Evaluation of digital dermoscopy in a pigmented lesion clinic: clinician versus computer assessment of malignancy risk

Abstract

Digital dermoscopy systems employ computer-based algorithms to quantitate features of pigmented skin lesions (PSLs) and provide an assessment of malignancy risk. We evaluated interobserver concordance of PSL malignancy risk between a pigmented lesion specialist and an artificial neural network (ANN)-based automated digital dermoscopy system. While digital dermoscopy provides a reliable means of image capture, storage, and comparison of PSLs over time, the ANN algorithm requires further training and validation before the malignancy risk assessment feature can be widely used in clinical practice.