The inverse approach for early diagnosis of lentigo maligna
The inverse approach is a new dermatoscopic method for the early diagnosis of lentigo maligna. It originated from previous evidence demonstrating that the classic pattern analysis results in poor diagnostic accuracy when it comes to differentiate early lentigo maligna from pigmented actinic keratosis and solar lentigo. The inverse approach suggests that lentigo maligna should be diagnosed by exclusion of the other 2 tumors, based on the absence of prevalent dermatoscopic features. We tested the effect of the inverse approach on the diagnostic performance of 78 clinicians participating in a dermatoscopy masterclass (Dermoscopy Excellence). We found that it significantly improved all measures of diagnostic accuracy and especially the sensitivity for lentigo maligna. We also used a previously tested convolutional neural network for comparison purposes. More than 50% of human readers outperformed the convolutional neural network when using the inverse approach. The latter study was published in 2020 in the Journal of the American Academy of Dermatology. The current dataset describes the details of the inverse approach and can be used an an educative material.