Diagnosis of Neurological Infections in Pediatric Patients from Cell-Free DNA Specimens by Using Metagenomic...

Central nervous system (CNS) infections can cause significant morbidity and mortality, especially in children. Rapid and accurate pathogenic detection in suspected CNS infections is essential for disease control at the early stage of infection. To evaluate the performance of metagenomic next-generation sequencing (mNGS) of cell-free DNA (cfDNA) in cerebrospinal fluid (CSF) in pediatric patients, we retrospectively analyzed the efficiency of cfDNA mNGS in children with CNS infections (n = 257) or noninfectious neurological disorders (n = 81). The CSF samples of 124 random subjects were used to evaluate the accuracy between mNGS of cfDNA and whole-cell DNA (wcDNA). In total, cfDNA mNGS detected a wide range of microbes with a detection rate of 71.0%, and the sensitivity and total coincidence rate with clinical diagnosis reached 68.9% and 67.5%, respectively. Compared with wcDNA mNGS, cfDNA mNGS had a higher efficacy in detecting viruses (66 versus 13) and Mycobacterium (7 versus 1), with significantly higher reads per million. The dominant causative pathogens were bacteria and viruses in CNS infections, but these presented with different pathogen spectra in different age categories. The best timing for the mNGS test ranged from 1 to 6 days after the start of anti-infection therapy, and the earlier mNGS started, the better was identification of bacterial CNS infections. This study emphasized that cfDNA mNGS had overall superiority to conventional methods on causative pathogen detection in pediatric CNS infections, and it was even better than wcDNA mNGS. Furthermore, research needs to be better validated in large-scale clinical trials to improve the clinical applications of cfDNA mNGS. IMPORTANCE Our study emphasized that cfDNA mNGS had overall superiority to conventional methods on causative pathogen detection in CNS-infected children, and it was even better than wcDNA mNGS. cfDNA mNGS detected a wide range of pathogens with a high total coincidence rate (67.5%) against clinical diagnosis. The best timing for cfDNA mNGS detection ranged from 1 to 6 days, rather than 0 days, after the start of empirical anti-infection therapy. The earlier mNGS started, the better the identifications of bacterial CNS infections. To the best of our knowledge, this research is the first report evaluating the clinical utility of mNGS with different methods (cfDNA versus wcDNA) of extracting DNA from CSF specimens in diagnosing pediatric CNS infections. Meanwhile, this is the largest cohort study that has evaluated the performance of mNGS using cfDNA from CSF specimens in pediatric patients with CNS infections.