Performance evaluation of ImmunoCAP® ISAC 112: a multi-site study

Abstract

Abstract Background: After the re-introduction of ImmunoCAP Methods: The study was carried out at 22 European and one South African site. Microarrays from different batches, eight specific IgE (sIgE) positive, three sIgE negative serum samples and a calibration sample were sent to participating laboratories where assays were performed according to the manufacturer’s instructions. Results: For both the negative and positive samples results were consistent between sites, with a very low frequency of false positive results (0.014%). A similar pattern of results for each of the samples was observed across the 23 sites. Homogeneity analysis of all measurements for each sample were well clustered, indicating good reproducibility; unsupervised hierarchical clustering and classification via random forests, showed clustering of identical samples independent of the assay site. Analysis of raw continuous data confirmed the good accuracy across the study sites; averaged standardized, site-specific ISU-E values fell close to the center of the distribution of measurements from all sites. After outlier filtering, variability across the whole study was estimated at 25.5%, with values of 22%, 27.1% and 22.4% for the ‘Low’, ‘Moderate to High’ and ‘Very High’ concentration categories, respectively. Conclusions: The study shows a robust performance of the ImmunoCAP