Clonorchis sinensis and Metagonimus spp. are prevalent parasites in Korea, and accurate diagnosis is essential because treatment dosages differ between infections. However, their eggs are morphologically similar under light microscopy, making differentiation difficult and dependent on examiner expertise. To address this limitation, we evaluated an artificial intelligence (AI)-based automated microscope solution for the simultaneous detection and discrimination of both parasites. Microscopic images from 170 stool samples were analyzed using an AI system based on You Only Look Once version 5. The annotated dataset comprised 9455 egg images (6494 C. sinensis and 2961 Metagonimus spp.), randomly divided at the slide/patient level into training (6862), validation (1301), and test (1292) sets. Diagnostic performance was evaluated using mean average precision, confusion matrix analysis, and correlation with conventional microscopy. The model achieved a classification accuracy of up to 97.8%. C. sinensis showed higher recall and F1 scores, whereas Metagonimus spp. showed higher precision and specificity. Species identification showed complete concordance with conventional microscopy, and egg quantification was strongly correlated. These results indicate that the proposed AI system may serve as a supportive diagnostic tool comparable to conventional microscopy.
Shin et al. (Wed,) studied this question.