3D presentation of the nuclear cell features in quantitative cytometry

The use of an image cytometer in analysis of smears and needle aspirates provides valuable information to a cytologist. It allows to combine the overall impression, formed by visually inspecting the cells, with measured and numerically expressed nuclear cell features. Both types of information can be used efficiently only if presented to the expert in an appropriate way. Cell images (as they are seen with a microscope) are easily analysed by the experts. However, measured nuclear features can not be presented as a list of numerical values. Instead, an user interface should be developed, providing graphical presentation of the nuclear features. It should show as much information as possible and provide a comprehensive link between nuclear features and cell images. The user interface described in this paper shows nuclear features in three dimensions. It is based on a perspective projection of the three dimensional feature space onto a two dimensional surface. It allows the user to dynamically change the perspective, i.e., to look at the virtual three dimensional structure from different viewpoints. Each nucleus is represented by a single object in the three dimensional space. When an object in the three dimensional feature space is selected, the image (or the visual appearance) of the corresponding cell is shown. When a nucleus image is selected, its position in the feature space is highlighted. This provides an interconnection between nuclear cell features and cell images allowing simultaneous analysis of both types of information.

Discrimination of respiratory dysfunction in yeast mutants by confocal microscopy, image, and flow cytometry

Living yeast cells can be selectively stained with the lipophilic cationic cyanine dye DiOC6(3) in a mitochondrial membrane potential-dependent manner. Our study extends the use of flow cytometric analysis and sorting to DiOC6(3)-stained yeast cells. Experimental conditions were developed that prevented the toxic side effect of the probe and gave a quantitative correlation between fluorescence and mitochondrial membrane potential, without any staining of other membranes. The localization of the fluorochrome was checked by confocal microscopy and image cytometry. The mitochondrial membrane alterations were also tested through cardiolipin staining with nonyl acridine orange. Differences in light scattering and in fluorescence were detected in mutants (rho-, rho degrees, mit-, or pet-) and wild-type (rho+mit+) populations of yeast. The dye uptake of respiratory-deficient yeast strains was significantly reduced as compared to that of the wild-type. Application of an uncoupler (mCICCP), which collapsed the mitochondrial membrane potential (alphapsi(m)), led to a drastic reduction of the dye uptake. It was observed that a decrease in deltapsi(m), was usually correlated with a decrease in cardiolipin stainability by nonyl acridine orange (NAO). Quantitative flow cytometry is a fast and reproducible technique for rapid screening of yeast strains that might be suspected of respiratory dysfunction and/or mitochondrial structural changes. We give evidence that it is an adequate method to characterize and isolate respiratory mutants through sorting procedure, with selective enrichment of the population studied in respiring or non-respiring yeast cells. Confocal microscopy and image cytometry corroborate the flow cytometry results.

Hematoxylin staining in quantitative DNA cytometry: an image analysis study

The suitability of commercial and pure aluminum-hematein for quantitative DNA image cytometry was investigated. Cervical smears, breast cancer aspiration biopsies, and rabbit liver tissue imprints were stained with Mayer's and Harris' al-hematein with variable staining times and dye concentrations. Pure and commercial hematoxylin was used. Nucleic acids were removed by enzyme digestion or by HCl-hydrolysis. A standard Feulgen stain served as control. DNA-polyacrylamide films were used as staining models. Absorption was measured using a VIDAS image analyzer. DNA in liver cell nuclei was not stained in a stoichiometric dye-DNA ratio. Sequential staining of cervical smears with hematein followed by the Feulgen reaction gave a covariance between 0.77 and 0.88 for IOD. Photometric errors due to unspecific RNA or protein staining were remarkable. Harris' and Mayer's hematein gave comparable results. Pure hematein gave slightly better results than commercial batches. DNA staining in model films was not quantitative with hematein. Al-hematein should therefore not be used for quantitative DNA cytometry.

Bleaching of melanin before image cytometry of the DNA content of pigmented skin tumors

OBJECTIVE

To study the effect of bleaching of melanin with KMnO4 on the results of DNA image cytometry in pigmented skin tumors.

STUDY DESIGN

Image cytometry of nuclear DNA content was performed on sections from 14 melanocytic skin tumors stained with Feulgen stain both with and without prior bleaching with KMnO4.

RESULTS

The nuclear staining intensity of Feulgen stain was lower in the bleached sections, but this did not significantly affect the evaluation of ploidy. Heavy pigmentation caused some false peaks in the histograms (4 of 28 measurements made on unbleached slides).

CONCLUSION

Bleaching of sections with KMnO4 can be useful when heavy melanin pigmentation would make DNA measurements impossible or difficult in image analysis cytometry. Bleaching is not advisable when only lightly pigmented tumors are analyzed if nuclei obscured by any pigment granules are to be avoided. In large series containing both bleached and nonbleached specimens, statistical analysis of these groups should be separated.

Disease-free survival of node-positive breast cancer patients. Improved prognostication by cytometrical parameters

Feulgen stained cytologic samples from 225 node-positive breast cancers were investigated by means of an image analysis system. From each tumor sample, 100 cells were scanned and several DNA, morphometrical and textural parameters were evaluated. The meaning of the cytometric parameters for prediction of distant metastases within five years was investigated by the stepwise Cox regression analysis. Most of the investigated DNA- and morphometrical parameters, as well as one textural feature, showed a significant univariate correlation with the clinical course. In the multivariate approach, the lymph node status (pN) was the strongest prognostic factor, followed by the histogram type, the tumor size (pT) and a textural parameter (heterochromatin area). By the linear combination of these selected variables a multivariate prognostic factor was calculated for each individual patient. Using this factor, the patients could be splitted into four groups according to their risk for distant metastases. For this, the continuous range of the multivariate factor was subdivided so that about 35% of the patients were in the middle groups and about 15% of the patients in each of the border groups with highest and lowest factors, respectively. Thus a low risk group (lowest factors) of node-positive patients could be identified with a 5-year distant recurrence-rate of only 6.5%, as well as a group of patients with a considerably worse prognosis (highest factors) and a distant recurrence-rate of 67%. Therefore, DNA, morphometrical and textural parameters can provide powerful prognostic information in node-positive breast carcinomas. Using the multivariate combination of clinical and relevant cytometrical parameters may allow a more appropriate selection of patients for adjuvant therapy.

Comparison of flow and image cytometry for DNA content analysis of fresh and formalin-fixed, paraffin-embedded tissue in breast carcinoma

DNA ploidy and S-phase fraction are considered to be prognostic variables in breast carcinoma. DNA content of 35 cases of breast carcinoma of varying histologic types and nuclear grades was analyzed by flow cytometry and image analysis in both fresh and formalin-fixed, paraffin-embedded tissue. Fresh cell and deparaffinized nuclear suspensions were used for flow cytometry. Fresh and deparaffinized tumor tissue samples were used for image analysis. The results of analysis for DNA ploidy, DNA index of DNA aneuploid Go/G1 peaks, and S-phase fraction were compared in different tissue preparations for both techniques. The two techniques produced comparable DNA ploidy results with both fresh and formalin-fixed, paraffin-embedded tissue. Sensitivity for detection of DNA aneuploidy was somewhat greater by image analysis, particularly in deparaffinized tissue. There was 89% agreement in detection of DNA aneuploidy by flow cytometry in fresh and paraffin-embedded, formalin-fixed tissue; the coefficients of variation of the DNA diploid Go/G1 peaks were much wider in the latter. In image analysis there was 91% agreement between fresh and fixed specimens. Agreement between the flow cytometry and image analysis in fresh specimens was 91%; in deparaffinized nuclear suspensions it was 94%. There is a high degree of correlation between the values of DNA index of DNA aneuploid Go/G1 peaks; the estimates of S-phase fraction are much more variable. Results also show a good correlation of the DNA ploidy with the nuclear grades.(ABSTRACT TRUNCATED AT 250 WORDS)

Impact of the preparation technique on DNA content measured by image analysis in early stage human testis cancer

Current clinical staging which includes serum tumour markers and imaging techniques fails to identify 30-40% of clinical stage I nonseminomatous germ cell testicular tumour (NSGCT) patients who have occult metastatic disease at time of orchiectomy and who will, therefore, develop clinically evident metastases, usually within two years of follow-up. Therefore, there is a real clinical need to evaluate new biological parameters of the primary tumour which might be useful as predictors for occult metastatic disease. Some investigators have described that DNA content measured by image cytometry is of prognostic value in early stage NSGCT to detect patients at risk for occult metastatic disease. However, optimal preparatory techniques are mandatory in establishing new tumour biological markers in order to obtain reliable and reproducible results. This study has analyzed the impact of the sedimentation technique in comparison to the cytocentrifugation technique on DNA measurement in early stage NSGCT obtained by image cytometry. Different tissue blocks of formalin fixed, paraffin embedded primary testicular tumours (NSGCT) of 50 clinical stage I patients, who underwent retroperitoneal lymph node dissection between 1985 and 1989, were analyzed. Thirty (60%) patients had histologically proven lymph node involvement (pathological stage B), whereas 20 (40%) patients (pathological stage A) had neither lymph node metastases nor tumour recurrence during follow-up. The samples were prepared according to a modified Hedley technique: Individual tissue digestion times were monitored closely to avoid overdigestion. The times varied from 30 to 60 min depending on the constituents of the tissue section. Prolonged digestion times did not correlate with poor quality of the preparations and brief digestion times did not always yield optimal specimens. The impact of two different techniques (cytocentrifugation and gravity sedimentation) on the Feulgen staining results were compared. Cytocentrifuged samples consistently provided larger and paler nuclei with less well defined borders compared to slides from the same cell suspension processed by the sedimentation technique. Nuclei from pathologic stage II samples were more vulnerable to cytocentrifuge alteration than those of stage I. According to the results obtained in this study, the sedimentation slide preparation technique should be preferred for DNA ICM in NSGCT, and possibly in other human malignancies as well.

Quantitative DNA analysis in renal cell carcinoma. Comparison of flow and image cytometry

In 70 renal cell carcinomas, nuclear DNA content was determined by means of flow cytometry (FCM) and image cytometry (ICM). The two methods produced comparable results as to DNA tumor ploidy (DNA tumor stemlines, DNA index): 14 of the tumors were tetraploid or aneuploid and 56 diploid. Results with the two methods were also comparable in a comparison of DNA ploidy with degree of tumor malignancy (tumor grade G1-3) and local tumor spread stage (pT stage). As a consequence, both methods appear suitable as means of determining DNA tumor ploidy and thus of formulating a prognosis in renal cell carcinoma. Renal cell carcinomas with diploid stemlines tend to be characterized by local growth, whereas tetraploid or aneuploid tumors show a tendency toward perirenal spread and venous invasion.

Comparison between image and flow cytometry. A priori factors that influence technique

Comparisons between flow cytometry (FCM) and image cytometry (ICM) have found a high concordance rate in pancreatic tissue, with some discrepancies between the two procedures. This study utilized 40 cases of chronic pancreatitis, primary pancreatic adenocarcinoma and metastatic pancreatic adenocarcinoma to determine the concordance rate between the two procedures. The reasons for discrepancies were identified and subsequently used to establish methods for a priori determination of which procedure to use. Using the FACSCAN flow cytometer and the CAS 200 on appropriately stained specimens that were disaggregated from 50-micron sections, we achieved a concordance rate of r = .878 (P < .01) after removing outliers. Thirty-one of 40 cases matched DNA content, and 9 cases had discrepant results. These discrepant cases were evaluated with factor analysis, in part because initial observations suggested that the variables evaluated could be combined into unifying concepts. The nine measured variables were compressed into three factors, which accounted for 68% of the variation observed between the two methods. Readily evaluated features, on a case-by-case basis, including tumor/nontumor ratios, accounted for the largest proportion of this variation. These findings suggest that tumor/nontumor cell ratios in hematoxylin-eosin-stained sections may provide adequate a priori information to direct the choice of either FCM or ICM to measure DNA ploidy in pancreatic tissue.

Imaging of L-type Ca2+ channels in olfactory bulb neurones using fluorescent dihydropyridine and a styryl dye

We have imaged the fluorescence of dihydropyridine-Bodipy (fDHP) in cultured olfactory bulb neurones in order to investigate the subcellular distribution of L-type calcium channels in these neurones. The neurones were stained with both fDHP and the voltage-sensitive styryl dye RH414. The fluorescence emission maxima of these dyes were in the green and red ranges of the spectrum, respectively, and were recorded by the 2 photomultiplier channels of a laser scanning microscope. The fDHP images were ratioed with the RH414 images taken simultaneously. The resulting ratio images revealed the spatial distribution of the surface density of L-type calcium channels. This density was highest on somata, in particular at the base of dendrites, and decreased with the distance from this maximum. Two classes of dendritic L-type channel distributions were observed: one with a homogeneously low density and another one with a characteristic gradient of the L-type channel density along proximal dendrites. This subcellular localization of L-type calcium channels is discussed in the light of specific functional roles.

Nuclear texture measurements in image cytometry

DNA image cytometry is widely used in cytopathology as a means to obtain objective information concerning the diagnosis and prognosis of human cancer. Using specially designed devices, the high resolution spatial and photometric information is available in the images of a microscopic field. If quantitative DNA specific stains are used the chromatin distribution in the cell nuclei can be measured, which is one of the critical features for cytopathological analysis. In normal cells, changes in the chromatin appearance reflect changes in the activation patterns of genes. In tumors, dramatic changes in the nuclear chromatin appearance are common and have been associated with the progression of the disease. Features describing the chromatin distribution pattern are referred to as texture features. Nuclear texture features are sensitive to the differences between the various descriptive classes of chromatin patterns. In this paper we discuss the main categories of nuclear texture measurements. Texture features can be roughly divided into the following categories: 1) descriptive statistics of chromatin distribution; 2) discrete texture features; 3) range extreme; 4) markovian; 5) run length and 6) fractal texture features. Representative features of each of the above categories are discussed together with mathematical formulas, simple figures for explanation as well as images of typical cells which differ significantly in some texture features. Key references are also provided.

Syntactic structure analysis

Syntactic structure analysis is a relatively new technique in diagnostic quantitative pathology that, based on the graph theory, can provide quantitative information on tissue architecture. This article gives an introduction to syntactic structure analysis and discusses the potential of syntactic structure analysis in diagnostic quantitative pathology. In addition, some examples of its application in histo- and cytopathology are presented and it is speculated which other applications the user of syntactic structure may find for quantitation of architecture in diagnostic quantitative pathology.

Unbiased stereologic techniques for practical use in diagnostic histopathology

Grading of malignancy by the examination of morphologic and cytologic details in histologic sections from malignant neoplasms is based exclusively on qualitative features, associated with significant subjectivity, and thus rather poor reproducibility. The traditional way of malignancy grading may therefore not be optimal regarding the scientific and clinical value in selecting specific therapy and making prognostic forecast for the individual patient with a solid cancer. A lot of examples in the international literature document the urgent need for replacing morphologically based techniques by introducing quantitative techniques in the histopathologic discipline of malignancy grading. Unbiased stereologic methods, especially based on measurements of nuclear three-dimensional mean size, have during the last decade proved their value in this regard. In this survey, the methods are reviewed regarding the basic technique involved, sampling, efficiency, and reproducibility. Various types of cancers, where stereologic grading of malignancy has been used, are reviewed and discussed with regard to the development of a new objective and reproducible basis for carrying out prognosis-related malignancy grading of solid tumors. This new, unbiased attitude to malignancy grading is associated with excellent virtues, which ultimately may help the clinician in the choice of optimal treatment of the individual patient suffering from cancer. Stereologic methods are not solely applicable to the field of malignancy grading. Surgical pathologists can use a wide range of these techniques in the routine of diagnostic and scientific work. It is not the intention, however, to give an extensive review of the methodology, but two other techniques will be covered: Unbiased estimation of volume of, and total tumor cell number within, solid neoplasms. Three-dimensional tumor size may be of relevance in staging cancer, whereas total tumor cell number may be of interest in the scientific work with malignancies and treatment of cancer.