To the Editor I applaud the authors of the study “Performance of the First Step of the 2-Step Dermoscopy Algorithm”1 and the accompanying editorial2 for handling the subject in a balanced way. It is difficult to establish and promote a method and still remain critical. When we tested the first step of the 2-step algorithm, the specificity was lower and the sensitivity was higher than in the study by Chen et al.3 An explanation is that we used a different threshold. We evaluated the lesions literally “by the book.” Even if there were obvious clues to seborrheic keratosis on one side but reticular lines (pigment network) on the other, we classified the lesion as melanocytic because, according to the algorithm, the presence of reticular lines trumps everything else. We followed the rules of the first step strictly. Experts, however, usually handle this in a different way. If they know that the lesion is a seborrheic keratosis, they probably disregard a network in small parts of the lesion or call it something else, like pseudonetwork. Experts sometimes violate their own rules, and exactly this makes them experts. They know when it is appropriate to violate the rules. I proposed pattern analysis instead of the 2-step method because no algorithm can ever capture the complexity of the real world sufficiently. Algorithms are good to start with, but if you want to become an expert, you need to create your own algorithm. The 2-step method and the chaos and clues algorithm4 are 2 of many possible interpretations of pattern analysis.
Kittler H. The 2-Step Method and the Recognition Process in Dermoscopy. JAMA Dermatol. 2015;151(9):1037–1038. doi:10.1001/jamadermatol.2015.2000
Customize your JAMA Network experience by selecting one or more topics from the list below.
Create a personal account or sign in to: