Assessment of the molecular identification algorithm and its impact on antifungal susceptibilities against clinical Fusarium isolates: a multicentre study in Taiwan, 2011–2023

Abstract Objectives Invasive fusariosis is rising and poses challenges due to species complexity and antifungal resistance. In vitro susceptibilities of new antifungals against Fusarium isolates are seldomly evaluated in Asia. This study aimed to evaluate a two-step molecular identification algorithm and to perform in vitro antifungal susceptibility with correlation of species and susceptibility patterns. Methods Fusarium clinical isolates collected at three hospitals in Taiwan (2011–2023) were identified to species level using sequential ITS and TEF1α sequencing (step I), followed by RBP2 sequencing (step II) for inconclusive isolates. Minimum effective/inhibitory concentrations (MECs/MICs) of manogepix, olorofim, amphotericin B and voriconazole were determined by EUCAST method (E.Def 9.4). Results Of 103 isolates (37 blood and 66 cornea isolates) evaluated, the two-step algorithm achieved 90% to species level. Fusarium solani species complex (FSSC) was predominant, especially in blood isolates (86.5% versus 65.2% in cornea isolates; P = 0.02). The rest 28 isolates belonged to 12 species within six species complexes (SCs). Manogepix exhibited potent activity against all isolates (MEC ≤0.015 mg/L), while olorofim activities varied by SCs, with MIC ≤0.25 mg/L against Fusarium fujikuroi SC. FSSC displayed higher voriconazole and amphotericin B MICs compared with other SCs, with Neocosmospora keratoplastica displaying a highest amphotericin B modal MIC of 4 mg/L. Four major Neocosmospora species showed voriconazole MIC ≥16 mg/L. Conclusions Our findings indicated the two-step molecular algorithm accurately identifies Fusarium to species level. Further, we underscored the significance of considering both Fusarium SCs and species for predicting antifungal susceptibility, particularly to olorofim and amphotericin B.