Artificial intelligence powered andrology diagnostics offers total solution for semen analysis

Published at the 2023 American Reproductive Medicine (ASRM) Scientific Congress and Expo

L. CHUN-I^1,2,3,4, H. CHUN-CHIA^1,4, H. CHENG-TENG⁵, C. HUI-CHEN⁵, W. TSE-EN⁵, A. ASHOK^6,7, L. MAW-SHENG^1,4
[1] Division of Infertility Clinic, Lee Women's Hospital, Taichung, Taiwan; [2] Institute of Medicine, Chung Shan Medical University, Taichung, Taiwan; [3] Department of Obstetrics and Gynecology, Chung Shan Medical University, Taichung, Taiwan; [4] Department of Post-Baccalaureate Medicine, National Chung Hsing University, Taichung, Taiwan;  [5] Center for Research and Development, Bonraybio Co., Ltd, Taichung, Taiwan;  [6] Global Andrology Forum, American Center for Reproductive Medicine, Moreland Hills, Ohio, USA; [7] Emeritus Staff, Cleveland Clinic Foundation, Cleveland, Ohio, Us
 

Purpose & Objectives

Computer-aided semen analysis (CASA) system improves the subjectivity manual-based semen evaluation; however, it still lacks reproducibility and accuracy due to systemic errors associated with sperm clumping and/or cloudy background. Moreover, it is difficult for physician to make a clinical decision based solely on sperm numbers, sperm motility, and sperm morphology. To overcome the limitations, we aimed to:
1) develop an artificial intelligence optical microscopic (AIOM)-based semen analyzer for basic semen parameters and sperm DNA fragmentation evaluation, and 2) demonstrate the system accuracy and precision in comparison with manual assessment.
 

Material & Methods

In this cross-sectional study design, total 492 semen samples were collected and allocated to 3 bodies of this study.

1. Liquefied semen aliquots (n=77) were analyzed for sperm concentration, total motility, and progressive motility by a trained andrology technician based on 10-µm Leja chamber slide and 10-µm LensHooke® CS3 slide, respectively.
2. Liquefied semen aliquots (n=98) were analyzed for sperm concentration, total motility, and progressive motility by CS3 slides and those slides were interpreted manually and automatically by LensHooke® X12 semen analysis system.
3. Liquefied semen aliquots (n=317) were evaluated LensHooke® R10 sperm chromatin dispersion assay, and the stained slides were interpreted manually and automatically by X12.

We performed linear regression and Spearman's rank analysis to validate the correlation between new advanced method(s) and reference method(s). Performance metrics were used to indicate the result accuracy.


▲Fully-automated semen analyzer serves as a revolutionary tool for diagnosing male infertility


▲AIOM-based semen analyzer for basic semen parameters and sperm DNA fragmentation
 

Result

1. When Leja slide as reference, newly-developed CS3 slide shows 100% accuracy for three semen parameters.
2. When manual method as reference, X12 shows >90% accuracy for three semen parameters according to WHO 5th and 6th lower reference limit.
3. X12 produced nearly-identical diagnostic results as manual method (accuracy:92%).

▲When Leja slide as reference, newly-developed CS3 slide shows 100% accuracy for three semen parameters.


▲When manual method as reference, X12 shows >90% accuracy for three semen parameters according to WHO 5th and 6th lower reference limit.


▲X12 produced nearly-identical diagnostic results as manual method (accuracy:92%).
 

Conclusions

By virtue of AIOM technology,

1. X12 provides an objective, automated and accurate evaluations for semen analysis.
2. X12 is capable of automatically conducting both basic and advanced semen analysis.
3. X12 can reduce work burdens and allows a better laboratory management.

References

1. Hsu CT, Lee CI, Huang CC, et al. Development and integration of LensHooke® R10 for automatic and standardized diagnosis for sperm DNA fragmentation [published online ahead of print, 2023 Mar 3]. Andrology. 2023;10.1111/andr.13419. doi:10.1111/andr.13419
2. Kuroda S, Karna KK, Kaiyal RS, et al. Novel sperm chromatin dispersion test with artificial intelligence-aided halo evaluation: A comparison study with existing modalities [published online ahead of print, 2023 Mar 31]. Andrology. 2023;10.1111/andr.13436. doi:10.1111/andr.13436

Full article：
Lee, C. I., Huang, C. C., Lee, M. S., Hsu, C. T., Chang, H. C., Wang, T. E., & Agarwal, A. (2023). ARTIFICIAL INTELLIGENCE POWERED ANDROLOGY DIAGNOSTICS OFFERS TOTAL SOLUTION FOR SEMEN ANALYSIS. Fertility and Sterility, 120(4), e302-e303. DOI:https://doi.org/10.1016/j.fertnstert.2023.08.878