In-house validation and small-scale collaborative study to evaluate analytical performances of multimycotoxin screening methods based on liquid chromatography-high-resolution mass spectrometry: Case study on Fusarium toxins in wheat.

A strong trend toward using highly selective mass spectrometry technologies for screening of multiple mycotoxins has been observed in recent years. In the present study, the process of validation of a multimycotoxin screening method based on liquid chromatography-high-resolution mass spectrometry method is presented. The method was intended for the simultaneous screening of the major Fusarium toxins (deoxynivalenol, 3- and 15-acetyl deoxynivalenol, T-2 and HT-2 toxins, zearalenone, enniatins A, A1, B, and B1, and beauvericin) in wheat. The sample preparation protocol was based on a double extraction (methanol followed by acetonitrile/water mixture) and purification through solid-phase extraction C18 column. To provide insights for full exploitation of the potential of the double-stage high-resolution mass spectrometry detection, a full-scan acquisition event followed by a sequence of 5 fragmentation events (variable data-independent acquisition) was set for mycotoxin detection, the latter to be exploited for confirmatory purposes. Method analytical performances were evaluated through in-house validation and small-scale interlaboratory study, designed according to Commission Regulation 519/2014/EU, setting performance requirements for screening methods for mycotoxins. Screening target concentrations were close to European Union maximum permitted or indicative levels. The in-house validation provided the precision of the response under repeatability conditions and the intermediate precision (both resulting lower than 30%), the cutoff value, and the rate of false suspect results for negative (free of the mycotoxin of interest) samples, which resulted lower than 0.1% in all cases. The collaborative study provided reproducibility and laboratory independent cutoff values. Analysis of reference materials proved method trueness and suitability for screening of the major Fusarium mycotoxins in wheat. Finally, the applicability of the full-scan/variable data-independent acquisition detection approach was successfully tested on a set of naturally contaminated wheat samples, where 2 characteristic product ions could be detected for all identified mycotoxins even at levels in the low μg/kg range.