Table 3. Results from the metabolomics studies that did not assess biomarker utility.

Authors	n. metabolites	Results	Validated?	Conclusions
Halliday et al. (1988)	49 metabolites	Lactate levels and lipid signals are higher in tumor tissue, citrate levels are decreased.	No	13C NMR spectroscopy can be used to differentiate neoplastic and non-neoplastic prostate tissue
Schiebler et al. (1993)	13 metabolite peaks	Citrate, Acetate, inositol and lactate differed significantly between adenocarcinoma and normal peripheral zone tissue, however there were a number of similarities between the spectra of BPH and adenocarcinoma	No	The citrate standardized peak area alone cannot be used to diagnose prostate adenocarcinoma
Swanson et al. (2003)	8 metabolites	Glandular tissue: healthy tissue had significantly higher levels of citrate, Taurine, myo-inositol, and scyllo-inositol and polyamines, and lower levels of choline, phosphocholine (PC), and glycerophosphocholine (GPC) relative to tumor tissue. Stromal tissue: healthy tissue had lower levels of choline compounds and higher levels of Taurine, myo-inositol, and scyllo-inositol.	Cross validation	Distinctive metabolic patterns were identified for tumor and healthy tissues and for cancers of varying grade; however, tissue type may affect the findings.
Cho et al. (2009)	21 corticosteroids	Urinary cortisol levels were significantly higher in prostate cancer patients than in healthy controls and BHP patients	Currently ongoing in a larger population	The results indicate dysregulated cortisol metabolism in prostate cancer patients and suggest that metabolic profiling may be the optimal way to measure this.
Thysell et al. (2010)	Bone 123 metabolites of which 49 could be identified. Tissue: 157 metabolites of which 59 could be identified. Plasma: 179 metabolites of which 50 could be identified	Bone: 58% of metabolites differed between normal and metastatic bone. Amino acid synthesis and metabolism were upregulated in metastatic bone, and high levels of cholesterol, myo-inositol-1-phosphate, citric acid, fumarate, glycerol-3-phosphate, and fatty acids detected. Tissue: 8% of metabolites differed between tissue from patients with and without metastases including four of those significantly increased in metastatic bone: asparagine, threonine, fumaric acid, and linoleic acid. Plasma: 15% of metabolites differed in the plasma of patients with and without bone metastases including four identified in metastatic bone; glutamic acid, taurine, and phenylalanine and stearic acid. Sarcosine was found to be increased in the bone of men with metastatic disease but not in their tumor tissue or plasma	No	Cholesterol is a possible therapeutic target for advanced PCa.
Komoroski et al. (2011)	4 phospholipid metabolites	Phosphoethanolamine and glycerophosphoethanolamine levels differed significantly between the cancer and BPH groups. None of the four metabolites were associated with Gleason score	No	These metabolites may be useful in the diagnosis of prostate cancer and may help to explain the high choline resonance identified in other studies.
Shuster et al. (2011)	260 metabolites	83 (32%) of metabolites were significantly different between cancer and benign tissue, with 82 at higher levels in cancer tissue. Common amino acids, long chain fatty acids and phospholipids were increased. Higher levels of uracil, kynurenine, glycerol-3-phosphate, leucine and proline were reported in agreement with Sreekumar's study but sarcosine was below the limit of detection.	No	Molecular biomarkers could augment histology in the characterization of disease.
Brown et al. (2012)	260 metabolites	Ala, Ile, Orn and Lys were downregulated in prostate cancer patients compared to healthy controls and Gln, Val, Trp and Arg were upregulated	No	It is possible to determine a metabolomic signature of prostate cancer
Gamagedara et al. (2012)	4 metabolites	No difference in the mean levels of proline, kynurine, uracil or Glycerol-3-phosphate between prostate cancer cases and healthy controls	Validating Sreekumar's findings	The explored biomarkers had no diagnostic or prognostic potential in prostate cancer, and could not distinguish prostate cancer from other malignancies
Saylor et al. (2012)	292 identified metabolites	56 metabolites changed significantly from baseline to three months: Multiple steroids, Markers of lipid beta-oxidation, markers of omega-oxidation and markers of insulin resistance (2-hydroxybutyrate, branch chain keto-acid dehydrogenase complex products) were lower. Most bile acids and their metabolites were higher.	Cross validation	Identified novel and clinically important ADT-induced metabolic changes
Kami K et al. (2013)	86 (of which 39 could be absolutely quantified)	TCA cycle intermediates, Succinate, fumarate, malate, pyruvate and lactate levels were higher, and ADP and phosphoenolypyruvate lower in tumor than in normal tissue.	Cross validation	Tumor metabolic profiles can help to distinguish normal from tumor tissue, and tumor stage
Li et al. (2013)	626 metabolites	53 metabolites associated with metastatic prostate cancer, 16 significant pathways including amino acid metabolism, tryptophan metabolism, cysteine and methionine metabolism, arachidonic acid metabolism and histidine metabolism	No	Identified novel disease relevant pathways using an alternative statistical method on Sreekumar et al.' s data
Huang et al. (2014)	8232 signals	In patients who did not develop castration-resistant prostate cancer (CRPC) for at least 2 years, serum deoxycholic acid (DCA), glycochenodeoxycholate (GCDC), L -tryptophan, docosapentaenoic acid (DPA), arachidonic acid, deoxycytidine triphosphate, and pyridinoline levels reverted to near healthy control levels during endocrine therapy. In contrast, the metabolite levels remained abnormal in patients who developed CRPC within 1 year	Validation currently ongoing	DCA, GCDC, L -tryptophan, DPA, arachidonic acid, deoxycytidine tri-phosphate, and pyridinoline represent potential biomarkers for evaluating patient response to endocrine therapy. These results suggest a role for cholesterol in PCA progression
Mondul et al. (2014)	420 metabolites	Circulating 1-stearoylglycerol was inversely associated with the risk of developing prostate cancer up to 23 years after blood collection. The magnitude of this association did not differ by disease aggressiveness. There was also suggestive inverse associations for glycerol and alpha-ketoglutarate.	Only the association between alpha-ketoglutarate and aggressive prostate cancer was replicated in a subsequent study including different participants from the same population.	The results support a role for dysregulation of lipid metabolism in the development of prostate cancer
Mondul et al. (2015)	626 metabolites	Strong inverse associations between energy and lipid metabolites particularly glycerophospholipids and fatty acids and aggressive cancer were observed with aggressive disease risk. Thyroxine and trimethylamine oxide were assocaited with aggressive disease risk while Alpha-ketoglutarate and citrate were inversely associated. Metabolites associated with nonaggressive cancers included 2′-deoxyuridine, adenosine 50 -monophosphate (AMP), 11-dehydrocorticosterone, 21-hydroxypregnenolone monosulfate, cotinine and hydroxycotinine.	Meta-analyses with the findings of a previous study confirmed a role for glycerophospholipids and long chain fatty acids	Prospective study data indicate that several circulating glycerophospholipid, fatty acid, energy and related metabolites are inversely associated with aggressive prostate cancer up to 20 years prior to diagnosis. Metabolite associations vary by cancer aggressiveness.