Performance of an AI-based automated feces analyzer compared with the formalin-ether concentration technique for intestinal parasite detection: a diagnostic accuracy study

다국어 초록 (Multilingual Abstract)

Background: To overcome the labor-intensive and technician-dependent nature of the formalin–ether concentration technique (FECT), artificial intelligence (AI)-based automated feces analyzers have been developed for detecting fecal parasites. This study compared the clinical performance of the KU-F40 AI-based automated system with that of FECT in a routine diagnostic environment.


Methods: In total, 1,011 fecal samples were prospectively collected and analyzed concurrently via both FECT and the KU-F40 system. The diagnostic performance of the KU-F40 system was evaluated using FECT as the reference method. The clinical performance parameters included accuracy, Cohen’s kappa coefficient, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).


Results: The KU-F40 system exhibited an accuracy of 98.5% and Cohen’s kappa of 0.608 relative to FECT. The sensitivity and specificity of the KU-F40 system were 57.1% (95% confidence interval [CI], 34.0–78.2) and 99.4% (95% CI, 98.7–99.8), respectively. When adjusted for an institutional prevalence of 0.45%, its PPV and NPV were 29.9% (95% CI, 15.0–50.7) and 99.8% (95% CI, 99.7–99.9), respectively. Among the 27 parasite-positive samples identified using at least one method, the methods achieved concordant species identification for six samples (22.2%). The KU-F40 system offers superior workflow efficiency by reducing labor time and hazard exposure.


Conclusion: The KU-F40 system demonstrated high specificity and negative predictive value in a low-prevalence setting; however, its limited sensitivity indicates it should not replace conventional examination when clinical suspicion is high. Despite these limitations, it may serve as a useful adjunctive tool that improves laboratory workflow efficiency and minimizes exposure to hazardous reagents.