Introduction:Skin Prick Tests (SPTs) are essential in diagnosing allergies, yet the optimal age for these tests in paediatric patients remains undetermined .

The study aimed to establish the ideal age for conducting SPTs in children with the assistance of AI.

Methods:

In this 30-year retrospective study, we included children aged 1 to 15 years who presented with allergic symptoms. Subjects underwent SPTs for common aeroallergens.Our deep neural network model for predicting SPT outcomes utilizes ReLU activation, L2 regularization, and dropout layers to prevent overfitting. It leverages the Adam optimizer to minimize binary cross-entropy loss and incorporates early stopping after 10 epochs of no validation loss improvement, ensuring the model's generalizability without overfitting (fig1). An age-analysis function was then applied to predict SPT positivity across ages, identifying the optimal age for initial testing. 

Results: A total of 4313 children were included, with an average age of 7.69 years (±3.5). Sex ratio was 1.44. A family history of atopy was observed in 59% of children. Rhinitis was the most common manifestation (76.4%), followed by asthma (63.9%). SPTs yielded positive results in 57.4% of cases, with positivity rates increasing with age (Fig. 2). Utilizing our model for deep learning, the optimal age for initial SPTs performance was determined to be 6 years (accuracy 95%).

Conclusion : Six years emerges as the optimal age for SPTs in pediatric patients. Our findings, supported by deep learning analysis and visual data representation, pave the way for more informed clinical decisions in allergy testing.