JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI) aims to create programs that reproduce human cognition and processes involved in interpreting complex data. Dermatology relies on morphological features and is ideal for applying AI image recognition for assisted diagnosis. Tibot® is an AI application that analyses skin conditions and works on the principle of convolutional neural network. Appropriate research analyzing the accuracy of such applications is necessary.

Objective:

To analyze the predictability of Tibot® AI application in the identification of dermatological diseases as compared to a dermatologist.

Methods:

This is a cross-sectional study. After taking informed consent, photographs of lesions of patients having different skin conditions were uploaded to the application. In every condition, AI predicted three diagnoses based on probability, and these were compared with that by a dermatologist. Accuracy, sensitivity, specificity and positive predictive value were used to assess the application's performance. Chi-square test was employed to contrast categorical variables. P<.05 was considered statistically significant.

Results:

Six hundred patients were included. Clinical conditions included alopecia, acne, eczema, immunological disorders, pigmentary disorders, psoriasis, infestation, tumours and infections. In the anticipated top three diagnoses, the application’s mean prediction accuracy was 96.1%, while for the exact diagnosis, it was 80.6%. Prediction accuracy for alopecia, eczema and tumours was 100%. The sensitivity and specificity of the application was 97% and 98%, respectively. There is a statistically significant association between clinical and AI-predicted diagnoses in all the conditions (P<.001).

Conclusions:

AI application has shown promising results in diagnosing various dermatological conditions, and there is great potential for practical applicability.