[Artificial intelligence-based fluorescence method versus traditional flow cytometry in detection of sperm DNA fragmentation index].

Objective: To compare the result of the artificial intelligence (AI) recognition-based fluorescence method and that of traditional flow cytometry in the examination of the sperm DNA fragmentation index (DFI) and assess the reliability of the AI-based fluorescence detection.

Methods: Using flow cytometry and the AI-based fluorescence method, we examined the sperm DFI in the semen samples collected from 338 outpatients. We analyzed the correlation between the results and compared the positive rates detected by the two methods. We repeated the AI-based fluorescence method twice for each semen sample to observe its technical stability in the detection of sperm DFI.

Results: The result of flow cytometry was well correlated with that of the AI-based fluorescence method in the detection of sperm DFI (R2 = 0.7131), but poorly correlated for low-concentration, sticky semen and some other extreme samples (R2 = 0.2065). No statistically significant difference was found between the two methods in the positive rate of detection. The AI-based fluorescence method exhibited an excellent technical stability (R2 = 0.9671).

Conclusion: The AI-based fluorescence method has an excellent technical stability in the detection of sperm DFI and the result is not significantly different from that of traditional flow cytometry.