4.2.3. Homogeneity of Test Materials

A homogeneity study was carried out according to the procedure described by ISO guide 35:2006 [28] on randomly selected units: specifically, 16 units (for blank test materials) and 11 units (for contaminated test materials) of about 25 g were taken at systematic intervals from the filling sequence. Each unit of 25 g was divided into two aliquots and analyzed in duplicate under repeatability conditions. The analytical method used for homogeneity testing was the one described in this protocol, keeping the same ratio of test portion to extraction solvent.

Homogeneity was statistically evaluated according to ISO 13528:2015 [29] and F-test. The parameters considered for the homogeneity test were the analytical precision (within bottle standard deviation, *sw*–analytical SD) and the heterogeneity standard deviation (between bottle standard deviation, sb–heterogeneity SD). The F-test was used to determine whether the observed sb deviated significantly from the *sw*.

The heterogeneity (*sb*) was then compared to the target standard deviation (*s*). The *s* values were obtained using the truncated Horwitz equation corrected by Thompson—i.e., if the relative target standard deviation according to Horwitz was greater than 22%, it was truncated to 22%. The samples were considered to be adequately homogenous if sb ≤0.3 s [28,29]. Data were processed using the ProLab software (ProLab Software—QuoData, Drezden—www.quodata.de).

All the test materials passed the homogeneity test and turned out to be appropriate for the collaborative study.

#### 4.2.4. Stability of Test Materials

Randomly selected units of the test materials were submitted to accelerated ageing at different temperatures (4 ◦C, 20 ◦C, and 60 ◦C) over a total period of 1.5 months according to the so-called isochronous short-term stability study [28–30]. A total of 26 bottles for each material were stored at −20 ◦C (reference temperature), then 2 random bottles per time were moved to the different temperatures after 0.25, 0.50, 1 and 1.5 month for a total of 24 bottles. All the units were analyzed at the end of month 1.5, under repeatability conditions, together with 2 reference samples which were kept at −20 ◦C over the whole period of the short-term stability study. The analytical method used for stability testing was the one described in this protocol. The stability study was performed on each contamination level per material.

Statistical results assessment was performed according to ISO guide 35:2006 using the *t-*test to test the regression for significance [28,29]. Specifically, the evaluation of data was carried out by performing a linear regression on the experimentally determined concentrations of each mycotoxin (mean values) versus time (days). For a stable material, it is expected that the intercept is equal to the reference value, whereas the slope does not differ significantly from zero.

The evaluation of data from the short-term stability study indicated that no significant trend was observed for the test samples at all temperature conditions (4 ◦C, 20 ◦C, and 60 ◦C) for the time span of the collaborative study. It was concluded that the three test materials were stable for at least 1.5 months following their preparation.

#### *4.3. Collaborative Study*

#### 4.3.1. Study Layout

Prior to the full validation study, laboratories had to participate in a pre-trial study to let them become familiarized with the correct execution of the method protocol and optimize the LC-MS/MS conditions for mycotoxin detection. To this scope, they received:


The full validation study was planned according to the requirements of the IUPAC/AOAC international harmonized protocol [22,31]. The main purpose of the collaborative study was to estimate the precision of the candidate method under repeatability and reproducibility conditions. The method accuracy was evaluated by spiking experiments. Fifteen laboratories were involved in the trial, representing a cross-section of academia, official control, and private laboratories.

Participants received the following materials:


The results had to be expressed in micrograms per kilogram (μg/kg). Each laboratory was free to use its own LC-MS/MS set-up, and the optimization of settings for the LC and MS/MS detection was left to the participants. Mycotoxin detection was requested to be performed in Selected Reaction Monitoring (SRM) in case of MS/MS analyzers or Parallel Reaction Monitoring (PRM) in case of MS/high-resolution mass spectrometry (HRMS) analyzers. The full chromatographic separation of 3-AcDON and 15-AcDON was mandatory, and the participants were asked to provide individual data for the two toxins. The participants were requested to make available chromatograms for samples and calibration standards and to provide the following details on the applied LC-MS/MS instrumentation and method settings: LC column characteristics, mobile phase composition and gradient elution, flow rate, injection volume, MS ion source, SRM or PRM ions.
