An international team of scientists from Sweden and Finland has developed an artificial intelligence (AI) tool to detect parasitic infections in patients’ stool samples. The study was published in the scientific journal magazine PLOS Neglected Tropical Diseases (PLOS NTD).
Currently, the job of searching for parasitic worm eggs in stool samples falls to highly trained laboratory technicians. However, people who live in places where there are no laboratories nearby or who do not have money for such tests are often not tested.
Researchers trained a neural network on 1.3 thousand stool samples from Kenyan children. The biomaterials were scanned using an electron microscope and then uploaded to a database to train the algorithm.
Experts focused on three main types of intestinal parasites: hookworms, roundworms, and whipworms. Infections were diagnosed by detecting eggs in the stool.
The accuracy of the AI ranged from 76% to 96%, depending on the egg type. It turns out that the neural network does not tend to make false positive diagnoses, producing only 1-2% of such errors.
According to the team, the test results showed that the system is highly reliable and ready for practical use.
previous doctors described A rare case of eye infection caused by a parasitic worm in Congo.