Winkler JK, Fink C, Toberer F, et al. - In this cross-sectional study of 130 melanocytic skin lesions (107 benign nevi and 23 melanomas) carried out from August 1, 2018, to November 30, 2018, experts examined the correlation between gentian violet surgical skin markings in dermoscopic images and the diagnostic performance of a convolutional neural networks (CNNs) certified for use as a medical device in the European market. Heat maps generated by vanilla gradient descent backpropagation designated that the blue markings were related to the heightened false-positive rate. The findings imply that by progressing the melanoma likelihood scores and subsequently the false-positive rate, skin markings considerably intervened with CNN’s correct diagnosis of nevi. Moreover, the prevalence of skin markings could have provoked the CNN’s relationship of markings with a melanoma diagnosis in melanoma training images. Consequently, these findings propose that skin marking should be circumvented in dermoscopic images designed for analysis by CNN.