AI Integration in Cohort Research on Childhood Asthma and Allergies

Improving the treatment of childhood asthma and allergies and potentially even preventing them in the future: that's the ambition of Prof. Dr. Gerard Koppelman. Over 25 years ago, the national PIAMA birth cohort started in Groningen along with other UMCs. An initiative that, in 2024, is employing innovative methods by applying AI to a vast amount of valuable data.
Prof. dr. Gerard Koppelman

The PIAMA cohort is a national collaboration between various UMCs (Utrecht, Groningen, and formerly Rotterdam) and the RIVM. It is a birth cohort, meaning it collects data and bodily materials from birth onwards, and was established in 1996/1997. PIAMA stands for "Prevention and Incidence of Asthma and Mite Allergy," with the cohort aiming to analyse the development of asthma and allergies in children. PIAMA investigated the idea that allergy could be prevented by impermeable  matrass covers for house dust mites; this turned out not to be the case. 
The PIAMA continued its work in 2008, where adolescent and the now-adult participants returned for follow-up studies. Alongside questionnaires, researchers examined blood and nasal DNA and lung function of the participants, yielding a plethora of new data for research.

AI implementation: applying business AI to the medical sector

Gerard Koppelman has been involved with the PIAMA cohort and its development from 2008. He observed the rapid growth of data from the cohort and began contemplating a smarter and less time-consuming approach to data analysis. This solution came from a unexpected sector.

During one of his lectures, Gerard connected with an AI company based in Amsterdam. Initially active in the cultural sector, Rewire (formerly MiCompany) sought to pivot towards the medical sector within  their not-for-profit program ‘Doing Good’. The company chose to collaborate with the PIAMA cohort in Groningen, providing an opportunity to undertake groundbreaking work with a team of experts from various fields specializing in asthma and allergies in children.

Discovery of a new algorithm using AI to identify allergies in children

Collaboration across various expertise has yielded significant results. Several publications have already emerged, including one of the most recent ones in Nature Communications, which discusses a new algorithm capable of identifying allergies in children. The algorithm accomplishes this, based on DNA from nasal cells collected via a nasal swab. By examining just three data-points in the nasal DNA, the algorithm can determine if a child has an allergy. Further details on the research can be found here.
Additionally, in collaboration with the Groningen Research Institute for Asthma and COPD (GRIAC) a follow-up study on the image analysis of COPD biopsies has been initiated.

The importance of cohort studies

Gerard emphasizes that cohort studies are crucial not only for rare diseases but also for common (chronic) diseases such as asthma and allergies. With the help of AI, earlier diagnosis, improved treatment, and potentially even prevention of asthma and allergies are becoming increasingly feasible. Due to his passion for this field, Gerard provides training in AI and its application in the medical sector through the exquAIro foundation, which was set up by UMCG and AI experts from Rewire. Last March during the UMCG Cohort Symposium, he co-organized an insightful knowledge session with his Rewire colleague and PhD student Merlijn van Breugel on implementing AI in cohort studies, sharing his experiences with the PIAMA cohort and AI in one of the use cases. 

For more information on PIAMA cohort data, please visit the UMCG Research Data Catalogue