The world journal of men's health · 2025
Machine Learning Evaluation of Semen Analysis Could Reveal New Infertility-Related Markers: A Pilot Study.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Santi Daniele, Pozza Carlotta, Spaggiari Giorgia, Gianfrilli Daniele, Sbardella Emilia, Paoli Donatella, Roli Laura, De Santis Maria Cristina, Bonomi Marco, Trenti Tommaso, Isidori Andrea M, Simoni Manuela
Аннотация
PURPOSE: To perform a pilot study aiming at evaluating whether machine learning could be a useful model to evaluate semen analysis, improving the diagnostic work-up of male partner of infertile couples.
MATERIALS AND METHODS: A retrospective observational study was conducted using real-world data on male evaluated in routine andrological clinical practice at two Italian tertiary centers. The study utilized two distinct datasets: the first (UNIROMA) encompassed three distinct variables, including semen analysis, sex hormones, and testicular ultrasound parameters. The second dataset (UNIMORE) was constructed incorporating semen analysis, sex hormones, biochemical examinations, and parameters related to environmental pollution. The XGBoost analysis, as part of machine learning techniques, was applied separately to each dataset, as the two datasets did not share a significant overlap in terms of variables.
RESULTS: The UNIROMA dataset comprised 2,334 male subjects. The XGBoost analysis exhibited the highest accuracy (area under the curve [AUC], 0.987) in predicting patients with azoospermia compared to other categories. Remarkably, our analysis revealed that among the most influential predictive variables, follicle-stimulating hormone serum levels (F-score=492.0), inhibin B serum levels (F-score=261), and bitesticular volume (F-score=253.0) stood out. The UNIMORE dataset consisted of 11,981 records. The XGBoost analysis demonstrated a good predictive accuracy (AUC, 0.668), especially for identifying the azoospermia group. Notably, the most crucial predictive variables were environmental pollution parameters (PM10, F-score=361; NO₂, F-score=299) and biochemical data (white blood cells, F-score=326; red blood cells, F-score=299).
CONCLUSIONS: This pilot study applies machine learning to two extensive datasets, suggesting that changes in semen analysis may be linked to other variables, such as testicular ultrasound characteristics, red blood cell count, and environmental pollution.
MATERIALS AND METHODS: A retrospective observational study was conducted using real-world data on male evaluated in routine andrological clinical practice at two Italian tertiary centers. The study utilized two distinct datasets: the first (UNIROMA) encompassed three distinct variables, including semen analysis, sex hormones, and testicular ultrasound parameters. The second dataset (UNIMORE) was constructed incorporating semen analysis, sex hormones, biochemical examinations, and parameters related to environmental pollution. The XGBoost analysis, as part of machine learning techniques, was applied separately to each dataset, as the two datasets did not share a significant overlap in terms of variables.
RESULTS: The UNIROMA dataset comprised 2,334 male subjects. The XGBoost analysis exhibited the highest accuracy (area under the curve [AUC], 0.987) in predicting patients with azoospermia compared to other categories. Remarkably, our analysis revealed that among the most influential predictive variables, follicle-stimulating hormone serum levels (F-score=492.0), inhibin B serum levels (F-score=261), and bitesticular volume (F-score=253.0) stood out. The UNIMORE dataset consisted of 11,981 records. The XGBoost analysis demonstrated a good predictive accuracy (AUC, 0.668), especially for identifying the azoospermia group. Notably, the most crucial predictive variables were environmental pollution parameters (PM10, F-score=361; NO₂, F-score=299) and biochemical data (white blood cells, F-score=326; red blood cells, F-score=299).
CONCLUSIONS: This pilot study applies machine learning to two extensive datasets, suggesting that changes in semen analysis may be linked to other variables, such as testicular ultrasound characteristics, red blood cell count, and environmental pollution.