Tislelizumab-induced Vogt-Koyanagi-Harada-like uveitis: a case report

Abstract

Background

While immune checkpoint inhibitors like the anti-PD-1 agent tislelizumab have revolutionized the treatment of cancer, they pose a risk of immune-related adverse events (irAEs). Vogt-Koyanagi-Harada (VKH)-like uveitis is a rare, vision-threatening ocular irAE that poses significant diagnostic and therapeutic challenges, often mimicking its idiopathic counterpart.

Case presentation

Following ten cycles of adjuvant tislelizumab treatment, a 52-year-old female with a past medical history of esophageal cancer presented with acute monocular vision loss. Comprehensive ophthalmic evaluation revealed left optic disc edema, exudative retinal detachment, and characteristic serous retinal detachments on optical coherence tomography (OCT). Extensive systemic workup excluded infectious, neoplastic, and primary autoimmune etiologies. Notably, cerebrospinal fluid (CSF) analysis revealed a lymphocyte-dominant profile but no pleocytosis, a finding atypical for classic VKH disease. The uveitis responded markedly to systemic corticosteroid pulse therapy but flared unequivocally upon two subsequent tislelizumab infusions, establishing a clear drug-effect temporal relationship. After completing the planned 12 cycles of tislelizumab alongside a prolonged, carefully tapered oral corticosteroid regimen, the patient’s ocular inflammation resolved completely. At one-year follow-up, visual acuity remained stable with no recurrence, demonstrating sustained remission after cessation of the inciting agent.

Conclusion

This case strongly implicates tislelizumab as a trigger for VKH-like uveitis and highlights two critical learning points: the potential lack of CSF pleocytosis as a differentiating feature from primary VKH, and the risk of recurrence upon drug rechallenge. It underscores the necessity for close collaboration between oncologists and ophthalmologists to navigate the dual imperatives of cancer control and vision preservation.

Clinical trial number

Not applicable.

Background

The integration of immune checkpoint inhibitors (ICIs) into oncologic practice, particularly antibodies targeting the programmed cell death-1 (PD-1) pathway like tislelizumab, represents a paradigm shift in cancer therapy. By disrupting the PD-1/PD-L1 axis, these agents potentiate T-cell-mediated antitumor immunity, yielding durable responses in multiple malignancies. Although tumor immunotherapy is generally regarded to have a favorable safety profile, immune-related adverse events (irAEs) still occur intermittently in clinical practice. Vogt-Koyanagi-Harada (VKH) syndrome is a disease entity with well-established diagnostic criteria and standardized treatment protocols; however, its exact etiological mechanism remains incompletely elucidated. Drug-induced uveitis, which establishes a definitive causal relationship with the use of specific medications, has emerged as a rare yet clinically significant etiological factor in the spectrum of uveitis [1]. Previous research studies have documented that novel tumor-targeted agents, such as anti-PD-1 antibodies, can induce adverse effects including bilateral anterior uveitis, panuveitis, or multiple serous retinal detachments (VKH-like manifestations) [2, 3].

The emergence of VKH-like uveitis blurs traditional diagnostic boundaries. While the ocular phenotype may be indistinguishable, the etiological driver—a pharmacologically induced immune dysregulation rather than a primary autoimmune process—carries distinct implications for management, particularly regarding the continuation of life-saving cancer therapy. We present a detailed case of tislelizumab-induced VKH-like uveitis that vividly illustrates this diagnostic dilemma. Our report is punctuated by two pivotal observations: an atypical CSF profile lacking pleocytosis, and a clear recrudescence of inflammation with each drug rechallenge. These features not only bolster the diagnosis of a drug-induced entity but also provide clinically relevant insights for distinguishing it from classical VKH disease.

Case presentation

Clinical history

A 52-year-old female patient suffered from decreased visual acuity in the left eye for half a month, accompanied by temporal headache. No symptoms such as alopecia, tinnitus, or hearing loss were reported. Her past medical history was significant for esophageal carcinoma, diagnosed ten months prior (October 2023), for which she underwent surgical resection. As part of her adjuvant treatment regimen, she had been receiving intravenous tislelizumab (200 mg every month) and had completed ten cycles at symptom onset. Her oncologists had planned a total of twelve cycles. She had no personal history of hypertension, diabetes, autoimmune disorders, or infectious diseases such as tuberculosis, syphilis, or hepatitis.

Ophthalmic, systemic and laboratory Investigations

Right eye (OD) best-corrected visual acuity (BCVA) was 0.10 logMAR, while the left eye (OS) had a BCVA of hand motion (HM). Intraocular pressure (IOP) was recorded as 14.6 mmHg (OD) and 10.7 mmHg (OS). Slit-lamp evaluation demonstrated clear corneas in both eyes, quiet anterior chambers, radial opacities in the lens cortex, and vitreous flocculent changes without inflammatory cells. Fundus examination (Optos photography, Fig. 1) showed a normal right fundus and left optic disc hyperemia with inferior retinal detachment. Left eye optical coherence tomography (OCT; Fig. 2) demonstrated diffuse edema of the macula, multiple neurosensory detachments, intraretinal cysts, and retinal pigment epithelium (RPE) undulations, with foveal thickness recorded at 1290 μm. A confirmatory ocular B-scan (Fig. 3) verified left-sided retinal detachment.

Fig. 1.

Optos fundus photography at admission (August 21, 2024). The right fundus is unremarkable; the left eye shows optic disc hyperemia, edema with blurred margins, and inferior retinal detachment

Fig. 2.

Optical coherence tomography (OCT) of macular area at admission (August 21, 2024). Right eye macular structure shows no significant abnormalities; the left eye presents diffuse macular edema, multiple neuroepithelial detachments, intraretinal cysts, wavy RPE changes, and foveal thickness of 1290 μm

Fig. 3.

Ocular B-scan at admission (August 21, 2024). Axial length of right eye is 21.78 mm and left eye is 21.12 mm; scattered medium-low echo spots in bilateral vitreous cavity, scleral thickening, and band-like medium-low echo in left scleral region suggesting retinal detachment

Fluorescein angiography (FFA, Fig. 4) demonstrated late-phase hyperfluorescence of the optic disc in the right eye and scattered punctate hyperfluorescence in the early phase and extensive peripapillary hyperfluorescence with patchy pooling in the late phase in the left eye. Chest, orbital, and cranial CT scans revealed no significant space-occupying lesions. Lumbar puncture indicated normal intracranial pressure. Blood test results: Complete blood count, urinalysis, liver and renal function evaluations, anti‑streptolysin O (ASO) titer, rheumatoid factor (RF), C‑reactive protein (CRP), anti‑neutrophil cytoplasmic antibodies (ANCA), and anti‑nuclear antibody (ANA) profiles were determined to be within the normal physiological limits. The result of the T-cell spot test for tuberculosis infection (T-SPOT.TB) was negative. Serology screening for hepatitis B, hepatitis C, syphilis, and HIV yielded negative findings.

Fig. 4.

Fluorescein angiography (FFA) on August 22, 2024. Right eye: late peripapillary fluorescence staining; left eye: scattered punctate hyperfluorescence in early phase, extensive peripapillary fluorescence staining and patchy fluorescence pooling in late phase

On the fourth day of hospitalization, bilateral shallowing of the anterior chamber was observed, more pronounced in the right eye. Ultrasound biomicroscopy (UBM, Fig. 5) revealed that the central anterior chamber depth (ACD) of the right eye was approximately 1.59 mm. The angle at the 12 o’clock position was closed, while the angles in the remaining quadrants were narrow, with circumferential cyclodialysis cleft noted. For the left eye, the central ACD measured about 2.30 mm; the angle at the 12 o’clock position was narrow, whereas the angles in the other quadrants were open, also with circumferential cyclodialysis cleft.

Fig. 5.

Ultrasound biomicroscopy (UBM) on August 26, 2024. Right eye: anterior chamber depth 1.59 mm, angle closure at 12 o’clock, narrow angles in other quadrants, generalized cyclodialysis; left eye: anterior chamber depth 2.30 mm, narrow angle at 12 o’clock, open angles in other quadrants, generalized cyclodialysis

The intraocular fluid viral PCR panel returned negative results, and intraocular fluid cytokine analysis showed an IL-10/IL-6 ratio of < 1, which further ruled out infectious uveitis and intraocular lymphoma. CSF cytology indicated a lymphocyte proportion of 61.3%, but critically, no pleocytosis was present.

Indocyanine green angiography (ICG) and fundus autofluorescence (FAF) were not performed at the baseline evaluation or follow-up evaluations.

Diagnosis, treatment and follow-up

Based on the classic ocular findings of bilateral (though asymmetrical) panuveitis with exudative retinal detachments, a preliminary diagnosis of VKH syndrome was made. The patient immediately commenced high-dose intravenous methylprednisolone pulse therapy (500 mg daily for 5 consecutive days). The response was robust: left BCVA improved to 0.22 logMAR, the cyclodialysis clefts resolved, and OCT showed significant resolution of the subretinal fluid (Figs. 6 and 7). She was discharged on a tapering course of oral prednisone acetate (1.2 mg/kg/day). Following the completion of systemic corticosteroid therapy and based on sustained clinical improvement monitored via OCT, follow-up FFA and ICG were not routinely repeated.

Fig. 6.

Optos fundus photography after glucocorticoid pulse therapy on September 2, 2024. Left exudative retinal detachment is significantly ameliorated

Fig. 7.

OCT of macular area after glucocorticoid pulse therapy on September 2, 2024. Macular morphology is improved compared with pre-treatment

The clinical course, however, took an instructive turn. The patient received her 11th scheduled cycle of tislelizumab shortly after discharge. At a one-week follow-up, her left BCVA had deteriorated to 0.70 logMAR. She subsequently received the 12th and final cycle. At a one-month follow-up post-final cycle, vision remained at 0.70 logMAR, though OCT (Fig. 8) showed no significant new fluid accumulation. This temporal correlation between tislelizumab infusion and visual decline strongly implicated tislelizumab as the causative agent, leading to a revised diagnosis of probable tislelizumab-induced VKH-like uveitis. Oral corticosteroid therapy was meticulously tapered over the ensuing year. At the final one-year follow-up (August 2025), her ocular condition was quiescent and stable. BCVA was 0.10 logMAR (OD) and 0.22 logMAR (OS), with normal intraocular pressures. OCT imaging (Fig. 9) confirmed complete absorption of all subretinal fluid and a persistent irregularity of the ellipsoid zone in the left eye. Clinically, this structural finding correlates with the patient’s final stable best-corrected visual acuity of 0.22 logMAR, which represents a mild permanent visual deficit. The absence of symptoms like metamorphopsia suggests the foveal center is relatively spared. The stabilized defect indicates the injury has reached a chronic, non-progressive stage, making significant future visual improvement unlikely, but the current vision offers a good functional prognosis for daily life.

Fig. 8.

OCT of macular area at 1-month follow-up after the 12th cycle of tislelizumab on October 9, 2024. Subretinal fluid in the left macular area was further absorbed

Fig. 9.

OCT of macular area during the 1-year follow-up on August 9, 2025. Complete absorption of all subretinal fluid and a persistent irregularity of the ellipsoid zone in the left eye. Right eye macular structure shows no significant abnormalities

Discussion

In recent years, tumor immunotherapy has gradually become one of the most successful approaches in the field of cancer treatment. Within this class, anti-PD-1/PD-L1 monoclonal antibodies are the most widely adopted agents.Tislelizumab [4], an anti-PD-1 monoclonal antibody, received marketing approval from the National Medical Products Administration (NMPA) on December 27, 2019. With the widespread use of PD-1/PD-L1 inhibitors in cancer patients, reports of ocular toxicities have continued to increase; severe manifestations include uveitis, macular edema, choroidal neovascularization, and serous retinal detachment. Despite the rarity of drug-induced uveitis, numerous drugs have been identified as causal factors underlying this condition. An increasing number of cases of drug-induced uveitis have been associated with the use of anti-tumor agents. Although only 0.4%-7.4% of patients develop ocular side effects following anti-PD-1/PD-L1 monoclonal antibody administration [5, 6], these side effects can severely threaten visual function, reduce quality of life, and even affect patient adherence to treatment. Currently, ophthalmologists and oncologists are still not fully familiar with the ocular toxicities of these drugs, and patients also pay insufficient attention to these adverse reactions.

While the exact pathogenic mechanism underlying Vogt-Koyanagi-Harada (VKH) disease remains elusive, it is hypothesized that a T cell-mediated autoimmune reaction targeting melanocyte-related antigens is implicated [7]. Theoretically, this reaction may trigger autoimmunity against other pigmented structures (particularly uveal, cutaneous, otic, or meningeal melanocytes) that share antigenic components with tumors. The precise pathogenesis of VKH-like uveitis is also incompletely understood, with two proposed mechanisms (direct and indirect). The direct mechanism involves drugs directly entering intraocular tissues to induce toxic reactions, damaging the blood-ocular barrier and causing ocular inflammation. The indirect mechanism primarily refers to drug-induced autoimmune responses, which may be associated with disrupted immune tolerance, immune dysregulation, sequestered antigen exposure, or drug-autoantigen cross-reactivity [8]. A study found that 82.35% of patients who developed VKH-like uveitis after cancer targeted therapy or immunotherapy had cutaneous melanoma [9]. The mitogen-activated protein kinase (MEK) signaling pathway plays a critical role in maintaining retinal pigment epithelium (RPE) integrity by combating oxidative stress, photoinduced injury, and inflammation [10]. Preclinical studies have shown that MEK inhibition elicits acute RPE toxicity, accompanied by increased RPE permeability and blood-retinal barrier impairment [11].

During the patient’s clinical course, the occurrence of transient bilateral ciliary body detachment led to a diagnostic consideration of VKH syndrome. Numerous clinical case reports have consistently identified definitive ciliary body detachment via UBM in VKH syndrome patients presenting with shallow anterior chambers [12]. The shallowing of the anterior chamber is attributed to uveitis-associated choroidal effusion [12]. Histopathological examinations in animal models have revealed inflammatory cell infiltration within uveal tissues, including the ciliary body and choroid [13]. This intense inflammatory response disrupts both the blood-aqueous barrier and the blood-retinal barrier, leading to increased vascular permeability. Consequently, plasma components exude and accumulate in the highly vascularized suprachoroidal space, resulting in ciliary body-choroidal detachment and exudative retinal detachment [12, 13]. Therefore, ciliary body detachment represents a direct consequence and an imaging manifestation of active uveal inflammation in VKH syndrome. The cornerstone of management lies in the aggressive control of uveal inflammation. Systemic corticosteroids, serving as the first-line therapy, effectively suppress the immune-inflammatory response and reduce vascular permeability, thereby promoting the resolution of suprachoroidal fluid [12, 14]. As the effusion is absorbed, the ciliary body repositions posteriorly. This alleviates its forward push against the iris, allowing for the restoration of anterior chamber depth and the reopening of the anterior chamber angle. This therapeutic response pathway further corroborates the causative relationship between ciliary body detachment and underlying inflammation.

The diagnostic challenge in this case was distinguishing it from primary VKH disease.

First, the diagnostic pivot hinges on meticulous exclusion and observed drug-response dynamics. Our patient’s presentation—unilateral onset with rapid progression to bilateral anterior chamber involvement, optic disc edema, and multifocal serous detachments—was highly suggestive of VKH syndrome. The comprehensive negative workup for infection, malignancy, and other autoimmune diseases satisfied the essential criteria for both “probable VKH” per revised diagnostic criteria [15] and for drug-induced VKH-like uveitis as proposed by recent case series [9]. The latter diagnosis was ultimately cemented not by a single test, but by the unequivocal recurrence of inflammation following two consecutive tislelizumab rechallenges. This pattern of “flare-on-infusion” provides a powerful, real-world evidence for causality that is often difficult to capture and is a critical piece of information for oncologists weighing the risks and benefits of continuing therapy. Second, the absence of CSF pleocytosis in our patient may be a meaningful differentiating clue. A study reported that the incidence of cerebrospinal fluid (CSF) pleocytosis in VKH patients is 82.7% [15]. On account of its superior sensitivity, cerebrospinal fluid (CSF) examination is recommended as a diagnostic tool for VKH disease. In this case, CSF examination was performed—a practice rarely reported in cases of VKH-like uveitis. Unexpectedly, the patient had no CSF pleocytosis. It is rare in VKH disease to have no neurological and/or auditory symptoms, nor pleocytosis. A previous study reported negative CSF cytology results in patients with VKH-like uveitis [15]. This divergence suggests potential differences in the primary site or mechanism of immune dysregulation. Therefore, VKH-like uveitis may have a different etiology from VKH disease, and CSF cytology results may be one of the key differential points between the two conditions. In drug-induced cases, the trigger may directly activate T-cells specific for choroidal melanocytes without invoking the same degree of meningeal involvement characteristic of the primary disease. While HLA-DRB1*04:05 positivity is a shared risk factor [16, 17], its absence in our patient (though untested) does not rule out the diagnosis but highlights that other genetic or pharmacologic factors are at play.

Atypical manifestations of rare conditions, including VKH-like uveitis induced by anticancer therapy, pose difficulties for clinicians in formulating disease definitions and treatment standards. Some researchers have proposed diagnostic criteria [9]. It has also been proposed that establishing a specific time interval between anticancer therapy and the onset of ocular manifestations is inappropriate; according to previous reports, VKH-like uveitis can occur within 1 month to more than 1 year after anticancer therapy initiation, with 64.71% of cases developing ocular symptoms within 12 weeks of treatment [9]. Furthermore, immune-related side effects have been documented to manifest within days to months of starting anticancer therapy, and can even develop subsequent to treatment cessation [18]. For this patient, an initial diagnosis of VKH disease was made, and they responded positively to systemic corticosteroid treatment. However, the condition relapsed after resumption of tislelizumab treatment. Ultimately, the patient achieved a favorable outcome following discontinuation of chemotherapy and adherence to corticosteroid therapy. Combined with the aforementioned diagnostic criteria, this case was appropriately diagnosed as VKH-like uveitis.

Due to the multifaceted impacts of cancer and ocular disease on patients, the management of VKH-like uveitis (e.g., discontinuation of anticancer therapy, switching to alternative agents, and anti-inflammatory treatment) is significantly complex. Currently, there is no consensus on this issue. The American Society of Clinical Oncology (ASCO) recommends temporarily discontinuing anticancer therapy and administering systemic corticosteroids in such case [19]. In this case, the possibility of drug-related uveitis was not recognized in the early stage, so the patient was not advised to discontinue chemotherapy. Even though the ocular condition relapsed after the last two cycles of chemotherapy, the patient ultimately achieved stable improvement with persistent corticosteroid therapy. In previous case series of VKH-like uveitis, most patients received high-dose systemic corticosteroids while discontinuing anticancer therapy; two patients required long-term corticosteroid therapy combined with immunomodulatory treatment because anticancer therapy could not be suspended due to the nature of their tumors [9]. These findings suggest that corticosteroid pulse therapy may be the optimal first-line treatment for both VKH disease and VKH-like uveitis, particularly in cases of severe posterior uveitis.

With the increasing incidence of cancer and the continuous emergence of cancer targeted therapy and immunotherapy agents, we should fully recognize the prevalence and harm of drug-induced uveitis. For ophthalmologists, it is essential to pay attention to potential ocular complications of systemic medications. For all patients with uveitis of unknown etiology, a detailed medical history should be obtained to achieve early detection, diagnosis, and treatment, thereby minimizing ocular damage caused by uveitis. For oncologists, it is important to understand and inform patients of the potential risk of severe ocular adverse reactions during anti-tumor treatment, and advise patients to undergo regular ophthalmic follow-up.

The absence of baseline ICG to fully assess choroidal involvement and the lack of follow-up angiographic studies represent limitations in our multimodal imaging assessment, and the lack of FAF imaging limits our ability to comment on RPE changes throughout the disease course. Furthermore, this study did not include analysis of choroidal thickness. Future investigations utilizing enhanced-depth imaging (EDI) OCT or swept-source OCT protocols would be valuable to characterize choroidal changes in this context.

Conclusions

In conclusion, as the use of tislelizumab and other ICIs expands, awareness of their potential to induce VKH-like uveitis must heighten among both oncologists and ophthalmologists. This case demonstrates that diagnosis relies on a high index of suspicion, a rigorous exclusion process, and close attention to the temporal relationship with drug administration. The atypical finding of absent CSF pleocytosis may aid in differentiation. Management requires seamless collaboration to tailor a strategy that safeguards vision without unduly compromising anti-cancer treatment, often involving a delicate balance of corticosteroids and careful timing of ICI administration.

Abbreviations

PD-1

Programmed death-1

VKH

Vogt-Koyanagi-Harada

BCVA

Best-corrected visual acuity

IOP

Intraocular pressure

OCT

Optical coherence tomography

FFA

Fluorescein angiography

UBM

Ultrasound biomicroscopy

CSF

Cerebrospinal fluid

RPE

Retinal pigment epithelium

irAEs

Immune-related adverse events

ICIs

Immune checkpoint inhibitors

ACD

Anterior chamber depth

Author contributions

All authors contributed substantially to the conception and design of the study, data acquisition, data analysis and interpretation, drafting the manuscript or critically revising it for important intellectual content, as well as the final approval of the published version.

Funding

No funding was received for this study.

Data availability

The data supporting the findings of this study can be obtained from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval was not required for this case report in accordance with the local ethical review regulations.

Consent for publication

The patient provided written informed consent for publication of this case report and accompanying images.

Competing interests

The authors declare no competing interests.