Assessment of systemic metabolic biomarkers to predict prostate cancer progression and mortality
Session type: Poster / e-Poster / Silent Theatre session
Overall survival for localised Prostate Cancer (PCa) is nearly 100% at 5 years, however a more reliable and less invasive predictor of PCa survival is needed to reduce unnecessary treatment in indolent cases. Recent evidence suggested that metabolic traits can predict the aggressiveness of PCa. We investigated the levels of metabolic traits in relation to the prospective development of metastases and PCa mortality, in the 10-year median follow-up of the Prostate Testing for Cancer Treatment (ProtecT) cohort of incident localised PCa.
We analysed the baseline levels of 159 serum metabolic traits of 1,245 men diagnosed with localised PCa who were eligible for randomisation and were followed-up for a mean 10-year period in the ProtecT trial. We used Cox multivariable regression to assess survival and PCa progression for each metabolic trait, adjusting for clinical stage, BMI and age. We used bootstrapping to develop and test the performance of our multi-metabolic trait prediction model.
Of the 1,245 patients, 12 died, 47 developed metastatic disease and 141 had clinical progression. There were no strong differences in the hazard ratios for any of the metabolites in relation to PCa death and clinical progression. We observed that small particle lipoproteins (HR 0.66, CI: 0.50- 0.85 per standard deviation (SD)) and the mean size of low-density lipoproteins (HR: 1.56, CI: 1.17-2.07 per SD) were related to developing metastases. The predictive model developed using PCa metastasis data, showed potential in predicting PCa metastases (area under the receiver operating characteristic curve (AUC): 0.73, CI:0.65-0.82) and PCa specific mortality (AUC: 0.76, CI:0.66-0.85).
We found metabolic traits that are associated with PCa outcomes and that may have a predictive role in distinguishing between high and low-risk localised PCa.