Computerized interactive morphometry. An expert system for the diagnosis of lymphoid-rich effusions

Pleural effusions in acute and chronic leukemia and myelodysplastic syndrome

PURPOSE OF REVIEW

Pulmonary manifestations have been well described in leukemia, but pleural disease is less common. This review highlights pleural effusions in acute and chronic leukemia and myelodysplastic syndrome (MDS) based on the evidence to date. Diagnostic workup and recommendations for the management of these effusions are also outlined.

RECENT FINDINGS

Pleural effusions in patients with leukemia are most often due to infection and to a lesser extent leukemic infiltration of the pleura. The prognostic implications of these effusions are unclear, but survival is most likely determined by the underlying malignancy and its response to treatment. New therapies have changed survival in these patients, and some of these treatments, such as tyrosine kinase inhibitors, have emerged as important causes for these effusions. Pleural interventions may be accomplished with few complications.

SUMMARY

Pleural effusions may occur with acute and chronic leukemia and MDS. Infection remains the most common cause. Malignant pleural effusions tend to occur in advanced disease in chronic leukemia, but they can be seen at any time with acute leukemia and MDS. With standard precautions, pleural procedures may be performed safely in this population. In cases of unclear cause, pleural and bone marrow biopsy should be considered.

Malignant pleural effusions in lymphoproliferative disorders

In order to determine variables that correlate with malignant pleural effusion and mortality in patients with lymphoproliferative disorders and pleural effusion, a retrospective study was performed. Clinical data of hospitalized patients with a lymphoid malignancy and pleural effusion who underwent thoracentesis from January 1993 to December 2002 were collected. A logistic regression analysis was carried out to determine prognostic variables that predict malignant pleural effusion and hospital mortality. There were 86 patients who were admitted on 91 occasions. The median age was 70 years (range 4 - 92) and the male:female ratio was 44:42. Sixty-four patients (74%) had advanced disease, 43 (50%) had received prior chemotherapy and 9 (10%) were in remission. Of 91 cases of pleural effusions, 44 (48%) were bilateral, 80 (88%) were exudates and 48 (53%) were due to malignant involvement of pleura. In multivariate analysis, symptomatic pleural effusion (odds ratio 10.3, 95% confidence interval 1.7 - 98.3), pleural fluid mesothelial cell count < 5% (odds ratio 8.0, 95% confidence interval 1.4 - 58.2), pleural fluid:serum lactate dehydrogenase (LDH) > or =1 (odds ratio 6.4, 95% confidence interval 1.2 - 45.6) and pleural fluid lymphocyte percentage > or =50 (odds ratio 6.4, 95% confidence interval 1.2 - 50) were significantly correlated with malignant effusion. A secondary cancer (odds ratio 11.9, 95% confidence interval 2.3 - 88.8), pleural fluid:serum LDH > or =1 (odds ratio 10.9, 95% confidence interval 2.6 - 64.9), and pneumonia (odds ratio 6.4, 95% confidence interval 1.7 - 28.6) were significantly correlated with hospital mortality. In conclusion, malignant pleural effusion is the common etiology of pleural effusion in patients with lymphoid malignancy. Many clinical and cytochemical markers have discriminatory values in identifying malignant effusion. A high pleural fluid to serum LDH level correlates with malignant pleural involvement and hospital mortality.

Serous effusions in malignant lymphomas: A review

Serous effusions are a common complication of lymphomas. Although the frequency of pleural effusion is 20-30% in non-Hodgkin's lymphoma (NHL) and Hodgkin's disease (HD), the involvement of peritoneal and pericardial cavities is uncommon. Among lymphoma subtypes, T-cell neoplasms, especially the lymphoblastic lymphomas, more frequently involve the serous fluids. The thoracic duct obstruction and impaired lymphatic drainage appear to be the primary mechanism for pathogenesis of pleural effusion in HD and direct pleural infiltration is the predominant cause in NHL. There is wide variation in rate of positive cytologic findings of NHL in pleural effusion (22.2-94.1%). Cytologic features of specific lymphoma subtypes such as lymphoblastic lymphoma, follicular center cell lymphoma, including Burkitt-type lymphoma, marginal zone lymphoma, MALT lymphoma, and anaplastic large-cell lymphoma, etc., have been described in the literature. The differential diagnostic problems of lymphomas in serous effusions include reactive lymphocytoses, early involvement by lymphomatous process, small round-cell tumors (SRCT), and presence of look-alike of Reed-Sternberg cells. To overcome these difficulties, various ancillary studies, including immunocytochemistry (ICC), morphometry, flow cytometry (FCM), and cytogenetics/molecular genetics (PCR, in-situ hybridization, and Southern blotting), have been performed on effusion specimens. ICC not only distinguishes lymphomas from reactive lymphocytoses and SRCTs, it significantly modifies the morphologic diagnosis to achieve a better classification of lymphomas. Combined morphology and immunophenotyping by FCM, has a sensitivity as well as specificity of 100%. Morphometry also distinguishes reactive lymphocytoses from malignant lymphoma with a high degree of sensitivity (>85%) and specificity (>95%). Limitations of individual ancillary techniques can be overcome by using multiple parameters. Although lymphomas rarely present as serous effusions without the involvement of other thoracic and extrathoracic sites, a small group of lymphomas called primary effusion lymphomas (PEL) exhibit exclusive or dominant involvement of serous cavities, without a detectable solid tumor mass. This body cavity based lymphoma (BCBL) is a distinct clinicopathologic entity and is found predominantly in AIDS patients with preexisting Kaposi sarcoma. In the absence of obstructive or infiltrative tumor mass, its pathogenesis has been attributed to stimulation by vascular endothelial growth factor (VEGF)/vascular permeability factor (VPF), leading to vascular leakage. Cytomorphologically, PEL is usually a large-cell lymphoma, which appears to bridge features of large-cell immunoblastic and anaplastic large-cell lymphoma (ALCL). Most of these cases comprise a unique subgroup of B-cell lymphoma, with features of both high-grade anaplastic and B-immunoblastic lymphoma, but T-cell and/or natural killer cell immunophenotypes are described. Its association with various viral DNAs has been studied in detail by molecular techniques. Pleural effusion due to lymphomas, either primary or otherwise, is considered as one of the factors adversely influencing overall survival. The presence of pleural effusion at the time of presentation is not only associated with extremely poor outcome of lymphomas, it is also a predictor of disease relapse after chemotherapy and decreased survival. When the patients of lymphomatous pleural effusions with and without mediastinal mass present in respiratory distress, thoracocentesis is the initial diagnostic and therapeutic choice in these patients. In such situations, cytology along with ancillary studies not only gives a quick diagnosis of lymphoma, but also offers prognostically significant information such as classification of lymphomas, its grade and immunophenotype, and presence/absence of viral DNAs and tumor lysis syndrome.

Classification of individual lung cancer cell lines based on DNA methylation markers: use of linear discriminant analysis and artificial neural networks

The classification of small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC) can pose diagnostic problems due to inter-observer variability and other limitations of histopathology. There is an interest in developing classificatory models of lung neoplasms based on the analysis of multivariate molecular data with statistical methods and/or neural networks. DNA methylation levels at 20 loci were measured in 41 SCLC and 46 NSCLC cell lines with the quantitative real-time PCR method MethyLight. The data were analyzed with artificial neural networks (ANN) and linear discriminant analysis (LDA) to classify the cell lines into SCLC or into NSCLC. Models used either data from all 20 loci, or from five significant DNA methylation loci that were selected by a step-wise back-propagation procedure (PTGS2, CALCA, MTHFR, ESR1, and CDKN2A). The data were sorted randomly by cell line into 10 different data sets, each with training and testing subsets composed of 71 and 16 of the cases, respectively. Ten ANN models were trained using the 10 data sets: five using 20 variables, and five using the five variables selected by step-wise back-propagation. The ANN models with 20 input variables correctly classified 100% of the cell lines, while the models with only five variables correctly classified 87 to 100% of cases. For comparison, 10 different LDA models were trained and tested using the same data sets with either the original data or with logarithmically transformed data. Again, half of the models used all 20 variables while the others used only the five significant variables. LDA models provided correct classifications in 62.5% to 87.5% of cases. The classifications provided by all of the different models were compared with kappa statistics, yielding kappa values ranging from 0.25 to 1.0. We conclude that ANN models based on DNA methylation profiles can objectively classify SCLC and NSCLC cells lines with substantial to perfect concordance, while LDA models based on DNA methylation profiles provide poor to substantial concordance. Our work supports the promise of ANN analysis of DNA methylation data as a powerful approach for the development of automated methods for lung cancer classification.

Pulmonary complications in chronic lymphocytic leukemia

BACKGROUND

Although pulmonary complications account for significant morbidity and mortality in patients with chronic lymphocytic leukemia (CLL), to the authors' knowledge there are sparse data available in published literature. The authors evaluated pulmonary complications in patients with CLL and identified prognostic variables that predict hospital mortality in these patients.

METHODS

Clinical data were analyzed retrospectively from patients with CLL who required hospitalization for a respiratory illness at a tertiary care institution from January 1993 to December 2001. A logistic regression analysis with a backward elimination procedure was carried out to determine prognostic variables that predict hospital mortality.

RESULTS

There were 110 patients who were admitted on 142 occasions with a pulmonary complication. The median age was 75 years (range, 43-97 years), and the male:female ratio was 1.7:1.0. Among 142 admissions, 68% were high risk according to the Rai criteria, 68% of patients admitted had received prior therapy for CLL, and 35% had received treatment within 3 months of admission. The most common pulmonary complications were pneumonias (75%), malignant pleural effusion/and or lung infiltrate due to CLL (9%), pulmonary leukostasis (4%), Richter transformation or nonsmall cell lung carcinoma (3%), and upper airway obstruction (2%). Forty-four of 110 patients (40%) died. In multivariate analysis, admission absolute neutrophil counts </= 0.5 x 10(9)/L (odds ratio, 4.6; 95% confidence interval [95% CI], 1.3-16.6) and blood urea nitrogen (BUN) levels >/= 20 mg/dL (odds ratio, 3.0; 95% CI, 1.1-8.3) were correlated significantly with mortality.

CONCLUSIONS

Pneumonia was the major pulmonary complication in hospitalized patients with CLL. Severe neutropenia and high BUN levels were correlated significantly with increased mortality.

Evaluation of diagnostic efficiency of computerized image analysis based quantitative nuclear parameters in papillary and follicular thyroid tumors using paraffin-embedded tissue sections

Computerized image analysis (IA) system has emerged in recent years as a very powerful tool for objective and reproducible quantification of histological features. It has shown considerable potential for diagnostic application in diverse histological situations. The objectives of the present study were to evaluate the discriminatory diagnostic efficiency of computerized image analysis based quantitative subvisual nuclear parameters in papillary and follicular neoplasms of thyroid. A total of 60 cases were studied. Forty-four cases belonged to training set and 16 cases belonged to a test set. A minimum of 100 nuclei was analyzed in each case using uniform 5 m mm thick hematoxylin stained sections. The IA workstation comprised of an Olympus microscope, a 10 bit digital video camera, an image grabber card and a pentium 120 MHz computer. Optimas 5.2 software was utilized for data collection on 8 morphometric and 8 densitometric parameters. Multivariate stepwise discriminant statistical analysis of data was done with the help of BMDP statistical software release 7.0. Results from a training set revealed correct classification rates of 98.0%, 84.5% and 61.2% for the histological groups of hyperplastic papillae versus papillae of papillary carcinoma (group I), follicular variant of papillary carcinoma versus the broad category of follicular neoplasms consisting of both follicular adenoma and follicular carcinoma (group II) and follicular adenoma versus follicular carcinoma (group III), respectively. Results of test set revealed correct classification rates of 100%, 80% and 50% for groups I, II and III respectively. It was concluded that computerized nuclear IA parameters have potential usefulness for discriminating benign versus malignant papillary lesions of thyroid, follicular variant of papillary carcinoma versus follicular adenoma and/or follicular carcinoma but are of no value in discriminating between follicular adenoma and follicular carcinoma.

Immunologic and molecular biologic characterization of pleural involvement in a case of T-chronic lymphocytic leukemia

Pleural involvement is a rare complication of chronic lymphocytic leukemia (CLL). We report a CLL case of T-cell origin (documented by cell surface marker as well as DNA rearrangement studies) where the lymphoid cells of the pleural fluid were found to belong to the same monoclonal population of T cells as those of the peripheral blood.

Colour thresholding and objective quantification in bioimaging

Computer imaging is rapidly becoming an indispensable tool for the quantification of variables in research and medicine. Whilst its use in medicine has largely been limited to qualitative observations, imaging in applied basic sciences, medical research and biotechnology demands objective quantification of the variables in question. In black and white densitometry (0-256 levels of intensity) the separation of subtle differences between closely related hues from stains is sometimes very difficult. True-colour and real-time video microscopy analysis offer choices not previously available with monochrome systems. In this paper we demonstrate the usefulness of colour thresholding, which has so far proven indispensable for proper objective quantification of the products of histochemical reactions and/or subtle differences in tissue and cells. In addition, we provide interested, but untrained readers with basic information that may assist decisions regarding the most suitable set-up for a project under consideration. Data from projects in progress at Tulane are shown to illustrate the advantage of colour thresholding over monochrome densitometry and for objective quantification of subtle colour differences between experimental and control samples.

Ultrastructural morphometry of human leucocytes in health and disease

In this review, the literature on ultrastructural morphometry of each of the main types of human blood leucocytes has been considered, together with the technical and numerical procedures essential for valid analysis. Quantitative data have been reported for these cell types in health and comparisons have been made with those in disease states. In monocytes, and in macrophages developing from them, subtle ultrastructural differences have been detected and quantitated in malignant lymphoma; as the mononuclear phagocytes were not themselves neoplastic, the changes may have related to defects in host defence. Change in the ultrastructural characteristics of leukaemic monoblasts have also been reported. Lymphocytes and malignant lymphoid cells have been extensively investigated: differences between different types and subsets have been shown to be present in both normal lymphocytes and their malignant counterparts in leukemias and lymphomas. Particular attention has been paid to morphometric assessment of nuclear shape and size in these disorders and to its possible value as a diagnostic tool. Granulocytes have so far been the subject of few morphometric studies, although in hypereosinophilic syndrome, cellular changes have been defined and have thrown light on the abnormal pattern of degranulation. There have also been scattered reports on the cells of acute myelogenous leukaemia. The use of computers and sophisticated statistical packages has greatly facilitated the application of multiple comparison procedures and has permitted discriminant analysis to be carried out where appropriate. This review shows that ultrastructural morphometry of leucocytes will have an increasing application in clinical pathology.