Urinary Volatile Organic Compounds As Early Diagnostic Biomarkers For Prostate Cancer In Obese Men Using Artificial Intelligence

Abstract

Background: Prostate cancer is one of the most frequent types of malignancy in men across the globe, and early diagnosis has not been easy, especially in obese men. There are also metabolic changes linked to obesity that have the potential to alter the biology of prostate cancer, as well as lowering the accuracy of standard prostate cancer markers like prostate-specific antigen (PSA). The urinary volatile organic compounds (VOCs) are promising candidates for non-invasive biomarkers since such compounds can indicate underlying metabolic and pathological alterations.

Objective: The paper was designed to assess the methodological and statistical viability of exploring urinary volatile organic compounds as an early diagnostic biomarker of prostate cancer in men with obesity using artificial intelligence-based predictive models.

Methods: The plastic cross-sectional analytical study of the treatment of obese men was performed with participants aged 50 years and older. The data on demographics, anthropometrics, and clinical data were gathered through a structured questionnaire. The Shapiro-Wilk test was used to check the normality of the continuous variables. The reliability was tested based on the alpha of Cronbach, whereas construct validity was tested based on the Kaiser- Meyer-Olkin (KMO) measure and Bartlett test of sphericity. Independent Samples t-test, one-way ANOVA, Kruskal-Wallis test, and Chi-square test of independence were also done as part of inferential analyses. The methods used were Pearson correlation and multiple linear regression analysis to determine the relationship between variables and to predict them. Artificial intelligence-based predictive modeling was applied to integrate urinary VOCs with clinical and anthropometric variables to assess their diagnostic performance for early prostate cancer detection.

Results: Normality analysis revealed that there had been mixed patterns of distribution, and therefore both parametric and non-parametric tests were used. The urinary symptom severity scale was found to have high reliability (Cronbach's 0.80) and fair construct validity (KMO 0.60; Bartlett test p < 0.001). Statistically significant differences and changes were indicated in the inferential analyses according to major clinical and anthropometric variables. There was a strong relationship that was found between prostate cancer status and PSA levels. Correlation was found to be positive between age, height, weight, and BMI, and the regression analysis identified height and BMI as the positive predictors of body weight. The artificial intelligence model demonstrated promising performance in identifying patients at higher risk of prostate cancer based on urinary VOCs and clinical variables.

Conclusion: The results prove that the data and the analysis system are sound, valid, and statistically strong to explore urinary VOCs as the initial diagnostic biomarkers of prostate cancer in obese men. The research confirms the fact that the possible use of incorporating artificial intelligence-assisted urinary-based non-invasive biomarkers in conjunction with clinical parameters could enhance early detection methods. Secondary research that includes direct VOC profiling and longitudinal designs will be channeled in the future to ascertain the accuracy of diagnosis and clinical applicability.