Traditional cytotoxic chemotherapy has historically been the primary approach for treating various malignant tumors. However, in recent years, remarkable advancements in cancer management strategies have emerged with the introduction of immunotherapies, signaling a new era in anti-neoplastic therapy [1]. Immune checkpoint inhibitors (ICIs), predominantly composed of programmed cell death protein 1, programmed cell death 1 ligand 1 (PD-L1), and cytotoxic T-lymphocyte antigen-4 monoclonal antibodies, constitute a class of immunotherapy that enhances antitumor immune responses by upregulating T-cell activity [1-3]. While ICIs have demonstrated high response rates in patients with various advanced malignancies, they can also be associated with several toxicities affecting any organ system including the nervous system [4]. Neurotoxicity triggered by ICIs can impact various components of the nervous system, including the central nervous system (CNS), the peripheral nervous system (PNS), and the neuroendocrine system [1]. Although neurological toxicities of ICIs are rare accounting for approximately 2% to 4% of all adverse effects, they can exhibit increased severity compared to other complications and pose life-threatening risks if left undiagnosed or poorly managed [1,2]. Previous research has suggested that neurologic adverse effects have been implicated in nearly half of all deaths associated with ICIs [4]. Atezolizumab, an immune checkpoint inhibitor that selectively binds to PD-L1 [5], is approved for the treatment of non-small-cell lung cancer (NSCLC), small-cell lung cancer (SCLC), advanced triple-negative breast cancer, hepatocellular carcinoma (HCC) and urothelial carcinoma and is currently under study for the treatment of lymphoma, melanoma, gynecological and colorectal malignancies [6,7]. Atezolizumab exhibits a side effect profile comparable to other ICIs, commonly manifesting as fatigue, rash, and gastrointestinal symptoms [8,9].

The incidence of neurological adverse events associated with atezolizumab is relatively lower compared to other ICIs. However, several serious cases of nervous system toxicities have been reported following atezolizumab therapy [7]. Diagnosing neurological adverse events poses a significant challenge due to often atypical clinical symptoms and laboratory findings, coupled with limited practical experience in managing ICI-related toxicities. With the increasing use of ICIs in cancer therapy, there is an anticipated rise in the incidence of neurotoxicities. Delayed recognition of these adverse events as being drug-related can exacerbate patient vulnerability to further toxicity. Currently, the literature lacks a comprehensive characterization of the clinical course and specific symptoms linked to neurological manifestations associated with atezolizumab. Here, we review atezolizumab induced neurotoxicity, aiming to describe its clinical presentation, delay of onset and resolution of symptoms, diagnostic findings, treatment options, and patient outcomes.

Search strategy 

A medline search on atezolizumab induced neurotoxicity using PubMed, Science Direct, and Google Scholar databases was performed and completed on March 15, 2024.

The search strategy included MeSH terms and free-text words. Search terms included: atezolizumab, PD-L1, neurotoxicity, neurological adverse events, neurological complications, neurological immune related side effects, central nervous system, encephalitis, encephalopathy, seizures, coma, myoclonus, confusion, aphasia, ataxia, and peripheral neuropathy. Duplicates were removed and the references of the included articles were cross-checked. Studies that discussed neurotoxicity associated with all immune checkpoint inhibitors without explicitly mentioning atezolizumab were excluded. Articles studying anti-PD-L1 agents without specifying atezolizumab were also excluded. Additionally, paraneoplastic neurological manifestations induced by atezolizumab were ruled out from our review.

These exclusion criteria were implemented to ensure that our literature review focused specifically on neurotoxicity attributed to atezolizumab therapy.

Study selection

We included English-language publications encompassing all age groups, randomized clinical trials, observational studies, systematic reviews, case reports, and case series. Table 1 presents data from clinical trials and large scale retrospective studies identified in the literature. In Table 2, data from case studies and observational studies were compiled detailing patient demographics, cancer type, neurotoxic symptoms including onset and recovery timing, diagnostic procedures, co-administered chemotherapies, interventions, and clinical outcomes. Most case reports underwent thorough investigations to rule out infection, tumor progression, or other causes of neurotoxicity, attributing the majority of cases to atezolizumab. Variables were labeled as “unable to assess” if pertinent patient data were unavailable.

A total of 56 articles were identified. Among these, 13 were clinical trials and only one was a prospective cohort study, while the remainder were retrospective data. This included 32 single-case reports, 1 case series, and 9 retrospective studies.

Controlled data from clinical trials and meta-analysis [10-22]

An analysis of data from controlled clinical trials shows that atezolizumab has a favorable safety profile. The most common adverse reactions (≥ 20%) included fatigue, nausea, urinary tract infection, fever, and constipation. The risk of adverse effects with atezolizumab is comparable to other chemotherapeutic agents and aligns with the incidence rates observed for other approved immune checkpoint inhibitors like pembrolizumab and nivolumab [12,18]. Immune-related adverse events (irAEs), including neurotoxicity, observed in patients treated with atezolizumab were predominantly low grade and manageable, with only a small number necessitating dose interruption or discontinuation alongside corticosteroid treatment. Both the POPLAR and OAK trials demonstrated favorable tolerability of atezolizumab compared to docetaxel, with a lower proportion of patients experiencing grade 3 or 4 treatment-related side effects [14,17]. Specifically, only 2.1% of irAEs in the atezolizumab group required treatment discontinuation [15]. Another Japanese patient study also revealed comparable rates of all-grade treatment-related adverse events between atezolizumab and docetaxel groups, albeit with fewer grade 3-4 events in the atezolizumab cohort [13]. Notably, neurological immune-related adverse events (irAEs), particularly neuropathy, have emerged as significant concerns in most previous clinical trials [10]. Peripheral neuropathy occurred approximately in 7% of patients in the atezolizumab group vs 1% in the placebo group [11,12,14]. Conversely, in the Japanese study, no cases of peripheral sensory neuropathy were reported, while serious neurological adverse effects such as meningoencephalitis and Guillain-Barre syndrome occurred in the atezolizumab group but were absent in the docetaxel group [13]. Notably, encephalitis was not reported as an irAE in earlier phases of POPLAR trials [17] but occurred at a low rate in subsequent studies, 0.8% and 0.3% in the OAK trial and the Impower 150 study, respectively [12,14]. Mikami’s analysis indicates that meningoencephalitis was the most common neurological complication associated with atezolizumab, occurring in 30.3% of cases, followed by Guillain-Barre syndrome and demyelinating disorders, each accounting for about 10.5% of cases. The least frequent complication was hypophysitis, occurring in 2.5% of cases [23]. Overall, the controlled data from these trials indicate a manageable safety profile for atezolizumab, with distinct advantages over traditional chemotherapy agents.

Overall incidence and severity of atezolizumab-induced neurotoxicities

Neurological adverse events associated with atezolizumab exhibit diverse clinical presentations affecting various parts of the nervous system (Figure 1). The most frequently reported manifestations are grades 1-2 neurotoxicities, often presenting as nonspecific symptoms, such as asthenia, headaches, dizziness, paresthesias, or dysgeusia [26]. Grades 3-4 neurological syndromes are less commonly reported and encompass severe conditions such as encephalitis, encephalopathy, aseptic meningitis, myelitis, neuropathy, Guillain-Barre syndrome, myasthenia gravis, and demyelinating disorders [7,26].

Epidemiological and clinical characteristics of patient cases of atezolizumab-induced neurotoxicities reported in literature

Table 1 documented 39 studies detailing grade 3 and 4 neurotoxicities induced by atezolizumab, involving a total of 45 patients [1,5-9,26-58]. Conclusions regarding clinical characteristics can be drawn from the data provided for these patients. The demographic profile of patients experiencing atezolizumab-induced neurotoxicity revealed a male predominance (46.7%) with a median age of 58 years (range 37–87). The primary tumor localizations varied with lung being the most common (n = 26) followed by bladder (n = 9), breast (n=4), liver (n=3), kidney (n=1), colon (n=1) and cervix (n = 1). Atezolizumab dosage was consistent across cases, administered at 1200 mg every three weeks. Symptoms of neurological toxicity typically manifested after the first cycle of atezolizumab therapy (37.8% of cases), with a median onset occuring 2 weeks after the last dose (range: 1day-1year). Six documented cases had a history of neurological disorders, including three with brain metastases [9,32,49], and four with preexisting demyelinating diseases [1,36,53,54]. Co-administration of chemotherapeutic drugs, notably bevacizumab, carboplatin, paclitaxel, and cobimetinib, was common in nearly half of the cases. Atezolizumab was discontinued in all published cases and treatment of neurotoxicities varied, including corticosteroids, antiepileptic drugs, empiric antimicrobial therapy, intravenous immunoglobulin, and plasmapherisis. Symptoms typically resolved within a median of 10 days after cessation of atezolizumab (range: 2days-2years) with partial or complete recovery noted in the majority of cases (82.6%). Atezolizumab rechallenge was successful in three cases [8,27,33] while one case reported negative rechallenge with pembrolizumab [46]. Recurrence of symptoms despite withdrawal of atezolizumab after a period of remission occurred in two cases [29,38]. Fatal outcomes were observed in 10 cases [7,9,26,38,39,42,50,52,53,55], however, definitively attributing neurotoxicities as the cause of death was challenging due to initial symptom improvement upon drug discontinuation and incomplete exclusion of disease progression in some cases. The neurotoxicities underlying fatal outcomes included encephalitis (n=6), multiple sclerosis (n=2), and single cases of myasthenia gravis and cerebellar ataxia.

Published cases of atezolizumab induced neurotoxicity (Table 1) encompassed various neurological manifestations, with encephalitis being the most common (39.6%), followed by cerebellar ataxia (10.4%), meningitis (8.3%), optic neuritis (6.3%), multiple sclerosis flare (6.3%), posterior reversible encephalopathy syndrome (4.2%), peripheral neuropathy (4.2%), Guillain-Barre syndrome (4.2%), myasthenia gravis relapse (4.2%), myelitis (4.2%), facial palsy (4.2%) and neuromyelitis optica (2.1%). Additionally, other neurological adverse events were reported in clinical trials and large-scale retrospective pharmacovigilance studies, including aphasia, hypophisitis, and paralysis.

Types of neurological adverse effects associated with atezolizumab

Encephalitis: Among CNS neurotoxicities associated with atezolizumab, encephalitis stands out as a rare yet potentially fatal adverse reaction [7]. The incidence rate of encephalitis following atezolizumab therapy, as reported in the OAK trial, was only 0.8% in patients with NSCLC [5].

At the time of this article, our review included 19 published cases of encephalitis following atezolizumab therapy (Table 1), with a few additional cases mentioned in clinical trials and large scae retrospective studies (Table 2). The presentation of atezolizumab-induced encephalitis revealed typical features but was also demonstrated heterogeneity, encompassing symptoms such as fever, headache, confusion, gait instability, seizures, and in rare instances, meningeal signs. The onset of symptoms varied; with most cases occuring between 10 and 21 days after atezolizumab administration, though some cases presented much later, such as nine months post-administration [40]. Common CSF features include pleocytosis and elevated protein levels while MRI findings often revealed leptomeningeal enhancement or brain parenchymal lesions although pathological findings in imaging were absent in some cases. Management of encephalitis linked to ICI treatment remains uncertain; however favorable responses to steroid therapy were observed in 14 of our cases. Only a small number of patients required intravenous immunoglobulin [40,41,43,50,52] or plasmapheresis [42]. It's worth noting that in 2022, we documented a case involving a patient with SCLC who experienced recurrent seizures 21 days after completing the seventh cycle of atezolizumab treatment. Extensive investigations ruled out the diagnosis of encephalitis and the patient's symptoms were successfully managed with leviteracetam, leading to complete recovery within a week. Atezolizumab was subsequently reintroduced after a one-month period of remission without any recurring neurological symptoms [27].

Aseptic meningitis: Aseptic meningitis occurs in approximately 0.1–0.2% of patients treated with ICIs. Within our review, we identified four cases of aseptic meningitis induced by atezolizumab [44,45]. Additionally, two other studies have reported cases of meningitis associated with atezolizumab. Aseptic meningitis typically manifested between 11 to 14 days following the initial administration of atezolizumab in three cases, while in the fourth case, it occurred 11 days after the third dose. Fever could signal the onset of meningitis. Other common symptoms included altered consciousness and headache. CSF analysis revealed lymphocytic meningitis and high protein level accompanied by meningeal enhancement observed on MRI scans. All documented cases of aseptic meningitis are fully resolved with the administration of steroids and cessation of atezolizumab treatment.

Encephalopathy: Two cases of posterior reversible encephalopathy syndrome (PRES) occurring in patients receiving atezolizumab have been documented in the literature [33,34]. Symptoms manifested differently in each case: one patient experienced symptoms 24 hours after the first dose, while the other developed symptoms 6 months after the fourth cycle. Neurological manifestations included altered consciousness, visual disturbances, focal neurological deficits, seizures, and typical imaging alterations primarily affecting the posterior parietal and occipital lobes on MRI. Notably, both patients were presented with elevated blood pressure at the time of PRES diagnosis: 206/108 mmHg in a patient undergoing atezolizumab treatment for small cell lung cancer [33] and 169/81 mmHg in another patient treated for breast cancer [34]. In both case reports, there was marked neurological improvement following antihypertensive therapy in the subsequent days.

Cereballar ataxia: In our review, we identified five cases of acute cerebellar ataxia induced by atezolizumab. Furthermore, ataxia was previously reported in a phase 2 clinical trial [16]. The time lapse between the initiation of atezolizumab and the onset of ataxia was unspecified in most cases, except for one instance where ataxia appeared two weeks after starting atezolizumab. Symptoms of cerebellar syndrome observed in the documented cases included gait disturbances, ataxia, dysarthria, nystagmus, and nausea. Treatment typically involved corticosteroids, leading to complete recovery in two cases, partial improvement in one case and initial improvement within one week followed by eventual death due to metastatic disease in another case [55]. Unfortunately, no clinical improvement was observed in the remaining case [35].

Peripheral neuropathy: Peripheral neuropathy is a prominent aspect of the literature concerning ICI-associated neurotoxicity, although it has been described as a complication of atezolizumab primarily in clinical settings. Both sensory and motor peripheral neuropathies have been documented, presenting in acute or chronic forms. Within our review, we encountered a case report highlighting enteric plexus neuropathy; however, patient characteristics were inaccessible [36]. In another instance, a patient developed peripheral neuropathy associated with facial palsy two weeks after completing the last cycle of atezolizumab. Symptoms resolved within seven days, but residual paresthesia persisted in the toes [51].

Guillain Barre syndrome (GBS): GBS induced by ICI is a rare occurence, with only a few reported cases in the literature. In a prospective cohort study, two cases of Guillain-Barré syndrome induced by atezolizumab were documented [43]. The presentation was typical, characterized by limb weakness and facial palsy. Symptoms appeared 15 and 18 days, respectively, after receiving atezolizumab treatment. Both patients were treated with intravenous immunoglobulin and corticosteroids. Atezolizumab was discontinued in both instances, and complete recovery was achieved within a few days of initiating steroid therapy.

Paralysis: Our review identified two cases of paralysis induced by atezolizumab [8,51]. In both instances, patients developed peripheral facial palsy after two weeks of atezolizumab therapy. Symptoms resolved in both patients after discontinuation of atezolizumab. Interestingly, in one case, there was no recurrence of symptoms upon rechallenge. Additionally, two other patients with Guillain-Barre syndrome experienced cranial nerve palsy [43]. Paralysis was also observed in a previous clinical trial [18].

Multiple sclerosis (MS): In our review, we identified three MS patients who experienced relapse while undergoing treatment with atezolizumab. The history of MS was confirmed in all instances. Among these cases, two patients developed encephalopathic symptoms during their relapse [1,53], accompanied by blurred vision and weakness in one case [38]. The median onset of symptoms was 15 days. Imaging results were consistent with typical MS manifestations in all patients. Corticosteroids were administered in every case. Unfortunately, a fatal outcome was observed in two patients, while the remaining case achieved complete recovery.

Optic neuritis (ON): In the literature, ON has been reported following atezolizumab treatment. Three cases were included in our review. The onset of symptoms occurred three weeks after treatment initiation in two cases [48,49], while in the remaining case, ON manifested after 12 months [56]. Optic neuritis tended to be bilateral in most cases. MRI findings showed optic nerve enhancement abnormalities in only one case. Corticosteroids were administered to all three patients. Resolution of symptoms was observed in two cases, while the outcome of the remaining patient was not specified.

Myasthenia gravis (MG): The most frequently reported neuromuscular disorder associated with ICIs is MG. It can manifest either as de novo or as an exacerbation of pre-existing myasthenia. However, only two case reports of atezolizumab induced MG were found in the literature. Chae et al. reported a case of MG exacerbation emerging six weeks after initiating atezolizumab [54]. The progression was further complicated by hypercapnic respiratory failure, requiring intubation. However, stability was achieved following five sessions of plasmapheresis. Additionally, Thakolwiboon et al. documented a case of new onset MG following atezolizumab therapy, which resulted in fatal outcome due to cardiac arrest [50]. In addition to the neurotoxicities mentioned earlier, our review revealed one case each of longitudinal extensive transverse myelitis [37] and neuromyelitis optica [28].

Clinical understanding of the toxic effects of ICIs remains relatively limited [1]. Due to the relatively low occurrence of ICI-related neurologic adverse events, there is limited data available, with most of these adverse effects documented in case reports. The majority of published reviews and large-scale studies have examined immune checkpoint inhibitors collectively, rather than focusing solely on any particular agent. This systematic review is the first to describe various types of neurotoxicities induced by atezolizumab and detail the range of symptoms, diagnosis features, and timing of onset and resolution.

Characteristics of atezolizumab-induced neurotoxicity

Encephalitis emerges as the most extensively studied neurotoxicity associated with atezolizumab therapy [4]. However, findings from clinical trials indicate that peripheral neuropathies are the most prevalent among observed neurological adverse events. In patients receiving atezolizumab, neurologic irAEs were most commonly observed in those with lung cancer [23]. In our review, the mean age of patients who developed atezolizumab-related multiple sclerosis (MS) flare-ups and ataxia were the youngest (46.7 and 56.25 years, respectively), while patients with myasthenia gravis and encephalitis were the oldest (77 and 60 years, respectively) [23]. Symptoms of atezolizumab associated neurotoxicity often exhibit a delayed onset, typically appearing around 15 days after drug initiation. However, in some instances, neurological toxicities occur much later, with rare cases emerging beyond 2 months after the start of atezolizumab treatment. Nearly all cases occur shortly after the first dose, with no instances reported following drug cessation. Atezolizumab-related multiple sclerosis or meningitis occurred significantly earlier (median of 15 days) compared to other neurologic irAEs.The broad range of onset times for neurological toxicity may compound clinicians' challenges in identifying and diagnosing atezolizumab-related neurotoxicity. MRI abnormalities are observed in almost all patients; however, certain findings lack specificity and could potentially signify alternative underlying causes. Neurological events are progressive unless drug discontinuation or interruption is initiated. Management of severe neurological events involves temporary immunosuppression utilizing steroids or alternative agents such as intravenous immunoglobulins, plasmapheresis, or in some cases, rituximab. These interventions lead to clinical resolution or improvement of symptoms in the majority of cases.

Comparison of neurological adverse effects between atezolizumab and other immune checkpoint inhibitors

The reported incidence and time course of irAEs in clinical trials have varied depending on the type of ICIs used. A recent meta-analysis identified atezolizumab as having the best safety profile [10,20,39]. Other studies reported an incidence of neurologic irAEs up to 5% with PD-1/PD-L1 inhibitors, and 12.7% with CTLA-4 inhibitors [23]. Moreover, anti-CTLA-4 agents have been associated with higher severity of irAEs. Clinical trials and meta-analyses report grade 3 or 4 neurologic adverse events occurring in 0.3–0.8% of patients under anti-CTLA-4 (ipilimumab) therapy, 0.2–0.4% under anti-PD-1 (nivolumab or pembrolizumab) treatment, and 0.1–1% under anti-PDL-1 (atezolizumab) treatment. Combined ipilimumab and nivolumab treatment increases the incidence of grade 3 and 4 neurologic adverse events to 0.7%. Anti-PD-1/L1 therapy is more frequently associated with myasthenic syndromes and less common in meningitis and cranial neuropathies, while anti-CTLA-4 therapy is more frequent in meningitis and less common in encephalitis and myositis [59]. In patients on anti-PD-1 or anti-PD-L1 monotherapies, neurologic AEs were most commonly observed in those with non-small cell lung cancer. In contrast, in patients on anti-CTLA-4, neurologic AEs were most commonly observed in those with melanoma [23]. Notably, no cases of melanoma were found in our review. Anti-PD-L1 monotherapy, predominantly atezolizumab, was associated with an earlier onset of neurologic adverse events compared to anti-PD-1, anti-CTLA-4, and combination therapies [23]. Concerning ICI dosage, there is no clear relationship between the incidence of neurologic adverse events and drug dosage with anti-CTLA-4 antibodies. However, findings are inconsistent for anti-PD-1 agents, with increased neurological adverse events at 10 mg/kg nivolumab compared to lower doses, but the reverse observed with pembrolizumab [60]. Studies regarding anti-PDL1 inhibitors are lacking, with no available data on the correlation between atezolizumab and treatment dosage or schedule. Interestingly, age, sex, and metastatic status were not significant risk factors for overall neurologic ICI-related AEs [23].

Possible mechanisms of immune checkpoint inhibitors-induced neurotoxicity

The exact pathophysiology of ICI-associated neurotoxicity remains unclear, with multiple proposed mechanisms. First, increased T-cell activity against antigens shared by cancerous and healthy tissues potentially leads to an exaggerated inflammatory response and autoimmune neurologic damage due to unregulated T-cell activation against nerves [8]. Second, immunotherapeutic agents may elevate levels of inflammatory cytokines and trigger augmented complement-mediated inflammation by binding antibodies against PD1 and CTLA-4 expressed in normal tissue. Notably, studies have shown correlations between the presence of autoantibodies, particularly antineuronal antibodies, and improved survival but increased neurological toxicity in patients treated with checkpoint inhibitors [46]. Third, genetic susceptibility was suggested by a cohort study where the HLA-B27:05 genotype was over-represented in patients who developed autoimmune encephalitis after receiving atezolizumab [32].

Strengths: This study highlights uncommon but serious irAEs arising from immune checkpoint inhibitors (ICIs). It represents the largest review of patients who developed neurological irAEs from an anti-PD-L1 inhibitor; atezolizumab. This review provides a comprehensive analysis of atezolizumab-induced neurotoxicity by meticulously compiling epidemiological and clinical characteristics from case reports and retrospective studies. It covers the spectrum of neurological adverse effects, their time course, diagnostic investigations, management options, and outcomes. Additionally, it includes a comparative analysis of the safety profiles between atezolizumab and other immune checkpoint inhibitors, as well as discussions on potential mechanisms. The strength of this review lies in its thoroughness and the novel insights it offers, particularly in identifying patterns and providing a detailed comparison of adverse effects. This information is crucial for clinicians to better understand, anticipate, and manage neurotoxic effects in patients treated with atezolizumab.

Limitations: Several limiting features of this review deserve comment. With the exception of one prospective cohort study, published data available are limited to case series, single-case reports, and retrospective pharmacovigilance studies. The characteristics of these studies restrict our systematic review to descriptive reporting and preclude an examination of risk factors. The observational design of reports describing treatment for atezolizumab neurotoxicity precludes any statement about the efficacy of any specific strategy. Distinguishing atezolizumab-induced neurotoxicity from many conditions present in critically ill patients remains clinically challenging, and concurrent diagnoses may confound its identification. In fact, cancer patients commonly exhibit neurological complications such as brain metastasis, paraneoplastic syndrome and cerebrovascular disorders. Adding to this complexity, chemotherapeutic agents and radiation therapy may also be risk factors for neurotoxicity. Many questions remain regarding the true incidence and scope of atezolizumab-associated neurotoxicity. Further research evaluating atezolizumab adverse effects may provide vital information to determine key trends. Prospective evaluations with more standard and rigorous datasets are needed.

Recommendations: While our systematic review provides valuable insights into the neurotoxicity associated with atezolizumab, it is evident that further research is needed to address several important recommendations. Prospective registries collecting standardized clinical data, controlled trials comparing neurotoxicity profiles among different checkpoint inhibitors, and deeper investigations into the pathophysiology of neurotoxic syndromes are crucial steps to better understand and manage these adverse effects. Additionally, future case reports should aim to include severity grading of events and rigorously exclude alternative causes to strengthen causal conclusions. Larger retrospective analyses pooling detailed clinical data internationally could further elucidate risk factors and outcomes associated with specific treatment strategies. Embracing these recommendations will undoubtedly contribute to a more comprehensive understanding of this serious adverse drug reaction.

Given that case reports have documented atezolizumab-induced neurotoxicity across diverse cancer types, it is imperative to establish a comprehensive safety profile for this agent. Further investigations could enhance our understanding of which patients are at risk and how we can safely manage this serious adverse reaction.