Effect of hematoxylin and eosin staining on tissue linear birefringence imaging: a case study

Abstract

Hematoxylin-eosin staining is the gold standard for anatomic pathology as the general stain for histological examination of human tissue. Birefringent property is a valuable and commonly used metric for characterizing the anisotropic structure of tissue. When assessing the birefringent property of tissue slices, unstained slice is often preferred to avoid the potential effects of staining dyes. However, obtaining high-quality adjacent tissue samples is often difficult due to possible distortion, rotation, and uneven thickness of the tissue slices during sample preparation. In this study, we measure the Mueller matrix images of the unstained, hematoxylin stained, eosin stained, and hematoxylin-eosin stained tissue slices using the transmission Mueller matrix microscope developed in our previous study. Then the Mueller matrix polar decomposition parameter of linear phase retardance is calculated to describe the linear birefringent property of tissue. It is shown that the linear birefringent property is enhanced by the combination of dye molecules and birefringent structures, which may be due to the parallel attachment of the dye molecules to the fibers. To evaluate in detail the contrast consistency of tissue linear birefringence imaging from differently stained slices, we perform a similarity analysis using the Bhattacharyya coefficient, which demonstrated that hematoxylin, eosin, and hematoxylin-eosin staining do not generate linear birefringent property on region of tissue samples without birefringent structures. Therefore, it can be acceptable to prepare hematoxylin-eosin stained slice to obtain both bright field image and linear phase retardance image simultaneously for digital pathologic diagnosis.