Abstract: PD71-10
Sources of Funding: none

Introduction

There are several systemic agents for metastatic castration-resistant prostate cancer (CRPC) result in a complicated decision-making while selecting an appropriate treatment. Therefore, there is a need to identify more powerful predictive biomarkers of CRPC to decide on therapeutic strategy. In the present study, we performed serum N-glycomics between healthy volunteer (HLT) and various stages of PC patients and evaluated its potential as a predictive serum-based glycobiomarkers of CRPC.

Methods

N-glycomics in serum were performed total 850 patients in randomly selected 128 healthy volunteer (HLT) and 286 benign prostatic hyperplasia (BPH), 258 PC before radical prostatectomy (RP), 29 PC after RP, 41 hormone sensitive PC who treated androgen deprivation therapy (HSAPC with ADT) and 68 CRPC patients. Candidate N-glycans for prediction of CRPC were selected from those with the highest area-under-the-curve (AUC) and used to create an N-glycan score (NGScore) based on presence and amount. The validity of this score for prediction of CRPC was analyzed using a receiver operating characteristic (ROC) curve.

Results

N-glycomics reveled that tri- and tetra-antennary N-glycans level of CRPC was significantly higher than that of HSPC with ADT group, while high mannose, hybrid and bi-antennary typed N-glycan of CRPC was significantly lower than that of HSPC with ADT. We identified 5 candidate N-glycans (high mannose; m/z 1565, bi-antennary; m/z 1752 and 2058, tri-antennary; m/z 2744 and 3414, respectively) significantly associated with CRPC patients. The accuracy of the NGScore for prediction of CRPC was significant with an AUC value of 0.7588. The positive and negative predictive value of NGScore at cutoff ≥2.0 was 77.1% and 71.2% respectively. A high value NGScore was significantly associated with CRPC. Finally, we prospectively investigated NGScore, PSA and testosterone levels of 13 HSPC with ADT patients who treated androgen deprivation therapy (ADT) and revealed that 66.6% of NGScore ≥2 points patients (4/6) was developed CRPC or experienced PSA failure. NGScore less than 2 points patients (7/7) was not developed CRPC during follow-up period.

Conclusions

Our newly developed NGScore seems to be a practical predictive method for CRPC.

Funding

none