Diagnosing helminth infections in a large reference laboratory in the United States: a 6-month pre- and post-implementation analysis of AI-augmented screening of concentrated fecal wet mounts.

内容摘要

Helminth infections are uncommon in the United States, yet accurate detection remains critical. Conventional ova-and-parasite examinations are labor-intensive and prone to error in low-prevalence settings. We evaluated the impact of implementing an artificial intelligence (AI)-based wet mount screening platform (WM-AI) in a high-volume reference laboratory. Detection rates, turnaround times, and specimen characteristics were compared across two 6-month periods before and after WM-AI implementation. Post-implementation, helminth detection increased from 29 to 104 patients, and client-level positivity rose from 4.5% to 12.3% (P < 0.0001). Enterobius vermicularis showed a fivefold increase in positivity (0.03% vs 0.15%, P < 0.0001). Median turnaround time improved from 5.2 to 4.0 days (P < 0.0001). Most infections were identified in the first or only specimen submitted, and over half of positive specimens contained &#x2264;5 eggs or larvae, highlighting the challenge of detecting low-intensity infections. AI-assisted screening enhanced operational efficiency and diagnostic yield, suggesting improved sensitivity for low-burden infections. These findings support the integration of AI tools in parasitology workflows.IMPORTANCEThe data presented in our manuscript are observational after adopting a more sensitive assay in our parasitology workflow, specifically the use of artificial intelligence (AI) to pre-screen concentrated stool specimens for parasites. The notable increase in the positivity of helminth cases in our lab, which is a large reference laboratory, indicates that adopting AI as a screening tool in a parasitology laboratory can increase sensitivity of pathogenic parasites, especially helminths.