Artificial Intelligence (AI)-assisted readout method for the evaluation of skin prick automated test results.

The skin prick test (SPT) is the gold standard for diagnosing allergic sensitization to aeroallergies. The Skin Prick Automated Test (SPAT) device has previously demonstrated reduced variability and more consistent test results compared to manual SPT. The current study aims to develop and validate an artificial intelligence (AI) assisted readout method to support physicians in interpreting skin reactions following SPAT. To train the AI algorithm, 7812 wheals (651 patients) are manually labeled. To validate the AI measurement, the longest wheal diameter of 2604 wheals (217 patients) is measured by the treating physician and compared to the AI measurement. In addition, AI-assisted readout is validated on a separate test cohort of 95 patients (1140 wheals). We demonstrate that the AI measurements of the longest wheal diameter exhibit a strong correlation with the physician's measurements. The AI algorithm shows a specificity of 98·4% and sensitivity of 85·0% in determining positive or negative test results in the validation cohort. In the test cohort, physicians adjust 5·8% of AI measurements, leading to a change in the test interpretation for only 0·5% of cases. AI-assisted readout significantly reduces inter- and intra-observer variability and readout time compared to manual physician measurements. Altogether, the AI-assisted readout method demonstrates high accuracy, with minimal misclassification of test results. Adding AI to SPAT further improves standardization across the SPT process, significantly reducing observer variability and time to readout.