Hue-saturation-density (HSD) model for stain recognition in digital images from transmitted light microscopy

BACKGROUND: Transmitted light microscopy is used in pathology to examine stained tissues. Digital image analysis is gaining importance as a means to quantify alterations in tissues. A prerequisite for accurate and reproducible quantification is the possibility to recognise stains in a standardised manner, independently of variations in the staining density. METHODS: The usefulness of three colour models was studied using data from computer simulations and experimental data from an immuno-doublestained tissue section. Direct use of the three intensities obtained by a colour camera results in the red-green-blue (RGB) model. By decoupling the intensity from the RGB data, the hue-saturation-intensity (HSI) model is obtained. However, the major part of the variation in perceived intensities in transmitted light microscopy is caused by variations in staining density. Therefore, the hue-saturation-density (HSD) transform was defined as the RGB to HSI transform, applied to optical density values rather than intensities for the individual RGB channels. RESULTS: In the RGB model, the mixture of chromatic and intensity information hampers standardisation of stain recognition. In the HSI model, mixtures of stains that could be distinguished from other stains in the RGB model could not be separated. The HSD model enabled all possible distinctions in a two-dimensional, standardised data space. CONCLUSIONS: In the RGB model, standardised recognition is only possible by using complex and time-consuming algorithms. The HSI model is not suitable for stain recognition in transmitted light microscopy. The newly derived HSD model was found superior to the existing models for this purpose.     

Quantitative and qualitative analysis of stain color using RGB computer color specification

OBJECTIVE: To establish standardized Papanicolaou stain for cytology using RGB color specification. This new method was formerly used in DTP software application for computer color specification. STUDY DESIGN: RGB color specification was taken from a color film, optical constituents of which were made into computer software. Cell samples used in this study were from 100 sputum specimens stained with Papanicolaou stain. We analyzed the color tone of the cytoplasm of squamous cells in the smear. RESULTS: The R and B value of eosinophilic cells were demonstrated statistically by different values between normal, borderline atypical and malignant squamous cells. G and B values of light green-philic cells demonstrated a statistical difference between normal, borderline atypical and malignant squamous cells. No significant differences were found in RGB value between normal, borderline atypical and malignant squamous orangeophilic cells. CONCLUSION: Using our own method of analyzing Papanicolaou-stained sputum, a new quantitative and qualitative analysis of stain color for standardized Papanicolaou stain was introduced.     

Computerized image analysis as a tool to quantify infiltrating leukocytes: a comparison between high- and low-magnification images.

The purpose of the present study was to establish a rapid and reproducible method for quantification of tissue-infiltrating leukocytes using computerized image analysis. To achieve this, the staining procedure, the image acquisition, and the image analysis method were optimized. Because of the adaptive features of the human eye, computerized image analysis is more sensitive to variations in staining compared with manual image analysis. To minimize variations in staining, an automated immunostainer was used. With a digital scanner camera, low-magnification images could be sampled at high resolution, thus making it possible to analyze larger tissue sections. Image analysis was performed by color thresholding of the digital images based on values of hue, saturation, and intensity color mode, which we consider superior to the red, green, and blue color mode for analysis of most histological stains. To evaluate the method, we compared computerized analysis of images with a x100 or a x12.5 magnification to assess leukocytes infiltrating rat brain tumors after peripheral immunizations with tumor cells genetically modified to express rat interferon-gamma (IFN-gamma) or medium controls. The results generated by both methods correlated well and did not show any significant differences. The method allows efficient and reproducible processing of large tissue sections that is less time-consuming than conventional methods and can be performed with standard equipment and software.(J Histochem Cytochem 49:1073-1079, 2001)     

Sequential immunofluorescence staining and image analysis for detection of large numbers of antigens in individual cell nuclei.

BACKGROUND: Visualization of more than one antigen by multicolor immunostaining is often desirable or even necessary to explore spatial and temporal relationships of functional significance. Previously presented staining protocols have been limited to the visualization of three or four antigens. METHODS: Immunofluorescence staining was performed both on slices of formalin-fixed tissue and on cells in culture. Images of the stained material were recorded using digital imaging fluorescence microscopy. The primary and secondary antibodies, as well as the fluorophores, were thereafter removed using a combination of denaturation and elution techniques. After removal of the fluorescence stain, a new immunofluorescence staining was performed, visualizing a new set of antigens. The procedure was repeated up to three times. A method for image registration combined with segmentation, extraction of data, and cell classification was developed for efficient and objective analysis of the image data. RESULTS: The results show that immunofluorescence stains in many cases can be repeatedly removed without major effects on the antigenicity of the sample. CONCLUSIONS: The concentration of at least six different antigens in each cell can thus be measured semiquantitatively using sequential immunofluorescence staining and the described image analysis techniques. The number of antigens that can be visualized in a single sample is considerably increased by the presented protocol.     

Color image analysis for quantifying renal tumor angiogenesis

OBJECTIVE: To segment and quantify microvessels in renal tumor angiogenesis based on a color image analysis method and to improve the accuracy and reproducibility of quantifying microvessel density. STUDY DESIGN: The segmentation task was based on a supervised learning scheme. First, 12 color features (RGB, HSI, I1I2I3 and L*a*b*) were extracted from a training set. The feature selection procedure selected I2L*S features as the best color feature vector. Then we segmented microvessels using the discriminant function made using the minimum error rate classification rule of Bayesian decision theory. In the quantification step, after applying a connected component-labeling algorithm, microvessels with discontinuities were connected and touching microvessels separated. We tested the proposed method on 23 images. RESULTS: The results were evaluated by comparing them with manual quantification of the same images. The comparison revealed that our computerized microvessel counting correlated highly with manual counting by an expert (r = 0.95754). The association between the number of microvessels after the initial segmentation and manual quantification was also assessed using Pearson's correlation coefficient (r = 0.71187). The results indicate that our method is better than conventional computerized image analysis methods. CONCLUSION: Our method correlated highly with quantification by an expert and could become a way to improve the accuracy, feasibility and reproducibility of quantifying microvessel density. We anticipate that it will become a useful diagnostic tool for angiogenesis studies.     

A comparative study of quantitative stains for DNA in image cytometry.

In this study we examined the reproducibility of several stains used to measure nuclear DNA by image cytometry. The specimens were touch preparations of liver and testis from mouse and liver, intestine and brain from rat, fixed in either neutral formalin or Carnoy's solution. The tested stains included four Feulgen methods (pararosaniline, azure-A, thionin and acriflavine), the gallocyanine-chromalum stain and two fluorescent stains (acridine orange and propidium iodide). Absorbance measurements employed a video image analysis system; fluorescence measurements were from a scanning microspectrophotometer. The acriflavine-Feulgen stain was analyzed for both absorbance and fluorescence. All seven stains were quantitative for DNA and gave reproducible results. The absorbance measurements had a lower coefficient of variation (CV) than the fluorescence values. In a nested analysis of variance of the pararosaniline Feulgen stains, cell-to-cell variability accounted for 67% of the total variance; slide-to-slide, 9%; and batch-to-batch, 24%. These values did not change significantly when the staining was performed in an automatic staining machine. For DNA analysis using image cytometry, we conclude that the Feulgen staining technique is the most useful. In particular, acriflavine-Feulgen-stained cells fixed in Carnoy's fluid give the least variation between measurement values and the most accurate ratios between the separate ploidy groups. For fluorescence cytometry we recommend Carnoy's fixation and the acriflavine-Feulgen stain because of its narrow CV as compared to acridine orange and propidium iodide.     

Staining Mast Cells for Morphometric Evaluation on an Image Analysis System

Multiple skin sections from three nonhuman primates (Macaca mulatta) and three hairless guinea pigs (Cavia porcellus) were stained with 12 different histologic stains to determine whether mast cells could be selectively stained for morphometric analysis using an image analysis system (IAS). Sections were first evaluated with routine light microscopy for mast cell granule staining and the intensity of background staining. Methylene blue-basic fuchsin and Unna's method for mast cells (polychrome methylene blue with differentiation in glycerin-ether) stained mast cell granules more intensely than background in both species. Toluidine blue-stained sections in the guinea pig yielded similar results. Staining of the nuclei of dermal connective tissue was enhanced with the methylene blue-basic fuchsin and toluidine blue stains. These two stains, along with the Unna's stain, were further evaluated on an IAS with and without various interference filters (400.5-700.5 nm wavelengths). In both the methylene blue-basic fuchsin and toluidine blue stained sections, mast cell granules and other cell nuclei were detected together by the IAS. The use of interference filters with these two stains did not distinguish mast cell granules from stained nuclei. Unna's stain was the best of the 12 stains evaluated because mast cell granule staining was strong and background staining was faint. This contrast was further enhanced by interference filters (500.5-539.5 nm) and allowed morphometric measurements of mast cells to be taken on the IAS without background interference.     

The influence of Romanowsky-Giemsa type stains on nuclear and cytoplasmic features of cytological specimens

The aim of the present study was to compare the staining pattern of the standard azure B-eosin Y stain with commercial May-Grünwald-Giemsa (MGG) stains on cytological specimens by means of high resolution image analysis. Several cytological specimens (blood smears, abdominal serous effusions, bronchial scrape material) were air dried, methanol fixed and stained with the standard azure B-eosin Y stain and with commercial May-Grünwald-Giemsa stains. Integrated optical density (IOD) and colour intensities of cell nuclei and cytoplasm were measured with the IBAS 2000 image analyser. Commercial MGG stains gave much higher coefficients of variation for all parameters than the standard stain. Reproducibility of cell nuclei segmentation versus cytoplasm was significantly better for the standard stain. Contamination of the standard stain with methylene blue partly copied the staining pattern of commercial stains. The standard azure B-eosin Y stain is recommended for high resolution image analysis (HRIA) of cytological samples.     

Viability assessment of honey bee, Apis mellifera, sperm using dual fluorescent staining

Since the development of instrumental insemination of honey bee (Apis mellifera) queens in the 1930s, there has been interest in the evaluation and in vitro storage of semen. Several fluorescent stains, when used in combination, have been effectively used to assess sperm viability in mammalian and avian species. Our objectives were to test two combinations of living:dead fluorescent stains, SYBR-14 with propidium iodide (PI), or Calcein-AM with PI, and validate the use of these probes with honey bee sperm. SYBR-14 is a nuclear stain producing green fluorescence of the DNA in living sperm, Calcein-AM is a membrane-permeant esterase substrate staining entire sperm green, and PI is a traditional dead cell stain giving a contrasting red color. Both living stains fluoresced bee sperm, but the SYBR-14:PI produced a clearer distinction between the living and dead sperm. A graduated series of known living:dead sperm proportions was used to validate the accuracy of the stains for determining sperm viability in honey bees.     

The degradation of Romanowsky-type blood stains in methanol.

The oxidative demethylation of Romanowsky-type stains in methanol has been examined quantitatively with respect to its effect upon the staining of blood smears. Spectral changes in bound dye, observed through two color filters, have been measured for the nuclei and cytoplasm of segmented neutrophils and monocytes utilizing the LARC automated differential analyzer. Stain decomposition in methanol results in a large loss in staining intensity with little change in color. The loss in intensity has been correlated with the observed spectral changes in the degraded stain. High-performance liquid chromatographic analysis of degraded stain samples has shown the products of methanolic degradation to be different from those obtained in aqueous polychroming reactions. To maintain a stain of defined thiazine dye composition and thus defined staining properties, refrigeration is recommended.     

Analysis of stained objects in histological sections by spectral imaging and differential absorption.

We describe a new light microscopic imaging system and method to perform high through put color image analysis on histological tissue sections. The system features a computer-controlled, random-access liquid crystal tunable filter and high-resolution digital camera on a conventional brightfield microscope. For any combination of stains, the method determines the spectral transmittance of each stain on the slide and selects two or more wavelengths at which the differential absorption between stain and counterstain is greatest and the exposure time is reasonably short. Flatfield corrected digital images at these wavelengths are acquired and divided to produce a gray scale ratio image. The ratio image is calculated such that the stained features of interest are highlighted above a uniform background and the counterstained features are highlighted below background. Image threshold procedures using either visual inspection or a threshold value determined by the image mean intensity and standard deviation are used to segment the stained features of interest for subsequent morphometry. Results are presented for peroxidase-AEC-labeled tumor tissue and trichrome-stained biomaterial implant tissues. In principle, the method should work for any combination of colored stains. (J Histochem Cytochem 47:1307-1313, 1999)     

Protein extraction and comparison of stain protocols for analysis of two-dimensional electrophoresis gels

Protein extraction and the proteome of Lactobacillus delbrueckii subsp. bulgaricus were studied using different stains. The reversible silver staining technique was shown to be more sensitive than the irreversible silver stain. Coomassie colloidal was demonstrated to be as sensitive as reversible silver stain; however, the Coomassie colloidal blue solution developed a higher background and for sample preparation was more time-consuming.     

Graphical methods for quantifying macromolecules through bright field imaging

Bright field imaging of biological samples stained with antibodies and/or special stains provides a rapid protocol for visualizing various macromolecules. However, this method of sample staining and imaging is rarely employed for direct quantitative analysis due to variations in sample fixations, ambiguities introduced by color composition and the limited dynamic range of imaging instruments. We demonstrate that, through the decomposition of color signals, staining can be scored on a cell-by-cell basis. We have applied our method to fibroblasts grown from histologically normal breast tissue biopsies obtained from two distinct populations. Initially, nuclear regions are segmented through conversion of color images into gray scale, and detection of dark elliptic features. Subsequently, the strength of staining is quantified by a color decomposition model that is optimized by a graph cut algorithm. In rare cases where nuclear signal is significantly altered as a result of sample preparation, nuclear segmentation can be validated and corrected. Finally, segmented stained patterns are associated with each nuclear region following region-based tessellation. Compared to classical non-negative matrix factorization, proposed method: (i) improves color decomposition, (ii) has a better noise immunity, (iii) is more invariant to initial conditions and (iv) has a superior computing performance.     

The Degradation of Romanowsky-Type Blood Stains in Methanol

The oxidative demethylation of Romanowsky-type stains in methanol has been examined quantitatively with respect to its effect upon the staining of blood smears. Spectral changes in hound dye, observed through two color filters, have been measured for the nuclei and cytoplasm of segmented neutrophils and monocytes utilizing the LARC™ automated differential analyzer. Stain decomposition in methanol results in a large loss in staining intensity with little change in color. The loss in intensity has been correlated with the observed spectral changes in the degraded stain. High-performance liquid chromatographic analysis of degraded stain samples has shown the products of methanolic degradation to be different from those obtained in aqueous polychroming reactions. To maintain a stain of defined thiazine dye composition and thus defined staining properties, refrigeration is recommended.     

Differentiation of bone and soft tissues in formalin-fixed, paraffin-embedded tissue by using methylene blue/acid fuchsin stain

OBJECTIVE: To find a staining method for formalin-fixed, paraffin-embedded tissue that would distinguish bone from surrounding soft tissues, including muscle, periosteal tissue and bone marrow. STUDY DESIGN: A variety of stains were tested and compared with hematoxylin-eosin. The potential value of any given stain was evaluated based on its ability to stain bone and soft tissues different colors or shades that could be readily identified in photomicrographs. Stains were evaluated using both endochondral (tibia) and intramembranous bone (calvaria) samples. RESULTS: In contrast to standard hematoxylin-eosin stain, which stains both bone and soft tissues pink, the methylene blue/acid fuchsin stain demonstrates remarkable contrast between bone and other tissues. Methylene blue/acid fuchsin stained bone bright pink and the surrounding soft tissues blue-purple. CONCLUSION: In addition to the superior staining properties of methylene blue/acid fuchsin, other benefits of this stain include its stability, ease of use and low cost. This stain has many potential applications in the study of erosive bone disease in humans and also in animal models for research.     

Quantitative Determination of Murine Dermal Elastic Fibers by Color Image Analysis: Comparison of Three Staining Methods

We compared three different staining methods to determine if the dermal elastic fiber content of the HRS/Skh-1 hairless mouse could be accurately measured by color image analysis. Comparisons were made among Klig-man's modification of Luna's mast cell stain for elastin, Unna's orcein stain with or without potassium permanganate preoxidation, and Gomori's aldehyde fuchsin stain with potassium permanganate preoxidation. The color image analysis system could be used to identify and quantify murine dermal elastin fibers in sections stained by all three methods. Gomori's aldehyde fuchsin stain with preoxidation demonstrated twice the content of dermal elastic fibers demonstrated by either Kligman's modification of Luna's mast cell stain or Unna's orcein stain with or without preoxidation. Gomori's aldehyde fuchsin method with preoxidation should be considered the stain of choice for evaluating murine dermal elastic fiber content.     

Standardization of the Papanicolaou stain. I. A comparison of five nuclear stains.

The staining characteristics of five nuclear stains used in a Papanicolaou staining procedure were investigated. Alcohol-fixed cervical smears were stained with a modified Papanicolaou procedure using hematoxylin, alcoholic thionin bromide, alcoholic Victoria blue B, gallocyanin or the thionin Feulgen reagent (thionin-SO2) as the nuclear stain. The same anionic counterstain was used for all slides, and the optical densities of cell nuclei and cytoplasm were measured with the IBAS 2000 image analyzer. Alcoholic thionin gave the most intense nuclear stain, with a very high reproducibility of the staining pattern. Hematoxylin showed the highest coefficient of variation of the staining intensity. Both hematoxylin and gallocyanin gave some nonspecific cytoplasmic staining. Thionin-SO2 allowed a quantitative assessment of DNA, but gave a low staining intensity. Staining with the metal complex dyes interfered with subsequent staining with the pararosaniline Feulgen reagent. Alcoholic thioinin is thus recommended as a nuclear stain for cervical cytology in the Papanicolaou procedure, both for image analysis and for visual microscopy.