Erich K, Sammour DA, Marx A, Hopf C.
Biochim Biophys Acta. 2016 Sep 6. pii: S1570-9639(16)30182-0. doi: 10.1016/j.bbapap.2016.08.020.
On-slide digestion of formalin-fixed and paraffin-embedded human biopsy tissue followed by mass spectrometry imaging of resulting peptides may have the potential to become an additional analytical modality in future ePathology. Multiple workflows have been described for dewaxing, antigen retrieval, digestion and imaging in the past decade. However, little is known about suitable statistical scores for method comparison and systematic workflow standardization required for development of processes that would be robust enough to be compatible with clinical routine. To define scores for homogeneity of tissue processing and imaging as well as inter-day repeatability for five different processing methods, we used human liver and gastrointestinal stromal tumor tissue, both judged by an expert pathologist to be >98% histologically homogeneous. For mean spectra-based as well as pixel-wise data analysis, we propose the coefficient of determination R2, the natural fold-change (natFC) value and the digest efficiency DE% as readily accessible scores. Moreover, we introduce two scores derived from principal component analysis, the variance of the mean absolute deviation, MAD, and the interclass overlap, Joverlap, as computational scores that may help to avoid user bias during future workflow development. This article is part of a Special Issue entitled: MALDI Imaging.
Copyright © 2016 Elsevier B.V. All rights reserved.
FFPE tissue; Gastrointestinal stromal tumor; MALDI imaging; Multivariate statistics; On-tissue digestion; Repeatability; Standardization; ePathology