Enhanced pathogen identification in fungal endophthalmitis by metagenomic next-generation sequencing: a retrospective clinical evaluation.

Fungal endophthalmitis (FE) is a vision-threatening emergency that requires rapid pathogen identification. Conventional microbial culture demonstrates limited sensitivity in FE, warranting improved diagnostic approaches. We evaluated the detection performance of unbiased metagenomic next-generation sequencing (mNGS) in 31 clinically diagnosed FE cases, including 16 vitreous humor (VH) and 15 aqueous humor (AH) specimens. mNGS showed a positivity rate of 90.3% (28/31, 95% CI: 74.2%-98%), outperforming culture (9.1%, 2/22, 95% CI: 1.1%-29.2%). The positivity rates were 100% for endogenous FE and 85% for exogenous FE, while VH and AH specimens achieved 100% and 80% positivity, respectively. mNGS identified polymicrobial infections in 5 exogenous cases, and a total of 15 fungal species across 9 genera, dominated by Aspergillus flavus, Candida albicans, and Aspergillus niger. Candida albicans and Aspergillus flavus were the predominant pathogens in endogenous and exogenous FE, respectively. Notably, mNGS enabled detection of rare fungal species including Aspergillus niger, Aspergillus welwitschiae, Fusarium oxysporum, Memnoniella echinata, Rhizopus oryzae, Rhizopus microsporus, Chaetomium globosum, and Debaryomyces fabryi. Sequencing results were supported or supplemented by culture, beta-D-glucan, and galactomannan testing in selected cases. Among mNGS-positive cases, 82.1% (23/28) experienced clinical management changes guided by fungal identification. We further propose a laboratory workflow integrating mNGS with conventional assays, tailored to the obtained specimen volume of intraocular fluids.