TNM stage, immunohistology, syntactic structure analysis and survival in patients with small cell anaplastic carcinoma of the lung

Elevated Serum Level of Angiopoietin-2 as a Potential Marker for Poor Prognosis in Small Cell Lung Cancer

Small cell lung cancer (SCLC) is a fast-growing cancer with poor prognosis. Patients with extensive-stage SCLC are generally treated with chemotherapy. Thus, it is essential to identify a predictor of efficacy and prognosis for SCLC. Angiopoietin-2 promotes vascular remodeling and angiogenesis. Increasing evidence reveals that angiopoietin-2 is preferentially expressed in cancer cells, and elevated angiopoietin-2 expression is related to invasive and metastatic phenotypes in various cancers. However, serum angiopoietin-2 level and its prognostic potential in SCLC have not been investigated. The aim of this study was to determine the usefulness of angiopoietin-2 level as a predictor of efficacy and prognosis for SCLC. This study consisted of sixty patients with SCLC. Each patient received four cycles of cisplatin-etoposide chemotherapy, and was followed for 36 months. Serum angiopoietin-2 levels were measured by Enzyme-linked immunosorbent assays. The angiopoietin-2 levels were significantly higher in SCLC patients than those in healthy subjects (P < 0.001). The patients were divided into high-level group (32 patients, 2,923.9 ± 294.7 pg/ml) and low-level group (28 patients, 1,789.5 ± 355.1 pg/ml) according to the mean value of the angiopoietin-2 level (2,400 pg/ml). Compared with the patients in the high-level group, the patients in the low-level group showed remarkably survival advantage (P = 0.002). During chemotherapy, the patients in the low-level group showed better treatment response than the patients in the high-level group (P < 0.05). Therefore, angiopoietin-2 might be useful as a prognostic factor for SCLC and for predicting SCLC response to chemotherapy.

Quantitative pathology in virtual microscopy: history, applications, perspectives

With the emerging success of commercially available personal computers and the rapid progress in the development of information technologies, morphometric analyses of static histological images have been introduced to improve our understanding of the biology of diseases such as cancer. First applications have been quantifications of immunohistochemical expression patterns. In addition to object counting and feature extraction, laws of thermodynamics have been applied in morphometric calculations termed syntactic structure analysis. Here, one has to consider that the information of an image can be calculated for separate hierarchical layers such as single pixels, cluster of pixels, segmented small objects, clusters of small objects, objects of higher order composed of several small objects. Using syntactic structure analysis in histological images, functional states can be extracted and efficiency of labor in tissues can be quantified. Image standardization procedures, such as shading correction and color normalization, can overcome artifacts blurring clear thresholds. Morphometric techniques are not only useful to learn more about biological features of growth patterns, they can also be helpful in routine diagnostic pathology. In such cases, entropy calculations are applied in analogy to theoretical considerations concerning information content. Thus, regions with high information content can automatically be highlighted. Analysis of the "regions of high diagnostic value" can deliver in the context of clinical information, site of involvement and patient data (e.g. age, sex), support in histopathological differential diagnoses. It can be expected that quantitative virtual microscopy will open new possibilities for automated histological support. Automated integrated quantification of histological slides also serves for quality assurance. The development and theoretical background of morphometric analyses in histopathology are reviewed, as well as their application and potential future implementation in virtual microscopy.

To be at the right place at the right time

AIM

To analyze the hypothesis of events or neighborhood interactions that is based upon recognizable structures of systems which possess a surface in a four dimensional space-time constellation {x, y, z, t}. To include the theory of hierarchic order of structures and aspects of thermodynamically open systems, especially entropy, structural entropy and entropy flow.

HYPOTHESIS

Any structure is a space-time constellation that occupies a unique space in its environment. The environment can be a system too, and is assumed to be (nearly) constant. Structures can interact in their environment and create a new structure at a higher order level. Interacting structures that create a surface are called a system. Starting from the bottom, such a system is characterized by its inner structures, its surface function, and its neighborhood. Interaction with a neighboring system is called an event. An event can alter a system, create new systems or induce the decay of a system, dependent upon the surrounding lower level system (background).

RESULTS

The hypothesis results in a uniform theory about matter, life, diseases, or behavior. Concrete applications permit the estimation of duration of life in man, for example the effect of solid cancer in man, or appearance of protozoans in sexual or asexual reduplication. In addition, it can successfully describe the development of the universe (small exceed of matter above antimatter at the big bang), or the increase of structures (and systems) with increasing time (development of intelligent systems). The three dimensional space possesses the lowest number of mandatory dimensions to implement such a system.

Quantitative assessment of gastric atrophy using the syntactic structure analysis

AIM

To assess the topographical relation between gastric glands, using the minimum spanning tree (MST), to derive both a model of neighbourhood and quantitative representation of the tissue's architecture, to assess the characteristic features of gastric atrophy, and to assess the grades of gastric atrophy.

METHODS

Haematoxylin and eosin stained sections from corporal and antral biopsy specimens (n = 139) from normal patients and from patients with nonatrophic gastritis and atrophic gastritis of grades 1, 2, and 3 (Sydney system) were assessed by image analysis system (Prodit 5.2) and 11 syntactic structure features were derived. These included both line and connectivity features.

RESULTS

Syntactic structure analysis was correlated with the semiquantitative grading system of gastric atrophy. The study showed significant reductions in the number of points and the length of MST in both body and antrum. The standard deviation of the length of MST was significantly increased in all grades of atrophy. The connectivity to two glands was the highest and most affected by the increased grade of atrophy. The reciprocal values of the Wiener, Randic, and Balaban indices showed significant changes in the volume of gland, abnormality in the shape of glands, and changes in irregularity and branching of the glands in both types of gastric mucosa. There was a complete separation in the MST, connectivity, and index values between low grade and high grade gastric atrophy.

CONCLUSIONS

(1) Gastric atrophy was characterised by loss of the gland, variation in the volume, reduction in the neighbourhood, irregularity in spacing, and abnormality in the shape of the glands. (2) Syntactic structure analysis significantly differentiated minor changes in gastric gland (low grade atrophy) from high grade atrophy of clinical significance. (3) Syntactic structure analysis is a simple, fast, and highly reproducible technique and appears a promising method for quantitative assessment of atrophy.

Application of computer-assisted morphometry to the analysis of prenatal development of human lung

Morphometry is well-established in tumour pathology. To evaluate is potential usefulness for description of developmental processes, histological slides from paraffin-embedded specimens of 67 human fetal lungs were Feulgen-stained, and morphometric characteristics of nuclei of epithelial pulmonary cells were analysed with an automated image analysis system. The measured cytometric features comprised of integrated optical density (IOD), S-phase-related IOD fraction, IOD entropy and nuclear area. Histometric features of the specimens were based upon the minimum spanning tree (MST) and included distances between neighboring epithelial cells, between epithelial cells and neighboring lymphocytes, and assessment of MST entropy. Notably, certain parameters revealed a non-uniform level during prenatal development S-phase-related IOD fraction increased from 5% to 8% between 14 and 16 weeks of gestation, then declined to 6% until birth. The IOD entropy steadily increased during development, whereas the extent of nuclear area remained constant. In accordance with an increase of the S-phase-related fraction the MST entropy displayed a singular peak between 14 and 16 weeks of gestation, which is probably associated with development of glandular structures in the lung. Correlation of expression of binding sites for markers, presumably involved in functional aspects of development, with such alterations, is shown for binding capacities of biotinylated fucoidan and the S-phase-related fraction. This may be helpful to infer immuno- or ligando histochemically defined tissue sites with potential physiological significance in morphometrically distinguished periods of development.

Integrated optical density (IOD), syntactic structure analysis, and survival in operated lung carcinoma patients

Histological sections of formalin fixed, paraffin-embedded tissue comprising 195 specimens of human lung carcinomas were Feulgen stained. The nuclei of the histomorphological images were segmented using an automated image analyzing system, and the attributed minimum spanning trees (MST) were calculated. Features related to the DNA-content of the nuclei (integrated optical density (IOD), IOD-entropy, S-phase related fraction, percentage > 5C, etc.), and structural parameters (minimum distance between tumor cell nuclei, minimum distance between tumor cells and neighboring lymphocytes, MST-entropy, MST-current of entropy (entropiefluss), distance between neighboring proliferating tumor cells, etc.) were measured. The following results were obtained: the measured IOD and MST features showed significant differences between the primary carcinomas and metastatic carcinomas in the intrapulmonary lymph nodes. The survival of patients was remarkably improved if the carcinomas displayed a low S-phase related fraction, a low percentage of tumor cells > 5C, a low number of stem lines, and a low MST-current of entropy.

An approach based on two-dimensional graph theory for structural cluster detection and its histopathological application

An approach based on graph theory is described for detecting clusters of cells in tissue specimens (two-dimensional space). With a set of discrete basic elements (cell nuclei) having several measurable features (area, surface, main and minor axis of best-fitting ellipses) a graph is defined as having attributes associated with edges. Different minimum spanning trees (MSTs) can be constructed using different weight functions on the attributes (attributed MST). Analysis of the MST and of an attributed MST by use of a decomposition function allows detection of image areas with similar local properties. These clusters, which are then clusters of the tree, describe, for example, partial growth in different directions in a case of a human fibrosarcoma assuming that tumour cell nuclei are homogeneous with respect to their configuration and size. The model allows the separation of clusters of tumour cells growing in different directions and the approximation of the different growth angles. This decomposition also allows us to create new (higher) orders of structure (cluster tree).

DNA-content, inflammatory tissue response and tumour size in human lung carcinoma

Tumour imprints of 75 resection specimens with human lung carcinoma (lobe and lungs) were fixed with alcohol and Feulgen-stained. Resection specimens were cut into serial sections 6 mm thick and tumour mass, tumour involvement into intrapulmonary and extrapulmonary lymph nodes, detailed pTN-stage were determined. DNA-content, DNA-index, percentage of diploid/hyperploid tumour cells, and morphometric nuclear features were measured using an automated image analyzing system (VISIAC). Only 10% of the measured carcinomas were diploid. The DNA-index showed the broadest variance in large cell anaplastic carcinoma (1.2-3.3). Carcinomas growing predominantly within the alveolar space, i.e. without destroying the interstitial tissue showed a lower DNA-content above 3c and above 5c compared to carcinomas destroying the interstitial tissue. Carcinomas with severe stroma reaction were found to have 33%-48% of DNA above the 3c value whereas carcinomas without stromal reaction had a percentage above 3c ranging 46%-64% (confidence limits, p less than 0.05). DNA-index increases with increasing tumour volume and decreases for large tumours (greater than 100 ccm). Hyperploid and polyploid tumours were found more frequently in case of pT2 and pT3-stages compared to pT1-stages. No relation of DNA-content or ploidy was found to lymph node involvement and inflammatory response of host tissue.

Minimum spanning tree, Voronoi's tesselation and Johnson-Mehl diagrams in human lung carcinoma

A program written in Amba was developed for the automated construction of minimum spanning trees, Voronoi's tesselation and Johnson-Mehl diagrams, combined with quantitative measurements of nuclear parameters in histopathological specimens. Sections, 4 microns thick obtained from paraffin-embedded surgical lung specimens, were Feulgen-stained. The boundaries of the nuclei were calculated automatically after interactive identification of the relevant cells. The geometric centres of the nuclei were computed. They were used for constructing the minimum spanning tree, Voronoi's tesselation and the Diriclet cells. The direction of the main axis to the best fitting ellipsis according to the nuclear size was used for construction of the Johnson-Mehl diagrams. The following textural parameters were measured: 1) Minimum spanning tree: Distribution of neighbouring cells; minimum distance between neighbouring cells according to number of neighbours. 2) Voronoi's tesselation: Area and surface of obtained cells; Ratio of nuclear area/cell area. 3) Johnson-Mehl diagram: Area and surface of obtained cells; Ratio of nuclear area/cell area. The following nuclear parameters were measured: 1) Area, surface, absolute extinction (DNA-content); Similarity of quantitative nuclear parameters between neighbouring cells and in relation to number of neighbouring cells was analyzed. Twenty cases of primary lung carcinoma (five cases of each of epidermoid, adeno, large cell anaplastic, and small cell anaplastic carcinoma) were measured. Analysis of texture parameters in combination with morphometric nuclear parameters offers new possibilities in quantitative histopathology.

Stereological estimates of nuclear volume in primary lung cancer

The aim of this study is to provide unbiased, objective and reproducible morphometric data on primary lung carcinomas by using the stereological principle of estimating volume weighed mean volume of particles of arbitrary shape. The study includes 30 lung specimens; 10 from small cell carcinoma, 10 from squamous cell carcinoma, and 10 from adenocarcinoma. After standard fixation, embedding, sectioning and hematoxylin eosin staining, the mean nuclear volume was estimated using the formula: v upsilon = pi/3 . l0(3), where l0 is the length of the intercept through a test point hitting a nucleus measured in a random direction through the test point. The collection of the data took less than 10 min per specimen. The mean nuclear volume of small cell carcinomas varied from 109 to 238 microns 3 (group mean 152 microns 3), in squamous cell carcinomas from 407 to 826 microns 3, and in adenocarcinomas from 411 to 1046 microns 3 (group mean 736 microns 3). This simple and rapid method seems to provide unbiased, objective and reproducible data useful in the description and separation of primary lung carcinomas.

Is the profile of binding of a panel of neoglycoproteins useful as a diagnostic marker in human lung cancer?

Histomorphological evaluation of the profile of carbohydrate-binding proteins with specificities to alpha- and beta-glucosides (maltose, cellobiose), alpha-L-fucosides (L-fucose), alpha-D-mannosides (D-mannose), a N-acetylated sugar, present in natural glycoconjugates (N-acetyl-D-glucosamine), two negatively charged sugars (glucuronic acid and sialic acid), rhamnose (6-deoxy-L-mannose) or to heparin has been assessed. 133 cases of formalin fixed, paraffin embedded human lung carcinoma specimens (45 cases of epidermoid carcinoma, 24 cases of small cell anaplastic carcinoma, 48 cases of adenocarcinoma, and 16 cases of large cell anaplastic carcinoma) as well as 14 cases with epithelial or biphasic mesothelioma and 7 cases of pleuritis carcinomatosa were incubated with biotinylated, suitably modified neoglycoproteins or heparin for detecting the corresponding endogenous sugar receptors (lectins). Histochemical reaction was demonstrated by use of the avidin-biotin-method. Significant differences were obtained between the different markers used for each cell type as well as between the different cell types. For separating small cell anaplastic carcinoma from non small cell anaplastic carcinoma the markers carrying the histochemically crucial maltose, fucose, and mannose residues are useful, showing a positive reaction to only 4-8% of the small cell anaplastic carcinoma cases compared to 60-80% of the non small cell carcinoma cases. For separation of mesothelioma from pleuritis carcinomatosa a carrier, modified by attachment of N-acetylglucosamine moieties is the most useful marker being negative in all cases of the analyzed mesothelioma cases and being positive in 62-80% of the adenocarcinoma cases.(ABSTRACT TRUNCATED AT 250 WORDS)

Expression of neuroendocrine markers (neuronspecific enolase, synaptophysin and bombesin) in carcinoma of the lung

The value of immunoreactivity of antibodies against neuronspecific enolase (NSE), bombesin (GRP), and synaptophysin (SY 38) as markers for various human lung carcinoma has been assessed. One hundred-forty-two primary bronchus carcinomas (small cell anaplastic carcinoma, epidermoid carcinoma, adeno carcinoma, and large cell anaplastic carcinoma) were studied by the indirect immunoperoxidase method (PAP). SY 38 was found to react positively in 49/68 (79%) of the small cell anaplastic carcinoma (SCCL) and in 6/74 (8%) of the non-small cell carcinoma of the lung (NSCCL). Positive immunohistochemical data with antibody SY 38 showed in some cases an immunoreactive polypeptide of Mr = 40.000 obtained by immunoblotting similar in molecular weight as described for synaptophysin in other tumours. Reactivity of NSE was observed in 41/68 (61%) of the SCCL and in 8/74 (10%) of the NSCCL. Positive reactivity to GRP was similar to NSE in 42/68 (62%) of SCCL and in 7/74 (10%) of NSCCL. All cases of NSCCL reacting positively to SY 38 were found to react positively to NSE, and to GRP. Prognostic value of SY 38 was calculated vp = 0.71 for positive prediction and vn = 0.91 for negative prediction. The data indicate that SY 38 represents the broadest marker for neuroendocrine carcinoma of the lung since in addition to the majority of SCCL about 10% of NSCCL are recognized by the antibody SY 38.