Performance of metagenomic next-generation sequencing for microbiological diagnosis of infectious uveitis

<b>Introduction.</b> Diagnosis of uveitis is challenging due to the multitude of possible pathogenies. Identifying infectious and non-infectious uveitis is of great clinical significance. Recently, metagenomic next-generation sequencing (mNGS) was used to detect infectious and non-infectious uveitis, but its efficacy has not been widely evaluated.<b>Hypothesis.</b> Compared with routine diagnostic tests (RDTs), mNGS is more effective in identifying infectious and non-infectious uveitis.<b>Aim.</b> To describe the microbiological diagnostic performance of mNGS in detecting infectious and non-infectious uveitis.<b>Methodology.</b> Patients with suspected infectious uveitis of uncertain pathogenesis were tested by mNGS and RDTs. Infectious and non-infectious uveitis were grouped according to the final diagnosis based on comprehensive analysis of the test results and the effect of therapy. The test results were used to assess the performance of mNGS in actual clinical practice.<b>Results.</b> Fifty-eight cases were enrolled in this project, including 32 cases of infectious uveitis and 26 cases of non-infectious uveitis. The sensitivity of mNGS was 96.88%, which was much higher than that of RDTs. The detected pathogenic micro-organisms included bacteria, fungi, viruses, <i>Toxoplasma gondii</i> and <i>Bartonella</i>. Consequently, mNGS showed a high negative predictive value (NPV) of 94.74%, indicating that an mNGS negative should be a true negative result most of the time, but a low positive predictive value (PPV) of 79.49%.<b>Conclusions.</b> mNGS showed extremely high sensitivity but low specificity, which increased the detection rate of infectious uveitis pathogens but might result in false positives. The excellent NPV suggested that the identification of non-infectious uveitis is of considerable clinical importance.