Image morphometry of acute leukemias. Comparison between lymphoid and myeloid subtypes

OBJECTIVE: To determine if image morphometry has any role in distinguishing blasts of acute lymphoblastic leukemia (ALL-L2) from those of acute myeloid leukemia (AML-M1) and (AML-M2). STUDY DESIGN: Ten cases each of ALL-L2, AML-M1 and AML-M2 diagnosed according to the French-American-British criteria were studied. In all cases May-Grünwald-Giemsa-stained bone marrow aspiration smears were obtained. At least 100 blast cells from each case were subjected to analysis randomly with an image cytometer using Leica Quantimet 600 software (Cambridge, U.K.). The area, convex area, length, width, perimeter, convex perimeter, roundness, total optical density, average optical density and pixel grey value variance of the nuclei were measured by random selection of cells using a 40:1 objective (1 pixel = 0.446 micron). RESULTS: Mann Whitney's nonparametric test showed that there was considerable overlap of morphometric variables between the 3 subtypes. Though statistical significance was found in "roundness" between blasts of AML-M1 and ALL-L2, power analyses (sample size of 100 blasts of each subtype) did not show sufficient power for this variable. However, between blasts of ALL-L2 and AML-M2, "average optical density" and "pixel grey value variance" were statistically significant with full power using power analyses. CONCLUSION: Image morphometry may be helpful in differentiating blasts from lymphoid and myeloid leukemic subtypes.     

肿瘤核仁形成区形态定量诊断的误差及规范化分析

本文采用包括自动图象分析技术在内的ＡｇＮＯＲ形态定量学方法,以大肠肿瘤为模型,进行了ＡｇＮＯＲ定量形态学研究的误差分析,以探讨肿瘤ＡｇＮＯＲ定量诊断规范化的可能。结果表明,染色条件、视场目标选择和细胞计数量是引起ＡｇＮＯＲ定量诊断的主要误差;恒定染色环境,正确选择欲测细胞及测定足够量的细胞是使ＡｇＮＯＲ形态定量诊断规范化的途径。     

Glomerular morphometry in biopsy evaluation of minimal change disease, membranous glomerulonephritis, thin basement membrane disease and Alport's syndrome

OBJECTIVE: To assess the role of glomerular morphometry in biopsy evaluation in renal disorders in addition to conventional diagnostic procedures. STUDY DESIGN: The study includes 10 cases each of minimal change disease (MCD), idiopathic membranous glomerulonephritis (idiopathic MGN), thin basement membrane disease (TBMD) and Alport's syndrome. Renal biopsies for normal study were obtained from age- and sex-matched autopsy cases without any renal disorder, confirmed histologically and ultrastructurally. Glomerular morphometry was performed by semiautomatic procedure using Quantimet-600 image analysis system (Leica, Cambridge, United Kingdom). RESULTS: Morphometric findings revealed significant increase in glomerular "diameter and area" and "tuft diameter and area" in patients of idiopathic MGN, but no significant difference was found in patients of MCD, TBMD and Alport's syndrome. Evaluation of glomerular volume fractions revealed a decrease in capillary space volume fraction and an increase in "membranes and mesangial matrix" volume fraction in patients with idiopathic MGN. Significant decrease in capillary space volume fraction was also observed in patients of MCD. Patients with Alport's syndrome showed variable changes. CONCLUSION: Glomerular morphometry could be considered as an adjunct to the diagnostic armamentarium of light microscopy, immunofluorescence and electron microscopy because it provides deep insight into quantitative parameters.     

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)     

Stereology and morphometry in histopathology. Principles of application

The role of stereology and morphometry (including image analysis) in histopathology is considered. Diagnostic histopathology has its own character, which should be considered when quantitative methods are applied. Three different types of studies in histopathology can be distinguished: prospective, retrospective and routine diagnostic studies. The variation in these studies is due to differences in the methods and differences in how the methods are applied. The overall variation is smallest in prospective studies and largest in routine diagnostic studies. Statistical classification of the overall variation does not cover all relevant aspects of the situation in which we study the sample from one individual patient. In that situation, we do not get support from the samples of other patients. In fact, we should speak about group morphometry (statistical morphometry) when samples from several patients are studied and diagnostic morphometry when the sample from one individual patient is under investigation. In the former case, simple morphometric methods may be applicable (the researcher is interested in mean values). In the latter case, one would often like to analyze the sample in more detail. This is why the diagnostic situation may benefit from computerized image analysis, which allows collection of large amounts of data in a short time. On the other hand, the great variation between the pathologic and the normal can allow the use of simple methods. The above basic principles need be considered before morphometric and stereologic methods are applied. These methods are simple in practice and they should be included in the training of all pathologists.     

Clinical applications of image cytometry to human tumour analysis

Image cytometry (ICM) is widely applied to the automated screening, the detection, the diagnosis, the classification, the prognosis and the therapeutic follow-up of different types of cancers (breast, bladder, cervix,...). This review describes the analysis methods and the applications of nuclear image analysis, the determination of DNA content and the analysis of morphometry and of nuclear texture. DNA content analysis can contribute to a prognostic information in addition to other prognostic factors for breast, renal and prostate cancers. For ovarian cancer, aneuploidy seems to be related to prognosis. Bladder tumours with DNA aneuploidy were frequently of high malignancy while ploidy was significantly correlated to relapse risk. For digestive cancers, patients presenting DNA diploid tumours show a better survival than patients with aneuploid ones. Morphometry seems to be a more important criterion than other conventional prognostic factors of invasive breast and digestive carcinomas. A differential diagnosis between normal and neoplastic thyroids is more precise when based on a quantitative evaluation of texture associated to morphometry. Textural parameters permit the discrimination of two populations of patients having a different prognosis and could thus be an aid for prognosis in prostatic cancers. Morphonuclear parameters contribute to separate low and high grade bladder carcinomas. Although ICM was frequently reported, results from the reported examples were not always obvious. In conclusion, the measurements obtained with ICM could be helpful for a decision in several cancers but could not be a substitute for the classical approach of the pathologist.     

Automatic evaluation of ram sperm morphometry

This study was designed to develop a new method based on fluorescence microscopy and image analysis for the automatic assessment of sperm morphometry and to study separately the effect of drying and fixation on the parameters of head sperm morphometry in the ram. The study was divided into two experiments. In the first experiment, ejaculates from 25 adult males were collected using an artificial vagina, diluted and divided into four sample aliquots. The first was labeled directly with Hoechst 33342 (FRESH), and the others were processed as smears. Between smears, one group was directly labeled with Hoechst after air drying (DRIED), and the other were fixed either with glutaraldehyde (GLUT), or with methanol (MET), and labeled with Hoechst afterward. Digital images of the fluorescence-labeled sperm were recorded with a digital camera, and sperm heads were automatically captured and analyzed using the ImageJ program. The method used allowed a fast and automatic selection of most sperm heads for a given image with high precision. There was a general trend toward significant decrease in head length, width, area and perimeter of air-dried sperm compared with fresh sperm. On average, this decrease was of 4.1% in length, 4.3% in width, 9.1% in area, and 2.8% in perimeter. Between semen smears, fixation with glutaraldehyde significantly increased head sperm dimensions. The smears fixed with glutaraldehyde method is recommended for a more practical use than with fresh samples, providing better quality images than the other methods, and because the morphometric results obtained were more similar to the FRESH group than those of the DRIED and MET. In the second experiment, ejaculates from adult males were used to compare the sperm head morphometric results obtained with the new method developed (using the GLUT treatment as reference) with a more conventional CASMA method (semen smears stained with Hemacolor and processed with the ISAS commercial software, HEM). The GLUT method allowed the analysis of 100% of sperm, whereas only 93% of sperm could be analyzed using HEM. Spermatozoa displayed a bigger size when processed with HEM than with GLUT method in all primary sperm head morphometric parameters. A significant correlation was observed between the two methods used in this experiment for all morphometric size parameters. The new method developed allows automatic determination of sperm head morphometry in a reduced time, which facilitates its use in routine semen analysis. It was concluded that the automation of sperm morphometry is feasible using fluorescence microscopy and image analysis and that the effect of drying and fixation was less important than previously stated.     

The relationship between ram sperm head morphometry and fertility depends on the procedures of acquisition and analysis used

Sperm head morphometry is a parameter in the evaluation of semen that has been associated with fertility in two ways: comparing morphometric measures between predefined groups of fertility; or analyzing morphometric data by multivariate techniques to identify cell populations. We analyzed the morphometry of ram sperm head by three procedures and checked its relationship with male fertility. A Computer-Aided Sperm Morphometric Assessment procedure (CASMA), an image analysis software (NIS-Elements) in combination with an optical microscope (MO-NIS) and this image analysis software in combination with a scanning electron microscope (SEM-NIS) were used. Eight morphometric parameters were assessed: length, width, area, perimeter, ellipticity, form factor, elongation and regularity. We observed significant differences between the morphometric data of sperm head obtained with three study procedures. The CASMA procedure shows the highest values for all parameters and the SEM-NIS procedure the lowest. The analysis of a semen sample, when only the mean of morphometric parameters is used to describe the cell population, is too limited to interpret their fertilizing capacity. It is essential to analyze the complex structure of the samples by defining subpopulations by multivariate methods. With few exceptions, the means of each morphometric parameter differ between the three subpopulations analyzed in each procedure. Only the subpopulations obtained with the MO-NIS procedure showed a significant correlation with male fertility. In short, it is necessary to establish an instrumental standard for the analysis of sperm morphometry to obtain reliable results and we believe that the MO-NIS system presents these basic requirements.     

Light microscopic morphometric analysis of peroxisomes by automatic image analysis: advantages of immunostaining over the alkaline DAB method.

The feasibility of light microscopic post-embedding immunocytochemistry for morphometry of peroxisomes using automatic image analysis was investigated and compared with the classical alkaline DAB method. Perfusion-fixed rat liver tissue was either embedded in LR White or incubated in the alkaline diaminobenzidine (DAB) medium for cytochemical visualization of catalase. Sections from the LR White-embedded material were incubated with a monospecific antibody against catalase, followed by protein A-gold and silver intensification. Determination of peroxisomal volume density in sections of different thickness revealed that the values increased with section thickness in DAB-stained sections but were unaffected in immunostained preparations. Moreover, the absolute value for volume density of peroxisomes, as determined by light microscopy in immunostained sections, was quite close to the value obtained by analysis of electron microscopic preparations. Finally, morphometric analysis of bezafibrate-induced peroxisome proliferation revealed that the ratio of proliferation obtained by light microscopy in immunostained sections was very close to the results obtained by electron microscopic morphometry. The main advantage of post-embedding immunostaining for light microscopic morphometry is that it restricts the immunocytochemical reaction product to the surface of the section, thus making it independent of section thickness.     

Morphometric image analysis of follicular lesions of the thyroid

OBJECTIVE: To determine the role of image morphometry in distinguishing various follicular lesions of the thyroid in cytologic smears. STUDY DESIGN: Archival fine needle aspiration smears of 10 cases each of follicular hyperplasia, follicular adenoma, follicular carcinoma and follicular variant of papillary carcinoma were used for the study. All cases were histopathologically proven. At least 100 random nuclei from each case were subjected to analysis with an image cytometer. Area, convex area, length, width, perimeter, convex perimeter and roundness of nuclei were measured using a 40 x objective (1 pixel = 0.446 micron). RESULTS: ANOVA showed that all the nuclear variables studied were significantly different (P < .05) in follicular hyperplasia as compared to follicular carcinoma and papillary carcinoma. All nuclear variables except roundness were also significantly different (P < .05) between follicular hyperplasia and follicular adenoma. However, between follicular adenoma, follicular carcinoma and papillary carcinoma there was considerable overlap of nuclear morphometric parameters. CONCLUSION: Image morphometry may help to distinguish nonneoplastic follicular lesions (hyperplasia) from neoplastic lesions (adenomas and carcinomas). However, to distinguish benign from malignant follicular lesions, image morphometry might not improve the accuracy of standard cytologic examination.     

Computerized morphometric study on fine needle aspirates of cellular follicular lesions of the thyroid

OBJECTIVE: To investigate the value of computerized nuclear morphometry in the differential diagnosis of cellular follicular lesions of the thyroid cytologically diagnosed on fine needle aspiration (FNA) smears. STUDY DESIGN: Sixty cases of FNA thyroid smears were cytologically diagnosed and classified as follows: 30 cases of follicular carcinoma, 20 cases of cellular hyperplastic nodules and 10 cases of follicular adenoma. Using an image analysis system, two morphometric variables, nuclear area and major axis length of the nucleus, were measured for each case. RESULTS: For both nuclear morphometric variables, statistical differences were found between carcinomas and hyperplastic nodules as well as between carcinomas and adenomas. No statistical differences were found between the nuclear variables in either hyperplastic nodules or adenomas. CONCLUSION: The results confirm the aim of our study, to establish nuclear morphometry by computerized image analysis as an additional tool in the differential diagnosis of thyroid follicular lesions cytologically diagnosed on FNA smears.     

A comparative study of three variations of the learning vector quantizer in the discrimination of benign from malignant gastric cells

A prospective study was undertaken to investigate the potential value of morphometry and artificial neural networks (ANN) for the discrimination of benign and malignant gastric lesions. Two thousand five hundred cells from 23 cases of cancer, 19 cases of gastritis and 58 cases of ulcer were selected as a training set, and an additional 8524 cells from an equal number of cases of cancer, gastritis and ulcer were used as a test set. Images of routine processed gastric smears stained by the Papanicolaou technique were processed by a custom image analysis system. The application of the learning vector quantization (LVQ) classifier enabled correct classification of > 97% of benign cells and > 95% of malignant cells, obtaining an overall accuracy of > 97%. This study presents the capabilities of ANN, and also indicates that ANN and image morphometry may offer useful information on the potential of malignancy in gastric cells.     

Quantification of epithelial volume by image processing applied to ovarian tumors

This paper describes an image analysis technique for automated estimation of the percentages of epithelium and stroma in (tumor) tissue. The program is evaluated on ovarian tumors of the serous, mucinous, and endometrioid type. From standard paraffin sections, stained with pararosanilin Feulgen and naphthol yellow, a blue-yellow image pair was recorded. The blue image was used for the determination of the total tissue area and the yellow image for the epithelial area. For the latter the image processing method is based on the fact that epithelial nuclei are generally more tightly packed than stromal nuclei. A structural approach was applied, in which the segmentation of the nuclei was based on the image contrast range in the density domain. The method has been tested with 78 image pairs from 19 ovarian tumors with varying degrees of malignancy. The area percentages, as assessed with image processing, were strongly correlated to control percentages, established by interactive morphometry (r = 0.98).     

Staining of Muscle Tissue for Quantitative Studies and Automated Morphometry

Seventy-six staining tests were carried out on paraffin sections of human and animal muscle to find a suitable staining method for quantitative morphometry of muscle fibers. The results were evaluated under the light microscope, on black and white photomicrographs and on an image analysing computer, the Quantimet 720. A brilliant scarlet-phosphotungstic acid-tartrazine method is described and recommended for automated morphometry after additional testing on 140 sections of developing human muscle using the Quantimet 720.     

Staining of muscle tissue for quantitative studies and automated morphometry.

Seventy-six staining tests were carried out of paraffin sections of human and animal muscle to find a suitable staining method for quantitative morphometry of muscle fibers. The results were evaluated under the light microscope, on black and white photomicrographs and on an image analysing computer, the Quantimet 720.A brilliant scarlet-phosphotungstic acid-tartrazine method is described and recommended for automated morphometry after additional testing on 140 sections of developing human muscle using the Quantiment 720.     

Grading of ductal breast carcinoma by cytomorphology and image morphometry with histologic correlation

OBJECTIVE: To investigate the relevance of image analysis for grading breast carcinoma. STUDY DESIGN: Twenty-five ductal breast carcinoma cases were chosen randomly from routine fine needle aspiration clinics. The results of cytomorphologic grading and image morphometry were correlated with those of histologic grading. The five image morphometric parameters studied were nuclear diameter, nuclear area, nuclear roundness, nuclear perimeter and grey level to compare with chromatin texture. RESULTS: Cytologic grading alone had a high correlation with histologic grading. The lowest correlation was found in grade 2 tumors. When cytologic grading was supplemented with image morphometric parameters, the correlation was higher than that of cytologic grading alone. CONCLUSION: Cytologic grading has a high correlation with histologic grading. The correlation improves further on supplementation with image morphometric parameters.     

Advances in morphometric identification of fishery stocks

Geographic variation in morphometry has been used todiscriminate local forms of fish for over a century. The historical development of stock identificationmethods has paralleled the advancement of morphometrictechniques. The earliest analyses of morphometricvariables for stock identification were univariatecomparisons, but were soon followed by bivariateanalyses of relative growth to detect ontogeneticchanges and geographic variation among fishstocks. As the field of multivariatemorphometrics flourished, a suite of multivariatemethods was applied to quantify variation in growthand form among stocks. More recent advances have beenfacilitated by image processing techniques, morecomprehensive and precise data collection, moreefficient quantification of shape, and new analyticaltools. Many benchmark case studies and critiquesoffer guidelines for sampling morphometrics andinterpreting multivariate analyses for exploratorystock identification, stock discrimination, and stockdelineation. As examples of morphometric stockidentification based on life history differences,allometric patterns of crustacean secondary sexcharacters have been used to detect geographicvariation in size at maturity, and morphometriccorrelates to smoltification have been used todiscriminate salmon from different rivers. Morphometric analysis provides a powerful complementto genetic and environmental stock identificationapproaches. The challenge for the future ofmorphometric stock identification is to develop aconsensus on biological interpretations of geometricanalyses, similar to the conventional interpretationsof size and shape from traditional multivariatemorphometrics.     

Application of automatic image analysis for morphometric studies of peroxisomes stained cytochemically for catalase

Summary The feasibility of the application of a television-based image analyzer, the Texture Analysis System (TAS, Leitz Wetzlar, FRG) in conjunction with a light microscope for morphometric studies of hepatic peroxisomes has been investigated. Rat liver peroxisomes were stained with the alkaline-DAB method for localization of catalase and semi-thin (0.25 and 1 m) sections of plastic-embedded material were examined under an oil immersion objective. The TAS detected the peroxisomal profiles selectively and determined their morphometric parameters automatically. The same parameters were obtained also by morphometric analysis of electron micrographs from the same material. The volume density of peroxisomes determined by TAS in semithin sections of normal liver, after correction for section thickness, is quite close to the corresponding value obtained by morphometry of electron micrographs. The difference is approximately 20%. In animals treated with the hypolipidemic drug bezafibrate, which causes proliferation of peroxisomes, TAS detected readily the increase in volume density of peroxisomes in semithin sections. In comparison with electron microscopy, however, the light-microscopic approach seems to underestimate the proliferation. The lower resolution of the light microscope and overlapping of neighbouring particles in relatively thick sections used for lightmicroscopic analysis may account for the differences.The present study has demonstrated the usefulness of automatic image analysis in conjunction with selective cytochemical staining of peroxisomes for morphometry of this organelle in rat liver. The light-microscopic approach is not only faster but is also extremely economical by obviating the use of an electron microscope.     

Unravelling the stock structure of the Persian brown trout by otolith and scale shape

The simultaneous use of the scale and otolith morphometry was assessed as a potential tool for the identification of Persian brown trout Salmo trutta stocks of the Lar Lake and five rivers from Lar Basin, Iran. Fourier coefficients (FC) and circularity, rectangularity, roundness, ellipticity and form factor shape indices (SI) were calculated for otolith and scale. Several SIs were significantly different among sites for both structures. Permutational multivariate analysis of variance revealed significant differences between several pairwise comparisons for otolith and scale (FCs and indices separately). Discriminant analysis showed otolith FCs (cross-classification rates: 25–86%) and SI (20–45%) appear to be a relatively acceptable tool to discriminate between several locations. Comparatively, the scale morphometry showed lower discriminatory power (FC = 3–65%; SI = 15–34%), with the exception of SI for Elarm River (60%), Kamardasht River (56%) and Lar Lake (75%). Cross-classification rates improved up to 100% when discriminate analysis incorporating all variables for otolith and scale was performed. The results showed a potential segregation between some water bodies, suggesting that the otolith and scale morphometry could be a useful tool to delimit S. trutta populations in relatively close freshwater environments.     

Neural network application in the discrimination of benign from malignant gastric cells

OBJECTIVE: To investigate the potential value of morphometry and neural networks for the discrimination of benign from malignant gastric lesions. STUDY DESIGN: One thousand cells from 19 cases of cancer, 19 cases of gastritis and 56 cases of ulcer were selected as a training set, and an additional 4,000 cells from the same cases of cancer, gastritis and ulcer were used as a test set. Images of routinely processed gastric smears stained by the Papanicolaou technique were analyzed by a custom-made image analysis system. RESULTS: Application of the neural network gave correct classification in 96% of benign cells and 89% of malignant cells. CONCLUSION: The results indicate that the use of neural networks and image morphometry may offer useful information concerning the potential of malignancy in gastric cells.