Metagenomic next-generation sequencing for cryptococcal meningitis diagnosis: a single-center experience

Cryptococcal meningitis (CM) is a devastating central nervous system infection with substantial mortality, particularly when diagnosis is delayed. This study aims to evaluate the diagnostic performance of metagenomic next-generation sequencing (mNGS) for CM detection in comparison with conventional tests.

We enrolled 23 consecutive patients with suspected CM at a tertiary center. All patients met a composite reference standard (CRS) based on CSF cryptococcal antigen (CrAg), CSF/sterile-site culture for Cryptococcus, or CNS histopathology; mNGS was excluded from the CRS. Primary outcomes were CRS-based sensitivity (computed only among CRS-positive patients who underwent each assay) and turnaround times (TATs); pairwise agreement metrics (PPA/NPA) between mNGS and conventional assays were estimated in co-tested subsets.

mNGS identified Cryptococcus in 18/23 (78.3%) cases and detected viral co-pathogens (EBV/CMV/HIV-1) in 5 patients. CRS-based sensitivities were: CrAg LFA (CSF) 83.3% (5/6), Alcian blue 72.7% (16/22), India ink 50.0% (3/6), and CSF culture 66.7% (8/12). Pairwise agreement favored mNGS against culture and CrAg (e.g., PPA 100% vs culture 8/8] and vs CSF CrAg [5/5]), with limited NPA where denominators were small. Median (IQR) TATs were 0.5 (0.5-0.5) days for CrAg LFA, 1 (0.5-1) day for India ink, 5 (3-8) days for first positive culture, and 2 (1-4) days for mNGS.

CSF mNGS complements CrAg, microscopy, and culture by increasing Cryptococcus detection and revealing mixed infections, with particular utility in atypical, pretreated, or complex hosts. Larger studies are warranted to validate clinical utility and define optimal integration with existing workflows.