High-throughput screening for central nervous system infection-associated pathogens using TaqMan array cards, metagenomic next-generation sequencing, and targeted next-generation sequencing: a retrospective study

Background Central nervous system (CNS) infections feature diverse pathogens and ambiguous clinical manifestations, which complicate rapid and accurate etiological diagnosis. This study aimed to characterize the pathogen spectrum in cerebrospinal fluid (CSF) specimens from patients with clinically suspected CNS infections and to evaluate the detection performance of metagenomic next-generation sequencing (mNGS) and targeted next-generation sequencing (tNGS) for pathogen identification in CSF. Methods This study employed three relatively independent workflows, namely TaqMan Array Cards (TAC) screening, Illumina-based unbiased mNGS, and Oxford Nanopore Technologies (ONT)-based tNGS, for the identification of CNS pathogens. Results In this retrospective study, 316 clinical CSF specimens were examined. The overall pathogen-positive rate of TAC-qPCR strategy was 24.68%, and polymicrobial positive samples accounted for 4.11% of all samples. We further compared the pathogen detection performance of mNGS and tNGS using 48 representative CSF samples (25 qPCR-positive, 23 qPCR-negative). The sensitivities of mNGS and tNGS for identifying CNS pathogens among qPCR-positive CSF samples were 68.57 and 65.71%, respectively, indicating comparable overall sensitivity; however, mNGS achieved higher specificity (82.35%) than tNGS (72.73%). Unbiased mNGS enables the identification of pathogens not covered by the qPCR panel; for instance, Human parvovirus B19 was detected by mNGS in several samples and further confirmed by qPCR. By contrast, benefiting from targeted amplification, tNGS yielded approximately 692-fold higher pathogen RPM and only 9.2% host reads (versus 89.0% in mNGS), as well as a substantially shortened sample-to-answer turnaround time (10 h versus 30 h). However, tNGS carried a risk of false positives caused by elevated signals from environmental contaminant bacteria. Our combined TAC-qPCR, Illumina-mNGS and Nanopore-tNGS testing identified a broad spectrum of CNS pathogens covering the major viral, fungal and bacterial etiological agents of central nervous system infections. Conclusion The TAC strategy enables high-throughput preliminary pathogen screening of large sample cohorts, while the unbiased mNGS workflow allows identification of rare or unexpected pathogens outside predefined panels, and tNGS achieves robust target enrichment and a greatly shortened testing turnaround time. Therefore, the optimal sequencing approach should be selected on a case-by-case basis according to clinical demands and suspected pathogens.