Quantification of histochemical stains using whole slide imaging: development of a method and demonstration of its usefulness in laboratory quality control.

AIMS: Histochemical staining of tissue is a fundamental technique in tissue diagnosis and research, but it suffers from significant variability. Efforts to address this include laboratory quality controls and quality assurance schemes, but these rely on subjective interpretation of stain quality, are laborious and have low reproducibility. We aimed (1) to develop a method for histochemical stain quantification using whole slide imaging and image analysis and (2) to demonstrate its usefulness in measuring staining variation. METHODS: A method to quantify the individual stain components of histochemical stains on virtual slides was developed. It was evaluated for repeatability and reproducibility, then applied to control sections of an appendix to quantify H&E staining (H/E intensities and H:E ratio) between automated staining machines and to measure differences between six regional diagnostic laboratories. RESULTS: The method was validated with <0.5% variation in H:E ratio measurement when using the same scanner for a batch of slides (ie, it was repeatable) but was not highly reproducible between scanners or over time, where variation of 7% was found. Application of the method showed H:E ratios between three staining machines varied from 0.69 to 0.93, H:E ratio variation over time was observed. Interlaboratory comparison demonstrated differences in H:E ratio between regional laboratories from 0.57 to 0.89. CONCLUSIONS: A simple method using whole slide imaging can be used to quantify and compare histochemical staining. This method could be deployed in routine quality assurance and quality control. Work is needed on whole slide imaging devices to improve reproducibility.

Effect of water softening on the colour intensity of routine haematoxylin and eosin stain.

Haematoxylin and eosin (H&E) is the most popular routine stain used in pathology laboratories for highlighting cellular structures. To study the effect of tap water'softening' (i.e. calcium extraction) on H&E stains, 5 sets of slides from 30 different paraffin-embedded human pathologic tissue blocks were prepared in the same way except for washing with 5 different types of water. Slides washed in untreated tap water showed the best results concerning differentiation and colour intensity, while slides washed with softened or other treated water showed poorer degrees of differentiation and colour intensity. The worst results were obtained from slides washed with water containing sodium bicarbonate. Low calcium and magnesium ions and high sodium ions in soft water adversely affect the results of routine H&E stain.

The compactness of populations of neutrophilic leucocytes as revealed by the new azure B-eosine stain, Pappenheim's stain and the peroxidase reaction.

The new azure B-eosine stain shows a greater inhomogeneity of the granule population of human neutrophil leucocytes than Pappenheim's stain. This difference seems to be dependent on the number of azure granules within the single neutrophil.

Standardization of the Romanowsky staining procedure: an overview.

Commerically available Romanowsky blood stains are variable mixtures of thiazein dyes and brominated fluorescein derivatives with varying degrees of metallic salt contamination in a number of different solvent systems. There is a need for standardized Romanowsky stains of constant composition, which, when used in conjunction with a carefully controlled specimen preparation technique, should give consistent performance. Such a preparation system would be of great value to hematologists in general and would be essential to the validity of data obtained by the digital processing of blood cell images. It is possible to prepare standardized Romanowsky stains as mixtures of two or three dye components, namely, eosin Y, azure B and methylene blue, although azure B has only recently become commercially available at an acceptable degree of purity. The logistic problems of stain standardization are discussed.