Imaging mass spectrometry (IMS) to discriminate renal oncocytoma from chromophobe RCC using formalin-fixed paraffin-embedded (FFPE) tissue

Background Chromophobe renal cell carcinomas (ChRCC) and renal oncocytomas (RO) are difficult to discriminate in some cases even if immunohistochemstry or molecular methods are applied. Histopathological diagnosis relies mainly on histology. The differentiation of both entities has major implications for therapy and thus for patient outcome. Imaging mass spectrometry (IMS) analysis has been performed in an attempt to characterize a peptide pattern able to discriminate ChRCCs from ROs, thus assisting the diagnosis of renal neoplasms. Methods Formalin fixed paraffin embedded tissues used in this study included ROs (n=22) and ChRCCs (n=40). The cases were reviewed by one pathologist and diagnosis was verified using immunohistochemistry. All samples were deparaffinized, antigen retrieved, and subjected to trypsin and matrix deposition using an automatic spray device (ImagePrep, Bruker, Bremen). Tissues were subsequently analyzed by IMS using a Bruker Autoflex Speed mass spectrometer. Mass spectra were exported from the image data set, defined according to their histology context, and two sample groups including ROs and ChRCCs were then created and subjected to spectral and statistical analysis using ClinProTools 3.0 software. Peak selection was evaluated through statistical comparison (Wilcoxon/Kruskal-Wallis test). Results A support vector machine algorithm incorporated 7 m/z species to generate a classification model able to discern ROs from ChRCCs with a recognition capability= 96.67% and a cross validation= 67.85%. The classification model was then applied to an independent test set (RO cases= 7, ChRCC cases= 22) which could classify ChRCC from RO with a sensitivity of 74.3%, and a specificity of 65.2%. Conclusions Our results indicate that IMS can be a useful supplementary technique for the differentiation of ChRCC and ROs. Further studies will involve a larger number of cases in order to provide greater statistical power.