Abstract
OBJECTIVES: Prostate cancer (PCa) is one of the most common malignancies in men worldwide and has caused increasing clinical morbidity and mortality, making timely diagnosis and accurate staging crucial. The authors introduced a novel approach based on mass spectrometry for precise diagnosis and stratification of PCa to facilitate clinical decision-making. METHODS: Matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry analysis of trace blood samples was combined with machine learning algorithms to construct diagnostic and stratification models. A total of 367 subjects, comprising 181 with PCa and 186 with non-PCa were enrolled. Additional 60 subjects, comprising 30 with PCa and 30 with non-PCa were enrolled as an external cohort for validation. Subsequent metabolomic analysis was carried out using Autoflex MALDI-TOF, and the mass spectra were introduced into various algorithms to construct different models. RESULTS: Serum metabolic fingerprints were successfully obtained from 181 patients with PCa and 186 patients with non-PCa. The diagnostic model based on the eight signals demonstrated a remarkable area under curve of 100% and was validated in the external cohort with the area under curve of 87.3%. Fifteen signals were selected for enrichment analysis, revealing the potential metabolic pathways that facilitate tumorigenesis. Furthermore, the stage prediction model with an overall accuracy of 85.9% precisely classified subjects with localized disease and those with metastasis. The risk stratification model, with an overall accuracy of 89.6%, precisely classified the subjects as low-risk and high-risk. CONCLUSIONS: Our study facilitated the timely diagnosis and risk stratification of PCa and provided new insights into the underlying mechanisms of metabolic alterations in PCa.