About 20 percent of the population of high-income countries are troubled by contact eczema, a disease often associated with exposure to chemicals in the working environment. There are two types of contact eczema, each with its own cause: allergic contact eczema, which is caused by an allergic reaction; and non-allergic irritant eczema, which is caused by chemical agents or physical factors.
Since the two types require different treatments, it is important that the correct diagnosis is made. This can prove difficult for dermatologists, as the diseases present similar clinical symptoms. Diagnoses are normally based on the results of a patch test, which are often difficult to interpret and can sometimes give false positive or false negative outcomes.
Gene expression differentiates between reactions
In this present study, researchers compared patch tests from 85 patients with contact eczema and healthy skin samples to examine the gene expression in the skin resulting from exposure to different allergens and irritants.
The findings, supporting the further development of an alternative to today’s diagnostic patch tests, were published in the journal Proceedings of the National Academy of Sciences (PNAS) on December 14th.
The study was done in collaboration between the Karolinska Institutet in Sweden, Institute of Biomedicine, University of Eastern Finland, Faculty of Medicine and Life Sciences, University of Tampere, Institute for Molecular Medicine Finland (FIMM) at the University of Helsinki, Medical University of Vienna in Austria, HUSLAB, Helsinki University Hospital and Finnish Institute of Occupational Health.
Using a biomarker discovery method entitled GARBO developed by the first author Vittorio Fortino and colleagues, which combines different AI techniques (Genetic Algorithm, Fuzzy Logic and Random Forest classifiers), the researchers identified sets of two or three genes that together could distinguish irritant from allergic skin reactions.
The results were confirmed in clinical samples using an image-based machine-learning technique developed at FIMM with tools provided by the FIMM spinoff company Aiforia Technologies. The results were replicable in an independent group of patients, and in external datasets. The external datasets included patients who were exposed to different substances than those of the first group, which laid the foundation for the new biomarkers.
Potential for new diagnostic method
“Our results show that there is considerable potential for the development of a new diagnostic method based on these biomarkers,” says corresponding author Nanna Fyhrquist, researcher and group leader at the Institute of Environmental Medicine, Karolinska Institutet.
“The next step in the project entails a more extensive clinical validation of the markers and technical optimisation of the method in order to attain sufficient cost-effectiveness and speed to clinical purposes.”
“The results show the power of deep learning-based algorithms aimed at evaluating clinical diagnostic samples for digital precision diagnostics,” says Nina Linder, the lead investigator of the study at FIMM, regarding the development of the image-based machine learning algorithms.
The study was financed by the Finnish Work Environment Fund and the Swedish Research Council for Health, Working Life and Welfare (FORTE).
Original publication: Machine learning driven biomarker discovery for the discrimination between allergic and irritant contact dermatitis. Vittorio Fortino, Lukas Wisgrill, Paulina Werner, Sari Suomela, Nina Linder, Erja Lehto, Alina Suomalainen, Veer Marwah, Mia Kero, Maria Pesonen, Johan Lundin, Antti Lauerma, Kristiina Aalto-Korte, Dario Greco, Harri Alenius, Nanna Fyhrquist. Proceedings of the National Academy of Sciences (PNAS), online December 14, 2020, doi: 10.1073/pnas.2009192117
For more information, please contact:
Nanna Fyhrquist, researcher
Institute of Environmental Medicine, Karolinska Institutet
Phone: +46 (0)73-654 1746
Nina Linder, MD, PhD, Adj Prof, Senior Researcher
Institute for Molecular Medicine, Finland - FIMM, University of Helsinki
Phone: +358 44 5555407