Paraneoplastic syndromes (PS) comprise a diverse group of disorders that affect different organ systems in cancer patients, arising independently of the primary tumor or its metastases. Their clinical manifestations are often distant from the tumor site and are thought to result from humeral factors produced by cancer cells or by immune system reactions. Due to their rarity and clinical heterogeneity, the reported incidence of PS varies widely, ranging from 7% to 33% of cancer patients (1, 2). These syndromes typically manifest within ±3 years of cancer diagnosis and are more frequently identified after cancer is detected (3).

The mechanisms underlying PS are incompletely understood and diverse. Non-autoimmune-mediated PS are often driven by endocrine imbalances (4), such as hyponatremia, hypercalcemia, or Cushing’s syndrome, as well as cytokine dysregulation (5), including neutrophilia, Trousseau’s syndrome, and thrombocytosis. These cases are associated with poor patient outcomes (5). In contrast, autoimmune-mediated paraneoplastic syndromes are linked to better cancer prognoses, suggesting that autoimmune flares are a sign of enhanced antitumor responses.

Autoimmune-mediated paraneoplastic syndromes (AMPS) most commonly affect the central nervous system (CNS) but can also manifest in the peripheral nervous system (PNS), skin, or blood (6). A hallmark of AMPS is the presence of autoantibodies, which have gained attention for their role as disease biomarkers (7). These autoantibodies are important in the diagnosis of paraneoplastic syndromes and may support early cancer detection in this group of patients. For instance, individual autoantibodies, such as p53 (8), HER2 (9), and NY-ESO-1 (10), as well as autoantibody panels, have shown predictive potential for cancer diagnosis (11, 12). However, these biomarkers still fail to meet clinical standards (13). Important observations have been made linking cancer alterations with commonly associated autoantibodies and the development of AMPS ( Table 1 ).

Despite advances in autoantibody research, the cellular and molecular mechanisms underlying AMPS pathogenesis remain poorly understood. This review aims to summarize the current state of research on AMPS exploring the mechanisms driving autoimmune disorders in cancer. Specifically, it will address how ectopic expression, protein structural alterations, and overexpression of self-proteins may trigger autoimmune responses. Overlapping clinical presentations between AMPS and immune-related adverse events (irAEs) of immune checkpoint inhibitor (ICI) treatment will be discussed, highlighting the need to understand autoimmunity in cancer. Further, the roles of chronic inflammation within the tumor microenvironment (TME) and localized immune tolerance in sustaining these responses will be examined. With this discussion, we aim to bridge existing knowledge gaps and provide a framework to better understand AMPS pathogenesis.

Ectopic expression refers to the production of a self-protein in atypical tissues. Malignant cells frequently exhibit ectopic expression of otherwise tissue-restricted proteins as a consequence of genomic instability. Genomic instability may also trigger missense mutations that may lead to changes in protein structure. Similarly, an antigen may be overexpressed due to chromosome amplifications. In cancer, any of these alterations can induce immune recognition and targeting of malignant cells. In the following AMPS cases, evidence has been provided of how genetic or proteomic alterations can lead to immune recognition and the development of autoimmunity.

In the previous section, we discussed cases of AMPS that arise independently of cancer treatment. Immune-modulating treatments such as ICIs can trigger irAEs, some of which share similarities with AMPS. This has become increasingly relevant given the quick adaptation of ICIs into the clinic, driven by their ability to improve overall survival while causing fewer general adverse effects than traditional therapies (78–80).

Immune checkpoints prevent overactivation of cytotoxic T cells, ensuring that self-tolerance is maintained. PD-L1 for instance, dampens T cell cytotoxicity through interaction with PD-1 expressed by T cells. This regulatory function is controlled by immune cells but can also be mediated in immune-privileged tissues such as the lung, liver, and placenta (81). Cancer cells exploit immune checkpoints by upregulating inhibitory signals such as PD-L1 (interacting with PD-1 (82) and CD80 (interacting with CTLA-4 (83)) to evade immune destruction and create an immunosuppressive tumor microenvironment. By inhibiting these pathways, ICIs enhance T cell recognition of cancer cells, but also increase the risk of autoimmunity, particularly in individuals with a history of autoimmune disorders (84, 85).

Several irAEs observed in ICI-treated patients overlap with AMPS; disorders like myositis, encephalitis, SSc, and MG are among the most prevalent overlapping conditions. While clinical similarities suggest that ICI treatment could increase the risk of developing AMPS, there are many contrasting features between them. For example, a systematic review by Buckley et al. (2025) on ICI-related encephalitis, found that only 46.9% of cases were autoantibody-positive (86), different from AMPS cases, where autoantibodies are significantly more frequent (100% in SCLC paraneoplastic encephalitis (42), 75% in general paraneoplastic encephalitides (87)). Similarly, in Hamada et al. (2021), only 67% of ICI-related MG cases were positive for any MG-related autoantibody (88) compared to over 95% in thymoma MG (89). A similar trend of seronegative cases has been noted in ICI-related SSc (90).

While seropositive irAEs seem to be a case where ICIs facilitate development of AMPS, a significant proportion of irAE cases are seronegative and appear to have mechanistic differences, where T cell cytotoxicity is the main driver of autoimmunity. Whether these autoreactive T cells are expanded through antitumor immunity, as observed in AMPS, remains to be explored.

Given the high prevalence of severe irAEs of ICI-treated patients (20% (91)), predictive screening tools could help physicians tailor treatment strategies to individual patients. Biomarkers such as blood counts, autoantibodies, HLA genotype, and microRNA expression profiles have been identified as prognostic and diagnostic tools in autoimmune diseases (92). Several studies suggest that pre-existing autoantibodies may indicate an increased risk of developing irAEs following ICI therapy (93, 94). Screening for autoantibodies before initiating ICI therapy could help classify patients based on their likelihood of developing severe irAEs, enabling safer and personalized treatment approaches.

While the presence of autoantibodies is a hallmark of AMPS, not all patients with detectable autoantibodies develop autoimmune symptoms. This discrepancy suggests that additional factors, such as chronic inflammation and localized immune responses within the TME, may determine whether autoimmunity manifests systemically. Understanding these factors is critical for understanding the connection between antitumor immunity and autoimmune pathology.

The disorders summarized earlier demonstrate that AMPS patients harbor autoantibodies capable of inducing immune responses with both cancer-resolving and autoimmune consequences. However, it remains unclear why some patients with autoantibodies benefit from a stronger antitumor response without developing systemic autoimmunity. This observation points to the importance of additional immune triggers beyond humoral activity in the development of AMPS.

Strong antitumor responses, a common feature of AMPS patients, are often accompanied by the presence of tertiary lymphoid structures (TLS) within the chronically inflamed TME (95). TLS are organized immune cell aggregates comprising T cells, B cells, dendritic cells, and macrophages. Functionally resembling secondary lymphoid organs (SLO), TLS promote antigen presentation, B cell priming, and T cell activation (96). In some AMPS cases, the autoimmune response is mediated entirely within the tumor. For example, Al-Diwani et al. (2022) observed that ex vivo cultures of B cells dissociated from ovarian teratomas generated anti-NMDAR antibodies, while B cells from cervical lymph nodes did not produce autoantibodies after tumor resection. This finding supports the idea that autoimmunity is initiated within the tumor-associated TLS (61). Similarly, Mazor et al. (2022) demonstrated that tumor-specific B cells are primed and clonally expanded within the tumor, often originating from naturally occurring self-reactive B cells (97). T cells can also expand and generate antitumor responses within the TLS (98). These findings suggest that TLS facilitate localized antitumor immunity, allowing B and T cells to target self-antigens without eliciting systemic autoimmunity.

This localized immune activity aligns with the observed improved patient prognosis in the presence of TLS (99, 100). Many AMPS cases reviewed here, such as anti-NMDAR encephalitis and anti-Yo cerebellar degeneration, involve highly infiltrated tumors (60, 68, 69). TLS-mediated immunity represents a controlled break in immune tolerance, enabling effective clearance of malignant cells. While circulating autoantibodies are detectable in patient serum, T cells are typically confined to the tumor. A breach in this localization (where T cells migrate from the tumor into circulation) could be the defining trigger for systemic AMPS.

Activated T cells tend to accumulate in inflamed tissues due to retention signals such as the chemokine CXCL10 (101). In the absence of systemic inflammation or additional chronic inflammatory events, these T cells are unlikely to enter the bloodstream. However, in cases of invasive cancers with high levels of circulating tumor cells or early metastatic spread, inflammation at the tumor periphery may promote signals that induce T cell migration. This process could extend the anti-self immune response to healthy tissues, contributing to the onset of paraneoplastic syndromes ( Figure 1 ).

Recent studies suggest that a humoral immune response alone is insufficient to trigger autoimmune symptoms but rather serves as a marker of broken self-antigen tolerance. For instance, in paraneoplastic SSc (66) and anti-Yo-related paraneoplastic cerebellar degeneration (68, 69), the majority of patients exhibited highly invasive or early metastatic malignancies. This suggests that the expansion of an immune response beyond the tumor, driven by T cell migration, could trigger autoimmune disorders. In contrast, patients with autoantibodies but no paraneoplastic symptoms often have a better prognosis, indicating that localized breaches in self-tolerance, mediated by TLS in the TME, serve as a controlled mechanism for combating malignancy.

A break in localized self-tolerance may be the defining trigger for paraneoplastic syndromes. This could explain the high specificity but low sensitivity of autoantibody markers; only a subset of autoantibody-positive patients, presumably those with T cell extravasation into the bloodstream due to extended inflammation, develop paraneoplastic features. Tumor-associated TLS are linked with better prognosis and higher immune infiltration, underscoring their role in cancer immunity. Treatment with ICIs can also facilitate the development of autoimmunity through mechanisms that are often independent of autoantibodies.

The roles of autoantibodies and T cell autoreactivity in cancer-related autoimmunity remain incompletely understood, as preclinical and clinical data are still limited. A deeper understanding of AMPS could help us clarify the intersection of TLS biology, immune tolerance, and cancer immunity. Resolving how localized immune responses within the tumor microenvironment contribute to either paraneoplastic syndromes or effective tumor resolution may uncover novel therapeutic targets and biomarkers. The similarities between AMPS and ICI-induced irAEs evidence the complex balance between antitumor activity and autoimmunity, highlighting the need for personalized approaches to cancer immunotherapy. Future research focused on unraveling the mechanisms underlying TLS formation, T cell migration, and systemic immune tolerance breaks, will enable a deeper understanding of these processes and will improve both cancer treatment and the management of autoimmune complications.