Ocular Microbiology and Immunology Group
Back to OMIG Main Page

< Previous | 2026 Agenda and Abstracts | Next >

2026 OMIG Abstract

Clinical Assessment of a Metagenomic Sequencing Approach for Diagnosis of Infectious Uveitis

Ananya Mukundan¹, Vridhi Vinaykiya¹, Sudeep Mehrotra¹, Katie Huynh¹, Renee Liu¹, Jennifer Hu¹, Matt McKay¹, Eduardo Ferracioli-Oda², Joyce Hisae Yamamoto², Lucia Sobrin¹, and Paulo J. M. Bispo¹

¹Department of Ophthalmology, Massachusetts Eye and Ear, Harvard Medical School, Boston, Massachusetts; ²Department of Ophthalmology, School of Medicine, University of Sao Paulo, Sao Paulo, Brazil

Purpose: Infectious uveitis is a vision-threatening intraocular infection caused by a wide range of pathogens from bacteria, fungi, and protozoa. Diagnosis and management typically rely on laboratory testing, such as cultures and targeted PCR, which are time-consuming, limited by the small volume of intraocular fluid that can be safely obtained, and sometimes nondiagnostic. We evaluated the clinical performance and utility of metagenomic next-generation sequencing (mNGS) for the diagnosis of infectious uveitis.

Methods: A mNGS workflow was systematically applied to aqueous and vitreous fluid specimens from infectious uveitis cases (n = 78) and non-infectious uveitis controls (n = 16). Pathogen load was quantified as reads per million (RPM) and expressed as an RPM ratio (RPM-r = RPM_sample / RPM_non-template control). Pathogen detection thresholds were derived using receiver operating characteristic (ROC) curve analysis and the Youden Index. Clinical performance was compared with targeted PCR.

Results: RPM-r values correlated with PCR Ct values (Spearman rho = -0.5286; 95% CI, -0.7485 to -0.2040; P = 0.0022), supporting semiquantitative assessment of pathogen load. Using RPM-r thresholds for viral uveitis and ocular toxoplasmosis, the assay showed good overall performance (AUC = 0.817; 95% CI, 0.721-0.914), with sensitivity of 53.8% (95% CI, 42.2%-65.1%) and specificity of 93.8% (95% CI, 67.7%-99.7%). Performance was higher for viral uveitis (sensitivity 63.8%; 95% CI, 46.2%-78.6%; AUC = 0.806; 95% CI, 0.691-0.921) than for ocular toxoplasmosis (sensitivity 35.1%; 95% CI, 21.0%-52.6%; AUC = 0.665; 95% CI, 0.540-0.791). Median RPM-r values were higher for herpesviruses (409,562; interquartile range [IQR], 11,911-145,664) than for Toxoplasma gondii (median RPM-r = 3,502; IQR, 1,837-19,568). mNGS identified likely etiologies in 11 specimens (10 cases) that were PCR-negative, with 70% concordance between the detected pathogen and the suspected clinical diagnosis.

Conclusions: Our mNGS pipeline detects infectious uveitis cases with good overall accuracy, adds diagnostic yield beyond extensive targeted PCR testing, and provides semiquantitative metrics that correlate with Ct values and support reliable pathogen calls, successfully establishing and etiology for all PCR-positive cases as well as several PCR-negative ones. These results highlight mNGS as a promising tool to advance precision diagnosis in uveitis.