Deep Learning With Radiogenomics Towards Personalized Management of Gliomas.

A state-of-the-art interdisciplinary survey on multi-modal radiogenomic approaches is presented involving applications to the diagnosis and personalized management of gliomas a common kind of brain tumors through noninvasive imaging integrated with genomic information. It encompasses mining tumor radioimages employing deep learning for the automated extraction of relevant features from the segmented volume of interest (VOI). Gene expression values from surgically extracted tumor tissues are often simultaneously analyzed to determine patient specific features. Association between genomic and radiomic features are also explored in some cases to determine the imaging surrogates. Deep learning and transfer learning are typically exploited for efficient knowledge discovery and decision-making. Some studies on survival prediction ensemble learning and interactive learning are also included. The literature mainly focuses on magnetic resonance imaging (MRI) data of the brain for learning and validation and generally involves the NIH TCIA and TCGA repositories as well as the BraTS Challenge databases.

Segmentation in Diabetic Retinopathy using Deeply-Supervised Multiscalar Attention.

Diabetic Retinopathy (DR) is a progressive chronic eye disease that leads to irreversible blindness. Detection of DR at an early stage of the disease is crucial and requires proper detection of minute DR pathologies. A novel Deeply-Supervised Multiscale Attention U-Net (Mult-Attn-U-Net) is proposed for segmentation of different DR pathologies viz. Microaneurysms (MA), Hemorrhages (HE), Soft and Hard Exudates (SE and EX). A publicly available dataset (IDRiD) is used to evaluate the performance. Comparative study with four state-of-the-art models establishes its superiority. The best segmentation accuracy obtained by the model for MA, HE, SE are 0.65, 0.70, 0.72, respectively.

Novel Volumetric Sub-region Segmentation in Brain Tumors.

A novel deep learning based model called Multi-Planar Spatial Convolutional Neural Network (MPS-CNN) is proposed for effective, automated segmentation of different sub-regions viz. peritumoral edema (ED), necrotic core (NCR), enhancing and non-enhancing tumor core (ET/NET), from multi-modal MR images of the brain. An encoder-decoder type CNN model is designed for pixel-wise segmentation of the tumor along three anatomical planes (axial, sagittal, and coronal) at the slice level. These are then combined, by incorporating a consensus fusion strategy with a fully connected Conditional Random Field (CRF) based post-refinement, to produce the final volumetric segmentation of the tumor and its constituent sub-regions. Concepts, such as spatial-pooling and unpooling are used to preserve the spatial locations of the edge pixels, for reducing segmentation error around the boundaries. A new aggregated loss function is also developed for effectively handling data imbalance. The MPS-CNN is trained and validated on the recent Multimodal Brain Tumor Segmentation Challenge (BraTS) 2018 dataset. The Dice scores obtained for the validation set for whole tumor (WT :NCR/NE +ET +ED), tumor core (TC:NCR/NET +ET), and enhancing tumor (ET) are 0.90216, 0.87247, and 0.82445. The proposed MPS-CNN is found to perform the best (based on leaderboard scores) for ET and TC segmentation tasks, in terms of both the quantitative measures (viz. Dice and Hausdorff). In case of the WT segmentation it also achieved the second highest accuracy, with a score which was only 1% less than that of the best performing method.

Feature Selection Through Message Passing.

A novel similarity-based feature selection algorithm is developed, using the concept of distance correlation. A feature subset is selected in terms of this similarity measure between pairs of features, without assuming any underlying distribution of the data. The pair-wise similarity is then employed, in a message passing framework, to select a set of exemplars features involving minimum redundancy and reduced parameter tuning. The algorithm does not need an exhaustive traversal of the search space. The methodology is next extended to handle large data, using an inherent property of distance correlation. The effectiveness of the algorithm is demonstrated on nine sets of publicly-available data.

A Novel GBM Saliency Detection Model Using Multi-Channel MRI.

The automatic computerized detection of regions of interest (ROI) is an important step in the process of medical image processing and analysis. The reasons are many, and include an increasing amount of available medical imaging data, existence of inter-observer and inter-scanner variability, and to improve the accuracy in automatic detection in order to assist doctors in diagnosing faster and on time. A novel algorithm, based on visual saliency, is developed here for the identification of tumor regions from MR images of the brain. The GBM saliency detection model is designed by taking cue from the concept of visual saliency in natural scenes. A visually salient region is typically rare in an image, and contains highly discriminating information, with attention getting immediately focused upon it. Although color is typically considered as the most important feature in a bottom-up saliency detection model, we circumvent this issue in the inherently gray scale MR framework. We develop a novel pseudo-coloring scheme, based on the three MRI sequences, viz. FLAIR, T2 and T1C (contrast enhanced with Gadolinium). A bottom-up strategy, based on a new pseudo-color distance and spatial distance between image patches, is defined for highlighting the salient regions in the image. This multi-channel representation of the image and saliency detection model help in automatically and quickly isolating the tumor region, for subsequent delineation, as is necessary in medical diagnosis. The effectiveness of the proposed model is evaluated on MRI of 80 subjects from the BRATS database in terms of the saliency map values. Using ground truth of the tumor regions for both high- and low- grade gliomas, the results are compared with four highly referred saliency detection models from literature. In all cases the AUC scores from the ROC analysis are found to be more than 0.999 ± 0.001 over different tumor grades, sizes and positions.

Robust Radiomics feature quantification using semiautomatic volumetric segmentation.

Due to advances in the acquisition and analysis of medical imaging, it is currently possible to quantify the tumor phenotype. The emerging field of Radiomics addresses this issue by converting medical images into minable data by extracting a large number of quantitative imaging features. One of the main challenges of Radiomics is tumor segmentation. Where manual delineation is time consuming and prone to inter-observer variability, it has been shown that semi-automated approaches are fast and reduce inter-observer variability. In this study, a semiautomatic region growing volumetric segmentation algorithm, implemented in the free and publicly available 3D-Slicer platform, was investigated in terms of its robustness for quantitative imaging feature extraction. Fifty-six 3D-radiomic features, quantifying phenotypic differences based on tumor intensity, shape and texture, were extracted from the computed tomography images of twenty lung cancer patients. These radiomic features were derived from the 3D-tumor volumes defined by three independent observers twice using 3D-Slicer, and compared to manual slice-by-slice delineations of five independent physicians in terms of intra-class correlation coefficient (ICC) and feature range. Radiomic features extracted from 3D-Slicer segmentations had significantly higher reproducibility (ICC = 0.85±0.15, p = 0.0009) compared to the features extracted from the manual segmentations (ICC = 0.77±0.17). Furthermore, we found that features extracted from 3D-Slicer segmentations were more robust, as the range was significantly smaller across observers (p = 3.819e-07), and overlapping with the feature ranges extracted from manual contouring (boundary lower: p = 0.007, higher: p = 5.863e-06). Our results show that 3D-Slicer segmented tumor volumes provide a better alternative to the manual delineation for feature quantification, as they yield more reproducible imaging descriptors. Therefore, 3D-Slicer can be employed for quantitative image feature extraction and image data mining research in large patient cohorts.

Extracting gene-gene interactions through curve fitting.

This paper presents a simple and novel curve fitting approach for generating simple gene regulatory subnetworks from time series gene expression data. Microarray experiments simultaneously generate massive data sets and help immensely in the large-scale study of gene expression patterns. Initial biclustering reduces the search space in the high-dimensional microarray data. The least-squares error between fitting of gene pairs is minimized to extract a set of gene-gene interactions, involving transcriptional regulation of genes. The higher error values are eliminated to retain only the strong interacting gene pairs in the resultant gene regulatory subnetwork. Next the algorithm is extended to a generalized framework to enhance its capability. The methodology takes care of the higher-order dependencies involving multiple genes co-regulating a single gene, while eliminating the need for user-defined parameters. It has been applied to the time-series Yeast data, and the experimental results biologically validated using standard databases and literature.

Genetic networks and soft computing.

The analysis of gene regulatory networks provides enormous information on various fundamental cellular processes involving growth, development, hormone secretion, and cellular communication. Their extraction from available gene expression profiles is a challenging problem. Such reverse engineering of genetic networks offers insight into cellular activity toward prediction of adverse effects of new drugs or possible identification of new drug targets. Tasks such as classification, clustering, and feature selection enable efficient mining of knowledge about gene interactions in the form of networks. It is known that biological data is prone to different kinds of noise and ambiguity. Soft computing tools, such as fuzzy sets, evolutionary strategies, and neurocomputing, have been found to be helpful in providing low-cost, acceptable solutions in the presence of various types of uncertainties. In this paper, we survey the role of these soft methodologies and their hybridizations, for the purpose of generating genetic networks.

Rough-fuzzy collaborative clustering.

In this study, we introduce a novel clustering architecture, in which several subsets of patterns can be processed together with an objective of finding a common structure. The structure revealed at the global level is determined by exchanging prototypes of the subsets of data and by moving prototypes of the corresponding clusters toward each other. Thereby, the required communication links are established at the level of cluster prototypes and partition matrices, without hampering the security concerns. A detailed clustering algorithm is developed by integrating the advantages of both fuzzy sets and rough sets, and a measure of quantitative analysis of the experimental results is provided for synthetic and real-world data.

Hidden Markov Models, grammars, and biology: a tutorial.

Biological sequences and structures have been modelled using various machine learning techniques and abstract mathematical concepts. This article surveys methods using Hidden Markov Model and functional grammars for this purpose. We provide a formal introduction to Hidden Markov Model and grammars, stressing on a comprehensive mathematical description of the methods and their natural continuity. The basic algorithms and their application to analyzing biological sequences and modelling structures of bio-molecules like proteins and nucleic acids are discussed. A comparison of the different approaches is discussed, and possible areas of work and problems are highlighted. Related databases and softwares, available on the internet, are also mentioned.