[Comparative analysis of two assaysin detection of sperm DNA fragmentation index, flow cytometry and AI-based fluorescence microscopy, based on AO staining: A multicentre study].

Objective: To study the correlation, consistency, and variations between two assays of DNA fragmentation index based on acridine orange (AO) staining via AI-based fluorescence microscopy(AI-DFI), and flow cytometry (FCM-DFI) across multiple centers.

Methods: We selected 421 male patients from Nanjing Drum Tower hospital ( Hospital G) (226 cases), Eastern Theatre General Hospital (Hospital J) (89 cases) and Jiangsu Province Hospital (Hospital S) (106 cases) . Semen samples from each patient were analyzed for routine semen parameters and for DFI using both AI fluorescence microscopy and flow cytometry. We studied the two methods' stability as well as the correlation, consistency, and variation between the two methods' results in various centers.

Results: The two replicate studies' results of AI-DFI and the three centers' FCM-DFI for linear regression analysis indicated strong stability (R2>0.9).Overall(Group A), the AI-DFI results demonstrated good correlation and consistency with the FCM-DFI results of three centers (r>0.85；ICC>0.9).The semen specimens were categorized into two groups: normal specimen group (group B) and abnormal specimen group (group C) (including asthenozoospermia, oligospermia, and semen samples with high impurities).Group C's results showed a decline in correlation and consistency when compared to group A and group B, whereas group B's results showed a little rise in correlation and consistency when compared to group A. Although the consistency and correlation between the results of the two DFI testing methods in the three centers were good, there was still a significant difference between Groups A and C (P<0.05), and in Group B there was a significant difference between the two DFI testing methods only in Hospital G (p＝0.02), with no significant difference in Hospitals J and S (P> 0.05).

Conclusion: The two detection methods exhibit good stability and correlation. However, significant differences are observed in the DFI detection methods in samples with abnormal semen parameters and high complexity. The main reason for these significant differences may lie in the variations in detection principles. Each detection method has its own advantages, allowing clinical or research settings to choose between them based on laboratory conditions or specific requirements.