Evaluation of a Commercially Developed Semiautomated PCR–Surface-Enhanced Raman Scattering Assay for Diagnosis of Invasive Fungal Disease

2014 
Nonculture-based tests are gaining popularity in the diagnosis of invasive fungal disease (IFD), but PCR is excluded from disease-defining criteria because of limited standardization and a lack of commercial assays. Commercial PCR assays may have a standardized methodology while providing quality assurance. The detection of PCR products by a surface-enhanced Raman scattering (SERS) assay potentially provides superior analytical sensitivity and multiplexing capacity compared to that of real-time PCR. Using this approach, the RenDx Fungiplex assay was developed to detect Candida and Aspergillus. Analytical and clinical evaluations of the assay were undertaken using extraction methods according to European Aspergillus PCR Initiative (EAPCRI) recommendations. A total of 195 previously extracted samples (133 plasma, 49 serum, and 13 whole blood) from 112 patients (29 with proven/probable IFD) were tested. The 95% limit of detection of Candida and Aspergillus was 200 copies per reaction, with an overall reproducibility of 92.1% for detecting 20 input copies per PCR, and 89.8% for the nucleic acid extraction–PCR-SERS process for detecting fungal burdens of <20 genome equivalents per sample. A clinical evaluation showed that assay positivity significantly correlated with IFD (P < 0.0001). The sensitivity of the assay was 82.8% and was similar for both Candida (80.0%) and Aspergillus (85.7%). The specificity was 87.5% and was increased (97.5%) by using a multiple (≥2 samples) PCR-positive threshold. In summary, the RenDx Fungiplex assay is a PCR-SERS assay for diagnosing IFD and demonstrates promising clinical performance on a variety of samples. This was a retrospective clinical evaluation, and performance is likely to be enhanced through a prospective analysis of clinical validity and by determining clinical utility.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    24
    References
    35
    Citations
    NaN
    KQI
    []