Immune checkpoint inhibitors (ICIs) have transformed the therapeutic landscape of numerous cancers by activating the immune system to target tumor cells. However, this nonspecific immune activation frequently leads to off-target immune-related adverse events (irAEs). Patients with preexisting autoimmune diseases (PADs) were initially largely excluded from ICI clinical trials due to concerns about autoimmune disease flares and elevated irAE incidence (1). Real-world evidence has shown that certain PAD patients—such as those with rheumatoid arthritis or systemic sclerosis—exhibit relatively mild irAE profiles during ICI treatment (2–4). In contrast, other preexisting inflammatory conditions are associated with a higher risk of severe irAEs (5).

As a prototypical antibody-mediated, T-cell-dependent autoimmune disorder, myasthenia gravis (MG) causes neuromuscular junction dysfunction. Approximately 85% of patients with MG have detectable AChR antibodies. Notably, up to 10–15% of AChR antibody-positive MG patients harbor thymomas, and 30–45% of thymoma patients develop MG (6, 7). Existing studies on ICI therapy in patients with preexisting MG have primarily focused on melanoma (8), and MG has been linked to a higher incidence of severe irAEs following ICI treatment compared to other preexisting autoimmune neurological disorders (9). Yet, the safety of ICIs and management of irAEs in thymoma patients with preexisting MG remain poorly characterized.

Refractory or recurrent thymomas show limited responsiveness to radiotherapy and chemotherapy (10). Additionally, corticosteroid-sparing immunosuppressive (CSIS) medications used for MG management increase the risk of extrathymic malignancies, including cancers of the digestive tract, male genital organs, and skin (11). While ICIs represent a critical therapeutic option for unresectable tumors, current guidelines caution against their use in patients with MG and/or thymoma due to the high risk of irAEs (12).

A recent study highlighted the susceptibility of thymoma patients to irAEs. However, pretreatment data on the association between AChR Ab status and myocarditis in this population are scarce. While most MG patients with thymoma (MGT) test positive for AChR Abs, more than half of AChR Ab-positive individuals do not have clinical manifestations of MG (13), leaving the mechanism underlying the predisposition to myocarditis in thymoma without MG (TOMA) patients unclear. Importantly, prior research has not separately or systematically analyzed clinical outcomes in patients with preexisting MG and thymoma initiating ICI therapy (9, 14).

MGT is characterized by ectopic germinal centers (GCs) in thymoma tissue and adjacent adipose tissue (15), with approximately 33.3% of GC-positive thymoma patients developing MG, whereas only 2.3% of GC-negative patients develop MG (16). These ectopic GCs, which are composed of B cells and T follicular helper cells, produce pathogenic AChR Abs that drive MG pathogenesis. Although AChR Abs have been proposed as biomarkers for irAEs in thymoma, the specific impact of preexisting MG on irAE severity during ICI therapy remains uninvestigated.

Given thymoma patients’ inherent predisposition to irAEs, we included a control cohort of thymoma patients without MG (TOMA group) to isolate the additional risk specifically attributable to preexisting MG, while controlling for the background risk associated with thymoma and ICI therapy. In this study, we aimed to 1) compare the incidence of severe irAEs between MGT and TOMA patients receiving ICIs; 2) characterize management strategies for MGT and longitudinal changes in AChR Ab levels during the acute and chronic irAE phases; and 3) explore the relationship between thymoma histopathology (particularly the presence of GCs) and irAE manifestations. These investigations aim to refine monitoring strategies and elucidate the mechanisms underlying irAEs in MGT and TOMA patients.

MG is a rare, prototypical autoimmune neuromuscular junction disease that is frequently comorbid with thymoma. This study focuses specifically on the management of irAEs in thymoma patients with preexisting MG receiving ICI therapy. Our findings highlight an elevated risk of severe irAEs (particularly myocarditis and myasthenic crisis) in this population. The maintenance of immunosuppressants prior to ICI initiation, combined with steroid and IVIG combination therapy following irAE onset, may be critical for improving clinical outcomes in MG patients. The exacerbation of MG by ICIs is not primarily mediated by the activation of B cells but may be directly associated with the activation of AChR-specific T cells. The presence of ectopic GCs is closely linked to the occurrence of irAEs in thymoma patients with or without MG.

ICIs are administered to patients with MG, mainly those with unresectable thymoma. In a retrospective study of low-dose nivolumab administered for metastatic thymic epithelial tumors, 2 out of 8 patients had preexisting MG, and 1 patient was positive for AChR Ab but had no MG symptoms. The two MG patients experienced severe (grade 4) irAEs after 1 cycle, and the AChR Ab-positive patient showed a grade 1 response after 3 cycles (19). In another study, 3 patients with preexisting MG, all of whom had MGFA type 1, had not received immunosuppressants for at least 1 year before ICI therapy. All three patients with MG developed grade 3 or higher irAEs after the first dose of ICIs (10). However, these studies lacked comparative analyses between thymoma-associated MG and thymoma-only cohorts, and the sample sizes were limited (<3 MG cases).

Compared with other autoimmune diseases, MG is rare, with an annual incidence of approximately 0.68 per 100,000 individuals in China (20). In our study, 6 MG patients with thymomas who had received ICI drugs were included, making this the largest single-center study to date. During the same period, 10 patients with thymoma alone were included as the control group. All 6 MG patients experienced an exacerbation of MG symptoms and myocarditis, with grade 4 or above irAEs. In the thymoma group, 4 out of 10 patients (40%) experienced grade 4 or higher irAEs, which is in line with other reports in the literature (14). Therefore, despite the limited sample size of 6 MG patients with thymoma and 10 patients with thymoma, our data support the notion that MG patients with thymoma are at high risk of irAEs when they are receiving ICIs.

ICI-induced myocarditis occurs in less than 1% of patient cases, excluding thymoma patients, and the mortality rate is 30–50% (21). For patients with thymoma, the incidence of ICI-induced myocarditis ranges from 16% to 29%, with a mortality rate of up to 64% (14). Myocarditis, a rare but fatal irAE (21), was also the primary cause of mortality in our cohort. In murine models, infiltrating CD8+ T cells recognize α-myosin antigens, and α-myosin-expanded TCRs are present in inflamed cardiac and skeletal muscle in patients with ICI-induced myocarditis (22, 23). Since alpha-myosin is not expressed in the thymus, the specific T cells involved in thymoma-induced autoreactive myocarditis need to be further determined. One study reported that 11–13% of MG patients with thymoma also had an autoimmune disease; notably, neurological autoimmune diseases were frequently paired with thymoma or thymic hyperplasia (24). These findings indicate that GCs in the thymoma are closely related to autoimmune disease. Therefore, ectopic GCs in MGT patients may contain autoreactive T cells and B cells that target other tissues (non-AChR) and represent abnormalities in the immune microenvironment in thymoma tissue. Consistent with this, both MGT patients and TOMA patients with GCs developed myocarditis in our study, indicating abnormalities in the immune microenvironment of thymomas with GCs. We propose that ectopic GCs in thymomas harbor autoreactive T/B cells that target non-AChR tissues and that ICI-induced T-cell activation may affect cardiac/skeletal muscle.

The specific mechanism underlying ICI-induced exacerbation of MG remains incompletely understood, and this exacerbation is not solely dependent on B-cell activation, as patients show no significant elevation in autoantibody levels during the occurrence of irAEs. Instead, it is more likely attributed to the extensive disruption of immune tolerance by ICIs. By blocking the CTLA-4 and PD-1/PD-L1 pathways, ICIs nonspecifically abrogate the inhibition of CD4+ and CD8+ T-cell subsets, leading to multidimensional immune dysregulation. On the one hand, they can directly activate preexisting acetylcholine receptor (AChR)-specific CD8+ cytotoxic T cells, which then target and attack neuromuscular junctions. This mechanism is supported by studies using the experimental autoimmune myasthenia gravis (EAMG) model, in which depletion of CD8+ T cells significantly alleviated MG symptoms (25). Additionally, research has confirmed that the proportion of central memory CD8+ T cells (Tcm) is increased while the proportion of effector memory CD8+ T cells (Tem) is decreased in MG patients (26). On the other hand, ICIs concurrently activate CD4+ T helper cells, particularly Th1 and Th17 subsets, which secrete cytokines, including IL-2 and IFN-γ (27). These cytokines not only increase the cytotoxicity of CD8+ T cells but also drive B-cell proliferation, differentiation, and autoantibody production (10, 27). Furthermore, ectopic germinal centers in thymoma tissues are enriched in autoreactive T and B cells (15), and ICIs can significantly increase the activation efficiency of these cells, further exacerbating immune microenvironment dysfunction. In summary, ICI-related MG exacerbation is a multidimensional synergistic outcome. These findings underscore the need for further mechanistic investigations, with a particular focus on the effects of ICIs on ectopic germinal centers within thymomas.

In the case of ICI drug-induced symptom exacerbation in patients with preexisting MG, those treated with combined IVIG and steroids had more favorable outcomes than those receiving steroids or IVIG monotherapy did. Similarly, for ICI drug- induced MG, IVIG or PLEX may lead to better outcomes than steroids alone. Two patients with preexisting MG did not experience worsening symptoms, and both were given 10 mg of corticosteroids before ICI therapy. Six patients with ICI-induced MG received ICI treatment again, and the symptoms of MG did not worsen (8). This may have occurred because all 6 patients received immunosuppressive treatment before the readministration of ICI treatment. On the basis of our findings, we suggest that even with stable MG symptoms, maintenance of low-dose prednisone and/or immunosuppressive agents prior to ICI therapy may significantly reduce ICI-induced mortality. Given the high risk of irAEs, consultation with a neurologist is recommended for preexisting MG patients before ICI therapy is initiated.

Currently, considering the high incidence and severity of irAEs in thymoma patients, ICI therapy for these patients is not recommended. However, some thymoma patients do not experience irAEs after receiving ICIs, and some may even experience a partial response (PR) (10). Similarly, in our study, we found that 5 TOMA patients (50%) did not experience significant irAEs. Therefore, identifying the appropriate population is crucial for providing ICI treatment to thymoma patients. Fenioux et al. proposed that the AChR Ab level could be a biomarker for irAEs in thymoma patients receiving ICI therapy (14). However, the origin of AChR Abs (thymus, peripheral blood, or other sources) and the timing of their production remain unclear. In patients with MG, the levels of AChR Ab are closely associated with the number of GCs. TOMA patients are also positive for AChR Abs (24%-49%) and GCs (10.7–15.6%) (15, 16). We found that there is a close relationship between the presence of GCs and the incidence of irAEs in TOMA patients. We propose that GCs may serve as biomarkers for irAEs in thymoma patients without MG. In addition, some autoimmune diseases, such as Hashimoto’s thyroiditis, systemic lupus, and Isaac’s syndrome, are closely related to thymoma (28). To date, whether thymic GCs exist in these thymoma-related AIDs other than MG remains unclear. The autoimmune regulator (AIRE), a key transcriptional regulator in medullary thymic epithelial cells (mTECs), is critical for central immune tolerance; its deficiency impairs the effective elimination of autoreactive T cells, relevant to multiple autoimmune diseases, including MG (29). As the core organ for T cell development and maturation, thymic immune microenvironment dysfunction is likely the main pathogenic basis for such autoimmune abnormalities. Based on this, we hypothesize that TOMA patients who are AChR Ab negative and have no other autoimmune diseases may represent a favorable population for ICI immunotherapy. The core value of this hypothesis lies in its ability to assess ICI treatment risk in thymoma patients based on clinical features (AChR Ab status, AID history) without the need for thymic GC detection. However, this hypothesis requires further validation through comprehensive autoantibody testing, systematic autoimmune disease screening, and large-scale cohort studies.

This study has several limitations because of its retrospective nature. MG is a rare autoimmune disease, and MGT patients with unresectable thymoma are exceptionally rare, leading to a modest sample size that limits the generalizability of some conclusions. MG flares and irAE types were documented on the basis of electronic medical records rather than standardized assessments by neurology specialists, potentially introducing data inaccuracies. Furthermore, given the potential heterogeneity of GC distribution in thymoma tissue, we recommend that future prospective studies prioritize intact thymic tissue from surgical resection to more comprehensively evaluate the association between GC distribution and irAEs. In addition, this was a single-center study, and the results may not be generalizable to other medical centers. We did not include thymoma patients with AIDs other than MG, which prevents us from verifying the cross-AID universality of the association between thymic immune dysregulation and irAEs. Integrating antibody titers, thymic histopathology, and changes in immune cell subsets may more effectively elucidate the mechanisms underlying ICI-induced MG exacerbation. Long-term follow-up after the occurrence of irAEs is necessary to further identify chronic irAEs and explore better treatment options.

For MG patients with thymoma who receive one session of ICI therapy, close monitoring is recommended because of the high incidence of severe myocarditis and MC. Prophylactic maintenance of immunosuppressants prior to ICI initiation, combined with early steroid-IVIG dual therapy during irAE onset, appears critical for improving clinical outcomes. We propose that ectopic GCs may serve as a significant marker for selecting thymoma patients who are suitable candidates for ICI therapy. Further multicenter studies with standardized neurological assessments are warranted to validate these findings and refine patient selection criteria for immunotherapy in patients with thymomas.