Andrology-Open Access
Open Access

Abstract

Impact of Lifestyle and Mental Health on Sperm DNA Fragmentation Rate in Infertile Men and Construction of a Predictive Model: Based on Machine Learning Algorithms

Ke Wang, Jing Hu, Ningxin Qin*, Jie Bai, You Zhang, Xin Huang and Yan Xu

Objective: This study aims to investigate the current status of sperm DNA Fragmentation rate (DFI) in infertile men, analyze the impact of lifestyle and mental health on it, and construct a predictive model for DFI using machine learning algorithms.

Methods: A cross-sectional survey was conducted from March, 2023 to August, 2023 involving 592 infertile male patients receiving Assisted Reproductive Technology (ART) treatment at the Reproductive Medicine Center of Tongji University Affiliated Obstetrics and Gynecology Hospital. Patients completed a general information questionnaire, the Self-Rating Anxiety Scale (SAS), the Self-Rating Depression Scale (SDS), and the Athens Insomnia Scale (AIS). DFI data were collected, and binary logistic regression was utilized to analyze independent risk factors affecting DFI quality. Predictive models for DFI quality were constructed using decision tree, random forest, and support vector machine algorithms based on machine learning, with 10-fold cross-validation used to search for optimal parameters. The models were evaluated using Receiver Operating Characteristic (ROC) curves, accuracy, precision, recall and F1 score to identify the best model and rank feature importance

Results: Among the surveyed infertile male patients, 434 (73.3%) had poor sperm DNA fragmentation rates. Binary logistic regression indicated that smoking status, alcohol consumption, regular exercise habits, depression, insomnia, and anxiety were independent risk factors for poor DFI (p<0.05). All three machine learning algorithms achieved an area under the curve (AUC) greater than 0.80, with the random forest-based predictive model demonstrating the best performance (AUC=0.95), making it the optimal model. Feature importance ranking from the optimal model revealed that insomnia was the most significant factor affecting DFI quality, followed by anxiety, alcohol consumption, depression, smoking and regular exercise.

Conclusion: In this study, the random forest predictive model exhibited the best performance and served as the optimal model. The high detection rate of abnormal DFI quality in infertile men is primarily influenced by lifestyle habits, negative emotions and insomnia. The establishment of the predictive model provides a convenient tool for male reproductive health professionals to assist clinical decision-making and adjust treatment plans. Healthcare providers should consider guiding patients to reduce smoking and alcohol intake, increase physical activity and improve negative emotions to enhance DFI quality in infertile men.

Published Date: 2024-11-21; Received Date: 2024-10-22