Login to Access Video or Poster Abstract: MP38-20
Sources of Funding: none

Introduction

One of the most important features of prostate cancer (PC) is the diverse biological activity attributed to heterogeneity within individual cancers. We hypothesize that metabolic difference identified by proton MR-spectroscopy (MRS) might be related to diverse biological potential resulting from heterogeneous cancer cell populations and well reflect the biological aggressiveness of PC.

Methods

In this study, 113 cases that underwent RALP were included. All patients performed 1.5T MRS examination prior to RALP. On MRS, each peak of the chemical shift between 0.5 and 4.2 ppm was determined by the computer-assisted objective measurement of MRS, and evaluated by cluster analysis or principal component analysis for measuring heterogeneity within individual cancers. The 1st principal component was considered as an index related to heterogeneity within individual cancers. The mean value, standard deviation (SD), coefficient of variation (CV) of the target shifts such as choline (3.2 ppm), citrate (2.6 ppm), N-acetyl-L-aspartate (NAA) (2.0 ppm) and alanine (1.48 ppm) were also calculated and compared with pathological findings as well as biochemical recurrence (BCR).

Results

Representative heatmaps of 2 different PC cases were shown in Figure. In this study, a correlation between the mean value and SD was positive in choline and NAA, (p<0.0001, both), while it was weak in alanine (p=0.0479). Based on this, the SD was used as an index of variation for choline and NAA, while the CV was used for alanine. The measured mean heterogeneity of our series predicted by the 1st principal component was 21.4%, which showed a significant correlation with SD for choline and NAA (p=0.0011and p<0.0001, respectively). On the other hand, the 1st principal component did not correlate with the CV for alanine, PSA or Gleason score. Although 8 out of 113 patients progressed into BCR, multivariate analysis showed that the SD for choline, followed by PSA and the SD for NAA, was an independent predictor of BCR (p=0.0354, p=0.0624 and p=0.0952, respectively).

Conclusions

To our knowledge, this is the first report of comprehensive analysis of MRS pattern to validate _x000D_ intra-tumor heterogeneity in PC, which can surely develop diagnostic and therapeutic strategy of PC._x000D_

Funding

none