A Point-of-Care, Real Time Artificial Intelligence System to Support Clinician Diagnosis of a Wide Range of Skin Diseases.

Dermatological diagnosis remains challenging for non-specialists as the morphologies of primary skins lesions widely varies from patient to patient. Although previous studies have used AI to classify lesions as benign or malignant, there haven't been extensive studies examining the use of AI on identifying and categorizing a primary skin lesion. In order to improve the diagnostic accuracy of primary care physicians at this task, we introduced a visual schematic system that improved their accuracies from 36% to 68% (p < 0.001). Additionally, an AI system was tested on the same image set and achieved a similar accuracy of 68%. When the AI's top prediction was broadened to its top 3 most likely predictions, its accuracy improved to 80%. The AI was subsequently tested on a larger set of 222 heterogeneous images of varying Fitzpatrick skin types and achieved an overall accuracy of 70% in the Fitzpatrick I-III skin type group and 68% in the Fitzpatrick IV-VI skin type group (p=0.79). Augmented intelligence is a powerful tool to assist physicians in the diagnosis of skin lesions while still requiring the user to critically consider other possible diagnoses.