Correlation between interleukins in aqueous humor and vitreous humor of vitreoretinal lymphoma patients

Yuli Liu, Xinyi Zhou, K Zhang, Shixue Liu, Ruiwen Li, Yifan Gong, Zhujian Wang, Tingting Jiang, Ting Zhang, Gezhi Xu, Junxiang Gu, Qing ChangEye 2025 · DOI

This study, published in Eye and Vision, investigates the diagnostic reliability of testing eye fluids for vitreoretinal lymphoma (VRL), a rare but aggressive "masquerade" cancer that often mimics common eye inflammation. By comparing paired samples of the easy-to-access fluid from the front of the eye (aqueous humor) with the harder-to-reach fluid from the back (vitreous humor), the researchers identify critical clinical markers—specifically intraretinal infiltration—that determine when the easier test might fail.

TL;DR

While testing aqueous humor (AqH) is a less invasive way to screen for vitreoretinal lymphoma, it is significantly less sensitive than testing vitreous humor (VH), especially in patients without intraretinal infiltration. The study reveals that interleukin levels in these two fluids correlate via a power function, and clinicians should be wary of false negatives in AqH when retinal involvement is absent.

What Problem Does This Paper Address?

Vitreoretinal lymphoma (VRL) is notoriously difficult to diagnose because its symptoms overlap with uveitis (inflammation). The "gold standard" for diagnosis involves a vitreous biopsy, which is invasive and requires surgery.

Clinicians have turned to measuring interleukins (specifically IL-10 and IL-6) in the aqueous humor (AqH) as a "liquid biopsy" because AqH can be sampled in an office setting. However, AqH levels are often lower than those in the vitreous humor (VH), leading to potential underdiagnosis. This paper seeks to quantify the correlation between these two fluids and identify which clinical factors (like age, lens status, or specific retinal lesions) cause AqH testing to lose its diagnostic edge.

Methods

The researchers conducted a retrospective case series of 74 eyes (64 patients) with biopsy-proven B-cell VRL. Key methodological points include:

  • Paired Sampling: For every eye, they took undiluted samples of both AqH and VH.
  • Cytokine Analysis: They measured IL-10 and IL-6 using a cytometric bead assay and calculated the ISOLD score (Interleukin Score for Intraocular Lymphoma Diagnosis).
  • Advanced Imaging: Optical Coherence Tomography (OCT) was used to categorize retinal involvement into two types: Retinal Pigment Epithelium (RPE) involvement vs. deeper intraretinal infiltration.
  • Statistical Modeling: They used General Estimating Equations (GEE) to account for patients where both eyes were studied and applied curve-fitting to determine the mathematical relationship between AqH and VH cytokine concentrations.

Datasets

The study utilized a clinical dataset from consecutive patients at the Fudan University Eye & ENT Hospital.

  • Cohort: 74 eyes from 64 patients.
  • Demographics: 69% female, average age 60.
  • Molecular Markers: 73% of cases were positive for the MYD88 L265P mutation, a common genetic marker for this type of lymphoma.

Key Insights and Ideas

  • The Power Function Relationship: The correlation of IL-10 and IL-6 between the two fluids is best described by a power function ($R^2 \approx 0.52$), rather than a linear one. This suggests that as the tumor load in the back of the eye increases, the concentration in the front doesn't rise at the same rate, potentially due to complex diffusion limits or metabolic clearance.
  • The "Infiltration" Threshold: The most significant finding was that intraretinal infiltration (cancer cells invading the neural retina) is the primary driver of high cytokine levels. If the lymphoma is limited to the sub-RPE space, IL-10 levels in the AqH remain low, making the "easier" test much more likely to return a false negative.
  • Diffusion Independence: Surprisingly, factors that one might assume would affect fluid movement—such as whether the patient had cataract surgery (lens status) or the length of the eye (axial length)—had no significant impact on the cytokine concentration gradient.

Key Results

  • Concentration Gap: IL-10 levels in the VH were significantly higher than in the AqH (Median: 1159.77 pg/mL vs. 225.74 pg/mL).
  • Sensitivity Disparity: The positive rate for the IL-10/IL-6 ratio in AqH was only 77%, compared to 89-91% for various parameters in the VH.
  • OCT as a Predictor: In eyes without intraretinal infiltration, the median IL-10 in AqH was only 141.08 pg/mL, compared to 449.10 pg/mL in eyes with infiltration ($P=0.001$).
  • Diagnostic Failure: Eyes without intraretinal involvement were significantly less likely to trigger a positive AqH ISOLD score (76% vs 97%).

Limitations and Future Work

The authors acknowledge several limitations:

  • Retrospective Design: The study looks back at existing cases, which can introduce bias.
  • Lack of Uveitis Control: Because diagnostic vitrectomies are rarely performed on standard uveitis patients, the study could not calculate the specificity (the ability to correctly identify those without the disease) of these markers.
  • Sample Size: While large for a rare disease, the number of pseudophakic (artificial lens) patients was small, which might have masked the influence of lens status on cytokine diffusion.

Future Work: The team is currently conducting a prospective cohort study to include more uveitis cases and further refine the predictive models for abnormally low AqH interleukin levels.

Summary updated

Abstract

IL-10/IL-6 in AqH may not be as sensitive as the parameters (including IL-10, IL-10/IL-6 and ISOLD) in VH for VRL screening. Cases without intraretinal involvement were less likely to be positive for IL-10 > 30 pg/mL and ISOLD > 0 in AqH; the possibility of VRL should be ruled out more cautiously in these cases.
BibTeX
@inproceedings{liu2025correlation,
  author = {Yuli Liu and Xinyi Zhou and K Zhang and Shixue Liu and Ruiwen Li and Yifan Gong and Zhujian Wang and Tingting Jiang and Ting Zhang and Gezhi Xu and Junxiang Gu and Qing Chang},
  title = {{Correlation between interleukins in aqueous humor and vitreous humor of vitreoretinal lymphoma patients}},
  booktitle = {Eye},
  year = {2025},
  doi = {10.1186/s40662-025-00438-0},
}
Figure