Metagenomic next-generation sequencing of cerebrospinal fluid for diagnosis of central nervous system infections: 7-year performance of a clinically validated test

Abstract Studies have shown that metagenomic next-generation sequencing (mNGS) testing of cerebrospinal fluid (CSF) in central nervous system (CNS) infections can improve diagnostic yields and provide actionable information. We analyzed the results of all CSF mNGS tests (n=4,828) performed at the University of California, San Francisco (UCSF) clinical microbiology laboratory from June 2016 to April 2023. We also assessed clinical metadata from a subset of samples that corresponded to a cohort of UCSF patients (n=1,164) who received CSF mNGS testing, and retrospectively evaluated performance compared to conventional microbiologic testing and adjudicated clinical diagnosis. Overall, 14.4% of CSF mNGS tests were positive for any microorganism. DNA viruses (7% of all samples) were detected most often, followed by RNA viruses (4.3%), bacteria (2.7%), fungi (1.4%), and parasites (0.5%). Using a composite gold standard obtained from clinical adjudication and all microbiological test results, sensitivity, specificity, and accuracy of CSF mNGS in the UCSF cohort who had clinically diagnosed infections were 56.5%, 98.8%, and 90.5%, respectively. The sensitivity of CSF mNGS testing (56.5%) was statistically higher than that from all direct detection testing from CSF (44.8%, p = 0.004), direct detection testing from samples other than CSF (15.2%, p<0.001), and indirect serologic testing (34%, p<0.001). When only considering diagnoses made by direct detection of pathogens on CSF, sensitivity of mNGS was 80.7%. These results justify the incorporation of CSF mNGS testing as part of the routine diagnostic workup in hospitalized patients presenting with potential CNS infection.