Clinical Snippets

Reflectance confocal microscopy (RCM) used in concert with dermoscopy provides a noninvasive option for dermatology diagnostics with good specificity and sensitivity. However, quality assurance during image acquisition is critical. Kose et  al. applied machine-learning techniques to objectively and accurately (82% sensitivity and 93% specificity) identify and quantify areas that were uninformative within a dataset of 117 RCM mosaics. As the location of the uninformative areas is important for diagnosis, a combination of the visual ov erlay of the diagnostically uninformative areas with the percentage of such areas within the lesion across the mosaic may indicate image quality at the time of image acquisition in the clinic, precluding patient call-backs for reimaging.
Source: Journal of Investigative Dermatology - Category: Dermatology Tags: Clinical Snippets Source Type: research