Depth-Sensitive Raman Spectroscopy of Intact Formalin-Fixed and Paraffin-Embedded Tissue Blocks for Objective Diagnosis of Cancer- An Exploratory Study

Abstract

Histopathology, the current “gold standard”, is prone to human errors as it depends on expert interpretation of the microscopically derived cellular and sub-cellular information for tissue diagnosis. Further, this light microscope based approach requires preparation of appropriately stained specimens of micro-thin tissue sections from the formalin-fixed and paraffin-embedded (FFPE) blocks of tissue samples. We report a method that provides quantitative feedback about tissue diagnosis by measuring depth-sensitive Raman spectra from the intact FFPE tissue blocks without requiring preparation of any thin tissue sections or any other processing. The FFPE blocks of pathologically certified cancerous and normal breast tissues were used for validating the approach. The measured depth-sensitive Raman spectra were mathematically de-paraffinized for retrieving the characteristic tissue Raman signatures using scaled-subtraction. A multivariate analysis of the scaled-subtracted, depth-sensitive Raman spectra employing a probability-based diagnostic algorithm developed using the framework of sparse multinomial logistic regression (SMLR) provided a sensitivity and specificity of up to 100% towards cancer based on leave-one-block-out cross validation. The results of this exploratory study suggest that depth-sensitive Raman spectroscopy along with a multivariate statistical algorithm can provide a valuable alternate diagnostic modality in clinical pathology setting for discriminating cancerous from normal FFPE tissue blocks.