Introduction: Intestinal parasitic infections, caused by protozoa and helminths, can lead to malnutrition, anaemia, and impaired growth. While direct wet mount microscopy is the routine diagnostic method, it is limited by low sensitivity, labour intensiveness, and reliance on skilled personnel. Automated image-based systems using artificial intelligence (AI) present a promising alternative.
Methods: This study was conducted in the Department of Microbiology of a tertiary care hospital from January 1, 2022 to December 31, 2024. Microscopic findings from 7267 fecal samples were analyzed. Additionally, 275 samples were processed using both conventional microscopy and the MISPA-F30 fecal analyzer. Microscopic images analyzed by the inbuilt AI software was reviewed independently by two microbiologists. Results from direct microscopy, MISPA-F30, and expert user audit were compared.
Results: Over three years, 7267 stool samples were examined via wet mount microscopy, detecting parasitic elements in 204 (2.8%) cases, with Giardia lamblia most frequently identified. A subset of 275 samples was concurrently analyzed using the MISPA F-30 fecal analyzer and microscopy. The analyzer showed 83.3% agreement with microscopy and 84.7% with user audit, which revealed 35 misidentified cases and six missed detections. However, MISPA F-30 identified ten cases overlooked by microscopy. Using a composite reference standard, the analyzer's sensitivity and specificity were 66.7% and 86.2%, respectively.
Conclusion: MISPA-F30 demonstrated acceptable sensitivity and specificity, it additionally reduces manual workload and alleviates fatigue of sample processing and microscopy. When supplemented by expert audit, its diagnostic performance was excellent. With further AI software development and database expansion, automated stool microscopy has considerable potential as a reliable diagnostic aid, though cost-effectiveness has to be weighed in.
扫码关注我们
求助内容:
应助结果提醒方式:
