MRS measured fatty acid composition of periprostatic adipose tissue correlates with pathological measures of prostate cancer aggressiveness

2015 
Purpose To investigate the association between magnetic resonance (MR) spectroscopically measured fatty acid composition of periprostatic adipose tissue and pathological markers of prostate cancer aggressiveness. Materials and Methods Periprostatic adipose (PPA) and subcutaneous adipose (SQA) tissue from prostate cancer patients undergoing radical prostatectomy were examined ex vivo by proton MR spectroscopy at 14.1T (n = 31). Fractions of monounsaturated, polyunsaturated, total unsaturated, and saturated fatty acids, as well as T2 relaxation times were measured from the spectra. Univariate and multivariate analyses based on receiver operating characteristic (ROC) and support vector machines (SVM) were used to evaluate the association between differential measures of fatty acid levels in the PPA and SQA tissues and Gleason score and extracapsular extension (ECE), which are pathological measures of prostate cancer aggressiveness. Results Both pathological markers for aggressive prostate cancer have separable patterns in the MRS features space. The association between ECE and PPA tissue fatty acid composition is linear (area under receiver operating characteristic curve (AROC) and 95% confidence intervals [CIs]: 1.00, [1.00, 1.00]), along the Δ(fM/fS) measure, and is marked by elevated monounsaturated and reduced saturated fatty acids in the PPA tissue relative to SQA. In contrast, the association between Gleason score and PPA tissue fatty acid composition is nonlinear (classifier AROC and 95% CIs: 0.86, [0.71, 1.00]). Conclusion Fatty acid composition is altered in the PPA tissue of patients with aggressive prostate cancer. Ex vivo MR spectroscopy may be a useful tool in studying the altered fatty acid metabolism in prostate cancer. J. Magn. Reson. Imaging 2015;42:651–657.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    38
    References
    12
    Citations
    NaN
    KQI
    []