The machine learning model showed approximately 40% improvement in diagnostic accuracy compared to the traditional model.
Standard oral food challenges, skin prick tests, and allergen-specific IgE measurements
milwaukee, February 14, 2026 /PRNewswire/ — Both machine learning and deep learning AI models show significant improvements over existing clinical standards for food allergy diagnosis, according to new research to be presented at the 2026 AAAAI Annual Meeting.
“Current standard of care for food allergy diagnosis relies on skin prick tests, allergen-specific IgE and oral food challenges when results are inconclusive. The DL model further improved diagnostic performance compared to ML methods, with a 10-15% improvement in area under the curve. ML/DL has the potential to enhance diagnostic methods for food allergy, outperforming current strategies and improving standard of care,” said lead author Kirsch of Howard University. said STEM scholar Mackenzie J. Williams.
In this study, researchers trained machine learning (ML) and deep learning (DL) convolutional neural networks (CNNs) on skin prick test (SPT) measurements, allergen-specific IgE (sIgE), and serum component proteins including peanut (PN)-IgE rAra h 1, 2, 3, and 6. PN-IgG4 rAra h 1, 2, 3, 6. Baseline was collected at the time of 146 peanut oral food challenges (OFC) as part of the IMPACT trial. Children aged 1 to 4 years participated in the study.
Within the study, algorithm performance showed strong predictive value for PN-sIgE Ara h2 and PN-IgE/IgG4 (sensitivity: 88.9; specificity: 84.5; positive predictive value (PPV): 89). According to the researchers, the ML model showed significant improvement over existing clinical standards, increasing diagnostic accuracy by approximately 40%. Using more advanced DL models improved diagnostic performance over ML methods, improving the area under the curve by 10–15%. As a result, the DL model was trained on standard-of-care tests and was able to significantly improve sensitivity and PPV while being comparable to diagnostic methods used in practice.
Researchers suggest that this improvement in the diagnostic performance of OFC biomarker discovery could be used to develop alternative diagnostic methods for food allergy that are more scalable and more efficient than standard OFC, skin prick tests, and allergen-specific IgE (sIgE) measurements.
For more information about food allergies, visit aaaai.org. Research presented at the 2026 AAAAI Annual Meeting in Philadelphia, Pennsylvania, February 27-March 2, will be published in an online appendix. Journal of Allergy and Clinical Immunology (JACI).
The American Academy of Allergy, Asthma, and Immunology (AAAAI) is a leading membership organization of more than 7,100 allergists, asthma specialists, clinical immunologists, and other professionals with a special interest in the research and treatment of allergic and immune diseases. Founded in 1943, AAAAI is the go-to resource for patients with allergies, asthma, and immunodeficiency disorders.
Source American Academy of Allergy, Asthma, and Immunology (AAAAI)
