Cancer remains a leading cause of mortality worldwide, with prostate cancer being one of the most prevalent malignancies among men. Despite advances in diagnostic technologies, early detection remains challenging, and current procedures may be invasive or lack sufficient specificity. This thesis investigates the feasibility of a non-invasive diagnostic approach for prostate cancer detection using mass-spectrometry (MS) analysis of urine samples acquired from highly miniaturized devices developed by NanoTech Analysis (NTA).. Mass spectrometry enables the simultaneous detection of thousands of molecular features, making it a promising tool for biomarker discovery. However, the high dimensionality, noise, and variability inherent in MS data—combined with additional signal stability challenges introduced by compact instrumentation—necessitate robust computational and statistical methodologies.