According to ISO 10993-1:2018, the skin sensitization potential of all medical devices must be evaluated, and for this endpoint ISO 10993-10:2010 recommends the use of in vivo assays. The goal of the present study was to determine if the in vitro SENS-IS assay could be a suitable alternative to the current in vivo assays. The SENS-IS assay uses the Episkin Large and SkinEthic RHE reconstructed human epidermis models to evaluate marker genes. In our study, the SENS-IS assay correctly identified 13 sensitizers spiked in a non-polar solvent. In a subsequent analysis six medical device silicone samples previously impregnated with sensitizers were extracted with polar and non-polar solvents. The SENS-IS assay correctly identified five of these extracts, while a sixth extract, which contained the weak sensitizer phenyl benzoate, was classified as negative. However, when this extract was concentrated, or a longer exposure time was used, the assay was able to detect phenyl benzoate. The SENS-IS assay was transferred to a naïve laboratory which correctly identified sensitizers in six blinded silicone samples, including the one containing phenyl benzoate. In light of these results, we conclude that the SENS-IS assay is able to correctly identify the presence of sensitizers in medical devices extracts.