Abstract
Background: Invasive filamentous fungal infections carry substantial mortality among immunocompromised hosts, and diagnostic delay associated with morphology-based mould identification is a recognised contributor to poor outcomes. The Mycelium Transfer (MyT) method, implemented on the Bruker MALDI Biotyper® platform, allows direct mass spectrometric profiling of leading-edge mycelium sampled from routine solid agar without solvent extraction. We conducted a single-centre analytical validation of an optimised, triplicate-spot MyT workflow as a standalone identification platform at a Canadian tertiary care academic centre. Methods: Fifty-six archived, pre-characterised clinical isolates spanning five fungal categories—Aspergillus species (n = 21), other hyaline non-Aspergillus moulds (n = 14), Mucorales (n = 8), dermatophytes (n = 8), and dematiaceous fungi (n = 5). These isolates were selected to represent our local isolate population and tested by MyT on a Bruker Microflex LT/SH system, with acquisition and scoring performed prospectively by technologists blinded to the archived reference identification. Spectra were interrogated against the MBT Myco 4.1 Filamentous Fungi Library. Each isolate was spotted in triplicate; a final identification required ≥2/3 concordant spots at a score ≥ 1.8, consistent with CLSI M58 guidance. A parallel formic acid–ethanol extraction was performed for a subset of 11 isolates as an exploratory, within-study comparator. Primary culture media were Brain Heart Infusion (BHI) agar and Sabouraud Dextrose Agar with chloramphenicol and gentamicin (SAB-CG). Agreement was assessed against CLSI M52 criteria, with an a priori target of ≥90% overall concordance. Results: Applying the triplicate-spot consensus rule at the ≥1.8 threshold, species- or complex-level identification was achieved for 43/56 isolates (76.8%; 95% CI 64.2–85.9%), and overall acceptable identification (species/complex plus acceptable genus-level calls) for 50/56 isolates (89.3%; 95% CI 78.5–95.0%). No misidentifications were observed. Aspergillus spp. and dermatophytes each achieved 100% overall identification; performance was lowest for dematiaceous fungi (2/5, 40.0%). In a post hoc analysis restricted to the categories in which discordant single-spot results occurred (other hyaline non-Aspergillus moulds and Mucorales, combined n = 22), triplicate consensus corrected three isolates in which a single high-scoring spot would have produced a false identification, and enabled identification of four further isolates that a single spot would have missed, corresponding to an estimated single-spot equivalent identification rate of 12/22 (54.5%) versus 19/22 (86.4%) with triplicate consensus, which is an approximately 32 percentage point absolute difference restricted to this 22-isolate subgroup and not an effect demonstrated across the complete 56-isolate panel. Growth sufficient for MyT sampling at 72 h was numerically higher on BHI agar than SAB-CG for hyaline moulds at 72 h (88.6% vs. 77.1%), although this difference was not statistically significant (McNemar’s test, p = 0.29). In a smaller exploratory subset of 11 isolates tested in parallel by both methods, MyT showed similar performance to formic acid–ethanol extraction (equivalent or higher scores in 8/11 isolates tested in parallel); this was not a formally powered non-inferiority comparison. Conclusions: A triplicate-spot MyT workflow at a ≥1.8 score threshold achieved 89.3% overall concordance, with reference identification and zero misidentifications across a 56-isolate panel, approaching the CLSI M52 ≥ 90% benchmark overall and meeting it for Aspergillus species and dermatophytes specifically. To our knowledge, this is among the first Canadian evaluations of the MyT method for routine clinical mould identification. Dematiaceous fungi and several individual species within other categories remained underpowered and of limited performance, reflecting both small subgroup sizes and known gaps in current spectral library coverage. These findings support MyT as a practical frontline identification method for Aspergillus, Mucorales, and dermatophytes within existing laboratory infrastructure, with chemical extraction or molecular methods retained as adjuncts for isolates that do not achieve threshold identification; broader implementation and clinical outcome benefits remain to be evaluated prospectively.