Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment by harnessing the body's immune system to target tumors, leading to significant improvements in survival rates for many patients. However, their use is accompanied by a unique spectrum of immune - related adverse events (irAEs), which can range from mild to life-threatening and impact various organ systems[1]. Understanding these AEs is crucial for optimizing patient care, as they can not only affect treatment adherence and quality of life but also, in severe cases, necessitate treatment discontinuation or result in irreversible damage[2].

The advent of large-scale real-world data sources, such as the U.S. Food and Drug Administration's (FDA) Adverse Event Reporting System (FAERS), has provided an invaluable resource for comprehensively evaluating the safety profiles of medications, including ICIs[3]. FAERS collects spontaneous reports of adverse events associated with drugs and medical products from healthcare professionals, patients, and other sources. By analyzing data from this database, researchers can identify signals of potential AEs that may not have been fully elucidated in pre-marketing clinical trials, which often have limited sample sizes, short follow-up periods, and strict inclusion and exclusion criteria.

In the context of ICIs, while numerous studies have investigated general irAEs like dermatological, gastrointestinal, and endocrine manifestations, the focus on neuropsychiatric adverse events has been relatively scarce[4]. Neuropsychiatric symptoms associated with ICIs can include depression, anxiety, insomnia, cognitive impairment, and in some cases, more severe manifestations such as suicidal ideation. These symptoms can have a profound impact on a patient's mental well-being, social functioning, and overall prognosis. For example, depression can lead to decreased motivation for treatment and self - care, potentially compromising the effectiveness of cancer therapy.

The period from 2011–2025 has seen a substantial increase in the use of ICIs, paralleled by a growing body of evidence regarding their associated AEs. During this time, more data has become available in the FAERS database, allowing for a more in-depth and contemporary analysis of the neuropsychiatric risks linked to ICIs. By leveraging the ROR-based quantitative evaluation and visual representation, we aim to systematically analyze the neuropsychiatric AEs of ICIs across multiple dimensions. This approach will enable us to identify potential risk factors, such as demographic characteristics (sex and age), geographical variations, temporal patterns, and the severity of these adverse events. Understanding these factors is essential for developing targeted strategies for early detection, prevention, and management of neuropsychiatric AEs in patients receiving ICIs, ultimately improving their overall treatment outcomes and quality of life.

Immune checkpoint inhibitors (ICIs) have significantly improved the treatment landscape for various malignancies, while also being associated with immune-related adverse reactions (irAEs) affecting multiple organ systems. Their widespread occurrence poses challenges for diagnosis and management in future clinical practice. A study by Fang Wu demonstrated that emotional distress is associated with poor clinical outcomes in patients with advanced non-small cell lung cancer (NSCLC) receiving ICI treatment, highlighting the potential importance of addressing emotional distress in cancer management[9]. Previous studies have summarized ICI-related mental adverse reactions from the FAERS database between January 2012 and December 2021[10], whereas the present study focuses on the period from 2004 to 2025. Currently, most proposed mechanisms rely on retrospective studies or case reports, lacking support from large-scale prospective data. Additionally, the difficulty in obtaining central nervous system (CNS) tissue samples further limits in-depth exploration of the underlying pathological mechanisms[11], resulting in insufficient evidence to support the real-world impact of these adverse reactions.

The potential mechanisms underlying immune checkpoint inhibitor (ICI)-related neuropsychiatric effects involve cytokine dysregulation and immune-mediated processes[12]. Research indicates that elevated serum levels of proinflammatory cytokines like interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α) following ICI treatment correlate with increased depressive symptoms[13]. These cytokines may cross the compromised blood-brain barrier or activate central nervous system microglia, leading to local cytokine production[14][15]. Such inflammatory mediators can disrupt hypothalamic-pituitary-adrenal axis function and alter neurotransmitter metabolism, particularly serotonin and dopamine[16], mirroring the pathophysiology of primary mood disorders but with immune activation as the initiating factor[17].

Regarding mild-to-moderate neuropsychiatric adverse events like anxiety and insomnia, while CD8 + T-cell infiltration has been observed in severe cases like ICI-associated encephalitis[18], this finding hasn't been replicated in patients with isolated neuropsychiatric symptoms. Instead, non-cytotoxic immune activation appears more relevant, particularly microglial activation triggered by peripheral inflammatory signals[19][20]. Activated microglia can impair synaptic plasticity and neurotransmitter release, contributing to mood and sleep disturbances[21][22][23]. Supporting this, studies demonstrate that dual CTLA-4/PD-1 blockade destabilizes the neuroimmune network, inducing microglial activation that may lead to persistent neurodegeneration and cognitive deficits[24].

Other potential contributors to neuroadverse reactions involve mechanisms such as molecular mimicry, where tumor antigens resemble those in healthy neural tissues, leading to immune cross-reactivity; epitope spreading, which occurs when tissue damage releases antigens that broaden immune targeting; and the influence of pre-existing neural antibodies or gut microbiota in modulating immune responses[11].

The temporal and geographic trends observed in FAERS data offer important real-world insights, though these findings should be interpreted carefully due to inherent database biases and clinical contexts to prevent overestimation or misattribution. From a temporal standpoint, the rise in reports of mental health-related immune-related adverse events (irAEs) aligns with the broader use of immune checkpoint inhibitors (ICIs), including newer agents like tislelizumab, as well as increased clinician awareness of neuropsychiatric toxicity[25][11]. However, this trend does not necessarily indicate a higher inherent risk. As ICIs have become standard therapy for a wider range of cancers such as non-small cell lung cancer (NSCLC), melanoma, and renal cell carcinoma the treated population has grown more diverse, encompassing more elderly patients and those with pre-existing mental health conditions, both of whom are more prone to experiencing and reporting such symptoms[26]. Additionally, the release of clinical guidelines, like the 2022 ASCO recommendations for irAE management, has encouraged more structured reporting of mental health events, which were previously often misclassified as cancer-related distress rather than drug-induced toxicity.

The study retrieved Preferred Terms (PTs) for mental disorders from the Medical Dictionary for Regulatory Activities (MedDRA), specifically screening reports related to immune checkpoint inhibitors (ICIs). The Reporting Odds Ratio (ROR) served as the primary signal detection method, revealing that delirium exhibited the highest reporting frequency across all stratifications including gender, age, onset time, and reporting country while the most significant ROR values were associated with dysphoria and laziness. Delirium, an acute and fluctuating state of consciousness impairment, primarily manifests as diminished attention and altered cognitive function. Its pathogenesis is multifactorial, involving acute cerebral dysfunction triggered by various endogenous and exogenous factors affecting the body, particularly the central nervous system (CNS). Post-ICI administration complications such as metabolic disturbances[27] (e.g., thyroid dysfunction, dyslipidemia, electrolyte imbalances), cardiotoxicity[28][29], pneumotoxicity[30], and hepatotoxicity[31] may collectively contribute to delirium's high incidence.The significantly elevated reporting odds ratios (RORs) for dysphoria and laziness suggest a notable association between these symptoms and immune checkpoint inhibitor (ICI)-related mental adverse reactions. ICIs can trigger organ-specific immune-related adverse events (irAEs)—such as rash, diarrhea[32], hepatotoxicity[33], and pneumonitis[30]—which often lead to physical discomfort including pain, fatigue, and shortness of breath. When these physical symptoms persist or become severe, they may initiate or worsen dysphoric mood through physiological-psychological interactions. Furthermore, the immune activation driven by ICIs demands substantial energy expenditure, contributing directly to fatigue. Concurrently, irAEs affecting systems such as hematologic[33] (e.g., anemia, neutropenia), endocrine[27] (e.g., thyroid dysfunction, adrenal insufficiency), and digestive[32] (e.g., diarrhea, nausea, vomiting) can also induce fatigue through multiple biological pathways.

Reports on immune checkpoint inhibitors (ICIs) are predominantly documented in North America and Europe, largely due to better treatment accessibility and stronger pharmacovigilance systems in these regions. However, the voluntary nature of the U.S. FDA's Adverse Event Reporting System (FAERS) contributes to underreporting, particularly in resource-limited areas, where the extent of this issue remains unclear[34]. This implies that the seemingly "low incidence" of adverse events in these regions may not reflect actual risk but rather reporting gaps, potentially obscuring a rising trend in mental health-related immune-related adverse events (irAEs) and highlighting the necessity for tailored pharmacovigilance programs. Furthermore, interpreting drug-specific report counts requires contextual understanding—drugs like pembrolizumab and nivolumab have more reports than newer agents such as tislelizumab, primarily because of their longer market presence (over a decade versus less than five years) and wider approved uses. Another challenge is the absence of standardized diagnostic criteria for mental health irAEs in FAERS. For instance, vague terms like "insomnia" are recorded without differentiating whether they stem from ICI toxicity, cancer-related fatigue, or pre-existing sleep disorders. This lack of precision may inflate reports of mild symptoms while underrepresenting severe, subtype-specific conditions like ICI-induced major depressive disorder.

In addressing the clinical management of mental adverse reactions, Pablo Gajate’s review emphasizes a strategy centered on risk stratification and tailored interventions, focusing on two core elements: the temporary or permanent discontinuation of immune checkpoint inhibitors (ICIs) and the application of immunomodulatory therapies[10]. Building on this framework and supported by Luo Peng’s related findings[9], our study highlights key high-risk groups for ICI-associated mental adverse reactions. These include patients aged 75 and above, those diagnosed with lung cancer, skin cancer (such as melanoma), or renal cancer, as well as individuals undergoing combination ICI therapy. For these high-risk patients, early baseline assessments using validated tools are essential—specifically, the Patient Health Questionnaire-9 (PHQ-9)(Table S10)[35] to evaluate depressive symptoms, with closer monitoring advised for scores of 5 or higher, and the Generalized Anxiety Disorder-7 (GAD-7) [36] scale to assess anxiety symptoms. It is also important to screen for concomitant use of medications that could worsen mood-related symptoms, such as glucocorticoids. For low-risk patients, a more streamlined, symptom-triggered monitoring approach can be adopted, where evaluations are only initiated when patients report specific issues like delirium (mainly marked by reduced attention, disorientation, or altered consciousness), sleep disturbances, mood changes, or cognitive difficulties. Based on the timing patterns observed in this study, adverse reactions most frequently occur within the first 30 days and between 31 to 60 days of treatment (excluding unspecified or overall categories). Accordingly, we recommend that high-risk patients be monitored every two weeks during the initial two months of therapy to allow for prompt intervention, while low-risk patients may be assessed on a monthly basis over the same period.

This study leverages an extended and updated temporal scope of FAERS data (2004–2025) to capture insights into ICI-related mental irAEs, evolving clinical use patterns that prior studies (limited to 2012–2021) missed. It employs systematic methodology using MedDRA for standardized extraction of mental disorder terms and ROR for signal detection, with stratification by gender, age, and region to enhance the specificity of findings (e.g., identifying delirium as the most frequently reported event across subgroups). Additionally, the study synthesizes multiple potential mechanisms (cytokine dysregulation, microglial activation, molecular mimicry, etc.) linking immune processes to neuropsychiatric symptoms, and translates observational data into actionable clinical guidance by identifying high-risk groups (e.g., patients ≥ 75, those with lung, skin, and renal cancer, or on combination therapy) and proposing evidence-based monitoring strategies (PHQ-9/GAD-7 assessments, timed follow-ups). It also critically contextualizes trends by acknowledging database biases, strengthening the validity of real-world conclusions. Overall, this work addresses key gaps in understanding and managing ICI-related mental irAEs, offering valuable insights for clinical practice and future research.

Despite the strengths of this study, several limitations should be acknowledged. First, reliance on FAERS’ voluntary reporting system leads to underreporting, particularly in resource-limited regions, potentially underestimating the true incidence of mental irAEs and skewing geographic trends. Second, the absence of standardized diagnostic criteria for mental irAEs in FAERS causes misclassification (e.g., conflating ICI-induced insomnia with cancer-related fatigue), impairing the accuracy of symptom-specific risk estimates. Third, as an observational study, it cannot establish definitive causality, with unmeasured confounders (e.g., pre-existing mental health conditions, concurrent glucocorticoid use) risking overattribution of symptoms to ICIs. Fourth, proposed mechanisms rely largely on retrospective studies or case reports, lacking support from large-scale prospective data, and the difficulty of obtaining CNS tissue samples hinders in-depth validation of pathological processes. Lastly, drug-specific report comparisons are confounded by market longevity and indication breadth (e.g., more reports for pembrolizumab/nivolumab reflect longer availability, not higher toxicity), making direct risk comparisons unreliable.

This study conducted a comprehensive analysis of 181,482 ICI-related mental adverse event reports from the FAERS database spanning 2004 to 2025, expanding the temporal scope beyond previous research and providing updated real-world evidence. Through disproportionality analysis and stratification by gender, age, geography, and onset time, we identified distinct patterns: nivolumab and pembrolizumab accounted for the majority of reports, delirium emerged as the most commonly reported mental adverse reaction across all subgroups, and tislelizumab showed notably high ROR values for dysphoria and listlessness. Demographically, males, patients aged ≥ 65 years, and those in North America constituted the populations with the highest reporting rates, while temporally, adverse events were most frequent within the first 60 days of ICI initiation. These findings align with and extend current mechanistic understanding rooted in cytokine dysregulation, microglial activation, and other immune-mediated processes by linking pathological hypotheses to clinical observational data.

Notably, the study underscores the need to interpret these trends cautiously, given FAERS’ inherent limitations such as voluntary reporting biases, lack of standardized diagnostic criteria, and confounding by drug market longevity. Nevertheless, our identification of high-risk populations (e.g., elderly patients, those on combination therapy) and key temporal windows of risk translates directly to actionable clinical guidance: we advocate for risk-stratified monitoring, with baseline assessments using validated tools (PHQ-9, GAD-7) and frequent follow-ups for high-risk patients in the initial two months of treatment. Overall, this research reinforces the clinical significance of ICI-related mental irAEs despite their low reported incidence and highlights the urgency of enhancing pharmacovigilance in underrepresented regions, standardizing diagnostic reporting, and conducting prospective studies to validate mechanisms and optimize management strategies, ultimately improving the safety and outcomes of cancer patients receiving ICI therapy.