Impact of early angiographic evaluation on the frequency of emergency reoperations after coronary bypass surgery.

BACKGROUND AND AIMS: early graft failure following coronary bypass surgery results in elevated morbidity and mortality. This study focused on the impact of angiographic graft evaluation. MATERIAL AND METHODS: of 5251 coronary artery bypass grafting (CABG) patients, 36 with postoperative persistent ischaemia underwent early angiography (23) or emergency resternotomy (13) 2000-2007 (Angiography era). Of the 23 patients, who underwent angiography, five were subsequently reoperated. Of 8807 CABG patients, 76 underwent postoperative emergency resternotomy 1988-1999 (Pre-angiography era) and served as controls. RESULTS: the angiography era patients were older (64.0 years vs. 58.2 years, P = 0.002) and the proportion of female patients (22% vs. 43%, P = 0.029) was smaller. The rate of emergency reoperations decreased (0.86% vs 0.34%, P < 0.001) during the Angiography era and graft repairs (P = 0.013) or additional grafts (P = 0.006) were less frequent, although occluded anastomoses were observed more often (P = 0.043). In 5 Angiography era patients graft complications were corrected with percutaneous coronary intervention. ICU stay (5.72 + 0.98 days vs. 5.53 + 0.68 days) and hospital stay (12.2 + 1.54 days vs. 13.1 + 1.63 days) did not differ between the groups, but the rate of myocardial infarction (63.8% vs. 92.1%, P < 0.001) and in-hospital death (22.2% vs. 46.1%, P = 0.015) decreased. CONCLUSION: after the introduction of early postoperative angiographic evaluation of CABG patients the rate of emergency reoperations and related morbidity and mortality decreased.

Physically principled reflection models applied to filtered camera imaging inversions in metal walled fusion machines.

Ray-tracing techniques are applied to filtered divertor imaging, a diagnostic that has long suffered from artifacts due to the polluting effect of reflected light in metal walled fusion machines. Physically realistic surface reflections were modeled using a Cook-Torrance micro-facet bi-directional reflection distribution function applied to a high resolution mesh of the vessel geometry. In the absence of gonioreflectometer measurements, a technique was developed to fit the free parameters of the Cook-Torrance model against images of the JET in-vessel light sources. By coupling this model with high fidelity plasma fluid simulations, photo-realistic renderings of a number of tokamak plasma emission scenarios were generated. Finally, a sensitivity matrix describing the optical coupling of a JET divertor camera and the emission profile of the plasma was obtained, including full reflection effects. These matrices are used to perform inversions on measured data and shown to reduce the level of artifacts in inverted emission profiles.

Irregular expression of hyaluronan and its CD44 receptor is associated with metastatic phenotype in laryngeal squamous cell carcinoma.

The distributions of hyaluronan (HA) and its CD44 receptor were studied in 24 normal, 27 dysplastic samples of laryngeal epithelium and in 172 squamous cell carcinomas (LSCC), using a specific probe prepared from cartilage proteoglycan (bHABC, biotinylated hyaluronan binding complex) and a monoclonal antibody (Hermes 3). HA and CD44 were expressed similarly in all normal and about 90% of dysplastic and neoplastic laryngeal epithelia. In the normal epithelium HA and CD44 were homogeneously distributed throughout the epithelium, whereas the most superficial layers were negative. This was in contrast to the picture in dysplastic epithelium and well-differentiated invasive carcinomas, which were entirely HA and CD44 positive. Local areas with a low signal for HA and CD44 were present in 11% and 22% of the samples with dysplasia, and in 27% and 28% of those with carcinoma, respectively. The presence of this staining irregularity was associated with poor differentiation of the carcinoma, a significantly elevated mitotic index and a high frequency of nodal spreading and metastases. Furthermore, the irregular staining showed a trend towards poor disease-free survival, suggesting that an altered metabolism of HA is a common feature in LSCC and is associated with an aggressive growth pattern.

A class of neural networks for independent component analysis.

Independent component analysis (ICA) is a recently developed, useful extension of standard principal component analysis (PCA). The ICA model is utilized mainly in blind separation of unknown source signals from their linear mixtures. In this application only the source signals which correspond to the coefficients of the ICA expansion are of interest. In this paper, we propose neural structures related to multilayer feedforward networks for performing complete ICA. The basic ICA network consists of whitening, separation, and basis vector estimation layers. It can be used for both blind source separation and estimation of the basis vectors of ICA. We consider learning algorithms for each layer, and modify our previous nonlinear PCA type algorithms so that their separation capabilities are greatly improved. The proposed class of networks yields good results in test examples with both artificial and real-world data.

A unified neural bigradient algorithm for robust PCA and MCA.

A new instantaneous-gradient search algorithm for computing a principal component or minor component type solution is proposed. The algorithm can use normalized Hebbian or anti-Hebbian learning in a unified formula. Starting from one-unit rule, a multi-unit algorithm is developed which can simultaneously extract several robust counterparts of the principal or minor eigenvectors of the data covariance matrix. Standard principal or minor components emerge as special cases from the general non-quadratic criterion. The learning rule is analyzed mathematically, and the theoretical results are verified by simulations. The proposed bigradient approach can be applied to blind separation of independent source signals from their linear mixtures.

Least-squares methods for blind source separation based on nonlinear PCA.

In standard blind source separation, one tries to extract unknown source signals from their instantaneous linear mixtures by using a minimum of a priori information. We have recently shown that certain nonlinear extensions of principal component type neural algorithms can be successfully applied to this problem. In this paper, we show that a nonlinear PCA criterion can be minimized using least-squares approaches, leading to computationally efficient and fast converging algorithms. Several versions of this approach are developed and studied, some of which can be regarded as neural learning algorithms. A connection to the nonlinear PCA subspace rule is also shown. Experimental results are given, showing that the least-squares methods usually converge clearly faster than stochastic gradient algorithms in blind separation problems.

Local linear independent component analysis based on clustering.

In standard Independent Component Analysis (ICA), a linear data model is used for a global description of the data. Even though linear ICA yields meaningful results in many cases, it can provide a crude approximation only for general nonlinear data distributions. In this paper a new structure is proposed, where local ICA models are used in connection with a suitable grouping algorithm clustering the data. The clustering part is responsible for an overall coarse nonlinear representation of the data, while linear ICA models of each cluster are used for describing local features of the data. The goal is to represent the data better than in linear ICA while avoiding computational difficulties related with nonlinear ICA. Several data grouping methods are considered, including standard K-means clustering, self-organizing maps, and neural gas. Connections to existing methods are discussed, and experimental results are given for artificial data and natural images. Furthermore, a general theoretical framework encompassing a large number of methods for representing data is introduced. These range from global, dense representation methods to local, very sparse coding methods. The proposed local ICA methods lie between these two extremes.

An experimental comparison of neural algorithms for independent component analysis and blind separation.

In this paper, we compare the performance of five prominent neural or adaptive algorithms designed for Independent Component Analysis (ICA) and blind source separation (BSS). In the first part of the study, we use artificial data for comparing the accuracy, convergence speed, computational load, and other relevant properties of the algorithms. In the second part, the algorithms are applied to three different real-world data sets. The task is either blind source separation or finding interesting directions in the data for visualisation purposes. We develop criteria for selecting the most meaningful basis vectors of ICA and measuring the quality of the results. The comparison reveals characteristic differences between the studied ICA algorithms. The most important conclusions of our comparison are robustness of the ICA algorithms with respect to modest modeling imperfections, and the superiority of fixed-point algorithms with respect to the computational load.

On neural blind separation with noise suppression and redundancy reduction.

Noise is an unavoidable factor in real sensor signals. We study how additive and convolutive noise can be reduced or even eliminated in the blind source separation (BSS) problem. Particular attention is paid to cases in which the number of sensors is larger than the number of sources. We propose various methods and associated adaptive learning algorithms for such an extended BSS problem. Performance and validity of the proposed approaches are demonstrated by extensive computer simulations.

Cul-de-sac hypernasality test with pattern recognition of LPC indices.

Acoustic differences between samples of [i], [u], and [a] uttered in nose-open and nose-obstructed condition were studied in 6 women with isolated cleft palate and pathological nasalance scores and 9 healthy women with normal nasalance scores. The speech samples were depicted by 14-component vocal tract area feature vectors obtained by linear prediction and the differences between the samples were studied with a self-organized feature map. Each location on the map corresponds to a certain signal pattern, neighboring locations to similar patterns. The group of healthy subjects differed from the patients for vowels [i] and [u] but not for [a]. In the patients the nose obstruction induced a significant change in the location of these vowel samples on the map. In healthy subjects no such changes were detected. The results agreed with perceived differences between the subjects.

Quantitation of staining for amyloid in histological sections: sources of variation, correlation analysis, and interstain reproducibility.

Five pathologists estimated with point counting the amount of amyloid in heart muscle autopsy samples from 19 patients who suffered from amyloidosis. The serial sections had been stained with 5 different methods. The different variation sources involved in a quantitative study on amyloid and the interstain reproducibility between the results given by these 5 stains were analyzed. The volume fractions of amyloid after different stains were subjected to correlation analysis. In our test system about a third (34%) of the variation was due to the point counting method. Sirius red had the lowest and Congo red the highest variation due to point counting, since Sirius red gave the most intense staining and highest volume fractions for amyloid. The volume fractions were lowest in Congo red. About 40% of the total variation was composed of the field variation. This variation includes the variation due to variation between fields, field selection and subjective intraobserver variation. Sirius red showed the smallest field variation in respect to the total variation. Congo red, on the other hand, had the highest variation. The variation caused by pure interobserver variation is presented by the difference between total variation and the calculated field variation. This was largest with Sirius red and smallest with Congo red. The pure interobserver variation accounted for about 26% of the total variation. The high reproducibility with Sirius stain is most probably due to the high volume fractions of positively stained material after this stain, not to low interobserver variation. Congo red seems to have low interobserver variation. The senior pathologists were more reliable in measuring the amyloid deposits than the junior pathologists as suggested by interstain reproducibility and variation.(ABSTRACT TRUNCATED AT 250 WORDS)

Reproducibility of morphometric measurements of amyloid after various staining methods.

Four histologic staining methods used for detecting amyloid (Congo red, viewed in both normal and polarized light, Sirius red, Crystal violet and Thioflavine T) were applied to heart muscle autopsy samples from 19 patients who suffered from amyloidosis. The amount of amyloid present was evaluated with morphometry (point counting) by five pathologists, and the interobserver reproducibility and variation of point counting in these staining methods were analyzed. The Sirius red method showed the least variation and was the most suitable stain for demonstrating amyloid with respect to reproducibility. Thioflavine T showed the greatest variation and was the least suitable stain with respect to reproducibility. The range of variation was considerable in all staining methods. The results show that stains differ in their specificity and sensitivity in staining amyloid, observers differ in their interpretation of staining results and certain stains result in more uniform interpretations than do others.

Reproducibility and variation of morphometric analysis of carcinoembryonic antigen staining in histopathology. The influence of standardized microscopic fields.

Using morphometric methods, five pathologists analyzed the positive staining for carcinoembryonic antigen (CEA) in sections from 17 ovarian tumors, with the intraclass correlation coefficient (ICC) and the mean values of the coefficients of variation (CV) used to assess reproducibility and variation. First, field and point scores for epithelium and mucin were estimated using 25 randomly selected square fields in sections from each of the tumors. The ICC range in the whole sample field was 0.53 to 0.81 (slight to substantial reproducibility) while the mean values of CV were 0.50 to 0.75. Second, the results of using random and standardized individual fields for the measurements were studied in three tumors. In random fields, the ICC was 0.57 to 0.71 (slight to moderate reproducibility) and the mean values of CV were 0.53 to 0.65. The corresponding values in standardized fields were 0.71 to 0.73 (moderate reproducibility) and 0.41 to 0.57, respectively. The results show that the variation is smaller and the degree of reproducibility higher in standardized fields. Considerable variation remains, however, revealing human factors as an important source of variation in practical morphometry.