Assessment of skin toxicity in an in vitro reconstituted human epidermis model using deep learning.

Skin toxicity is a common safety concern associated with drugs that inhibit epidermal growth factor receptors as well as other targets involved in epidermal growth and differentiation. Recently, the use of a 3-D reconstructed human epidermis model enabled large-scale drug screening and demonstrated potential for predicting skin toxicity. Although a decrease in epidermal thickness was often observed when the 3-D reconstructed tissues were exposed to drugs causing skin toxicity, the thickness evaluation of epidermal layers from a pathologist was subjective and not easily reproducible nor scalable. In addition, the subtle thickness differences among tissues as well as the large number of samples tested made cross-study comparison difficult when a manual evaluation strategy was employed. In the present study, we used deep learning and image processing algorithms to measure the viable epidermal thickness from multiple studies and demonstrated that the measured thickness was not only significantly correlated with a pathologist's semi-quantitative evaluation but in close agreement with the quantitative measurement performed by pathologists. Moreover, a sensitivity of 0.8 and specificity of 0.75 was achieved when predicting the toxicity of 18 compounds with clinical observations with these epidermal thickness algorithms. This approach is fully automated, reproducible and highly scalable, which not only demonstrates reasonable accuracy in predicting skin toxicity but enables cross-study comparison and high-throughput compound screening.