The difficulty of diagnosing prostate cancer (PC) with the available biomarkers frequently leads to over-diagnosis and overtreatment of PC, underscoring the need for novel molecular signatures. The purpose of this review is to provide a summary of the currently available cellular metabolomics for PC molecular signatures. A comprehensive search on PubMed was conducted to find studies published between January 2004 and August 2022 that reported biomarkers for PC detection, development, aggressiveness, recurrence and treatment response. Although potential studies have reported the presence of distinguishing molecules that can distinguish between benign and cancerous prostate tissue. However, there are few studies looking into signature molecules linked to disease development, therapy response or tumour recurrence. The majority of these studies use high-dimensional datasets, and the number of potential metabolites investigated frequently exceeds the size of the available samples. In light of this, pre-analytical, statistical, methodological and confounding factors such as antiandrogen therapy (NAT) may also be linked to the identified chemometric multivariate differences between PC and relevant control samples in the datasets. Despite the methodological and procedural challenges, a range of methodological groups and processes have consistently identified a number of signature metabolites and pathways that appear to imply a substantial involvement in the cellular metabolomics of PC for tumour formation and recurrence.