Evaluation of two MALDI-TOF MS systems and extraction methods for identification of filamentous fungi recovered from clinical specimens

Rapid and accurate identification of cultured molds is important to determine clinical significance and therapeutic decision-making. Conventional mold identification uses phenotypic macroscopic and microscopic characterization; however, this can take days or weeks for colony maturity and definitive microscopic structure formation, be limited to genus-level identification, and be misidentified due to morphologic mimics or similarities between closely related species. Matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI–TOF MS) revolutionized bacterial and yeast identification but remains uncommon for molds in part because of limited reference libraries. Here, a retrospective 5-year review at a large teaching hospital found that 88.6% of identified molds were in the Bruker Filamentous Fungi Library 3.0 and 91.5% in the VITEK Knowledge Base Library 3.2.0. A prospective evaluation was also performed on early growth from 205 consecutive, working clinical isolates. Each mold was processed using the VITEK chemical extraction method and modified NIH chemical plus bead-beating extraction method; both extractions were tested on both systems. When compared to conventional identification, more molds were identified using VITEK extractions over NIH extractions using the VITEK (65 and 59%) and Bruker (56 and 54%) systems, respectively, using the ≥1.5 log Bruker threshold. VITEK MS identified more molds, regardless of the extraction method. Isolates without consensus agreement (n = 116) underwent sequence-based identification, which demonstrated that conventional identification had the highest genus-level (84%) but lowest species-level (3%) identification rates compared to VITEK (59 and 52%, respectively) and Bruker (52 and 36%) using VITEK extractions. Taken together, our findings suggest both MALDI-TOF systems can supplement conventional mold identification to optimize identification rates with species-level distinction.