Identification of filamentous fungi isolates by MALDI-TOF mass spectrometry: clinical evaluation of an extended reference spectra library

Abstract

The identification of filamentous fungi by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) relies mainly on a robust and extensive database of reference spectra. To this end, a large in-house library containing 760 strains and representing 472 species was built and evaluated on 390 clinical isolates by comparing MALDI-TOF MS with the classical identification method based on morphological observations. The use of MALDI-TOF MS resulted in the correct identification of 95.4% of the isolates at species level, without considering LogScore values. Taking into account the Brukers' cutoff value for reliability (LogScore >1.70), 85.6% of the isolates were correctly identified. For a number of isolates, microscopic identification was limited to the genus, resulting in only 61.5% of the isolates correctly identified at species level while the correctness reached 94.6% at genus level. Using this extended in-house database, MALDI-TOF MS thus appears superior to morphology in order to obtain a robust and accurate identification of filamentous fungi. A continuous extension of the library is however necessary to further improve its reliability. Indeed, 15 isolates were still not represented while an additional three isolates were not recognized, probably because of a lack of intraspecific variability of the corresponding species in the database.

Evaluation of the Vitek MS Matrix-Assisted Laser Desorption Ionization–Time of Flight Mass Spectrometry System for Identification of Clinically Relevant Filamentous Fungi

“…The majority of published studies examining large study sets of filamentous fungi utilize the Bruker Biotyper MALDI-TOF MS and the associated library and/or laboratory-developed libraries (10,11,(15)(16)(17)(18). Prior studies have evaluated the accuracy of MALDI-TOF MS for the identification of only a few genera or groups of organisms, such as Aspergillus spp.…”

“…Despite this, a meta-analysis examining the use of MALDI-TOF MS for the identification of yeasts and filamentous fungi showed that while the analytical performance characteristics are not as desirable as for the yeasts, the method is able to provide accurate and rapid mold identification (17). Of note, investigations using expanded laboratory-developed databases have returned a much higher rate of mold identification (11,15). Studies examining MALDI-TOF MS for the identification of filamentous fungi have used different MALDI-TOF MS systems, databases (both manufacturer developed and laboratory developed), and extraction protocols.…”

Identification of Molds by Matrix-Assisted Laser Desorption Ionization–Time of Flight Mass Spectrometry

“…While the reference database provided with each commercial MALDI-TOF MS platform may not be sufficient for routine analyses (see below for specific examples), some authors noticed that increasing the number of mass spectra obtained from distinct subcultures of strains included in the reference spectrum library (i.e., the number of reference entries) would improve the accuracy of MALDI TOF MS-based mold identification. The most clinically comprehensive mold databases, developed recently by Lau et al (12), Gautier et al (14), and Becker et al (15), contained 152, 347, and 472 fungal species, respectively. Notably, the second of the abovedescribed in-house databases was constructed with four spectra for 708 isolates each, encompassing 2,832 reference spectra in total (14).…”

“…Finally, the database of Becker et al enabled the identification of 95.4% (372/390) of clinical isolates to the species level, whereas comparison of the isolates' spectra against the Bruker Biotyper library resulted in a rate of species-level identification of Ͻ65% (15). However, several isolates were still misidentified or not identified, essentially because they belonged to species that were insufficiently or not represented in the databases (12,14,15). Also, identification exclusively to the genus level was achieved for 18 (4.3%) and 11 (0.9%) isolates, respectively, in two studies (12,14).…”

“…Based on a procedure proposed previously by Cassagne et al (18) but using only in-house databases for molds (in combination with the Bruker Biotyper library), studies have shown that MALDI-TOF MS could be reliable for the identification of any species of clinically relevant mold, including the "cryptic" Aspergillus species (14,15,19). However, these studies applied a lower species cutoff (mostly 1.7 [value accepted for genus assignment]) to increase identification rates.…”