Pd37-09 a Novel Machine Learning-Based Nomogram to Predict Health Worsening in Infertile Men

You have accessJournal of UrologyInfertility: Epidemiology & Evaluation II (PD37)1 May 2024PD37-09 A NOVEL MACHINE LEARNING-BASED NOMOGRAM TO PREDICT HEALTH WORSENING IN INFERTILE MEN Federico Belladelli, Edoardo Pozzi, Christian Corsini, Fausto Negri, Alessandro Bertini, Massimiliano Raffo, Eugenio Ventimiglia, Giuseppe Fallara, Luca Boeri, Paolo Capogrosso, Federico Dehò, Alessia d'Arma, Francesco Montorsi, and Andrea Salonia Federico BelladelliFederico Belladelli , Edoardo PozziEdoardo Pozzi , Christian CorsiniChristian Corsini , Fausto NegriFausto Negri , Alessandro BertiniAlessandro Bertini , Massimiliano RaffoMassimiliano Raffo , Eugenio VentimigliaEugenio Ventimiglia , Giuseppe FallaraGiuseppe Fallara , Luca BoeriLuca Boeri , Paolo CapogrossoPaolo Capogrosso , Federico DehòFederico Dehò , Alessia d'ArmaAlessia d'Arma , Francesco MontorsiFrancesco Montorsi , and Andrea SaloniaAndrea Salonia View All Author Informationhttps://doi.org/10.1097/01.JU.0001009464.98066.03.09AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Increasing epidemiological data show that infertile men have a higher risk of more frequently developing early onset non communicable diseases (NCDs) compared to fertile men. We aimed to develop a novel nomogram for primary infertile men to predict general health worsening. METHODS: Complete data from the last 495 men seeking first medical help for infertility were analysed. Clinical and socio-demographic data were collected in all patients. Health-significant comorbidities were scored using the Charlson comorbidity index (CCI), with annual follow-ups until 2023. The Boruta machine learning algorithm was employed to select the relevant variables of interest among BMI, smoke, azoospermia, FSH, LH, total testosterone (tT), sperm concentration, rate of progressive sperm motility, rate of normal sperm morphology, and DNA fragmentation index (SDF). Cox regression model was developed to predict CCI score increases. Therefore, a nomogram was derived from the prediction model. The Harrel's concordance index (c-index) was used to test the accuracy of the nomogram. RESULTS: Median (IQR) age and BMI were 37 (34-41) years and 24.8 (23.1-26.8) kg/m2, respectively. Of 495, 63 (13%) and 107 (22%) patients presented with CCI ≥1 and were active smokers, respectively. At a median follow-up of 153 (119-194) months, 268 (54%) patients have had an increase of at least 1 point at CCI score. Overall, BMI, rate of normal sperm morphology, and SDF values emerged as the most relevant variables in predicting CCI score increases, when Boruta machine learning algorithm was applied. Figure 1 depicts the nomogram derived from the cox-regression model using the variables above mentioned. The nomogram's c-index was 72%. CONCLUSIONS: A novel nomogram based on infertile men's clinical parameters to predict general health worsening over time was developed by applying Boruta machine learning algorithm. Future studies will demonstrate the clinical effectiveness of the nomogram in terms of counseling primary infertile patients throughout their follow-up investigation and to develop tailored secondary preventive strategies. Download PPT Source of Funding: None © 2024 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 211Issue 5SMay 2024Page: e803 Advertisement Copyright & Permissions© 2024 by American Urological Association Education and Research, Inc.Metrics Author Information Federico Belladelli More articles by this author Edoardo Pozzi More articles by this author Christian Corsini More articles by this author Fausto Negri More articles by this author Alessandro Bertini More articles by this author Massimiliano Raffo More articles by this author Eugenio Ventimiglia More articles by this author Giuseppe Fallara More articles by this author Luca Boeri More articles by this author Paolo Capogrosso More articles by this author Federico Dehò More articles by this author Alessia d'Arma More articles by this author Francesco Montorsi More articles by this author Andrea Salonia More articles by this author Expand All Advertisement PDF downloadLoading ...