Bagging linear sparse Bayesian learning models for variable selection in cancer diagnosis

The Fluctuations of Blood Oxygen Level-Dependent Signals as a Method of Brain Tumor Characterization: A Preliminary Report

OBJECTIVE

In this study we present the nature and characteristic of the fluctuation of blood oxygen level-dependent (BOLD) signals measured from brain tumors.

METHODS

Supratentorial astrocytomas, which were neither operated nor previously managed with chemotherapy or radiotherapy, were segmented, and the time series of the BOLD signal fluctuations were extracted. The mean (across patients) power spectra were plotted for the different World Health Organization tumor grades. One-way analysis of variance (ANOVA) was performed to identify significant differences between the power spectra of different tumor grades. Results were considered significant at P < 0.05.

RESULTS

A total of 58 patients were included in the study. This group of patients included 1 patient with grade I glioma; 15 with grade II; 12 with grade III; and 30 with grade IV. The power spectra of the tumor time series were individually inspected, and all tumors exhibited high peaks at the lower frequency signals, but these were more pronounced in high-grade tumors. ANOVA showed a significant difference in power spectra between groups (P = 0.000). Post hoc analysis with Bonferroni correction showed a significant difference between grade II and grade III (P = 0.012) and grade IV (P = 0.000). There was no significant power spectra difference between grade III and IV tumors (P = 1).

CONCLUSIONS

The power spectra of BOLD signals from tumor tissue showed fluctuations in the low-frequency signals and were significantly correlated with tumor grade. These signals could have a misleading effect when analyzing resting state functional magnetic resonance imaging and could be also viewed as a potential method of tumor characterization.

Sparse Bayesian classification and feature selection for biological expression data with high correlations

Classification models built on biological expression data are increasingly used to predict distinct disease subtypes. Selected features that separate sample groups can be the candidates of biomarkers, helping us to discover biological functions/pathways. However, three challenges are associated with building a robust classification and feature selection model: 1) the number of significant biomarkers is much smaller than that of measured features for which the search will be exhaustive; 2) current biological expression data are big in both sample size and feature size which will worsen the scalability of any search algorithms; and 3) expression profiles of certain features are typically highly correlated which may prevent to distinguish the predominant features. Unfortunately, most of the existing algorithms are partially addressing part of these challenges but not as a whole. In this paper, we propose a unified framework to address the above challenges. The classification and feature selection problem is first formulated as a nonconvex optimisation problem. Then the problem is relaxed and solved iteratively by a sequence of convex optimisation procedures which can be distributed computed and therefore allows the efficient implementation on advanced infrastructures. To illustrate the competence of our method over others, we first analyse a randomly generated simulation dataset under various conditions. We then analyse a real gene expression dataset on embryonal tumour. Further downstream analysis, such as functional annotation and pathway analysis, are performed on the selected features which elucidate several biological findings.

Microarray gene expression: a study of between-platform association of Affymetrix and cDNA arrays

Microarrays technology has been expanding remarkably since its launch about 15 years ago. With its advancement along with the increase of popularity, the technology affords the luxury that gene expressions can be measured in any of its multiple platforms. However, the generated results from the microarray platforms remain incomparable. In this direction, we earlier developed and tested an approach to address the incomparability of the expression measures of Affymetrix®- and cDNA-platforms. The method was an exploit involving transformation of Affymetrix data, which brought the gene expressions of both cDNA and Affymetrix platforms to a common and comparable level. The encouraging outcome of that investigation has subsequently acted as a motivator to focus attention on examining further in the direction of defining the association between the two platforms. Accordingly, this paper takes on a novel exploration towards determining a precise association using a wide range of statistical and machine learning approaches, specifically the various models are elaborately trailed using-regression (linear, cubic-polynomial, LOESS, bootstrap aggregating) and artificial neural networks (self-organizing maps and feedforward networks). After careful comparison, the existing relationship between the data from the two platforms is found to be non-linear where feedforward neural network captures the best delineation of the association.

Investigating machine learning techniques for MRI-based classification of brain neoplasms

PURPOSE

Diagnosis and characterization of brain neoplasms appears of utmost importance for therapeutic management. The emerging of imaging techniques, such as Magnetic Resonance (MR) imaging, gives insight into pathology, while the combination of several sequences from conventional and advanced protocols (such as perfusion imaging) increases the diagnostic information. To optimally combine the multiple sources and summarize the information into a distinctive set of variables however remains difficult. The purpose of this study is to investigate machine learning algorithms that automatically identify the relevant attributes and are optimal for brain tumor differentiation.

METHODS

Different machine learning techniques are studied for brain tumor classification based on attributes extracted from conventional and perfusion MRI. The attributes, calculated from neoplastic, necrotic, and edematous regions of interest, include shape and intensity characteristics. Attributes subset selection is performed aiming to remove redundant attributes using two filtering methods and a wrapper approach, in combination with three different search algorithms (Best First, Greedy Stepwise and Scatter). The classification frameworks are implemented using the WEKA software.

RESULTS

The highest average classification accuracy assessed by leave-one-out (LOO) cross-validation on 101 brain neoplasms was achieved using the wrapper evaluator in combination with the Best First search algorithm and the KNN classifier and reached 96.9% when discriminating metastases from gliomas and 94.5% when discriminating high-grade from low-grade neoplasms.

CONCLUSIONS

A computer-assisted classification framework is developed and used for differential diagnosis of brain neoplasms based on MRI. The framework can achieve higher accuracy than most reported studies using MRI.

Automatic classification of lymphoma images with transform-based global features

We propose a report on automatic classification of three common types of malignant lymphoma: chronic lymphocytic leukemia, follicular lymphoma, and mantle cell lymphoma. The goal was to find patterns indicative of lymphoma malignancies and allowing classifying these malignancies by type. We used a computer vision approach for quantitative characterization of image content. A unique two-stage approach was employed in this study. At the outer level, raw pixels were transformed with a set of transforms into spectral planes. Simple (Fourier, Chebyshev, and wavelets) and compound transforms (Chebyshev of Fourier and wavelets of Fourier) were computed. Raw pixels and spectral planes were then routed to the second stage (the inner level). At the inner level, the set of multipurpose global features was computed on each spectral plane by the same feature bank. All computed features were fused into a single feature vector. The specimens were stained with hematoxylin (H) and eosin (E) stains. Several color spaces were used: RGB, gray, CIE-L*a*b*, and also the specific stain-attributed H&E space, and experiments on image classification were carried out for these sets. The best signal (98%-99% on earlier unseen images) was found for the HE, H, and E channels of the H&E data set.

Classification of brain tumor type and grade using MRI texture and shape in a machine learning scheme

The objective of this study is to investigate the use of pattern classification methods for distinguishing different types of brain tumors, such as primary gliomas from metastases, and also for grading of gliomas. The availability of an automated computer analysis tool that is more objective than human readers can potentially lead to more reliable and reproducible brain tumor diagnostic procedures. A computer-assisted classification method combining conventional MRI and perfusion MRI is developed and used for differential diagnosis. The proposed scheme consists of several steps including region-of-interest definition, feature extraction, feature selection, and classification. The extracted features include tumor shape and intensity characteristics, as well as rotation invariant texture features. Feature subset selection is performed using support vector machines with recursive feature elimination. The method was applied on a population of 102 brain tumors histologically diagnosed as metastasis (24), meningiomas (4), gliomas World Health Organization grade II (22), gliomas World Health Organization grade III (18), and glioblastomas (34). The binary support vector machine classification accuracy, sensitivity, and specificity, assessed by leave-one-out cross-validation, were, respectively, 85%, 87%, and 79% for discrimination of metastases from gliomas and 88%, 85%, and 96% for discrimination of high-grade (grades III and IV) from low-grade (grade II) neoplasms. Multiclass classification was also performed via a one-vs-all voting scheme.

Protein networks markedly improve prediction of subcellular localization in multiple eukaryotic species

The function of a protein is intimately tied to its subcellular localization. Although localizations have been measured for many yeast proteins through systematic GFP fusions, similar studies in other branches of life are still forthcoming. In the interim, various machine-learning methods have been proposed to predict localization using physical characteristics of a protein, such as amino acid content, hydrophobicity, side-chain mass and domain composition. However, there has been comparatively little work on predicting localization using protein networks. Here, we predict protein localizations by integrating an extensive set of protein physical characteristics over a protein's extended protein–protein interaction neighborhood, using a classification framework called ‘Divide and Conquer k-Nearest Neighbors’ (DC-kNN). These predictions achieve significantly higher accuracy than two well-known methods for predicting protein localization in yeast. Using new GFP imaging experiments, we show that the network-based approach can extend and revise previous annotations made from high-throughput studies. Finally, we show that our approach remains highly predictive in higher eukaryotes such as fly and human, in which most localizations are unknown and the protein network coverage is less substantial.

Using support vector machines to detect therapeutically incorrect measurements by the MiniMed CGMS

BACKGROUND

Current continuous glucose monitors have limited accuracy mainly in the low range of glucose measurements. This lack of accuracy is a limiting factor in their clinical use and in the development of the so-called artificial pancreas. The ability to detect incorrect readings provided by continuous glucose monitors from raw data and other information supplied by the monitor itself is of utmost clinical importance. In this study, support vector machines (SVMs), a powerful statistical learning technique, were used to detect therapeutically incorrect measurements made by the Medtronic MiniMed CGMS.

METHODS

Twenty patients were monitored for three days (first day at the hospital and two days at home) using the MiniMed CGMS. After the third day, the monitor data were downloaded to the physician's computer. During the first 12 hours, the patients stayed in the hospital, and blood samples were taken every 15 minutes for two hours after meals and every 30 minutes otherwise. Plasma glucose measurements were interpolated using a cubic method for time synchronization with simultaneous MiniMed CGMS measurements every five minutes, obtaining a total of 2281 samples. A Gaussian SVM classifier trained on the monitor's electrical signal and glucose estimation was tuned and validated using multiple runs of k-fold cross-validation. The classes considered were Clarke error grid zones A+B and C+D+E.

RESULTS

After ten runs of ten-fold cross-validation, an average specificity and sensitivity of 92.74% and 75.49%, respectively, were obtained (see Figure 4). The average correct rate was 91.67%.

CONCLUSIONS

Overall, the SVM performed well, in spite of the somewhat low sensitivity. The classifier was able to detect the time intervals when the monitor's glucose profile could not be trusted due to incorrect measurements. As a result, hypoglycemic episodes missed by the monitor were detected.