Combining artificial intelligence and human expertise for more accurate dermoscopic melanoma diagnosis: A 2-session retrospective reader study

To the Editor: Convolutional neural networks (CNNs) outperform dermatologists in classifying melanomas and nevi based on dermoscopic images.1,2 However, dermatologists consider more aspects than a dermoscopic image in clinical decision-making. Under less artificial conditions, dermatologists and CNNs show comparable performance.3 Recent findings emphasize the need for a collaborative utilization of CNNs alongside clinical expertise.4