Real-world diagnostic performance of artificial intelligence in skin cancer screening.
Artificial intelligence detected skin cancer with higher sensitivity than face-to-face dermatologists in real-world screening.
Real-world diagnostic performance of artificial intelligence in skin cancer screening.
The use of artificial intelligence (artificial intelligence) in many aspects of society is expanding rapidly.
To assess the real-world diagnostic performance, at the histopathological level, of artificial intelligence employed in the urgent skin cancer screening pathway.
Data were retrieved from medical records.
Artificial intelligence caught more true skin cancers than dermatologists did face to face
The positive predictive value of the artificial intelligence algorithm was lower, at 46.5% (95% CI 44.6-48.5).
This compared with 62.1% (95% CI 53.9-67.7) in dermatologists.
Artificial intelligence correctly identified the precise diagnosis 28.6% of the time, compared with dermatologists 61.6% of the time ( P < 0.001).
The correct tumour/lesion type was identified by artificial intelligence 51.4% of the time and by dermatologists 75.5% ( P < 0.001).
Artificial intelligence has high sensitivity in the detection of skin cancer.
However, the diagnostic accuracy of the information provided by artificial intelligence to clinicians is low and could be further optimized to reduce the risk of automation bias.
Furthermore, this study suggests the removal of human validation of artificial intelligence decisions may be premature due to the potential for missed cancer diagnoses.