“…To develop this algorithm, we followed different steps. First, in order to simplify the execution of the algorithm and lower the interobserver variability, the evaluation in each dermoscopy image was restricted to the presence or absence of melanoma‐specific structures with the highest interobserver agreement (Kappa coefficient ≥0.30) 14 . Those included: polychromy, defined as the presence of ≥3 colors (light brown, dark brown, black, blue/gray, red/pink, white); asymmetry of color and/or structures in one or two axes; and the presence of the following melanoma‐specific dermoscopic structures: atypical pigmented network, irregular globules, irregular blotches, streaks, negative pigment network, regression structures, blue‐white veil, shiny white streaks, milky red areas, pseudolacunae, rainbow pattern, and irregular vessels.…”