Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “classifier”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

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

At least 505 records · Page 28Linked to original sources
PubMed2006

Validation of pharmacogenomic biomarker classifiers for treatment selection.

Physicians need improved tools for selecting treatments for individual patients. Many syndromes traditionally viewed as individual diseases are heterogeneous in molecular pathogenesis and treatment responsiveness. This results in treatment of many patients with ineffective drugs and leads to the conduct of large clinical trials to identify small average treatment benefits for heterogeneous groups of patients. New genomic and proteomic technologies provide powerful tools for the selection of patients likely to benefit from a therapeutic without unacceptable adverse events. In spite of the large literature on developing predictive biomarkers and on statistical methodology for analysis of high dimensional data, there is considerable uncertainty about the validation of biomarker based diagnostic classifiers for treatment selection. In this paper we attempt to clarify these issues and to provide guidance on the design of clinical trials for evaluating the clinical utility and robustness of pharmacogenomic classifiers.

Biomarkers↗
PubMed2005

Classifying genotype F of hepatitis B virus into F1 and F2 subtypes.

AIM: To explore the propriety of providing hepatitis B virus (HBV) genotypes F and H with two distinct genotypes. METHODS: Eleven HBV isolates of genotype F (HBV/F) were recovered from patients living in San Francisco, Japan, Panama, and Venezuela, and their full-length sequences were determined. Phylogenetic analysis was carried out among them along with HBV isolates previously reported. RESULTS: Seven of them clustered with reported HBV/F isolates in the phylogenetic tree constructed on the entire genomic sequence. The remaining four flocked on another branch along with three HBV isolates formerly reported as genotype H. These seven HBV isolates, including the four in this study and the three reported, had a sequence divergence of 7.3-9.5% from the other HBV/F isolates, and differed by > 13.7% from HBV isolates of the other six genotypes (A-E and G). Based on a marked genomic divergence, falling just short of > 8% separating the seven genotypes, these seven HBV/F isolates were classified into F2 subtype and the former seven into F1 subtype provisionally. In a pairwise comparison of the S-gene sequences among the 7 HBV/F2 isolates and against 47 HBV/F1 isolates as well as 136 representing the other six genotypes (A-E and G), two clusters separated by distinct genetic distances emerged. CONCLUSION: Based on these analyses, classifying HBV/F isolates into two subtypes (F1 and F2) would be more appropriate than providing them with two distinct genotypes (F and H).

Amino Acid Sequence↗
PubMed1999

Induction with cross-classified categories.

One of the main functions of categories is to allow inferences about new objects. However, most objects are cross-classified, and it is not known whether and how people combine information from these different categories in making inferences. In six experiments, food categories, which are strongly cross-classified (e.g., a bagel is both a bread and a breakfast food), were studied. For each food, the subjects were told fictitious facts (e.g., 75% of breads are subject to spoilage from Aspergillus molds) about two of the categories to which it belonged and then were asked to make an inference about the food (e.g., how likely is a bagel to be subject to spoilage from Aspergillus molds?). We found no more use of multiple categories in these cases of cross-classification than in ambiguous classification, in which it is uncertain to which category an item belongs. However, some procedural manipulations did markedly increase the use of both categories in inferences, primarily those that focused the subjects' attention on the critical feature in both categories.

Adolescent↗
PubMed2006

The Wisconsin breast cancer problem: Diagnosis and TTR/DFS time prognosis using probabilistic and generalised regression information classifiers.

This study addresses the breast cancer diagnosis and prognosis problem by employing two neural network architectures with the Wisconsin diagnostic and prognostic breast cancer (WDBC/WPBC) datasets. A probabilistic approach is dedicated to solve the diagnosis problem, detecting malignancy among cases (instances) as derived from fine needle aspirate (FNA) tests, while the second architecture estimates the time interval that possibly contains the right endpoint of disease-free survival (DFS) of the patient. The accuracy of the neural classifiers reaches nearly 98% for the diagnosis and 93% for the prognosis problem, while the prognostic recurrence predictions were evaluated using survival analysis through the Kaplan-Meier approximation method. Both architectures were compared with other similar approaches. The robustness and real-time response of the proposed classifiers were further tested over the web as a potential integrated web-based decision support system.

Biopsy, Needle↗
PubMed2000

p53, ErbB2, and TAG-72 expression in the spectrum of ductal carcinoma in situ of the breast classified by the Van Nuys system.

CONTEXT: The Van Nuys (VN) classification system for ductal carcinoma in situ (DCIS) of the breast is a simplified morphology-based system that uses the presence of nuclear pleomorphism and comedo-type necrosis to stratify DCIS lesions into 3 prognostic groups. OBJECTIVE: To determine if there is an underlying biological basis that correlates with the morphologic aspects of the VN classification system. DESIGN: We evaluated the expression of markers implicated in the development of breast cancer (p53, ErbB2, and TAG-72) in DCIS classified with the VN system. Forty-five cases of pure DCIS were classified as 8 cases of VN1, 7 cases of VN2, and 30 cases of VN3. p53, ErbB2, and TAG-72 antigen expression was measured by immunohistologic means in each of the cases. RESULTS: Nuclear accumulation of p53 was only observed in VN3 (30%). ErbB2 overexpression was found only in VN2 (14%) and VN3 (43%). TAG-72 expression was observed in all categories of lesions but was more frequent in VN2 (71%) and VN3 (70%) compared with VN1 (25%). It appears that overexpression of ErbB2 and p53 are features associated with the high-grade lesions. CONCLUSION: The simplified VN classification system for DCIS has a clear underlying biological basis as evidenced by differential expression of tumor-associated antigens in each of the 3 morphologic categories. These differences may contribute to the differential clinical behavior of the separate groups.

Antigens, Neoplasm↗
PubMed2003

Nitrergic hyperinnervation in appendicitis and in appendices histologically classified as normal.

CONTEXT: The pathogenesis of appendicitis remains poorly understood. Despite new diagnostic techniques, appendices removed from patients with suspected appendicitis often appear histologically normal on conventional examination. There is increasing evidence of involvement of the enteric nervous system in immune regulation and in inflammatory responses in the gastrointestinal system. OBJECTIVE: To investigate the nitrergic innervation of (a) acutely inflamed appendices, (b) appendices classified as histologically normal from patients with a clinical diagnosis of appendicitis, and (c) normal control appendix specimens, using the whole-mount preparation technique. PATIENTS AND DESIGN: Full-thickness specimens were collected from 28 acutely inflamed appendices (age range, 3.2-13.4 years), 31 histologically normal appendices removed from patients (age range, 5.7-13.6 years) with suspected appendicitis, and 23 histologically normal appendices from patients (age range, newborn to 12.1 years) undergoing elective abdominal surgery (controls). Whole-mount preparation using nicotinamide adenine dinucleotide phosphate (NADPH) diaphorase histochemistry and neuronal nitric oxide synthase immunohistochemistry were performed. The density of myenteric plexus was measured with a computerized analysis system. RESULTS: The density of myenteric plexus in normal appendix specimens was similar to that of large bowel from the newborn period up to 3 years of age; this density decreased significantly thereafter. The myenteric plexus of normal appendix specimens from patients older than 4 years demonstrated smaller ganglia connected by thin nerve bundles, compared to larger ganglia and nerve bundles in large bowel. Significant neuronal hypertrophy was found in 55% of acutely inflamed and 41% of histologically classified normal appendix specimens. The myenteric plexus of these appendix specimens had even thicker nerve bundles connecting an increased number of ganglion cells. CONCLUSIONS: Differences in the architecture of the myenteric plexus in patients older than 3 years suggest an altered function and motility of appendix in the early years of life. The significant increase in neuronal components of the myenteric plexus in a high proportion of acutely inflamed and histologically normal appendix specimens is unlikely to have developed during a single acute inflammatory episode. This suggests an underlying chronic abnormality as a secondary response to chronic luminal obstruction or repeated inflammatory episodes in the histologically normal appendix.

Adolescent↗
PubMed1999

Reevaluation of discograms not classified into usual classifications.

Discograms of images that were eccentrically dyed because of insufficient infiltration of contrast medium are difficult to classify into the usual past discogram patterns. In this study, these types of images were detected in 40 discs of 36 patients with lumbar disc disease. We classified these images into the following three types, and analyzed the dye mechanisms in each case by computed tomography discographic findings: (1) type A (image of the annulus fibrosus only). Nine discs in nine cases. A part of the marginal annulus fibrosus was dyed. (2) type B (image of the right or left half of the nucleus pulposus). Eighteen discs in 15 cases. Unilateral dyeing was considered nucleus pulposus existing in the central region of the disc. (3) type C (partial image of the superior or inferior half of the nucleus pulposus). Thirteen discs in 12 cases. Only the superior or inferior half showing a cotton-ball pattern was dyed.

Adolescent↗
PubMed1999

[Risk of malignant degeneration of preoperatively-classified benign large sessile rectal polyps. A comparison with adenoma size].

From January 1986 to December 1995 307 patients with preoperatively as benign classified rectal polyps underwent transanal endoscopic microsurgery or transanal excision at the Hospital of General and Abdominal Surgery, Johannes Gutenberg-University Mainz. Mean polyp size was 3.9 cm in diameter. Postoperatively in 233 patients (75.9%) a benign rectal adenoma was found. In 69 patients (22.5%) with the preoperative diagnosis "benign rectal adenoma" the postoperative histologic result was a carcinoma. No residual polyps were encountered on snare excision in 5 patients (1.6%) with inconclusive evidence of surgical margin involvement. The mean size of malignant polyps was 3.4 cm and significantly below the mean size of benign polyps (4.1 cm, p = 0.009). Especially in polyps with a size to 1 cm and from 1 to 2 cm in diameter the part of malignant rectal polyps was unexpected high (8/15 polyps, 53.3%, respectively 16/37 polyps, 43.3%. Patients with preoperatively as benign classified large sessile rectal polyps had a high risk of malignancy even when polyp's size was small. Therefore in toto excision has to be done also in small polyps.

Adenocarcinoma↗
PubMed2000

Patients with abnormal ultrasound of the prostate but normal digital rectal examination should be classified as having clinical stage T2 tumors.

PURPOSE: The current TNM staging system classifies prostate tumors with abnormal transrectal ultrasound but normal digital rectal examination as clinical stage T2. However, most urologists consider these tumors as clinical stage T1c due to the perceived inaccuracy of transrectal ultrasound in clinical staging. To determine the role of transrectal ultrasound in the clinical staging of prostate cancer we evaluated the pathological stage and disease-free survival of patients undergoing radical prostatectomy who had tumor detected by needle biopsy because of elevated serum prostate specific antigen with or without transrectal ultrasound abnormalities. MATERIALS AND METHODS: Between 1991 and 1996, 738 patients underwent radical retropubic prostatectomy as monotherapy for clinically localized prostate cancer. Patients were classified into group 1-normal digital rectal examination and transrectal ultrasound (138), group 2-normal digital rectal examination but abnormal transrectal ultrasound (366) and group 3 -abnormal digital rectal examination (234). We compared pathological parameters and disease-free-survival among the 3 groups. RESULTS: Tumors were organ confined in 61%, 42% and 41% of patients in groups 1, 2 and 3, respectively (p = 0.0001). Overall disease-free survival was 80% with a mean followup of 68 months. Disease recurred in 8%, 22% and 25% of patients in groups 1, 2 and 3, respectively (p = 0.007). Group 1 had better disease-free survival compared to groups 2 and 3 (p = 0.003 and p = 0.002, respectively), and there was no difference in disease-free survival between groups 2 and 3 (p = 0.39). CONCLUSIONS: We provide evidence to support the use of transrectal ultrasound findings in the clinical staging system for prostate cancer. Patients with normal digital rectal examination, elevated serum prostate specific antigen and abnormal transrectal ultrasound should be considered as having clinical stage T2 disease.

Adult↗
PubMed2000

[Latent autoimmune (type 1) diabetes mellitus in patients originally classified as type 2. Divergence of etiologic markers].

BACKGROUND: To assess the prevalence of markers of autoimmune insulitis (AII) in patients classified originally as having Type-2 diabetes mellitus (Type-2 DM). 386 patients subdivided according to the BMI, C-peptide and type of treatment. METHODS AND RESULTS: Age, BMI, C-peptide, Glutamic acid decarboxylase autoantibodies (GADA), HLA-DR/,-DQ alleles. Prevalence of GADA varied from < 5% in obese patients with normal/increased C-peptide to > 30% in non-obese patients with low C-peptide. In majority of GADA positive patients, the Type-1 DM high-risk HLA-DRB1*, HLA-DQB1* alleles have been found. Among them HLA-DRB1*0302 and HLA-DRB1*0201 were more frequent than HLA-DRB1*040x and HL:A-DQB1*0302. CONCLUSIONS: Significant fraction of patients classified initially as Type-2 DM may have in fact Type-1 DM. Such patients can be recognized on the basis of assessment of serological (GADA) and immuno-genetical (HLA-DR/,-DQ alleles) markers. In some patients clinical, metabolic, immune, and immunogenetic markers may disagree. This divergence stresses multifactorial genesis of diabetes. Moreover, it can also suggest that both autoimmune insulitis and insulin resistance may coexist in parallel.

Adult↗
PubMed2000

Improving screening for melanoma by measuring similarity to pre-classified images.

A new method that can dramatically improve the sensitivity and especially the specificity of skin naevi screening for melanoma by physicians is introduced. The method is based on measuring the similarity to previously classified naevi images. As more naevi are being classified, the power of the method is enhanced. Thus physicians can benefit from experience accumulated by others and the screening can be performed effectively by physicians with less experience.

Dysplastic Nevus Syndrome↗
PubMed2001

A multivariate method for classifying third to seventh cervical newborn vertebrae using bone measurements.

This research provides a method to classify human newborn (9 to 10 lm) third to seventh cervical ossification centers. Nine linear measurements on the cervical neural arch were defined from 35 human neonates. Four discriminant functions were performed using the stepwise method. The model classifies 82.8% of grouped cases and 77.9% of cross-validated cases correctly. The model is useful in cases with isolated or commingled remains during anthropological or forensic investigations.

Cervical Vertebrae↗
PubMed2001

Classifying vertical facial deformity using supervised and unsupervised learning.

OBJECTIVES: To evaluate the potential for machine learning techniques to identify objective criteria for classifying vertical facial deformity. METHODS: 19 parameters were determined from 131 lateral skull radiographs. Classifications were induced from raw data with simple visualisation, C5.0 and Kohonen feature maps; and using a Point Distribution Model (PDM) of shape templates comprising points taken from digitised radiographs. RESULTS: The induced decision trees enable a direct comparison of clinicians' idiosyncrasies in classification. Unsupervised algorithms induce models that are potentially more objective, but their blackbox nature makes them unsuitable for clinical application. The PDM methodology gives dramatic visualisations of two modes separating horizontal and vertical facial growth. Kohonen feature maps favour one clinician and PDM the other. Clinical response suggests that while Clinician 1 places greater weight on 5 of 6 parameters, Clinician 2 relies on more parameters that capture facial shape. CONCLUSIONS: While machine learning and statistical analyses classify subjects for vertical facial height, they have limited application in their present form. The supervised learning algorithm C5.0 is effective for generating rules for individual clinicians but its inherent bias invalidates its use for objective classification of facial form for research purposes. On the other hand, promising results from unsupervised strategies (especially the PDM) suggest a potential use for objective classification and further identification and analysis of ambiguous cases. At present, such methodologies may be unsuitable for clinical application because of the invisibility of their underlying processes. Further study is required with additional patient data and a wider group of clinicians.

Algorithms↗
PubMed1992

Computer-aided classification of malignancy in astrocytomas. II. The value of categorically evaluated histologic and non-histologic features for a numerical classifier.

The present study was carried out in order to obtain a numerical classifier for the assessment of the malignancy in astrocytomas including glioblastomas ('astrocytomas grade 4'). The attempt resulted in 'TESTAST 268', a classifier based on a reference sample of 268 tumours, 67 in each of four malignancy classes. TESTAST 268 aids the identification of astrocytomas with one of four malignancy classes by means of eight classification variables, five histologic and three non-histologic. Identification is achieved with the aid of linear discriminant functions, both according to Bayes' decision rule (BAYTEST) and by canonical discriminant analysis (CANTEST) using the squared Mahalanobis distance. The discriminant functions with the calibration of the reference sample of the 268 tumours may be implemented on personal and even small pocket computers for practical application.

Astrocytoma↗
PubMed2003

A framework for classifying decision support systems.

BACKGROUND: Computer-based clinical decision support systems (CDSSs) vary greatly in design and function. A taxonomy for classifying CDSS structure and function would help efforts to describe and understand the variety of CDSSs in the literature, and to explore predictors of CDSS effectiveness and generalizability. OBJECTIVE: To define and test a taxonomy for characterizing the contextual, technical, and workflow features of CDSSs. METHODS: We retrieved and analyzed 150 English language articles published between 1975 and 2002 that described computer systems designed to assist physicians and/or patients with clinical decision making. We identified aspects of CDSS structure or function and iterated our taxonomy until additional article reviews did not result in any new descriptors or taxonomic modifications. RESULTS: Our taxonomy comprises 95 descriptors along 24 descriptive axes. These axes are in 5 categories: Context, Knowledge and Data Source, Decision Support, Information Delivery, and Workflow. The axes had an average of 3.96 coded choices each. 75% of the descriptors had an inter-rater agreement kappa of greater than 0.6. CONCLUSIONS: We have defined and tested a comprehensive, multi-faceted taxonomy of CDSSs that shows promising reliability for classifying CDSSs reported in the literature.

Classification↗
PubMed1992

[A clinical study on the therapeutic significance of classifying total amnesia].

Eight cases of total amnesia were reported. In common to all the cases, there were the definite histories of characteristic situations charged with depressive and isolated mood, although their modes of social adjustment appeared relatively good respectively. Four of the eight cases presented the clinical course of total amnesia as Yamada et al. had previously pointed out. The other four developed total amnesia in peculiar progress and presented rather involved clinical courses. Examining the clinical courses and psychodynamics, the author has suggested that total amnesia would be able to classify two types, such as "simple course type" and "unstable course type". The important difference between two types is as follows. "Simple course type": There were pressing situations after series of actual pending problems such as debts, troubles with a lover and mismanagement in business. Onset of them were reactive under these environmental situations. They presented calm and stable features during partial amnestic period. As the family relations and the actual environments were improved better, they recalled more parts of their personal histories without any difficulty. "Unstable course type": In these cases, the intrapsychic weakness were more important factors than the actual environmental problems. Unlike "simple course type", whenever they recalled some parts of their personal histories, they presented unstable fluctuating features with confusion. Therefore, they needed intensive therapeutic intervention and concern. Generally speaking, total amnesia can be considered not only as defense for the actual conflicts with a repression mechanism but also as "psychologic suicide" which Abeles et al. have mentioned. In addition, emotionally confused evolution in this type may be regarded as exposure of the intrapsychic weakness itself. By the way, the close relationship between total amnesia and suicide has been mentioned in some papers. In this type of total amnesia, attempted suicides were actually seen. So the author has stressed that the risk of suicide should be impressed in the treatment of total amnesia. Among the previously reported cases of total amnesia, in which the clinical courses can be read, 27 cases are "simple course type" and 22 cases are "unstable course type" according to the author's knowledge. And the number of total amnesia has been increasing with the times. So in future, total amnesia may increase in number. The author has emphasized the therapeutic significance of classifying total amnesia. It is useful for the treatment of total amnesia to recognize "unstable course type", in order to foresee the unstable clinical course.(ABSTRACT TRUNCATED AT 400 WORDS)

Adult↗
PubMed1992

[A nested case-control study of risk factors for intracerebral hemorrhage and cerebral infarction classified by computed tomographic findings].

Risk factors for intracerebral hemorrhage (ICH) and cerebral infarction (CI), were studied by a prospective study of 7,390 men and women aged 40-69 without a history of stroke living in three rural populations in Japan. Baseline examinations were done for populations in Akita-Ikawa and Akita-Ishizawa in 1975-1979, and for Ibaraki-Kyowa in 1981-1987, and followed until 1989 for Akita-Ikawa and Ibaraki-Kyowa and 1987 Akita-Ishizawa. There were 246 stroke cases diagnosed by clinical criteria during the follow-up period in which 74 percent (n = 181) had data from computed tomography (CT) performed within three weeks of the onset. According to these CT-findings, 181 stroke were classified as 48 with ICH, 50 with CI in penetrating artery regions (penetrating artery infarction), 33 with CI in cortical artery regions (cortical artery infarction), and 31 with subarachnoid hemorrhage while there were 19 with stroke without any evident CT abnormality. Cortical artery infarction was further classified as embolic type (n = 17) and thrombotic type (n = 9) according to clinical findings of the onset and presence of possible embolic sources such as atrial fibrillation, congenital heart disease, myocardial infarction and heart valve diseases. Using a nested case-control design, risk variables at baseline examination were compared between 131 stroke cases, 48 ICH and 83 CI, with 655 controls matched for sex, age (+/- 3), and the follow-up year. Univariate analysis showed that high blood pressure was associated with all types of stroke. From conditional logistic regression analysis significant risk variables were found to be high blood pressure for ICH and penetrating artery infarction, while atrial fibrillation and ST-T abnormality in electrocardiogram (ECG) were risk variables for cortical artery infarction. Associations with hypertensive or arteriosclerotic changes in ocular fundus were stronger for penetrating artery infarction than ICH and cortical artery infarction. ST-T abnormality in ECG was associated with embolic type cortical artery infarction and high blood pressure was associated with the thrombotic type, although the number of cases were small. Compared to controls, cortical artery infarction showed a higher mean value of serum total cholesterol for thrombotic type cortical infarction, and lower mean values for embolic type and ICH, but none of them reached statistical significance. The present study also suggests that duration of hypertension varied with type of stroke. ICH may develop due to acute effects of hypertension, while penetrating artery infarction and cortical artery infarction develop by chronic effects of hypertension.

Adult↗
PubMedDate not supplied

Structure of the O-polysaccharide of Proteus mirabilis O19 and reclassification of certain Proteus strains that were formerly classified in serogroup O19.

INTRODUCTION: Bacteria of the genus Proteus are a common cause of urinary tract infections. The O-polysaccharide chain of their LPS (O-antigen) defines the serological specificity of these bacteria. Based on the immunospecificity of the O-antigens, two species, P. mirabilis and P. vulgaris, were classified into 49 O-serogroups, and more O-serogroups for strains of these species and P. penneri have been subsequently proposed. MATERIAL AND METHODS: The lipopolysaccharide of P.mirabilis CCUG 19011 from serogroup O19 was degraded under mildly acidic and mildly alkaline conditions. Polysaccharides thus obtained were studied by chemical methods, including O -deacetylation, sugar and methylation analyses, and 1H- and 13C-NMR spectroscopy. Antisera were obtained by immunization of New Zealand white rabbits with heat-killed bacteria. In serological studies, enzyme immunosorbent assay, passive hemolysis test, and inhibition of passive hemolysis were used. RESULTS: The following structure of the O-polysaccharide repeating unit was established:-->3)- beta-D-GlcrhoNAc-(1-->3)- alpha-D-GalrhoNAc4,6(R-Pyr)-(1-->4)- a-D-GalrhoA-(1-->3) alpha-L-Rhap2Ac-(1-->where R-Pyr is (R)-1-carboxyethylidene (an acetal-linked pyruvic acid). This structure is significantly different from the O-polysaccharide structures of P. vulgaris, P.hauseri and P. penneri strains from the same Proteus serogroup O19. CONCLUSIONS: Based on immunochemical studies of the lipopolysaccharides, it is suggested 1) to keep P. vulgaris CCUG 4654 and P. penneri 31 in serogroup O19 as two subgroups, 2) to reclassify P. mirabilis CCUG 19011 into a new Proteus serogroup, O51, and 3) to classify serologically related strains, including P. vulgaris ATCC 49990, P. hauseri> 1732-80 and 1086-80, P. penneri 15, and some other P. penneri strains, in yet another Proteus serogroup, O52.

Animals↗