Comprehensive Summary
In this study, Sven F Seys et al looks at how Artificial Intelligence compares to physicians in interpreting the Skin Prick Automated Test (SPAT) used to diagnose aeroallergies. Several parameters were used to compare AI and physicians, including wheal diameter, area of inflammation, and speed of analysis. A total of 868 patients were tested and the correlation between AI and physicians was strong (Pearson r = 0.81, p<0.0001), with only a 0.2 mm median difference between their wheal diameter measurements. Additionally, the AI-assisted readout was around 3.7 times faster than the manual measurement done by a physician. These results show that the AI system reduces the time needed to interpret allergy tests without sacrificing accuracy.
Outcomes and Implications
Today, around 30% of the world’s population deal with airborne allergic reactions, whether it be from pollen, dust, etc. The Skin Prick Automated Test is a widely used method to administer allergen tests in a way that the results do not depend on the operator. Still, interpreting the results can vary from physician to physician or even between clinical centers. The AI-assisted SPAT readout offers a solution to this problem with a 98.4% specificity and 85% sensitivity in inflammation measurement, according to the study. This makes AI a promising method to administer standardized and reliable allergy diagnoses in clinical settings.