medical procedure where more patient medical data are taken into account. Moreover, only 15% of the works use direct explainability methods, meaning that the efforts for solving the trustworthiness issue of deep learning models were limited. In this context, this work reveals the gap between the...