Highlight
- The ISOLD score demonstrated high diagnostic accuracy for vitreoretinal lymphoma (VRL) using intraocular cytokine levels, especially in aqueous humor (AH) samples.
- The score outperformed the conventional IL-10/IL-6 ratio by providing a continuous probabilistic assessment.
- False-negative results primarily occurred in immunosuppressed patients receiving corticosteroids or chemotherapy, emphasizing clinical context in interpretation.
- Vitreous sample data, though limited, indicate excellent ISOLD performance but require further validation through larger studies.
Study Background
Vitreoretinal lymphoma (VRL) is a rare but aggressive intraocular malignancy, often associated with primary central nervous system lymphoma (PCNSL). It poses significant diagnostic challenges due to nonspecific clinical presentations and overlap with inflammatory conditions such as uveitis. Early and accurate diagnosis is crucial, as delayed or missed detection can lead to poor prognosis and inappropriate management. Traditional diagnostic gold standards rely on cytology from vitreous biopsies; however, cytological analysis suffers from limited sensitivity due to sample paucity and technical difficulties. Measurement of intraocular cytokines, particularly interleukin-10 (IL-10) and interleukin-6 (IL-6), has emerged as a valuable adjunct diagnostic tool because VRL typically exhibits elevated IL-10 and suppressed IL-6 levels compared to inflammatory etiologies. The IL-10/IL-6 ratio has been widely used but has limitations in sensitivity and specificity.
The ISOLD score (Interleukin Score for IntraOcular Lymphoma Diagnosis) is a recently developed probabilistic formula incorporating absolute IL-10 and IL-6 levels to provide a continuous estimate of VRL likelihood. While it shows promise in small research settings, its diagnostic performance in routine clinical practice, especially using aqueous humor (AH) samples obtained by anterior chamber paracentesis (ACP), remains under-evaluated.
Study Design
This retrospective diagnostic accuracy study analyzed data from 166 patients undergoing intraocular fluid sampling for IL-6 and IL-10 measurement at a tertiary VRL referral center from January 2017 to June 2019. A total of 183 samples were included: 156 aqueous humor (AH) samples from ACP and 27 vitreous samples. Among these, 25 patients had confirmed VRL (18 AH samples, 20 vitreous samples; 13 patients contributed both AH and vitreous samples). The comparator group consisted of 141 non-VRL patients, including 9 with PCNSL without ocular involvement and 132 without lymphomatous disease.
IL-6 and IL-10 concentrations were measured in undiluted samples using cytometric bead array (CBA) technology. The ISOLD score was calculated using published formulas adapted to sample type (AH or vitreous). Diagnostic performance metrics — sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) — were assessed for the ISOLD score and the traditional IL-10/IL-6 ratio at established thresholds. Receiver operating characteristic (ROC) curve analysis was conducted to establish optimal cutoffs for maximal diagnostic accuracy.
Key Findings
Overall, the ISOLD score at the threshold of 0 achieved excellent diagnostic performance across all samples (n=183), with sensitivity 86.8%, specificity 98.6%, PPV 94.3%, and NPV 96.6%. Detailed findings include:
Aqueous Humor Samples (n=156)
– At threshold 0: sensitivity 77.8%, specificity 98.6%, PPV 87.5%, NPV 97.1%
– ROC-optimized threshold of -5.2 enhanced sensitivity to 94.4% and specificity to 96.7%, with AUC 0.966
– PPV decreased to 73.9% at this cutoff, but NPV increased to 99.3%
Vitreous Samples (n=27)
– At threshold 0: sensitivity 95.0%, specificity 100%, PPV and NPV nearly 100%
– Optimal threshold of -2.55 yielded perfect sensitivity, specificity, PPV, and NPV, with an AUC of 1.0
IL-10/IL-6 Ratio Performance
– AH samples at conventional threshold 0.6: sensitivity 77.8%, specificity 93.9%, PPV 63.6%, NPV 96.9%
– ROC-optimized threshold 0.48 improved sensitivity to 83.3% and specificity to 93.9%, with AUC 0.942
– Vitreous samples at threshold 1: sensitivity 80%, specificity 100%, PPV 100%, NPV 63.6%
– ROC curve analysis identified an optimal threshold of 0.195 with 100% sensitivity and specificity
Additional Observations
– Two false positives with ISOLD occurred in patients with varicella-zoster virus (VZV)-associated ocular disease
– Five false negatives were mainly linked to prior systemic corticosteroid or chemotherapy exposure, which may suppress cytokine expression
– Clinical features such as bilateral ocular involvement and intermediate uveitis were significantly more frequent in VRL patients (p=0.008 and p=0.005)
Expert Commentary
The ISOLD score represents an advancement over the IL-10/IL-6 ratio by quantifying lymphoma probability on a continuous scale, enhancing nuanced clinical interpretation. This continuous probabilistic approach mitigates the limitations inherent in dichotomous thresholds and offers clinicians a more sensitive and specific diagnostic decision aid. Its superior accuracy in AH samples is particularly valuable since anterior chamber paracentesis is less invasive than vitreous biopsy, facilitating earlier and safer diagnosis.
However, caution is warranted when interpreting results in patients with immunosuppression or prior steroid therapy, as these factors may blunt cytokine levels and lead to false-negative assessments. Moreover, the small size of the vitreous sample cohort limits the robustness of conclusions for this specimen type; larger studies are needed to validate ISOLD’s vitreous performance and refine thresholds.
Despite the promising performance of ISOLD, vitreous cytology remains the definitive gold standard due to its direct visualization of malignant lymphoma cells. Therefore, ISOLD should complement, not replace, cytological analysis, particularly when clinical suspicion is high and biopsy is feasible.
Conclusion
The ISOLD score is a reliable and validated diagnostic tool for vitreoretinal lymphoma in routine clinical practice, especially using aqueous humor samples obtained by anterior chamber paracentesis. It offers improved sensitivity and specificity compared to the traditional IL-10/IL-6 ratio by providing a continuous probability estimate. Clinicians should be aware of potential false negatives in immunosuppressed patients and continue to employ cytological evaluation where possible. Future studies with larger vitreous sample cohorts are necessary to finalize the role of ISOLD in vitreous diagnostics and to optimize its integration into VRL diagnostic algorithms.
Funding and Clinicaltrials.gov
The study by Chevalier et al. was conducted at a tertiary VRL referral center, with no specific funding or clinical trial registration information reported in the source article.
References
1. Chevalier M, Sourdeau E, Garff-Tavernier ML, et al. Performance of the ISOLD Score (Interleukin Score for IntraOcular Lymphoma Diagnosis) in the Diagnosis of Vitreoretinal Lymphoma in Clinical Practice. Am J Ophthalmol. 2026 Jul 6. PMID: 42409156.
2. Chan CC, Wallace DJ. Diagnosis and Treatment of Intraocular Lymphoma. Cancer Control. 2010;17(3): 195-204.
3. Shen DF, Marcet MM, Schiffman J, et al. Interleukin-10 and Interleukin-6 Levels in Primary Intraocular Lymphoma and Uveitis. Ophthalmology. 2011;118(4): 776-780.
4. Buggage RR, Chan CC, Nussenblatt RB. Current concepts and approach in the diagnosis and management of ocular lymphoma. Curr Opin Ophthalmol. 2006;17(3): 227-232.
5. Tennenbaum R, Esquenazi S. Vitreoretinal lymphoma: diagnostic and management considerations. Curr Opin Ophthalmol. 2019;30(6): 499-505.