Pathogen Detection and Associated Patient Characteristics from Clinical Metagenomic Next-generation Sequencing Testing for Neurological Infections
Amy Wong1, Brian O'Donovan1, Noah Brazer2, Jenny Nguyen2, Miriam Oseguera2, Jack Tang2, Charles Omura2, Venice Servellita2, Patrick Benoit2, Emily Kelly2, Steve Miller1, Charles Chiu2
1Delve Bio, 2UCSF
Objective:
To analyze the performance of a metagenomic next-generation sequencing (mNGS) assay for diagnosis of infectious causes of meningitis and/or encephalitis when used in real-life clinical settings over 7 years.
Background:
Meningitis and encephalitis (ME) cases remain largely undiagnosed, with infectious pathogens representing 40-60% of cases. Metagenomic next-generation sequencing (mNGS) is a broad, agnostic detection technology that has been used to detect viruses, bacteria, fungi and parasites in CSF samples from patients with neurologic infection. Here we present the analysis of a large cohort of UCSF patients over 7 years who have received mNGS testing.
Design/Methods:
Electronic medical records were reviewed to identify results of mNGS testing on CSF samples from 2016-2023 (n=4849). Chart review was conducted for UCSF patients (n=968) to assess demographics, diagnoses (whether infectious or non-infectious), CSF profile, including cell count and differential, glucose, and protein, and other relevant laboratory and imaging results.
Results:
25.4% of patients (n=1231/4849) had a positive mNGS result with at least one organism detected and 74.6% had a negative result. Detected organisms were DNA virus (26%), RNA virus (16%), bacteria (44%), fungi (11%), and parasite (3%). Patient demographics (n=968) showed an average age of 46.4 years, 37.1% immunocompromised status, 39.8% admitted to the ICU, and average length of stay of 20.9 days. mNGS detected significantly more organisms than CSF culture and universal PCR. Average CSF protein, glucose and WBC counts did not differ between mNGS-positive and mNGS-negative samples, although variability in these parameters was high.
Conclusions:
Overall, mNGS detected at least one pathogen in CSF in 25% of cases, with 192 different species represented.  Positive rate was higher in immunocompromised patients (30%) versus non-immunocompromised (15%), although there were no significant differences in CSF cell counts between patients with positive versus negative mNGS results. Further analyses and clinical studies are required to maximize mNGS diagnostic utility.
10.1212/WNL.0000000000204795