Clinical performance evaluation of the KU-F50 fully automated feces analyzer with AI-based recognition of intestinal parasites

Abstract Intestinal parasitic infections represent a significant global public health issue, with accurate identification of parasites in stool examinations being a critical diagnostic step. This study aims to evaluate the diagnostic performance of the KU-F50 fully automated feces analyzer, which utilizes artificial intelligence (AI) technology, in recognizing intestinal parasites and to assess its clinical application value. A retrospective diagnostic comparative study design was adopted, using manual interpretation as the gold standard. A total of 1290 stool samples from patients at the hospital between December 10, 2025, and January 15, 2026, were included. By comparing AI recognition results with manual interpretation, we analyzed the AI’s sensitivity, specificity, overall agreement rate, and Kappa consistency for intestinal parasite identification, and calculated the false positive and false negative rates. The results showed that the instrument detected 10 types of intestinal parasites. The true positive rate (sensitivity) for AI recognition was 99.03% (95% confidence interval: [96.63%, 99.87%]), the true negative rate (specificity) was 96.31% (95% confidence interval: [95.15%, 97.26%]), the overall agreement rate reached 96.74% (95% confidence interval: [95.71%, 97.57%]), and the Kappa value was 0.8871, indicating good consistency. The false positive rate was 3.69% (40/1084), and the false negative rate was 0.97% (2/206). Analysis of the causes of AI misidentification revealed that the 40 false positives were primarily due to interference from impurities such as pollen, fat droplets, plant cells, and plant fibers, while the 2 false negatives were mainly caused by excessively deep image background interference resulting from an excessive fecal sample volume. All misidentified results were corrected upon manual review. The AI recognition function of the KU-F50 fully automated feces analyzer demonstrates high sensitivity and specificity in detecting intestinal parasites. It can serve as an auxiliary tool for screening intestinal parasitic infections and, following manual review, can further enhance diagnostic accuracy, indicating significant clinical application value.