Type 1a
Cross-reaction of antitumor T cells: Cross-reaction with melanin-containing cells induces VKH-like panuveitis.
Frequency: 5–14% of ICI-related ocular adverse reactions in melanoma patients are VKH-like reactions.
Immune checkpoint inhibitors (ICIs) are a class of drugs that block the “immune checkpoint pathways” used by cancer cells to evade immune attack, thereby reactivating the patient’s own immune system to attack tumors. On the other hand, they can cause immune-related adverse events (irAEs) in which normal tissues throughout the body are attacked due to overactivation of the immune system, and the optic nerve can also be a target.
ICIs are classified into four classes based on their target. The main drugs currently approved by the FDA are as follows:
| Class | Main Approved Drugs (Year of Approval) |
|---|---|
| CTLA-4 inhibitor | Ipilimumab (2011), Tremelimumab (2022) |
| PD-1 inhibitors | Pembrolizumab (2014), Nivolumab (2014), Cemiplimab (2018) |
| PD-L1 inhibitors | Atezolizumab (2016), Avelumab (2017), Durvalumab (2017) |
| LAG-3 inhibitors | Relatlimab (2022) |
The incidence of neuro-ophthalmic irAE after ICI treatment is reported to be up to 0.46%. Overall ocular irAEs occur in 1–3%, of which posterior segment inflammation (including optic neuropathy) accounts for 5–20%1). Optic neuropathy is considered a more severe ocular irAE.
The time from initiation of ICI to onset varies depending on the case and drug, ranging from several weeks to several months.
Characteristic clinical features reported for each drug are shown below.
The direct cause of optic neuropathy due to ICIs is that ICIs release immune checkpoints, triggering unintended autoimmune reactions that damage the optic nerve.
PD-1 inhibitors (e.g., pembrolizumab) are the most commonly used drugs and have many case reports, but optic neuropathy has also been reported with CTLA-4 inhibitors (ipilimumab) and PD-L1 inhibitors (atezolizumab, durvalumab). In particular, the combination of ipilimumab and nivolumab increases the risk of irAE 1).
There are no established biomarkers for ICI-related optic neuropathy, and at present it is a diagnosis of exclusion.
A complete ophthalmic evaluation including visual acuity testing, visual field testing, color vision testing, pupillary response (checking for RAPD), and fundus examination is recommended. The following additional tests are useful 1):
Head and orbital MRI (with and without contrast) is essential. It is performed to differentiate metastatic, radiation-induced, and paraneoplastic optic neuropathies. Contrast enhancement of the optic nerve has been confirmed in 2 out of 4 cases, and small vessel ischemic changes have also been reported.
HLA testing and anti-retinal autoantibody testing may aid in clinical diagnosis 1).
Severity is assessed using the 4-grade scale of the Common Terminology Criteria for Adverse Events (CTCAE) version 5 from the U.S. Department of Health and Human Services. Grade determination directly influences treatment decisions (see “Standard Treatment” section).
The first-line treatment for optic neuropathy is steroid pulse or half-pulse therapy. After 1 to 3 courses, switch to oral steroids. With appropriate treatment, approximately 70% or more of patients achieve visual function recovery.
| Grade | Criteria | Management |
|---|---|---|
| Grade 1 | Mild visual impairment | Usually no need for steroids or ICI discontinuation |
| Grade 2 | Affects daily life | Consider temporarily holding ICI and starting systemic steroids; consider resuming ICI after improvement |
| Grade 3 | Marked vision loss | Consider holding ICI; if no improvement in 4–6 weeks, consider discontinuing + systemic steroids |
| Grade 4 | Visual acuity 20/200 or worse | Usually discontinue ICI + administer systemic steroids |
The decision to continue or discontinue ICI therapy should be made on a case-by-case basis by a multidisciplinary team including ophthalmology and oncology, weighing the risks and benefits.
ICI discontinuation is not mandatory; the decision is based on the severity grade and the patient’s underlying disease status. For grade 1, ICI can usually be continued; for grade 2, resumption after temporary interruption may be considered. There are reports of improvement with steroid eye drops alone while continuing ICI. The final decision is made by a multidisciplinary team including ophthalmology and oncology.
The eye has ocular immune privilege due to anatomical barriers such as the blood-retinal barrier (BRB) and sparse ocular lymphatics. The normal BRB prevents peripheral T cells from entering the vitreous and choroidal space 1). CTLA-2α, PD-L1, and PD-L2 on RPE cells, and PD-L1 expression on retinal Müller glia and microglia, convert T cells into regulatory T cells (Tregs), minimizing T cell-mediated inflammation 1). ICIs target these defense mechanisms, potentially leading to immune attack on the eye.
Haliyur et al. (2025) classified the mechanisms of ICI-related posterior segment inflammation into the following three types 1).
Type 1a
Cross-reaction of antitumor T cells: Cross-reaction with melanin-containing cells induces VKH-like panuveitis.
Frequency: 5–14% of ICI-related ocular adverse reactions in melanoma patients are VKH-like reactions.
Type 1b
Auto-reactivity to eye-specific self-peptides: Involves HLA predisposition (e.g., HLA-DR4) in susceptible individuals.
Phenotypes: Granulomatous uveitis, multifocal geographic chorioretinitis, Birdshot-like chorioretinitis, MEWDS, etc.
Type 2
Nonspecific vasculitis: Nonspecific inflammation induced by ICI causes retinal vasculitis and vascular occlusion.
Mechanism: Lymphoplasmacytic infiltration with predominant CD4+ T cells and upregulation of cell adhesion molecules. More common with anti-PD-1 therapy.
Type 3
Autoantibody-mediated inflammation: CTLA-4 inhibition → Treg suppression + B cell activation. PD-1 inhibition → memory T cell activation → B cell clonal expansion.
Outcome: Progression to autoimmune retinopathy (AIR), CAR, MAR.
The eye normally has “immune privilege” due to the blood-retinal barrier (BRB), but ICIs target defense mechanisms such as PD-L1, making it easier for immune T cells to enter the eye 1). Furthermore, multiple mechanisms—including cross-reaction of antitumor T cells with ocular tissues (Type 1a), reaction to self-antigens in genetically predisposed individuals (Type 1b), nonspecific vasculitis (Type 2), and autoantibody-mediated inflammation (Type 3)—can damage ocular tissues including the optic nerve.
Rituximab (CD-20 monoclonal antibody) is attracting attention as an immune countermeasure against steroid-resistant irAEs. Significant improvement has been reported in 7 cases of steroid-resistant immune-related cutaneous adverse events, and another study showed improvement of irCAE by 2 or more grades in all patients. It is also being investigated in combination with plasma exchange for treatment-resistant neurological irAEs (autoimmune encephalitis, myasthenia gravis). However, direct evidence for optic neuropathy is still lacking and remains controversial.
Haliyur et al. (2025) stated that the type 3 classification framework for ICI-related posterior segment inflammation (see “Pathophysiology” section) is expected to improve prognosis prediction and treatment decision-making1). In Type 1a/1b and Type 3, HLA and anti-retinal autoantibody testing may support clinical diagnosis. Future mechanistic laboratory studies are expected to develop targeted treatment strategies that control ocular inflammation while avoiding ICI discontinuation.
Currently, there are no biomarkers for OirAE, and diagnosis relies on exclusion. The development of biomarkers is considered an important future task. With the expansion of indications for FDA-approved ICIs, an increase in the number of posterior segment irAE cases is predicted, and the need for long-term follow-up has also been indicated1).