Integrating central nervous system metagenomics and host response for diagnosis of tuberculosis meningitis and its mimics
Problem
Tuberculosis meningitis is notoriously hard to diagnose in sub-Saharan Africa, and causes of meningitis there are poorly understood.
Why it matters
Difficult diagnosis of TBM and its mimics in HIV-infected adults delays appropriate treatment.
What the AI did
CSF metagenomic sequencing was integrated with a host gene expression machine learning classifier to distinguish TBM from other infections, evaluated blinded on an independent dataset.
Results
368 HIV-infected Ugandan adults prospectively enrolled; classifier trained and blindly evaluated on independent data; numeric accuracy not stated in abstract.
Impact
mNGS identified treatable and vaccine-preventable TBM mimics and co-infections, enhancing diagnostic accuracy.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2022 |
| Work type | article |
| Authors | Prashanth Ramachandran, Akshaya Ramesh, Fiona V Cresswell, Anne E. Wapniarski, Rithwik Narendra, Carson M. Quinn, Edwina B. Tran, Morris K Rutakingirwa et al. |
| Citations | 82 (OpenAlex, at import time) |
| References | 54 |
| Model or method | Host gene expression-based machine learning classifier (MLC) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-022-29353-x |
| Open access | gold |
| Keywords | tuberculosis meningitis, metagenomics, machine learning, CSF, diagnostics |
Importance rating
significant — thousands to millions affected
Abstract states TB is a common cause of meningitis in sub-Saharan Africa and notoriously hard to diagnose; combined mNGS+MLC assay aids diagnosis of TBM and treatable/vaccine-preventable mimics in HIV-infected adults.