Symptoma’s Artificial Intelligence (AI) based approach appears to be a resource-efficient method for identifying patients with rare conditions, such as Pompe disease, as well as estimating its prevalence in a given region, according to a recently published study in Frontiers.
Rare diseases, defined as those with a prevalence lower than 200 people in the United States, represent a significant diagnosis challenge. They are often unknown to many physicians and are often mistaken for more common diseases, delaying diagnosis and, therefore, appropriate treatment.
Read more about Pompe disease diagnosis
“The biggest challenge is identifying suspicious patients and then routing them into the correct clinical lane for further diagnostic workup, especially in rare disease competence centers,” the authors wrote.
The authors aimed to identify patients with PD using electronic health records and AI and compare the results with other comparable screening studies. The AI developed by Symptoma was designed to identify patients likely to suffer from a rare disease. In cases with enough evidence to suggest that the patient might suffer from a rare disease, the patient is flagged for further evaluation by specialized physicians.
Researchers used data from over 300 patients from the University Hospital Salzburg, Salzburg, Austria. AI flagged 104 suspicious of having Pompe disease, equivalent to 1 suspicious patient for every 3300.
After further evaluation by general practitioners, the number of suspected patients was reduced to 22, of which 5 were classified as diagnoses, 10 as suspected, and the rest as reduced suspicion. Further evaluation by Pompe disease specialists classified patients into 2 probable, 5 definitive, and 12 inconclusive or possible. Based on the results, AI estimated the prevalence of Pompe disease among the Austrian population to be 1 in every 18 persons.
The authors highlighted the fact that by using an AI approach, physicians only had to personally examine 5 patients in order to detect 1 true case of Pompe disease.
“As such, we demonstrated both the efficiency of the approach and the potential of a scalable solution to the systematic identification of rare disease patients,” the authors wrote. “Thus, a similar implementation of this methodology should be encouraged to improve care for all rare disease patients.”
Reference
Lin S, Nateqi J, Weingartner-Ortner R, et al. An artificial intelligence-based approach for identifying rare disease patients using retrospective electronic health records applied for Pompe disease. Front Neurol. Published online April 21, 2023. doi:10.3389/fneur.2023.1108222
