Identifying prostate carcinoma by MALDI-Imaging

Abstract

Prostate cancer has become one of the most common malignancies worldwide. Although lacking in specificity its diagnosis is still based partially on the serum-based test for prostate-specific antigen. As its pathogenesis has not yet been deciphered, the ongoing search for new and more reliable biomarkers remains a challenge to stratify disease onset and progression. Matrix-assisted laser desorption/ionization (MALDI)-Imaging is a promising technique to assist in this endeavor. It delivers accurate mass spectrometric information of the sample's proteins and enables the visualization of the spatial distribution of protein expression profiles and correlation of the information with the histomorphological features of the same tissue section. This study describes the analysis of 22 prostate sections (11 with and 11 without prostate cancer) by MALDI-Imaging. Specific protein expression patterns were obtained for normal and cancerous regions within the tissue sections. Applying a 'support vector machine' algorithm to classify the cancerous from the non-cancerous regions, an overall cross-validation, a sensitivity and specificity of 88, 85.21 and 90.74%, respectively, was achieved. Additionally four distinctively overexpressed peaks were identified: 2,753 and 6,704 Da for non-cancerous glands, and 4,964 and 5,002 Da for cancerous glands. The results of this first clinical study utilizing the new technique of MALDI-Imaging underline its vast potential to identify candidates for more reliable prostate cancer tumor markers and to enlighten the pathogenesis of prostate cancer.