Use of an artificial intelligence (AI) decision support tool with a machine learning algorithm is associated with improved quality of skin disease images submitted for telemedicine, according to a study published online March 15 in JAMA Dermatology.

Kailas Vodrahalli, from Stanford University in California, and colleagues examined whether an AI decision support tool can improve the quality of images submitted for telemedicine by providing real-time feedback and explanations. The AI decision support tool was tested on 357 retrospectively collected telemedicine images from March 2020 to June 2021. Feasibility was subsequently assessed among 98 adults presenting for a skin condition and who were able to photograph their own skin with a smartphone in a clinical pilot study.

The researchers found that the machine learning algorithm effectively identified poor-quality images on retrospective telemedicine images, with a receiver operator characteristic area under the curve (ROC-AUC) of 0.78, and identified the reason for poor quality (blurry: ROC-AUC, 0.84; lighting issues: ROC-AUC, 0.70). Consistent performance was seen across age and sex. Patient use of the machine learning algorithm in the clinical pilot study was associated with improved image quality. Use of the AI algorithm was associated with a 68.0 percent reduction in the number of patients with a poor-quality image.

"Artificial intelligence support tools could assist patients in taking photos for telemedicine use and lead to a higher percentage of sufficient-quality photos submitted," the authors write.

Several authors disclosed financial ties to industry and reported holding related patents or having patents pending.

Abstract/Full Text (subscription or payment may be required)