SEARCH · PubMed Health

Results for “Classification”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 757 records · Page 42Linked to original sources

PubMed1994

Dukes's classification: a valid prognostic indicator for gastric cancer.

Dukes's classification for colorectal cancer is simple, and correlates well with survival. This study assessed whether the principles of Dukes's classification applied to gastric cancer resulted in the separation of prognostically useful groups. The cumulative survival curves determined by Dukes's, Astler-Coller's (another classification for colorectal cancer), and the Japanese classification for gastric cancer were evaluated using 476 curatively treated patients with gastric cancer. Of the three staging systems, both Dukes's and the Japanese classifications showed a step-wise relationship between the stage of the tumour and the survival of patients. The prognostic value of Dukes's classification was reinforced when Dukes's C cases were subdivided according to the number of positive nodes (1-6 v > or = 7) or the level of positive nodes (n1 v n2,n3) but not when the subdivision was made according to the depth of wall invasion. A modified Dukes's classification in which Dukes's cases are subdivided according to the number of metastatic nodes (Dukes's Ca = 1-6 nodes, Dukes's Cb > 6 nodes) could be both simply and accurately applied to gastric cancer.

Evaluation Studies as Topic↗

PubMedDate not supplied

Nursing classification systems: necessary but not sufficient for representing "what nurses do" for inclusion in computer-based patient record systems.

Our premise is that from the perspective of maximum flexibility of data usage by computer-based record (CPR) systems, existing nursing classification systems are necessary, but not sufficient, for representing important aspects of "what nurses do." In particular, we have focused our attention on those classification systems that represent nurses' clinical activities through the abstraction of activities into categories of nursing interventions. In this theoretical paper, we argue that taxonomic, combinatorial vocabularies capable of coding atomic-level nursing activities are required to effectively capture in a reproducible and reversible manner the clinical decisions and actions of nurses, and that, without such vocabularies and associated grammars, potentially important clinical process data is lost during the encoding process. Existing nursing intervention classification systems do not fulfill these criteria. As background to our argument, we first present an overview of the content, methods, and evaluation criteria used in previous studies whose focus has been to evaluate the effectiveness of existing coding and classification systems. Next, using the Ingenerf typology of taxonomic vocabularies, we categorize the formal type and structure of three existing nursing intervention classification system--Nursing Interventions Classification, Omaha System, and Home Health Care Classification. Third, we use records from home care patients to show examples of lossy data transformation, the loss of potentially significant atomic data, resulting from encoding using each of the three systems. Last, we provide an example of the application of a formal representation methodology (conceptual graphs) which we believe could be used as a model to build the required combinatorial, taxonomic vocabulary for representing nursing interventions.

Abstracting and Indexing↗

PubMedDate not supplied

Ad hoc classification of radiology reports.

OBJECTIVE: The task of ad hoc classification is to automatically place a large number of text documents into nonstandard categories that are determined by a user. The authors examine the use of statistical information retrieval techniques for ad hoc classification of dictated mammography reports. DESIGN: The authors' approach is the automated generation of a classification algorithm based on positive and negative evidence that is extracted from relevance-judged documents. Test documents are sorted into three conceptual bins: membership in a user-defined class, exclusion from the user-defined class, and uncertain. Documentation of absent findings through the use of negation and conjunction, a hallmark of interpretive test results, is managed by expansion and tokenization of these phrases. MEASUREMENTS: Classifier performance is evaluated using a single measure, the F measure, which provides a weighted combination of recall and precision of document sorting into true positive and true negative bins. RESULTS: Single terms are the most effective text feature in the classification profile, with some improvement provided by the addition of pairs of unordered terms to the profile. Excessive iterations of automated classifier enhancement degrade performance because of overtraining. Performance is best when the proportions of relevant and irrelevant documents in the training collection are close to equal. Special handling of negation phrases improves performance when the number of terms in the classification profile is limited. CONCLUSIONS: The ad hoc classifier system is a promising approach for the classification of large collections of medical documents. NegExpander can distinguish positive evidence from negative evidence when the negative evidence plays an important role in the classification.

Algorithms↗

PubMed1998

Reproducibility of new classification schemes for the pathology of ductal carcinoma in situ of the breast.

AIM: To compare the interobserver variation in the pathological classification of ductal carcinoma in situ of the breast using two recently proposed classification schemes. METHODS: 11 pathologists classified a set of 25 cases of ductal carcinoma in situ chosen to reflect a range of lesions, using the traditional architectural classification together with the modified cytonuclear grading scheme of Holland et al and the Van Nuys classification scheme. Participating pathologists received a standard tutorial, written information, and illustrative photomicrographs before their assessment of the cases. RESULTS: Interobserver agreement was poorest when using the architectural scheme (kappa = 0.44), largely owing to variations in classifying lesions with a mixed component of patterns (kappa = 0.13). Agreement was better using the modified cytonuclear grading scheme (kappa = 0.57), with most consistency achieved using the Van Nuys scheme (kappa = 0.66). Most discordant results using the later scheme were due to inconsistency in assessing the presence or absence of luminal necrosis. CONCLUSIONS: Both the new classification schemes assessed in this study were an improvement over the traditional architectural classification system for ductal carcinoma in situ, and resulted in more reproducible pathological assignment of cases. The Van Nuys classification scheme is easy to apply, even to small areas of carcinoma, resulting in acceptable interobserver agreement between reporting pathologists. Additional work will be required to arrive at a consensus definition of necrosis for cases in the non-high-grade group.

Breast Neoplasms↗

PubMed2000

How well does the Oxfordshire community stroke project classification predict the site and size of the infarct on brain imaging?

OBJECTIVES: The Oxfordshire Community Stroke Project (OCSP) classification is a simple clinical scheme for subdividing first ever acute stroke. Several small studies have shown that when an infarct is visible on CT or MRI, the classification predicts its site in about three quarters of patients. The aim was to further investigate this relation in a much larger cohort of patients in hospital with ischaemic stroke. METHODS: Between 1994 and 1997, inpatients and outpatients with ischaemic stroke were assessed by one of several stroke physicians who noted the OCSP classification. A neuroradiologist classified the site and extent of recent infarction on CT or MRI. RESULTS: Of 1012 patients with ischaemic stroke, 655 (65%) had recent visible infarcts. These radiological lesions were appropriate to the clinical classification in 69/87 (79%) patients with a total anterior circulation syndrome, 213/298 (71%) with a partial anterior circulation syndrome, 105/144 (73%) with a lacunar syndrome, and 105/126 (83%) with a posterior circulation syndrome. Overall, 75% of patients with visible infarcts were correctly classified clinically. If patients without a visible infarct did have an appropriate lesion in the brain (best case), the classification would have correctly predicted its site and size in 849/1012 (84%) patients, compared with only 492/1012 (49%) in the worst case scenario. CONCLUSION: The OCSP classification predicted the site of infarct in three quarters of patients. When an infarct is visible on brain imaging, the site of the infarct should guide the use of further investigations, but if an infarct is not seen, the OCSP classification could be used to predict its likely size and site.

Aged↗

PubMed1986

Classification of primate spinothalamic and somatosensory thalamic neurons based on cluster analysis.

Data analyzed in this study were derived from the responses of 128 spinothalamic tract (STT) cells and 110 thalamic neurons recorded in 75 anesthetized monkeys. A k-means cluster analysis, a nonhierarchical clustering technique, was performed using the relative magnitudes of responses to a graded series of innocuous and noxious mechanical stimuli applied to the receptive field. For comparison, a parallel analysis was performed based on definitions of low-threshold (LT), wide dynamic range (WDR), and high-threshold (HT) cells used by our laboratory. For 128 STT cells, a classification scheme with three clusters was found statistically to be the best. This yielded groups of 22, 57, and 49 cells in clusters 1, 2, and 3, respectively. Cluster 1 cells were activated best by low-intensity mechanical stimuli, whereas cluster 3 cells were activated primarily by nociceptive stimuli. Cluster 2 cells had intermediate characteristics. When the classification scheme based on the cluster analysis was compared with the classification of the same neurons as LT, WDR, and HT cells, cluster 1 cells were divided into LT and WDR cells, whereas cluster 2 and 3 cells included WDR and HT cells. For 110 thalamic neurons, a classification scheme with five clusters was found statistically to be the best. Clusters 1-5 contained 25, 34, 17, 10, and 24 cells, respectively. Response characteristics of cells in each group indicated a gradual change in sensitivity to higher intensities of peripheral input from cluster 1 to 5. When this classification scheme was compared with the classification scheme previously used by our laboratory, cluster 1 cells belonged to the LT group, clusters 2 and 3 split into LT and WDR cells, and clusters 4 and 5 included WDR and HT cells. It is concluded that a classification scheme based on a cluster analysis of the responses of neurons to standardized stimuli may provide an objective and functionally meaningful way to categorize somatosensory neurons.

Animals↗

PubMed1987

COMSTAT rule for vigilance classification based on spontaneous EEG activity.

For the classification of sleep stages, international standards based on visual EEG analysis have been established and are in common use, although we are well aware of their limitations. Several authors have suggested different procedures for classifying the stages of vigilance during the waking stages. No universally accepted paradigm, however, has yet been developed. The proposed vigilance classification procedures are based either on visual or automatic analysis procedures. Even though the EEG activity and patterns that reflect vigilance changes have been identified and described as indicators of the state of alertness, opinion is divided on how these should be combined in a vigilance classification rule. Automatic methods, on the other hand, have up to now used only part of the information available, the relationship of which to vigilance indicators has only been partially explored. The COMSTAT (Dept. of Computation and Statistics, AFB-Arzneimittelforschung, Berlin, FRG) rule combines visual and automatic analysis procedures. Different vigilance-dependent EEG patterns, such as the proportion of occipital background rhythm under resting conditions and its replacement by either faster or slower waves, the frequency range of the occipital rhythm and the anteriorization phenomena, have been used as information for a latent class analysis (LCA5) with 5 classes (stages of vigilance). There is a high correlation between the results of the LCA5 with visual classification rules made by experts. Using a robust discriminant analysis function which takes into account prior probabilities of the classes, and with a linear cost function for misclassification, an automatic rule with power spectrum variables was fitted to the results of the LCA5. Reclassification and split-half classification showed a high overlap between LCA5 and automatic classification. The result of this procedure is a new vigilance classification rule that is based on an objective mathematical rationale for the combination of different vigilance-indicative EEG activities and patterns but which can be applied to power-spectral estimators in an automatic EEG analysis procedure.

Aged↗

PubMedDate not supplied

Aerobic capacity and functional classification of patients with severe left-ventricular dysfunction.

Classes I/II and III of the classification systems of the New York Heart Association (NYHA), Canadian Cardiovascular Society (CCS) and American Medical Association (AMA) were compared with each other and with the Weber classification (O2 uptake, VO2/kg during treadmill walking) in 35 male patients with severe left ventricular dysfunction. Measured end points were ventilatory threshold (VT) and peak exercise. Also investigated was whether the CCS and AMA scales, due to their more stringent differentiation, are more precise than the NYHA system in determining a limited physical capacity and whether there are other differentiating factors useful in classification which may be derived from cardiopulmonary exercise testing. At the VT, the mean VO2/kg did not differ significantly in any classification system between classes I/II and III (12.8 +/- 2.5 vs. 11.1 +/- 2.3 ml/kg/min) and corresponded to Weber class B. At peak exercise, the mean VO2/kg only differed significantly within the NYHA classification; classes I/II (16.3 +/- 3.1 ml/kg/min) corresponded to Weber class B, and class III (13 +/- 3 ml/kg/min) to Weber class C. The individual values displayed a large scatter. Factors differing in classes I/II and III of all three systems at peak exercise were the ventilatory equivalent of O2 and CO2 as well as end-tidal partial pressure for O2 and CO2. At VT these factors showed a separating character only in the AMA classification. It is not possible to determine objective functional impairment by use of the NYHA, CCS and AMA systems because they are not analogous to the Weber system. Nevertheless, these classification systems can be used for clinical assessment and follow-up.

Exercise Test↗

PubMed1996

What classification is appropriate in renal trauma?

OBJECTIVE: The objective of this paper is to attain an adequate consideration of the relevant factors in conventional classifications of renal trauma and to work out a differentiated classification with a simple trauma formula suitable for both scientific purposes and for clinical decision-making. METHOD: Previous classifications of renal trauma were evaluated retrospectively to establish whether and to what extent all criteria have been considered which may be important for the outcome of kidney trauma. The pathogenesis, the locations and the symptoms of the injuries with their degrees served as principal parameters. RESULTS: From 1950 to 1991, a minimum of 24 classifications of renal trauma with different criteria has been used. In 54.2% of these, blunt and perforating injuries were not distinguished. Injuries of the renal pelvis or the vascular system were not considered in 62.9 and 8.3% of these, respectively. The degrees of severity ranged from 2 to 6 (on average, 3.5). In 87.5%, there was a combination with the injury location (renal pelvis, vascular system). Finally, of 29 definable criteria, an average of only 9.6 was specified, i.e. most were not evaluated at all. CONCLUSION: It is concluded that most of the previous classifications of renal trauma neither clearly distinguish between nor adequately differentiate their individual factors. This may be a major reason for the persistence of controversies with regard to concepts of therapy, since they are based on inadequate definitions of comparative studies. A more differentiated PLS classification is proposed comprising the following elements: the pathogenesis P1 (blunt injuries), P2 (perforating injuries), the locations of the injuries LA (parenchyma), LB (renal pelvis), LC (vascular system) with their degrees of injury LA 0-7, LB 0-2, LC 0-9, and the symptoms of injury SA (hemorrhage), SB (extravasation of urine), SC (kidney damage) with their degrees of severity SA 0-3, SB 0-3, SC 0-3 as prognostic factors. The new classification incorporated in a simple trauma formula may enable more precise scientific investigation and also facilitate clinical decisions, so that patients with kidney trauma can be treated more specifically.

Adult↗

PubMed1996

Variability in salt sensitivity classifications in black male versus female adolescents.

Salt sensitivity (changes in blood pressure in response to alterations in salt intake) may be a risk factor for hypertension. In the present study, we examined the prevalence of salt sensitivity based on two different classifications in healthy black male and female adolescents (aged 13 to 16 years). A total of 135 black adolescents participated in a 50 mmol/24 h low sodium diet for 5 days and a 150 mmol/24 h NaCl supplement for 10 days. Dietary compliance was defined as sodium excretion less than or equal to 50 mmol/24 h for the low sodium diet and greater than or equal to 165 mmol/24h for the high NaCl supplement. Salt sensitivity was defined by two classifications: (1) as a decrease in mean blood pressure greater than or equal to 5 mm Hg from baseline to the low sodium diet, and (2) as an increase in mean blood pressure greater than or equal to 5 mm Hg from the low sodium diet to the high NaCl supplement. With classification 1, 14% of boys were identified as salt sensitive compared with 22% of girls. With classification 2, however, 31% of boys were identified as salt sensitive compared with 18% of girls. Analyses based on changes in systolic pressure demonstrated similar findings across sex, although overall classifications based on systolic pressure yielded a greater percentage of salt-sensitive subjects. These sex differences in classification patterns were not due to differences in other important variables, such as changes in sodium excretion, potassium excretion, or Quetelet index. These results suggest that the prevalence of salt sensitivity differs by sex depending on the type of protocol used for the classification of salt sensitivity in a black pediatric population.

Adolescent↗

PubMed2002

Optimal methods for calculating classification images: weighted sums.

In signal detection theory, an observer's responses are often modeled as being based on a decision variable obtained by cross-correlating the stimulus with a template, possibly after corruption by external and internal noise. The response classification method estimates an observer's template by measuring the influence of each pixel of external noise on the observer's responses. A map that shows the influence of each pixel is called a classification image. Other authors have shown how to calculate classification images from external noise fields, but the optimal calculation has never been determined, and the quality of the resulting classification images has never been evaluated. Here we derive the optimal weighted sum of noise fields for calculating classification images in several experimental designs, and we derive the signal-to-noise ratio (SNR) of the resulting classification images. Using the expressions for the SNR, we show how to choose experimental parameters, such as the observer's performance level and the external noise power, to obtain classification images with a high SNR. We discuss two-alternative identification experiments in which the stimulus is presented at one or more contrast levels, in which each stimulus is presented twice so that we can estimate the power of the internal noise from the consistency of the observer's responses, and in which the observer rates the confidence of his responses. We illustrate these methods in a series of contrast increment detection experiments.

Contrast Sensitivity↗

PubMed2006

Interobserver and intraobserver reliability of two classification systems for intra-articular calcaneal fractures.

BACKGROUND: For a fracture classification to be useful it must provide prognostic significance, interobserver reliability, and intraobserver reproducibility. Most studies have found reliability and reproducibility to be poor for fracture classification schemes. The purpose of this study was to evaluate the interobserver and intraobserver reliability of the Sanders and Crosby-Fitzgibbons classification systems, two commonly used methods for classifying intra-articular calcaneal fractures. METHODS: Twenty-five CT scans of intra-articular calcaneal fractures occurring at one trauma center were reviewed. The CT images were presented to eight observers (two orthopaedic surgery chief residents, two foot and ankle fellows, two fellowship-trained orthopaedic trauma surgeons, and two fellowship-trained foot and ankle surgeons) on two separate occasions 8 weeks apart. On each viewing, observers were asked to classify the fractures according to both the Sanders and Crosby-Fitzgibbons systems. Interobserver reliability and intraobserver reproducibility were assessed with computer-generated kappa statistics (SAS software; SAS Institute Inc., Cary, North Carolina). RESULTS: Total unanimity (eight of eight observers assigned the same fracture classification) was achieved only 24% (six of 25) of the time with the Sanders system and 36% (nine of 25) of the time with the Crosby-Fitzgibbons scheme. Interobserver reliability for the Sanders classification method reached a moderate (kappa = 0.48, 0.50) level of agreement, when the subclasses were included. The agreement level increased but remained in the moderate (kappa = 0.55, 0.55) range when the subclasses were excluded. Interobserver agreement reached a substantial (kappa = 0.63, 0.63) level with the Crosby-Fitzgibbons system. Intraobserver reproducibility was better for both schemes. The Sanders system with subclasses included reached moderate (kappa = 0.57) agreement, while ignoring the subclasses brought agreement into the substantial (kappa = 0.77) range. The overall intraobserver agreement was substantial (kappa = 0.74) for the Crosby-Fitzgibbons system. CONCLUSIONS: Although intraobserver kappa values reached substantial levels and the Crosby-Fitzgibbons system generally showed greater agreement, we were unable to demonstrate excellent interobserver or intraobserver reliability with either classification scheme. While a system with perfect agreement would be impossible, our results indicate that these classifications lack the reproducibility to be considered ideal.

Ankle Injuries↗

PubMed2002

The World Health Organization (WHO) classification of the myeloid neoplasms.

A World Health Organization (WHO) classification of hematopoietic and lymphoid neoplasms has recently been published. This classification was developed through the collaborative efforts of the Society for Hematopathology, the European Association of Hematopathologists, and more than 100 clinical hematologists and scientists who are internationally recognized for their expertise in hematopoietic neoplasms. For the lymphoid neoplasms, this classification provides a refinement of the entities described in the Revised European-American Lymphoma (REAL) Classification-a system that is now used worldwide. To date, however, there has been no published explanation or rationale given for the WHO classification of the myeloid neoplasms. The purpose of this communication is to outline briefly the WHO classification of malignant myeloid diseases, to draw attention to major differences between it and antecedent classification schemes, and to provide the rationale for those differences.

Humans↗

PubMed2005

Evaluation of normalization methods for cDNA microarray data by k-NN classification.

BACKGROUND: Non-biological factors give rise to unwanted variations in cDNA microarray data. There are many normalization methods designed to remove such variations. However, to date there have been few published systematic evaluations of these techniques for removing variations arising from dye biases in the context of downstream, higher-order analytical tasks such as classification. RESULTS: Ten location normalization methods that adjust spatial- and/or intensity-dependent dye biases, and three scale methods that adjust scale differences were applied, individually and in combination, to five distinct, published, cancer biology-related cDNA microarray data sets. Leave-one-out cross-validation (LOOCV) classification error was employed as the quantitative end-point for assessing the effectiveness of a normalization method. In particular, a known classifier, k-nearest neighbor (k-NN), was estimated from data normalized using a given technique, and the LOOCV error rate of the ensuing model was computed. We found that k-NN classifiers are sensitive to dye biases in the data. Using NONRM and GMEDIAN as baseline methods, our results show that single-bias-removal techniques which remove either spatial-dependent dye bias (referred later as spatial effect) or intensity-dependent dye bias (referred later as intensity effect) moderately reduce LOOCV classification errors; whereas double-bias-removal techniques which remove both spatial- and intensity effect reduce LOOCV classification errors even further. Of the 41 different strategies examined, three two-step processes, IGLOESS-SLFILTERW7, ISTSPLINE-SLLOESS and IGLOESS-SLLOESS, all of which removed intensity effect globally and spatial effect locally, appear to reduce LOOCV classification errors most consistently and effectively across all data sets. We also found that the investigated scale normalization methods do not reduce LOOCV classification error. CONCLUSION: Using LOOCV error of k-NNs as the evaluation criterion, three double-bias-removal normalization strategies, IGLOESS-SLFILTERW7, ISTSPLINE-SLLOESS and IGLOESS-SLLOESS, outperform other strategies for removing spatial effect, intensity effect and scale differences from cDNA microarray data. The apparent sensitivity of k-NN LOOCV classification error to dye biases suggests that this criterion provides an informative measure for evaluating normalization methods. All the computational tools used in this study were implemented using the R language for statistical computing and graphics.

Algorithms↗

PubMed2005

Optimized between-group classification: a new jackknife-based gene selection procedure for genome-wide expression data.

BACKGROUND: A recent publication described a supervised classification method for microarray data: Between Group Analysis (BGA). This method which is based on performing multivariate ordination of groups proved to be very efficient for both classification of samples into pre-defined groups and disease class prediction of new unknown samples. Classification and prediction with BGA are classically performed using the whole set of genes and no variable selection is required. We hypothesize that an optimized selection of highly discriminating genes might improve the prediction power of BGA. RESULTS: We propose an optimized between-group classification (OBC) which uses a jackknife-based gene selection procedure. OBC emphasizes classification accuracy rather than feature selection. OBC is a backward optimization procedure that maximizes the percentage of between group inertia by removing the least influential genes one by one from the analysis. This selects a subset of highly discriminative genes which optimize disease class prediction. We apply OBC to four datasets and compared it to other classification methods. CONCLUSION: OBC considerably improved the classification and predictive accuracy of BGA, when assessed using independent data sets and leave-one-out cross-validation. AVAILABILITY: The R code is freely available [see Additional file 1] as well as supplementary information [see Additional file 2].

Algorithms↗

PubMed2006

Regularized binormal ROC method in disease classification using microarray data.

BACKGROUND: An important application of microarrays is to discover genomic biomarkers, among tens of thousands of genes assayed, for disease diagnosis and prognosis. Thus it is of interest to develop efficient statistical methods that can simultaneously identify important biomarkers from such high-throughput genomic data and construct appropriate classification rules. It is also of interest to develop methods for evaluation of classification performance and ranking of identified biomarkers. RESULTS: The ROC (receiver operating characteristic) technique has been widely used in disease classification with low dimensional biomarkers. Compared with the empirical ROC approach, the binormal ROC is computationally more affordable and robust in small sample size cases. We propose using the binormal AUC (area under the ROC curve) as the objective function for two-sample classification, and the scaled threshold gradient directed regularization method for regularized estimation and biomarker selection. Tuning parameter selection is based on V-fold cross validation. We develop Monte Carlo based methods for evaluating the stability of individual biomarkers and overall prediction performance. Extensive simulation studies show that the proposed approach can generate parsimonious models with excellent classification and prediction performance, under most simulated scenarios including model mis-specification. Application of the method to two cancer studies shows that the identified genes are reasonably stable with satisfactory prediction performance and biologically sound implications. The overall classification performance is satisfactory, with small classification errors and large AUCs. CONCLUSION: In comparison to existing methods, the proposed approach is computationally more affordable without losing the optimality possessed by the standard ROC method.

Biomarkers, Tumor↗

PubMed2004

Binary classification of dyslipidemia from the waist-to-hip ratio and body mass index: a comparison of linear, logistic, and CART models.

BACKGROUND: We sought to improve upon previously published statistical modeling strategies for binary classification of dyslipidemia for general population screening purposes based on the waist-to-hip circumference ratio and body mass index anthropometric measurements. METHODS: Study subjects were participants in WHO-MONICA population-based surveys conducted in two Swiss regions. Outcome variables were based on the total serum cholesterol to high density lipoprotein cholesterol ratio. The other potential predictor variables were gender, age, current cigarette smoking, and hypertension. The models investigated were: (i) linear regression; (ii) logistic classification; (iii) regression trees; (iv) classification trees (iii and iv are collectively known as "CART"). Binary classification performance of the region-specific models was externally validated by classifying the subjects from the other region. RESULTS: Waist-to-hip circumference ratio and body mass index remained modest predictors of dyslipidemia. Correct classification rates for all models were 60-80%, with marked gender differences. Gender-specific models provided only small gains in classification. The external validations provided assurance about the stability of the models. CONCLUSIONS: There were no striking differences between either the algebraic (i, ii) vs. non-algebraic (iii, iv), or the regression (i, iii) vs. classification (ii, iv) modeling approaches. Anticipated advantages of the CART vs. simple additive linear and logistic models were less than expected in this particular application with a relatively small set of predictor variables. CART models may be more useful when considering main effects and interactions between larger sets of predictor variables.

Adult↗

PubMed1999

World Health Organization classification of neoplastic diseases of the hematopoietic and lymphoid tissues: report of the Clinical Advisory Committee meeting-Airlie House, Virginia, November 1997.

PURPOSE: The European Association of Hematopathologists and the Society for Hematopathology have developed a new World Health Organization (WHO) classification of hematologic malignancies, including lymphoid, myeloid, histiocytic, and mast cell neoplasms. DESIGN: Ten committees of pathologists developed lists and definitions of disease entities. A clinical advisory committee (CAC) of international hematologists and oncologists was formed to ensure that the classification would be useful to clinicians. The CAC met in November 1997 to discuss clinical issues related to the classification. RESULTS: The WHO uses the Revised European-American Lymphoma (REAL) classification, published in 1994 by the International Lymphoma Study Group, to categorize lymphoid neoplasms. The REAL classification is based on the principle that a classification is a list of "real" disease entities, which are defined by a combination of morphology, immunophenotype, genetic features, and clinical features. The relative importance of each of these features varies among diseases, and there is no one gold standard. The WHO Neoplasms recognizes distinct entities defined by a combination of morphology and cytogenetic abnormalities. At the CAC meeting, which was organized around a series of clinical questions, participants reached a consensus on most of the questions posed. They concluded that clinical groupings of lymphoid neoplasms were neither necessary nor desirable. Patient treatment is determined by the specific type of lymphoma, with the addition of grade within the tumor type, if applicable, and clinical prognostic factors, such as the International Prognostic Index. CONCLUSION: The WHO classification has produced a new and exciting degree of cooperation and communication between oncologists and pathologists from around the world, which should facilitate progress in the understanding and treatment of hematologic malignancies.

Hematologic Neoplasms↗