High-throughput imaging cytometry paired with artificial intelligence identifies ultrarare sperm in men previously diagnosed with clinical azoospermia.

OBJECTIVE

To determine whether a novel diagnostic platform, which pairs high-throughput imaging cytometry with artificial intelligence (AI)-assisted image analysis, can identify ultrarare sperm in ejaculatory samples deemed clinically azoospermic.

DESIGN

A retrospective cohort review of remnant biobanked ejaculatory samples processed by high-throughput imaging cytometry and AI-assisted image analysis.

SUBJECTS: Remnant ejaculatory samples (n = 148) biobanked from December 2024 to September 2025 at a single high-volume male infertility clinic.

EXPOSURE: Image-based flow cytometry and convolutional neural network (CNN) driven image analysis for diagnostic classification.

MAIN OUTCOME MEASURES: Diagnostic reclassification rate of remnant ejaculatory samples initially deemed clinically azoospermic using CNN-based morphologic assessment and corollary sperm specific protein expression.

RESULTS

Within the 83 azoospermic samples analyzed, 40 samples (48%) were found to have sperm using this high-fidelity diagnostic platform.

CONCLUSION

We propose a new classification of sperm concentration, microzoospermia, a sperm concentration so low that only advanced image-based search technologies can reliably identify ultrarare sperm.