Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Multivariate statistics”

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 577 records · Page 32Linked to original sources
PubMed1990

Blood transfusion and prognosis in Dukes' B and C colorectal cancer.

To evaluate the prognostic influence of blood transfusion in cancer patients, transfusion data were reviewed on 468 radically operated patients (260 Dukes' B and 208 Dukes' C) with carcinoma of the rectum and the rectosigmoid. Data on whole blood and packed red blood cell transfusions were recorded together with a number of clinical, pathological and histochemical characteristics. The endpoint used was death with cancer. All patients were followed for 2-7 years or until time of death. Univariate statistical methods revealed a highly significant trend towards worsened prognosis with increasing volume of transfusion blood. However, this effect was insignificant when multivariate statistical methods were employed: patients receiving whole blood or packed red blood cell transfusions did no worse than expected from their clinico-pathological characteristics. It is concluded that in this series the observed association between transfusion status and prognosis is adequately explained by a multivariate prognostic model including well-established prognostic factors.

Aged↗
PubMed2005

Estimating third molar extraction difficulty: a comparison of subjective and objective factors.

PURPOSE: The purpose of this study was to compare and contrast subjective and objective assessments of variables associated with third molar (M3) extraction difficulty. MATERIALS AND METHODS: To address the research purpose we implemented a prospective cohort study and enrolled a sample of surgeons removing M3s in an ambulatory care setting. Predictor variables were categorized as demographic, anatomic, or operative. The outcome variables were subjective and objective rankings of the importance of the variables in terms of estimating M3 extraction difficulty. Subjective rankings were made by surveying the surgeons and asking them to rank each variable's importance on a scale ranging from 0 (not important) to 100 (extremely important). Objective rankings of each variable's importance were made using the absolute values of coefficients derived from a multivariate linear regression model with extraction time as the outcome. Appropriate uni-, bi-, and multivariate statistics were computed. RESULTS: The sample consisted of 14 surgeons who removed 450 M3s from 150 subjects from June 2002 to August 2003. Based on the multivariate linear regression model, variables associated with M3 extraction time were gender, arch location, Winter's classification, tooth morphology, number of teeth extracted, procedure type, and surgical experience. For these variables, there was a strong, statistically significant correlation (r = 0.86; P <.01) between the standardized coefficient absolute values and the surgeons' estimates of importance. CONCLUSION: There was a large, positive correlation between variables that surgeons consider most important in determining M3 extraction difficulty and those exhibiting the most influence over extraction times in a multivariate model.

Adult↗
PubMed2005

The early origins and development of the scatterplot.

Of all the graphic forms used today, the scatterplot is arguably the most versatile, polymorphic, and generally useful invention in the history of statistical graphics. Its use by Galton led to the discovery of correlation and regression, and ultimately to much of present multivariate statistics. So, it is perhaps surprising that there is no one widely credited with the invention of this idea. Even more surprising is that there are few contenders for this title, and this question seems not to have been raised before. This article traces some of the developments in the history of this graphical method, the origin of the term scatterplot, the role it has played in the history of science, and some of its modern descendants. We suggest that the origin of this method can be traced to its unique advantage: the possibility to discover regularity in empirical data by smoothing and other graphic annotations to enhance visual perception.

Animals↗
PubMed1995

Visual information processing: topography of brain electrical activity.

Multichannel recordings allow the non-invasive assessment of the electrical fields of the brain. Topographical analysis of EEG and EPs should not be restricted to the qualitative graphical display of maps at many time points instead of time series at many recording points. It is mandatory that quantitative methods are applied to EEG and evoked potential data in order to extract relevant information from such series of maps. For spontaneous EEG data the spatial analysis of frequency bands is of major interest while evoked potential fields are analyzed in terms of latencies and amplitudes of individual components. The computation of global field power, global dissimilarity as well as the location of potential maxima, minima or centroids enable a meaningful data reduction and the statistical evaluation of the effects of different experimental conditions and the comparison of various clinical populations. In a similar way multivariate statistical techniques may be employed in order to detect characteristic potential field configurations. Topographical methods to analyze brain electrical activity are illustrated with multichannel data evoked by localized visual stimuli and by stereoscopic stimulation as well as with evoked potential data obtained during perceptual learning experiments.

Brain↗
PubMed1988

Evaluation of preoperative blood tests for predicting deep vein thrombosis after total hip replacement.

Determinations of the total calcium content and aggregability of the platelets as well as tests of coagulation and fibrinolysis were carried out on 93 patients before undergoing total hip replacement. All patients received low dose heparin, solely or combined with dihydroergotamine. Twenty-six patients developed deep vein thrombosis (DVT) detected by the labelled fibrinogen uptake test and confirmed by ascending phlebography. Only a few tests, among them the total calcium content of platelets, showed a statistical difference between the patients who subsequently developed DVT and those who did not. Combination of several tests by a multivariate statistical analysis programme proved to have more predictive value than the analysis of single tests.

Adult↗
PubMed1977

PaO2 levels and retrolental fibroplasia: a report of the cooperative study.

The relation between PaO2 and retrolental fibroplasia (RLF) was studied prospectively in 719 premature infants born in or treated in the intensive care units of a group of university hospitals. Blood gas studies were performed on 589 of these infants, 66 of whom had a diagnosis of RLF; in 27 of these 66, some grade of mostly nonblinding cicatricial disease developed. The frequency of RLF was highest among infants of lowest birth weight. A multivariate statistical method was used to analyze simultaneously the effect of possible etiologic factors associated with RLF. The occurrence of RLF was found to be unrelated to PaO2, as determined by the limited information available from intermittent sampling. RLF is associated with concentration of oxygen administered in the lightest birth weight group, but the strongest association, aside from birth weight, was with time in oxygen. None of the other variables involving blood chemical values appeared to be associated with RLF. The severity of cicatricial RLF is clearly greater in infants weighing less than 1,200 g at birth. Conservative administration of oxygen may have been responsible for failure to demonstrate quantitative association between PaO2 levels and disease. Agreement between the observed and predicted numbers of infants with RLF demonstrate the strength of the multivariate technique employed in making the statistical analyses.

Birth Weight↗
PubMed1998

Alcohol consumption enhances liver metastasis in colorectal carcinoma patients.

BACKGROUND: It is important to identify risk factors for liver metastasis in patients with colorectal carcinoma because the liver is the most common site of recurrence. Alcohol consumption reportedly is associated with hematogenous metastasis in certain animal models. Furthermore, some studies have shown that carmofur, a derivative of 5-fluorouracil, is particularly effective as adjuvant chemotherapy for colorectal carcinoma, and may even suppress liver metastasis, although the mechanism by which this occurs remains unknown. In addition, carmofur is known to inhibit alcohol metabolism. To the authors' knowledge, the relation between liver metastasis in colorectal carcinoma and alcohol consumption has not been examined previously. Therefore, the authors studied the relations between liver metastasis in colorectal carcinoma and various clinicopathologic factors including alcohol consumption status. METHODS: This study was comprised of 133 colorectal carcinoma patients with invasion beyond the submucosal layer who had undergone surgical resection. The subjects were examined and divided into two groups according to the occurrence or absence of liver metastasis. The relations between liver metastasis and other clinicopathologic factors were analyzed by univariate and multivariate statistical methods. RESULTS: Univariate analysis showed alcohol consumption (P=0.0021) and blood vessel invasion (P=0.0045) were correlated with liver metastasis. Multivariate analysis showed both to be independent risk factors for liver metastasis. CONCLUSIONS: Alcohol consumption is an independent risk factor for liver metastasis in colorectal carcinoma patients. Therefore, patients with colorectal carcinoma who drink alcohol require intensive examination and follow-up with respect to liver metastasis. Further study is necessary to confirm the effect of adjuvant chemotherapy using carmofur in colorectal carcinoma patients.

Adult↗
PubMed1992

Morphometry and prognosis in cancer of the pancreatic head.

The histological slides of 39 cases of cancer of the pancreatic head were analysed using an interactive image analyser system. Some 14 cases were classified as periampullary, 25 as ductal pancreatic cancer. All cases had undergone radical tumor resection according to Whipple's procedure. Morphometric data, tumor size and metastatic nodal involvement were correlated to prognosis. Univariate statistical analysis showed that the classical differentiation between ductal and periampullary cancer was a weaker prognosticator than morphometric variables. In fact, multivariant statistical analysis showed that the morphometric variable irregularity was the best prognosticator (p = 0.0001). No other variable added significant prognostic information. Irregularity is a newly developed variable describing the nuclear shape corrected for roundness. We conclude that morphometry can be of essential prognostic information for the clinician in cancer of the pancreatic head.

Carcinoma, Intraductal, Noninfiltrating↗
PubMed1996

[Prognostic factors in patients with non-metastatic renal cell carcinoma].

BACKGROUND: The objective of this study is to evaluate prognostic factors of non-metastatic renal cell carcinoma using univariate and multivariate statistical methods. METHODS: 390 patients who were performed radical surgery in our hospital between 1960 and 1994 were entered to this study. Kaplan-Meier method with log-rank test and Cox's proportional hazard model were used for statistical analysis. RESULTS: Chief complaints, tumor localization on kidney, tumor size, fever, hemoglobin, C-reactive protein, erythrocyte sedimentation rate, immunosuppressive acidic protein, operative method, blood loss, histologic grade, cell type, pT, pN and pV revealed to have prognostic statistical significance in the univariate analysis. Whereas, sex, age, blood type, hypertension, affected side and preoperative embolization did not. In the multivariate analysis of the preoperatively evaluable factors, CRP, tumor size and tumor localization proved to be significant prognostic predictors in this order. When the pathological factors were included to the multivariate analysis, pV, historogic grade and pN revealed to be significant prognostic predictors in this order. CONCLUSION: These results suggest that pathological factors were significant prognostic predictors for non-metastatic renal cell carcinoma.

Carcinoma, Renal Cell↗
PubMedDate not supplied

High performance liquid chromatography analysis of tear protein patterns in diabetic and non-diabetic dry-eye patients.

PURPOSE: To analyze and compare the high performance liquid chromatography (HPLC) runs of tear proteins from diabetic (DIDRY) and non-diabetic (DRY) dry-eye patients, and healthy subjects (CTRL). The patterns were analyzed using multivariate statistical methods. METHODS: Tears (total 56 eyes: CTRL: n=16, DIDRY: n=21, and DRY: n=19) were analyzed by HPLC, using a size-exclusion column with an eluent of 0.5 M sodium phosphate buffer. The patients were primarily grouped according to the results of the basic secretory test (BST) in combination with subjective symptoms such as burning, foreign body sensations, tearing, and "dryness" of the eyes. Patients with BST values < or = 10 mm/5 min plus two subjective symptoms were grouped as dry-eye patients. Before statistical analysis, each HPLC run was quantitatively analyzed using ScanPacK software (ScanPacK, Göttingen, Germany), and a data set was created from each HPLC run. The data were then analyzed by multivariate analysis of discriminance. RESULTS: The HPLC patterns of CTRL, DIDRY and DRY were significantly different (Wilks' lambda: 0.0209; p<0.01). The area of the sigA peak was significantly smaller (p<0.05) in dry-eye tears than controls. There was a good correlation between the extent of separation in the multivariate analysis and the BST value (r = -0.71). Classification of all samples resulted in 98% correct assignments. CONCLUSIONS: The analysis of HPLC patterns and subsequent statistical evaluation are useful for the detection of dry eyes. The HPLC method and the statistical routines described allow a shorter analysis time than electrophoresis. HPLC analysis in combination with statistical analysis can be used as a diagnostic tool for the detection of dry eyes, and also improves the quality of analysis of disease-associated tear proteins in clinical research.

Chromatography, Gel↗
PubMed1990

Structure-activity relationships of antifilarial antimycin analogues: a multivariate pattern recognition study.

The structure-activity relationships of a series of novel antifilarial antimycin A1 analogues have been investigated by using computational chemistry and multivariate statistical techniques. The physiochemical descriptors calculated in this way contained information which was useful in the classification of compounds according to their in vitro antifilarial activity. This approach generated a 53 parameter descriptor set, which was reduced with a multivariate pattern recognition package, ARTHUR. Regression analysis of the reduced set yielded several statistically significant regression equations; e.g.-log in vitro activity = 0.017 mp + 0.65 log P - 0.81ESDL10-7.33 (R = 0.9). With use of this equation, it was possible to make predictions for further untested analogues. The analysis indicated that membrane or lipid solubility is an important determinant in biological activity agreeing with the proposed primary mode of action of the compounds as disrupters of cuticular glucose uptake.

Animals↗
PubMed1986

Predicting outcome in coronary disease. Statistical models versus expert clinicians.

To study the accuracy with which long-term prognosis can be predicted in patients with coronary artery disease, prognostic predictions from a data-based multivariable statistical model were compared with predictions from senior clinical cardiologists. Test samples of 100 patients each were selected from a large series of medically treated patients with significant coronary disease. Using detailed case summaries, five senior cardiologists each predicted one- and three-year survival and infarct-free survival probabilities for 100 patients. Fifty patients appeared in multiple samples for assessing interphysician variability. Cox regression models, developed using patients not in the test samples, predicted corresponding outcome probabilities for each test patient. Overall, model predictions correlated better with actual patient outcomes than did the doctors' predictions. For three-year survival, rank correlations were 0.61 (model) and 0.49 (doctors). For three-year infarct-free survival predictions, correlations with outcome were 0.48 (model) and 0.29 (doctors). Comparisons by individual doctor revealed Cox model three-year survival predictions were better than those of four of five doctors (model predictions added significant [p less than 0.05] prognostic information to the doctor's predictions, whereas the converse was not true). For infarct-free survival, the Cox model was superior to all five doctors. Where predictions were made by multiple doctors, the interphysician variability was substantial. In coronary artery disease, statistical models developed from carefully collected data can provide prognostic predictions that are more accurate than predictions of experienced clinicians made from detailed case summaries.

Coronary Disease↗
PubMed2004

Fourier transform infrared (FTIR) spectral mapping of the cervical transformation zone, and dysplastic squamous epithelium.

OBJECTIVES: This paper is aimed at establishing infrared spectral patterns for the different tissue types found in, and for different stages of disease of squamous cervical epithelium. Methods for the unsupervised distinction of these tissue types are discussed. METHODS: Fourier transform infrared (FTIR) maps of the squamous and glandular cervical epithelium, and of the cervical transformation zone, were obtained and analyzed by multivariate unsupervised hierarchical cluster methods. The resulting clusters are correlated to the corresponding stained histopathological features in the tissue sections. RESULTS: Multivariate statistical analysis of FTIR spectra collected for tissue sections permit an unsupervised method of distinguishing tissue types, and of differentiating between normal and diseased tissue. By analyzing different spectral windows and comparing the results with histology, we found the amide I and II region (1740-1470 cm(-1)) to be very important in correlating anatomical and histopathological features in tissue to spectral clusters. Since an unsupervised, rather than a diagnostic, algorithm was used in these efforts, no statistical analysis of false-positive/false-negative results is reported at this time. CONCLUSIONS: The combination of FTIR micro-spectroscopy and multivariate spectral processing provides important insights into the fundamental spectral signatures of individual cells and consequently shows potential as a diagnostic tool for cervical cancer.

Cell Transformation, Neoplastic↗
PubMed2001

Prognostic factors for response to sildenafil in patients with erectile dysfunction.

OBJECTIVE: To assess the clinical efficacy of sildenafil as treatment for erectile dysfunction (ED) the factors associated with treatment failure were investigated. METHODS: Open, prospective study including 244 patients suffering from ED who were evaluated by anamnesis, physical exploration, blood test, dynamic penile color duplex ultrasonography and Sexual Health Inventory for Male (SHIM). The efficacy of sildenafil was assessed by repeating the SHIM 2 months after therapy, independent of the final dose used. Side effects were also recorded. Factors influencing treatment outcome were evaluated by univariate and multivariate statistical analysis. RESULTS: Overall, sildenafil was effective in 56.8% of 213 eligible patients. When the etiologic diagnosis was not included in the multivariate analysis, antecedents of diabetes mellitus, non-nerve-sparing radical prostatectomy and SHIM basal score were selected as predictors of a poor response. In a second analysis including etiologic diagnosis, only SHIM basal score and etiological diagnosis proved to be of prognostic value. Side effects were noticed by 24.4% of patients, none of them being severe. CONCLUSIONS: Sildenafil is a rather effective and well-tolerated treatment for ED. The basal severity of ED and etiological diagnosis are the prognostic factors most significantly associated with treatment outcome.

3',5'-Cyclic-GMP Phosphodiesterases↗
PubMed2005

Des-gamma-carboxyprothrombin expression in cancer and/or non-cancer liver tissues: association with survival of patients with resectable hepatocellular carcinoma.

Des-gamma-carboxyprothrombin (DCP), also known as a protein induced by vitamin K absence or antagonist II, is an abnormal prothrombin. The level of serum DCP is used in clinical diagnosis and prognosis of patients with hepatocellular carcinoma (HCC). Our previous immunohistochemical study showed that DCP is expressed not only in cancer tissues but also in the surrounding non-cancer tissues of HCC patients. However, clinical significance of DCP expression in non-cancer tissues of HCC patients remains unclear. In this study, we examined the relationship between histochemical expression of DCP in cancer and/or non-cancer tissues and the clinical outcome of the HCC patients. A retrospective study was performed of 132 patients each with a single primary HCC nodule. Expression of DCP in tissues was evaluated with immunohistochemical staining using anti-DCP antibody (MU-3). Serum DCP levels were determined using enzyme immunoassay with a double antibody sandwich system. Experimental and clinical data were processed by univariate and multivariate statistical analyses. DCP expression, even if it was observed only in non-cancer tissues, was related to poorer prognosis. Univariate and multivariate analyses showed that DCP expression in cancer and/or non-cancer tissues was a significant prognostic factor. Furthermore, the combined evaluation of tissue DCP expression and serum DCP levels showed that prognosis was poorest for patients showing positive tissue DCP expression and high levels of serum DCP. These results suggest that immunohistochemical evaluation of DCP expression not only in cancer but also in non-cancer tissues serves as a valuable factor in the prognosis of HCC patients and that the combined evaluation of tissue DCP expression and serum DCP level will be more valuable than either factor alone in the effective treatment and prognosis.

Adolescent↗
PubMed2005

Pharmacokinetic comparison of two recombinant human granulocyte colony-stimulating factor after subcutaneous administration in rabbits.

A pharmacokinetic comparison between two formulations (LeukoCIM, CIMAB SA versus Neupogen, Hoffman-La Roche, licensed by Amgen) of recombinant human granulocyte colony-stimulating factor (rhG-CSF) using non-compartmental analysis was performed in male F1 rabbits after a single subcutaneous 11.5 microg/kg dose to help decide whether to conduct further comparability tests. Unlike the absorption phase, a statistical difference was not detected between Neupogen and LeukoCIM for clearance (18.69+/-11.83 versus 28.42+/-12.11 mL/h/kg, P=0.22). In addition, using a multivariate statistical analysis by independent samples test, a significant difference was not found between the two formulations (P=0.88). Finally, the results obtained in this study confirmed the pharmacokinetic comparability between both formulations, supporting the claim for further assessments following the current protocol on biogeneric equivalence.

Animals↗
PubMed1976

Implications of mass automated instruments on medical practice.

The clinical laboratory is in a stage of change which has been the direct result of an influx of technology and the current demand for lower costs and increased productivity. There is no significant trend anticipated in the future which would forestall this particular process. The practicing physician can anticipate a fairly marked reduction in the cost of individual tests and, likewise, a marked increase in the amount of available information. There will be a larger clerical function involved in the review of such information along with a series of relevant problems related to the actual size of the new data base. One of the most significant problems related to this increase in numerical reporting will be the statistical problem of the "abnormal-normal." Statistical misclassification must be considered with each abnormal test and a specific approach must be taken to these tests to minimize the inconvenience to the patient. It is recommended that any unsuspected abnormal test(s) that is discovered be repeated as a single test(s) through a reference method. Reordering of the test battery or screen will only offer another possibility for statistical misinterpretation. If the test on isolated evaluation does appear abnormal, then this should call for the routine evaluation of this particular condition. Multivariate statistical analysis could eliminate this probability dilemma, yet is presently experimental and is not in routine use within any clinical setting. We can look forward to the evolution of such a process through the mechanism of a hospital and laboratory communication system which will allow the physician to work with an evaluation process capable of handling these large volumes of numbers (22). Such programs are on the horizon and should be anticipated within the next five years in some of the larger centers that have this new instrumentation.

Analysis of Variance↗
PubMed2002

The influence of dendritic cell infiltration and vascular endothelial growth factor expression on the prognosis of non-small cell lung cancer.

Vascular endothelial growth factor (VEGF) is well known to be produced by many human tumors, and it also plays an important role in tumor neovasculature formation. In addition to angiogenesis promotion, recent basic research has shown that VEGF has another function that allows it to inhibit dendritic cell (DC) maturation. However, very little is known about VEGF-dependent DC inhibition in a clinical setting. In this study, we analyzed the immunohistochemical expression of VEGF, microvessel density (MVD), and intratumoral DC infiltration in 132 surgically resected lung cancer specimens. We also evaluated the influence of these factors on their survival by a multivariate statistical analysis. VEGF expression was positively related to MVD (P = 0.0003) and negatively related to the degree of DC infiltration (P = 0.0232). A multivariate analysis also showed the VEGF expression, MVD, and DC infiltration to be independent prognostic factors. Moreover, we also accurately analyzed patient prognoses using the double stratification method for determining VEGF expression and DC infiltration. The patient group with a high VEGF expression/low DC infiltration showed a worse prognosis (P < 0.0001), whereas the group with a low VEGF expression/high DC infiltration had a better prognosis (P = 0.0001).

Adult↗