Currently, the detection of mycotoxins deriving from fungal/microbial contamination is verified only in food and herbal medicines, for which a systematic surveillance over raw materials and final products is highly regulated. However, considering their adverse effects on human and animal health, but also the associated economic losses, mycotoxin detection could be expanded to non-sterile Active Pharmaceutical Ingredients (APIs). Actually, according to both European and US Pharmacopoeia, microorganism contamination is verified in the API, while only the presence of bacterial (not fungal) endotoxins is investigated. No testing is hence mandatory for determining the possible presence of mycotoxins in case of mould contamination. In the present study, two methods have been proposed specifically for the analysis of aflatoxins and ochratoxins in iopamidol, an iodinated API formulated as an injectable contrast agent. First, an optimized extraction procedure to avoid matrix interferences was developed and validated. The ELISA assay was proposed to be used for screening purposes. Two commercially available kits were adapted, demonstrating their effectiveness to detect selected mycotoxins at the acceptable levels of exposure. The UPLC-MS/MS method was, instead, proposed as the preferred method for detection. In fact, it was demonstrated to have high selectivity, accuracy and sensitivity, reaching very low limits of detection. It was also applied to confirm the complete removal of spiked mycotoxins in iopamidol after charcoal treatment. To the best of our knowledge, this paper represents the first example of investigation of mycotoxins in commercial APIs and it could pave the way in developing specific methods.
A bundled yarn cuboid device, suitable for chromatographic separations is described and discussed. The acrylic yarn-based stationary phase was assembled within a cuboid frame by a simple pull and trim methods, and this was further integrated with appropriate flow distribution and collections plates. The 24 mL (2 cm × 2 cm × 6 cm) device discussed in this paper had a high flow rate range and could be operated at relatively low backpressure. In the operating range examined in this study, the acrylic yarn media showed no signs of compression or compaction. Compared to conventional resin-based separation media, the height equivalent of a theoretical plate (HETP) for the acrylic bundled yarn cuboid chromatography device was found to be relatively insensitive to the mobile flow rate. At a flow rate of 20 mL/min, the backpressure was <0.2 MPa. The device was found to be suitable for carrying out hydrophobic interaction based bind and elute separations of model proteins. The feasibility of purifying monoclonal antibody and human IgG using the acrylic bundled yarn cuboid chromatography device was demonstrated. The acrylic-bound proteins eluted in less than one column volume due to the excellent hydrodynamic properties of the media which resulted in minimal solute dispersion. The proposed bundled yarn cuboid chromatography device could be developed further for applications such as fast, efficient and scalable purification of proteins and other biologics.
Gas chromatography–ion mobility spectrometry (GC–IMS) produces information-rich two-dimensional urinary volatile fingerprints. We evaluated a full-map workflow that uses the standardized separation map for machine-learning analysis and subsequently localizes signals for follow-up, without manual peak selection at model input. The development cohort included 46 biopsy-confirmed prostate cancer samples and 43 male control samples. Reactant ion peak-relative alignment and common-grid preprocessing generated 512 × 256 maps. Fold-wise Pareto scaling and principal component analysis reduced 131,072 intensities to 20 scores for classifier comparison. Random forest (RF) had the highest pooled out-of-fold area under the receiver operating characteristic curve (ROC–AUC) among six classifiers sharing this input: 0.835 (95% confidence interval [CI], 0.743–0.915); partial least squares discriminant analysis was evaluated separately on the Pareto-scaled full map. Applied without refitting to 39 prostate cancer and 17 prostate cancer-negative samples collected in 2022, the final RF pipeline achieved ROC–AUC 0.723 (95% CI, 0.551–0.879), sensitivity 0.949, and specificity 0.529. Temporal ROC–AUCs for the six comparator algorithms ranged from 0.427 to 0.646. Post hoc spatial localization defined 38 coordinate-based regions of interest; development-only RF models using all 38 regions or a fixed five-region subset achieved ROC–AUCs of 0.804 and 0.780. The workflow combines full-map input, classifier comparison, temporal evaluation, and coordinate-level signal localization. It provides a reproducible basis for investigating urinary volatile fingerprints and testing disease specificity in larger, clinically characterized urological cohorts.
Despite regulatory bans on the production of common per- and polyfluoroalkyl substances (PFAS), these compounds remain prevalent in environmental matrices such as soil and water. Consequently, PFAS are everywhere, including in process industry streams. Due to their adverse effects on human health, controlling further dissemination of PFAS is essential.One of the most commonly employed strategies for removing PFAS from waste streams is granular activated carbon (GAC). Therefore, precise quantification of PFAS adsorption on activated carbon post-application is essential. In this paper, the method development of a robust and reliable method for the quantification of PFAS loading on activated carbon is presented. During an optimization phase, suitable solvents were selected based on their recoveries. Solid-phase extraction (SPE) was applied prior to liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis. Method validation included assessment of detection limits, trueness, linearity, uncertainty, and intermediate precision.The best solvent combination for extraction was acetonitrile (ACN) followed by hexane. After extraction, a wash step with methanol and alkaline methanol was necessary. Validation results indicated reduced performance for long-chain PFAS compounds (> C11). The average limit of detection (LOQ), bias, intermediate precision (CVR) and uncertainty across all the molecules were 0.57 ± 0.45 µg/kg, 20% ± 43%, 22% ± 17% and 80% ± 74%, respectively. During the validation phase, differences in extraction yield were also observed between the various types of carbon.
