BACKGROUND: In vivo prostate cancer imaging is limited by the specificity and sensitivity of the current imaging technology. In order to provide the biochemical basis for improved in vivo MR spectroscopic (MRS) imaging of prostate cancer, we hypothesized that the metabolite processes (pathways), or metabolomics, are specifically altered in prostate cancer and studied 200 tissue samples from radical prostatectomy specimens. By analyzing the chemical shifts via MR spectroscopy, we were able to identify a specific metabolomic profile of prostate cancer. In the current study, we attempted to apply the metabolomic profile of prostate cancer in ex in vivo MRS imaging of prostate cancer.
METHODS: Five radical prostatectomy specimens from prostate cancer patients were analyzed within for this study. Two- and three-dimensional MRS chemical shift imaging (CSI) analyses were conducted at room temperature, on a 7.0 T scanner. The specimens were subsequently submitted for whole mount histological evaluation and mapping of the prostate cancer. Metabolite intensities (n=36) from CSI data for each voxel were processed individually, and used to construct prostate cancer specific metabolomic profiles based on the published prostate cancer results (Cancer Res. 2005;65:3030-3034). The values of the calculated profiles for each voxel were used as indices in the color-map to determine the color and displayed transparently on an overlay of the anatomic MRI image. The final metabolomic image is the anatomic map of the prostate with the voxel grid overlaid - each voxel’s color indicates that voxel’s status on the metabolomic cancer-index. The Metabolomic images were then compared to the histological mapping of the prostate cancer.
RESULTS: Metabolomic images demonstrated high intensities of cancer specific profiles in or around the regions where tumors are identified by histopathological mapping of the prostate cancer. Using the level of the tested metabolomic profiles, we were able to detect prostate cancer that extended into the voxel rim and with an overall size of more than twice the MRS voxel sizes (18 mm²), with an overall accuracy of 93%.
CONCLUSIONS: Our results demonstrate the potential of detecting cancer in human prostate by the novel concept of metabolomic imaging. With additional systemic developments, we consider the application of the concept in in vivo clinical diagnosis to be imminent.