCR, T-cell receptor; TYRP1, tyrosinase-related protein 1; UM, uveal melanoma. Table 2.Guidelines for grading and management of CRS and ICANS related to TCEs.Guideline sourceGrade 1Grade 2Grade 3Grade 4ASTCT consensus for CRS gradingSymptoms are not life-threatening and require symptomatic treatment only (fever, nausea, fatigue, headache, myalgias, malaise)Symptoms require and respond to moderate intervention:Symptoms require and respond to aggressive intervention:Life-threatening symptoms:- Oxygen requirement <40% FiO2 OR- Oxygen requirement ⩾40% FiO2 OR- Requirement for ventilator support OR- Hypotension responsive to i.v. fluids or low dose of one vasopressor OR- Hypotension requiring high-dose or multiple vasopressors OR- Grade 4 organ toxicity (excluding aminotransferase increased)- Grade 2 organ toxicity- Grade 3 organ toxicity or grade 4 aminotransferase increasedCRS treatment guideline from Gloucestershire hospitals (2024)- Treat symptoms. a - Hypotension: 0.9% saline fluid challenge in 500 mL boluses. If >1000 mL administered, consider vasopressor support.- Symptomatic management of organ toxicities.- Symptomatic management of organ toxicities.- Consider initial single dose Dexamethasone 10 mg IV and increase frequency up to 6-hourly if ongoing concern.- Hypoxia: administer oxygen.- Admit to intensive care unit for vasopressor and/or respiratory support.- Admit to intensive care unit for vasopressor and/or respiratory support.- If ongoing TCE infusion, suspend.- If ongoing TCE infusion, discontinue.- If ongoing TCE infusion, discontinue.- Administer Tocilizimab. b - Consider Tocilizumab b if ongoing clinical concern following Dexamethasone.- If fever, explore infection cause.- Administer Tocilizimab. b - Administer Dexamethasone 10 mg IV 6-hourly. If refractory grade 4 CRS consider Methylprednisolone 1 g once daily for 3 days.- If fever, explore infection cause.- Administer Tocilizimab. b - Give broad spectrum antibiotics.- Treat symptoms. a - Administer Dexamethasone 10 mg IV, consider frequency up to 6-hourly if ongoing concern.- Administer Dexamethasone 10 mg IV 6-hourly.- Give broad spectrum antibiotics and explore infection cause.- Permanently discontinue TCE therapy.ASTCT consensus for ICANS grading- ICE 7–9- ICE 3–6- ICE 0–2- ICE 0 (unrousable)- Awakens spontaneously.- Awakens to voice.- Awakens to tactile stimulus.- Unrousable- Focal seizure or rapidly resolved generalized seizure or non-convulsive seizure on EEG that resolves with intervention.- Life-threatening prolonged seizure (>5 min) or repetitive clinical seizures or EEG seizure without return to baseline.- Focal edema on neuroimaging.- Deep focal weakness as s hemiparesis or paraparesis.- Cranial nerve VI palsy.- Decerebrate or decorticate posturing.- Papilledema or Cushing’s triad.- Diffuse cerebral edema on neuroimaging or decorticate posturingCARETOX guidelines for ICANS management (2023)- Brain MRI- Brain MRI- Brain MRI, repeated every 2–3 days for persistent ⩾grade 3 encephalopathy.- Brain MRI, repeated every 2–3 days for persistent ⩾grade 3 encephalopathy. Spine MRI if motor weakness is observed.- Neurology consultation and EEG.- Neurology consultation and EEG.- Neurology consultation and EEG.- Neurology consultation and EEG.- ICE score every 6 h.- ICE score every 6 h.- ICE score every 6 h.- ICE score every 6 h.- Consider diagnostic lumbar puncture if other causes of encephalopathy are suspected.- Consider diagnostic lumbar puncture if other causes of encephalopathy are suspected.- Consider diagnostic lumbar puncture if other causes of encephalopathy are suspected.- Consider diagnostic lumbar puncture if other causes of encephalopathy are suspected.- Supportive cares. c - Dexamethasone 10 mg IV every 6–12 h.- Admit patient to intensive care unit.- Admit patient to intensive care unit and consider mechanical ventilation if patient is unarousable.- Dexamethasone 10 mg IV for one dose.- Consider Anakinra. d - Neurology consult.- Neurology consult.- If ongoing TCE infusion, suspend.- Supportive care. c - If focal edema is in brain stem or thalamus, Methylprednisolone 1000 mg/day in divided doses IV for 3 days followed by taper depending on clinical situation.- Treat focal or generalized convulsive or non-convulsive/EEG seizures.- If ongoing TCE infusion, suspend.- If focal edema is in other areas of brain, Methylprednisolone 1000 mg/day in divided doses IV for 1 day; assess daily and continue or taper depending on clinical situation.- Treat diffuse cerebral edema or signs of raised intracranial pressure.- Consider Anakinra. d - Methylprednisolone 1000 mg/day in divided doses IV for 3 days followed by taper as clinically indicated.- Supportive care. c - Consider Anakinra. d - Permanently discontinue TCE therapy.- If Grade 4 ICANS is refractory for >24 h or if patient is deteriorating rapidly, consider additional therapies.- Supportive care. c - Permanently discontinue TCE therapy.Guidelines for grading and management of CRS and ICANS have been developed for CAR-T cells therapies and TCE in hematology malignancies. No specific guidelines are available for TCE in solid tumors. The bolded medications represent the cornerstone drug therapies for the management of CRS and ICANS.aParacetamol 1 g IV and antihistamine* (e.g., Loratadine 10 mg PO).bOnly intravenous route should be used for the treatment of CRS. The recommended dose for treatment of CRS given as a 60-min intravenous infusion is 12 mg per kg for patients less than 30 kg weight and 8 mg per kg for patients at or above 30 kg weight.cAssess swallowing and withhold oral intakes if needed; Avoid medications that cause central nervous system depression; Low doses of lorazepam after EEG is performed (0.25–0.5 mg IV every 8 h) or haloperidol (0.5 mg IV every 6 h) for agitated patients; if no seizures on EEG, continue prophylactic Levetiracetam.dConsider Anakinra for patients who are refractory to corticosteroids (defined as ongoing grade 2–4 ICANS 6 h after initiation of corticosteroids), dependent on corticosteroids (relapse while tapering off), and/or intolerant to corticosteroids (e.g., psychosis). Anakinra should be administered subcutaneously or IV at the dose of 100 mg every 12 h until resolution.CAR, chimeric antigen receptors; CRS, cytokine release syndrome; EEG, electroencephalogram; ICANS, immune effector cell-associated neurotoxicity syndrome; ICE, immune effector cell encephalopathy; IV, intravenous; MRI, magnetic resonance imaging; TCE, T-cell engager. Guidelines for grading and management of CRS and ICANS related to TCEs. Guidelines for grading and management of CRS and ICANS have been developed for CAR-T cells therapies and TCE in hematology malignancies. No specific guidelines are available for TCE in solid tumors. The bolded medications represent the cornerstone drug therapies for the management of CRS and ICANS. Paracetamol 1 g IV and antihistamine* (e.g., Loratadine 10 mg PO). Only intravenous route should be used for the treatment of CRS. The recommended dose for treatment of CRS given as a 60-min intravenous infusion is 12 mg per kg for patients less than 30 kg weight and 8 mg per kg for patients at or above 30 kg weight. Assess swallowing and withhold oral intakes if needed; Avoid medications that cause central nervous system depression; Low doses of lorazepam after EEG is performed (0.25–0.5 mg IV every 8 h) or haloperidol (0.5 mg IV every 6 h) for agitated patients; if no seizures on EEG, continue prophylactic Levetiracetam. Consider Anakinra for patients who are refractory to corticosteroids (defined as ongoing grade 2–4 ICANS 6 h after initiation of corticosteroids), dependent on corticosteroids (relapse while tapering off), and/or intolerant to corticosteroids (e.g., psychosis). Anakinra should be administered subcutaneously or IV at the dose of 100 mg every 12 h until resolution. CAR, chimeric antigen receptors; CRS, cytokine release syndrome; EEG, electroencephalogram; ICANS, immune effector cell-associated neurotoxicity syndrome; ICE, immune effector cell encephalopathy; IV, intravenous; MRI, magnetic resonance imaging; TCE, T-cell engager. Cytokine release syndrome CRS represents the most frequent TRAE across TCEs, reported in 43%–97% of patients, appearing as a class effect related to the CD3-mediated T-cell activation. However, its occurrence is not systematic. Notably, Reozalimab, a bi-specific IgG-like engaging programmed death 1 (PD-1) and programmed death-ligand 1 (PD-L1), did not induce CRS. No CRS cases have been reported with Solitomab, a 1 + 1 format engaging epithelial cell adhesion molecule (EpCAM) × CD3. These observations suggest that CRS occurrence depends mainly on the T-cell engaging domain and is more prevalent with CD3-targeting TCEs. Severe CRS remains rare, except for Acapatamab (20% of grade 3, no grade 4) and manageable using intravenous fluids, systemic corticosteroids, and Tocilizumab (Table 2).27,36,37 No CRS-related toxic deaths or treatment discontinuations were reported. Across studies, CRS typically occurs during the initial administrations and weans with subsequent administrations, consistently with the immune adaptation. 38 In the phase I trial evaluating Acapatamab, fewer than 20% of patients developed CRS after cycle 3. Cytokine peaks were observed 24 h post-infusion. Interleukin-6 (IL-6) pathway inhibitors (Tocilizumab or Siltuximab) were used in 40% of patients and did not influence the median CRS duration (3 days). Corticosteroid and vasopressor use were, respectively, reported in 35%–44% and 8%–15% of patients. CRS incidence and severity are closely linked to dose intensity. In the TYRP1-TCB phase I, cytokines were induced in a dose-dependent manner. 32 Peak cytokine levels occurred 24 h post-infusion with flat dosing, decreasing upon subsequent infusions. In contrast, patients treated with step-up dosing exhibited detectable cytokine increases until the third or fourth infusion. Step-up dosing mitigated CRS and was associated with lower IL-6 and tumor necrosis factor-alpha levels. In MI-CP216, investigating MEDI-565, a carcinoembryonic antigen and CD3 TCE, dexamethasone premedication effectively prevented CRS and reduced 24 h post-infusion IL-6 plasma levels. Consistent with their underlying pathophysiology, no relationship between anti-drug antibody titer and CRS was observed. 32 Importantly, no ICANS have been reported for these seven TCEs. Importantly, no ICANS was reported. Skin toxicity Skin toxicity was reported across the seven TCEs, regardless of the targeted TAA, suggesting a class effect. The typical presentation consists of a confluent macular exanthema affecting the hands and feet, associated with pruritus, exfoliation, edema, and xerosis. The temporal association with CRS suggests cytokine-mediated mechanisms. However, TYRP1-TCB demonstrated more frequent skin toxicity, consistent with on-target-off-tumor TYRP1 expression in melanocytes. Because rash and pruritus may result from cytokine release during or after CRS, the use of the IL-4/IL-13R blocker Dupilumab may be considered in selected protocols, in collaboration with dermatologists, for severe cases or those refractory to systemic corticosteroids. Other immune-related AEs The safety profile varies depending on on-target-off-tumor TAA expression, with low reported incidences (⩽10%). For example, one uveitis case occurred with TYRP1-TCB, consistent with TYRP1 expression on uveal melanocytes. 39 One grade 4 pneumonitis was reported with Acapatamab, even though PSMA is not expressed in lung tissue. 40 Tarlatamab, another half-life extended 1 + 1 TCE, has also been associated with four pneumonitis cases, including one fatality. Reozalimab represents an exception, demonstrating dose-dependent immune toxicity resembling ICI-related AEs, consistent with PD-1/PD-L1 engagement. During dose-escalation, 4 immune-related hepatitis events and 1 myocarditis occurred; during dose-expansion, 11 hypothyroidism, 4 immune-related hepatitis, 2 myositis, and 1 diabetes mellitus were reported. Hematology Hematologic TRAE rates across trials were low, approximately 10%, and largely attributable to lymphocyte count decreases. Grade 3–4 events were uncommon (0%–13%). No infections attributable to hematotoxicity were reported, except for four device-related infections in the phase I Pasotuxizumab trial, with unclear treatment association. Pasotuxizumab, a PSMA- and CD3-targeting TCE, exhibited the highest incidence of hematologic TRAEs, mainly related to lymphocyte count decrease. 35 No disseminated intravascular coagulation (DIC) has been reported (Table 3). Table 3.Incidence of treatment-related toxicities according to organ class of interest in trials investigating T-cell engager as monotherapy.TCE and trialsCRS and AE related a SkinImmune-related AEHematologyICANSAny gradeGrade ⩾3Any gradeGrade ⩾3Any gradeGrade ⩾3Any gradeGrade ⩾3Any gradeGrade ⩾3Tebentafusp IMCgp100-0160%0%68%26%0%0%20%13%0%0% IMCgp100-10286%4%87%16%0%0%0%0%0%0% IMCgp100-20283%18%83%18%0%0%0%0%0%0%Tarlatamab DeLLphi-300 b 52%1%6%2%4%2%11%8%0%0% DeLLphi-301 c 55%NRNRNRNRNRNRNR15NR DeLLpro-30082%5%0%0%0%0%35%NR122Reozalimab NCT03936959 d 0%0%13%0%10%0%7%0%0%0%TYRP1-TCB NCT0455135260%0%55%0%10%0%10%10%0%0%Solitomab MT110-101 e 43%0%40%0%0%0%0%0%0%0%Acapatamab NCT0379284197%20%27%0%1%1%12%6%0%0%MEDI-565 MI-CP21623%NR0%NR0%NR21%NR0%0%Pasotuxizumab NCT017234756%NR0%NR0%NR44%NR0%0%Ertumaxomab ERTUSO78%14%14%0%0%0%14%0%0%0%The incidence is reported for the most frequent TRAEs in each organ class. TRAE are listed in the Supplemental Table 1.aCRS-related AEs were fever/pyrexia, chills, and hypotension.bIncluding one fatal pneumonitis related to treatment.cAEs were reported regardless of causality.dInfusion-related reactions were reported for 10% of patients with pyrexia. No CRS was reported.eIncluding one fatal diarrhea related to treatment.AE, adverse event; CRS, cytokine release syndrome; ICANS, immune effector cell-associated neurotoxicity syndrome; NR, not reported; TCE, T-cell engager; TRAE, treatment-related adverse events. Incidence of treatment-related toxicities according to organ class of interest in trials investigating T-cell engager as monotherapy. The incidence is reported for the most frequent TRAEs in each organ class. TRAE are listed in the Supplemental Table 1. CRS-related AEs were fever/pyrexia, chills, and hypotension. Including one fatal pneumonitis related to treatment. AEs were reported regardless of causality. Infusion-related reactions were reported for 10% of patients with pyrexia. No CRS was reported. Including one fatal diarrhea related to treatment. AE, adverse event; CRS, cytokine release syndrome; ICANS, immune effector cell-associated neurotoxicity syndrome; NR, not reported; TCE, T-cell engager; TRAE, treatment-related adverse events. Safety optimization strategies for TCEs Safety optimization strategies for TCEs. The efficacy of TCE relies on targeted T-cell activation mediated through the formation of an immunological synapse between CD3 and a TAA. This targeted activation should limit off-tumor immune AEs commonly observed with ICIs. However, CRS driven by cytokine release, primarily IL-6, IL-10, and IFN-γ, remains a major challenge during treatment initiation. 41 No CRS-related toxic deaths or treatment discontinuations were reported. Since CRS incidence and severity are closely linked to dose intensity, step-up dosing has emerged as a strategy to reduce both their frequency and severity. Nevertheless, it may prolong CRS occurrence in subsequent cycles. 32 Because excessive CD3-mediated T-cell activation is a major driver of CRS, next-generation trispecific and tetraspecific TCEs are being developed to better control T-cell activation. For example, combining a lower-affinity CD3-binding domain with a CD28 co-stimulatory domain may enhance T-cell persistence while reducing excessive cytokine release, thereby potentially improving the safety profile.42–44 Altered keratinization may explain the skin exfoliation observed in approximately 25% of patients. Beyond gp100, other TCEs also report rashes (0%–40%) during early dosing, and the temporal association with CRS suggests cytokine-mediated mechanisms. Immune-related AEs were mostly rare and non-severe, such as uveitis reported with TYRP1-TCB, targeting TYRP1 expressed in melanocytes including uveal melanocytes. 39 Reducing on-target-off-tumor activity remains a strategy to optimize immune-related AE profiles. However, intracellular TAA expression and significant intra/inter-patient heterogeneity present challenges. 45 Masked TCEs are under development to restrict activation to the tumor site: pH-sensitive TCEs are unmasked in the acidic tumor microenvironment, and protease-activated TCEs incorporate domains cleaved by tumor-associated proteases. 46 Fc-domain engineering, when present, is also critical to minimize immune activation via Fc-receptor including natural killer (NK)-cells, macrophages, and complement activation. 47 Infectious complications represent a major concern with TCEs in hematologic malignancies, with infections reported in approximately 15% of treated patients and severe cases occurring frequently. 48 In contrast, infections appear uncommon and generally less severe in solid tumor trials. This discrepancy likely reflects major differences in the underlying disease-related immune fitness, the targeted antigens, and the pattern of treatment-induced immunosuppression between hematologic and solid malignancies. In hematologic cancers, baseline immunosuppression related to the disease itself is often compounded by prior intensive therapies and by the immunologic effects of TCEs. Hematologic TCEs are commonly directed against CD19, CD20, B-cell maturation antigen, or G-protein coupled receptor family C group 5 member D, targets that are closely linked to normal hematopoietic and immune cell lineages. 46 Consequently, these agents may induce profound immune impairment through on-target/off-tumor effects, leading to profound hypogammaglobulinemia, humoral immunity defect, lymphopenia, and prolonged cytopenias.48–50 Cytopenias have been frequently reported with hematologic TCEs and may be severe in 10%–30% of patients. 51 Decreases in lymphocyte counts, alterations in CD4/CD8 ratios, and sustained T-cell activation may further contribute to immune dysfunction and susceptibility to opportunistic infections.49,50 Interestingly, the microbiological spectrum of infections also differs between hematologic and solid tumor settings, supporting the existence of distinct patterns and degrees of immunosuppression. Gram-negative bacteria, resistant gram-positive bacteria, and viral infections were more frequent with TCEs used in hematology. 48 Given the relatively low frequency and severity of infections reported with TCEs in solid tumors, routine anti-bacterial or anti-fungal prophylaxis has not been systematically recommended. However, as proposed in hematologic malignancies, prophylactic strategies may still be considered for selected high-risk patients or specific treatment combinations.52–54 An additional difference between hematologic and solid tumor TCEs is the absence, to date, of DIC reported in solid tumor TCE trials, whereas coagulation abnormalities and DIC have occasionally been described in hematologic settings. Pharmacokinetic modulation offers another strategy to improve safety. TCEs generally have short half-lives due to their small size and rapid catabolism into peptides; repeated dosing is required to maintain therapeutic exposure. For example, Tebentafusp has a terminal half-life ranging from 6.8 to 7.5 h, supporting weekly administration. 1 Various TCE formats have been designed to extend half-life. Addition of an Fc domain prolongs half-life and introduces potential interactions with complement proteins and Fc receptors on innate immune cells (Figure 1), though this may increase half-life variability. Tarlatamab, a 1 + 1 extended-half-life TCE, demonstrates terminal half-life ranges between 4.3 and 26.5 days (median 11.2 days), allowing dosing every 2 weeks. 2 Alternatively, incorporation of serum albumin-binding domains can extend half-life. HPN328 is a trispecific TCE containing a CD3 binder, an albumin binder, and a DLL3 binder, with a mean half-life of 78–187 h. 55 Although extended half-life formats provide more convenient administration schedules, their impact on safety remains uncertain. T-cell engager formats. (a) 1 + 1 format composed of two-single variable fragment (scFv). (b) HLE 1 + 1 format, in which two scFv are fused to a crystallizable fragment (Fc) domain to confer prolonged half-life and potential for interactions with complement proteins and Fc receptors on innate immune cells. (c) 2 + 1 format composed of two tumor-targeted binding domain and one immune-effector domain. (d) Soluble TCR-based bispecific proteins consisting of a scFv fused to a TCR domain. Soluble TCR allows engagement of intracellular tumor-associated antigens expressed at the cell surface by the HLA. (e) Full-length IgG-like format. HLA, human leukocyte antigen; HLE, half-life extended; IgG, immunoglobulin G; TCR, T-cell receptor. Conclusion TCEs are transforming the therapeutic landscape of solid malignancies, yet their safety profile remains incompletely characterized compared with hematologic indications. CRS and dermatologic toxicities emerge as class effects driven by systemic cytokine activation and, in some cases, on-target-off-tumor antigen expression. ICANS, immune-related and hematologic AEs remain generally infrequent. Optimizing safety will rely on refined dosing strategies, pharmacological engineering, and more selective tumor-target specificity. As these agents are expanding in solid tumors, prospective real-world data and harmonized toxicity management guidelines will be essential to ensure safe implementation in routine practice. Supplemental Material sj-docx-1-tam-10.1177_17588359261467651 – Supplemental material for A review on toxicities of T-cell engager-based therapies in solid tumors Supplemental material, sj-docx-1-tam-10.1177_17588359261467651 for A review on toxicities of T-cell engager-based therapies in solid tumors by Etienne Bastien, Clara Vacheret, Raphael Sanchez, Francesca Lucibello, Nicolas Girard and Emanuela Romano in Therapeutic Advances in Medical Oncology