Development and Evaluation of an Artificial Intelligence Model for the Diagnosis of Aeroallergies

Completed

• Conditions: Allergy
• Interventions: Device: Skin Prick Automated Test

• Registration Number: NCT05918354
• Lead Sponsor: Hippocreates

Brief Summary

To improve accuracy and take out human error, Hippo Dx previously developed S.P.A.T., an automated skin prick test. Now, Hippo Dx will develop and evaluate a S.P.A.T. artificial intelligence (AI) medical software to support physicians in the read out of the test results from an automated skin prick test.

In this study 700 images will be collected to develop the AI model, an additional 200 images will be collected to evaluate performance of the AI model. In a last independent cohort of 100 patients, performance of read-out by the physician will be compared with read out by the physician supported by the S.P.A.T. AI medical software.

The primary endpoint is the accuracy of the maximal wheal size measurement by S.P.A.T. AI versus the maximal wheal size measurement by the physician.

It is a prospective multicentric study that will include 1000 study participants from different countries.

Detailed Description

Not available

Study Timeline

• Study First Submitted: 2023-06-15
• Study First Submitted QC: 2023-06-23
• Study First Posted: 2023-06-26
• Study Start: 2023-07-10
• Primary Completion: 2024-09-11
• Study Completion: 2024-09-11
• Status Verified: 2024-10
• Last Update Submitted: 2024-10-14
• Last Update Posted: 2024-10-16

Recruitment & Eligibility

• Status: COMPLETED
• Sex: All
• Target Recruitment: 993

Inclusion Criteria

• Adults aged 18 years or above, consulting the outpatient clinic for an allergy diagnosis to aeroallergens will be included.

Exclusion Criteria

• Skin pathology like chronic or exuberant urticaria, dermographism, chronic dermatitis that needs daily treatment
• Use of antihistaminic medication < 7 days before the start of the study
• Use of tricyclic antidepressants (antihistamine activity) < 7 days before the start of the study
• Use of topical corticoids on the forearm < 7 days before the start of the study
• Use of Omalizumab < 6 months before the start of the study
• Pregnancy: there is a remote possibility of inducing a systemic allergic reaction that could induce uterine contractions or necessitate the use of epinephrine (thought to cause constriction of the umbilical artery)
• Incapacitated individuals
• Individuals who do not speak the local language (Dutch - French)
• Individuals who cannot read or write

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Arm && Interventions

Group	Intervention	Description
Allergy	Skin Prick Automated Test	Adults aged 18 years or above, consulting the outpatient clinic for an allergy diagnosis to aeroallergens with skin prick automated test device.

Primary Outcome Measures

Name	Time	Method
Equivalency of S.P.A.T. artificial intelligence versus physician	15 minutes after the diagnostic test	The equivalency of the maximal wheal diameter as measured by S.P.A.T. AI versus the maximal wheal diameter as assessed by the physician.

Secondary Outcome Measures

Name	Time	Method
Intra-reader variability of S.P.A.T. read out by physicians	15 minutes after the diagnostic test	Variation in maximal wheal size diameter of the same image by 3 independent physicians.
Precision of S.P.A.T. artificial intelligence versus physician	15 minutes after the diagnostic test	Likelihood that the S.P.A.T. AI maximal wheal diameter measurement lays within the range of the maximal wheal size measurement of 3 independent physicians.
Comparison of different methods (HEP index - maximal wheal diameter) of read out by S.P.A.T. artificial intelligence versus physician	15 minutes after the diagnostic test	Accuracy of HEP-index measurement by S.P.A.T. AI versus the maximal wheal diameter measurement by the physician.
Time impact of S.P.A.T. artificial intelligence supporting the physician versus physician	15 minutes after the diagnostic test	Total time required to perform read out of S.P.A.T. results by the physician supported by the S.P.A.T. AI software versus the physician alone.
Comparison of different methods (mean - maximal wheal diameter) of read out by S.P.A.T. artificial intelligence versus physician	15 minutes after the diagnostic test	Accuracy of mean wheal diameter by S.P.A.T. AI versus the maximal wheal diameter measurement by the physician.
Comparison of S.P.A.T. artificial intelligence supporting the physician versus physician	15 minutes after the diagnostic test	Inter-reader variability of read out of S.P.A.T. results by the physician supported by the S.P.A.T. AI software versus the physician alone.
Comparison of different methods (maximal wheal diameter) of read out by S.P.A.T. artificial intelligence versus physician	15 minutes after the diagnostic test	Accuracy of the maximal wheal diameter by S.P.A.T. AI versus the maximal wheal diameter measurement by the physician.
Comparison of different methods (wheal area - maximal wheal diameter) of read out by S.P.A.T. artificial intelligence versus physician	15 minutes after the diagnostic test	Accuracy of wheal area by S.P.A.T. AI versus the maximal wheal diameter measurement by the physician.

Trial Locations

Locations (6)

Clinique Universitaires Saint-Luc; 🇧🇪 Brussels, Belgium

Ziekenhuis Oost Limburg; 🇧🇪 Genk, Belgium

UZ Leuven; 🇧🇪 Leuven, Belgium

CHU de Liège; 🇧🇪 Liège, Belgium

GZA ziekenhuizen; 🇧🇪 Wilrijk, Belgium

Klinikum rechts der Isar; 🇩🇪 Munich, Germany