skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Learning-based Segmentation Framework for Tissue Images Containing Gene Expression Data

Abstract

Abstract Associating specific gene activity with functional locations in the brain results in a greater understanding of the role of the gene. To perform such an association for the over 20,000 genes in the mammalian genome, reliable automated methods that characterize the distribution of gene expression in relation to a standard anatomical model are required. In this paper, we propose a new automatic method that results in the segmentation of gene expression images into distinct anatomical regions in which the expression can be quantified and compared with other images. Our contribution is a novel hybrid atlas that utilizes a statistical shape model based on a subdivision mesh, texture differentiation at region boundaries, and features of anatomical landmarks to delineate boundaries of anatomical regions in gene expression images. This atlas, which provides a common coordinate system for internal brain data, was trained on 36 images manually annotated by neuroanatomists and tested on 64 images. Our framework has achieved a mean overlap ratio of up to 91 § 7% in this challenging dataset. This tool for large-scale annotation will help scientists interpret gene expression patterns more efficiently.

Authors:
; ; ; ; ;
Publication Date:
Research Org.:
Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
912493
Report Number(s):
PNNL-SA-52381
Journal ID: ISSN 0278-0062; ITMID4; TRN: US200801%%1049
DOE Contract Number:
AC05-76RL01830
Resource Type:
Journal Article
Resource Relation:
Journal Name: IEEE Transactions on Medical Imaging, 26(5):728-744; Journal Volume: 26; Journal Issue: 5
Country of Publication:
United States
Language:
English
Subject:
59 BASIC BIOLOGICAL SCIENCES; BRAIN; DISTRIBUTION; FUNCTIONALS; GENES; SHAPE; TEXTURE

Citation Formats

Bello, Musodiq, Ju, Tao, Carson, James P., Warren, Joe, Chiu, Wah, and Kakadiaris, Ioannis. Learning-based Segmentation Framework for Tissue Images Containing Gene Expression Data. United States: N. p., 2007. Web. doi:10.1109/TMI.2007.895462.
Bello, Musodiq, Ju, Tao, Carson, James P., Warren, Joe, Chiu, Wah, & Kakadiaris, Ioannis. Learning-based Segmentation Framework for Tissue Images Containing Gene Expression Data. United States. doi:10.1109/TMI.2007.895462.
Bello, Musodiq, Ju, Tao, Carson, James P., Warren, Joe, Chiu, Wah, and Kakadiaris, Ioannis. Tue . "Learning-based Segmentation Framework for Tissue Images Containing Gene Expression Data". United States. doi:10.1109/TMI.2007.895462.
@article{osti_912493,
title = {Learning-based Segmentation Framework for Tissue Images Containing Gene Expression Data},
author = {Bello, Musodiq and Ju, Tao and Carson, James P. and Warren, Joe and Chiu, Wah and Kakadiaris, Ioannis},
abstractNote = {Abstract Associating specific gene activity with functional locations in the brain results in a greater understanding of the role of the gene. To perform such an association for the over 20,000 genes in the mammalian genome, reliable automated methods that characterize the distribution of gene expression in relation to a standard anatomical model are required. In this paper, we propose a new automatic method that results in the segmentation of gene expression images into distinct anatomical regions in which the expression can be quantified and compared with other images. Our contribution is a novel hybrid atlas that utilizes a statistical shape model based on a subdivision mesh, texture differentiation at region boundaries, and features of anatomical landmarks to delineate boundaries of anatomical regions in gene expression images. This atlas, which provides a common coordinate system for internal brain data, was trained on 36 images manually annotated by neuroanatomists and tested on 64 images. Our framework has achieved a mean overlap ratio of up to 91 § 7% in this challenging dataset. This tool for large-scale annotation will help scientists interpret gene expression patterns more efficiently.},
doi = {10.1109/TMI.2007.895462},
journal = {IEEE Transactions on Medical Imaging, 26(5):728-744},
number = 5,
volume = 26,
place = {United States},
year = {Tue May 01 00:00:00 EDT 2007},
month = {Tue May 01 00:00:00 EDT 2007}
}
  • Purpose: In radiation treatment planning, delineation of gross tumor volume (GTV) is very important, because the GTVs affect the accuracies of radiation therapy procedure. To assist radiation oncologists in the delineation of GTV regions while treatment planning for lung cancer, we have proposed a machine-learning-based delineation framework of GTV regions of solid and ground glass opacity (GGO) lung tumors following by optimum contour selection (OCS) method. Methods: Our basic idea was to feed voxel-based image features around GTV contours determined by radiation oncologists into a machine learning classifier in the training step, after which the classifier produced the degree ofmore » GTV for each voxel in the testing step. Ten data sets of planning CT and PET/CT images were selected for this study. The support vector machine (SVM), which learned voxel-based features which include voxel value and magnitudes of image gradient vector that obtained from each voxel in the planning CT and PET/CT images, extracted initial GTV regions. The final GTV regions were determined using the OCS method that was able to select a global optimum object contour based on multiple active delineations with a level set method around the GTV. To evaluate the results of proposed framework for ten cases (solid:6, GGO:4), we used the three-dimensional Dice similarity coefficient (DSC), which denoted the degree of region similarity between the GTVs delineated by radiation oncologists and the proposed framework. Results: The proposed method achieved an average three-dimensional DSC of 0.81 for ten lung cancer patients, while a standardized uptake value-based method segmented GTV regions with the DSC of 0.43. The average DSCs for solid and GGO were 0.84 and 0.76, respectively, obtained by the proposed framework. Conclusion: The proposed framework with the support vector machine may be useful for assisting radiation oncologists in delineating solid and GGO lung tumors.« less
  • Purpose: Automatic prostate segmentation from MR images is an important task in various clinical applications such as prostate cancer staging and MR-guided radiotherapy planning. However, the large appearance and shape variations of the prostate in MR images make the segmentation problem difficult to solve. Traditional Active Shape/Appearance Model (ASM/AAM) has limited accuracy on this problem, since its basic assumption, i.e., both shape and appearance of the targeted organ follow Gaussian distributions, is invalid in prostate MR images. To this end, the authors propose a sparse dictionary learning method to model the image appearance in a nonparametric fashion and further integratemore » the appearance model into a deformable segmentation framework for prostate MR segmentation. Methods: To drive the deformable model for prostate segmentation, the authors propose nonparametric appearance and shape models. The nonparametric appearance model is based on a novel dictionary learning method, namely distributed discriminative dictionary (DDD) learning, which is able to capture fine distinctions in image appearance. To increase the differential power of traditional dictionary-based classification methods, the authors' DDD learning approach takes three strategies. First, two dictionaries for prostate and nonprostate tissues are built, respectively, using the discriminative features obtained from minimum redundancy maximum relevance feature selection. Second, linear discriminant analysis is employed as a linear classifier to boost the optimal separation between prostate and nonprostate tissues, based on the representation residuals from sparse representation. Third, to enhance the robustness of the authors' classification method, multiple local dictionaries are learned for local regions along the prostate boundary (each with small appearance variations), instead of learning one global classifier for the entire prostate. These discriminative dictionaries are located on different patches of the prostate surface and trained to adaptively capture the appearance in different prostate zones, thus achieving better local tissue differentiation. For each local region, multiple classifiers are trained based on the randomly selected samples and finally assembled by a specific fusion method. In addition to this nonparametric appearance model, a prostate shape model is learned from the shape statistics using a novel approach, sparse shape composition, which can model nonGaussian distributions of shape variation and regularize the 3D mesh deformation by constraining it within the observed shape subspace. Results: The proposed method has been evaluated on two datasets consisting of T2-weighted MR prostate images. For the first (internal) dataset, the classification effectiveness of the authors' improved dictionary learning has been validated by comparing it with three other variants of traditional dictionary learning methods. The experimental results show that the authors' method yields a Dice Ratio of 89.1% compared to the manual segmentation, which is more accurate than the three state-of-the-art MR prostate segmentation methods under comparison. For the second dataset, the MICCAI 2012 challenge dataset, the authors' proposed method yields a Dice Ratio of 87.4%, which also achieves better segmentation accuracy than other methods under comparison. Conclusions: A new magnetic resonance image prostate segmentation method is proposed based on the combination of deformable model and dictionary learning methods, which achieves more accurate segmentation performance on prostate T2 MR images.« less
  • Purpose: Accurate image segmentation is a crucial step during image guided radiation therapy. This work proposes multi-atlas machine learning (MAML) algorithm for automated segmentation of head-and-neck CT images. Methods: As the first step, the algorithm utilizes normalized mutual information as similarity metric, affine registration combined with multiresolution B-Spline registration, and then fuses together using the label fusion strategy via Plastimatch. As the second step, the following feature selection strategy is proposed to extract five feature components from reference or atlas images: intensity (I), distance map (D), box (B), center of gravity (C) and stable point (S). The box feature Bmore » is novel. It describes a relative position from each point to minimum inscribed rectangle of ROI. The center-of-gravity feature C is the 3D Euclidean distance from a sample point to the ROI center of gravity, and then S is the distance of the sample point to the landmarks. Then, we adopt random forest (RF) in Scikit-learn, a Python module integrating a wide range of state-of-the-art machine learning algorithms as classifier. Different feature and atlas strategies are used for different ROIs for improved performance, such as multi-atlas strategy with reference box for brainstem, and single-atlas strategy with reference landmark for optic chiasm. Results: The algorithm was validated on a set of 33 CT images with manual contours using a leave-one-out cross-validation strategy. Dice similarity coefficients between manual contours and automated contours were calculated: the proposed MAML method had an improvement from 0.79 to 0.83 for brainstem and 0.11 to 0.52 for optic chiasm with respect to multi-atlas segmentation method (MA). Conclusion: A MAML method has been proposed for automated segmentation of head-and-neck CT images with improved performance. It provides the comparable result in brainstem and the improved result in optic chiasm compared with MA. Xuhua Ren and Hao Gao were partially supported by the NSFC (#11405105), the 973 Program (#2015CB856000), and the Shanghai Pujiang Talent Program (#14PJ1404500).« less
  • Highlights: Black-Right-Pointing-Pointer Development of a quantitative high throughput in situ expression profiling method. Black-Right-Pointing-Pointer Application to a tissue microarray of 242 AML bone marrow samples. Black-Right-Pointing-Pointer Identification of HOXA4, HOXA9, Meis1 and DNMT3A as prognostic markers in AML. -- Abstract: Measurement and validation of microarray gene signatures in routine clinical samples is problematic and a rate limiting step in translational research. In order to facilitate measurement of microarray identified gene signatures in routine clinical tissue a novel method combining quantum dot based oligonucleotide in situ hybridisation (QD-ISH) and post-hybridisation spectral image analysis was used for multiplex in-situ transcript detection inmore » archival bone marrow trephine samples from patients with acute myeloid leukaemia (AML). Tissue-microarrays were prepared into which white cell pellets were spiked as a standard. Tissue microarrays were made using routinely processed bone marrow trephines from 242 patients with AML. QD-ISH was performed for six candidate prognostic genes using triplex QD-ISH for DNMT1, DNMT3A, DNMT3B, and for HOXA4, HOXA9, Meis1. Scrambled oligonucleotides were used to correct for background staining followed by normalisation of expression against the expression values for the white cell pellet standard. Survival analysis demonstrated that low expression of HOXA4 was associated with poorer overall survival (p = 0.009), whilst high expression of HOXA9 (p < 0.0001), Meis1 (p = 0.005) and DNMT3A (p = 0.04) were associated with early treatment failure. These results demonstrate application of a standardised, quantitative multiplex QD-ISH method for identification of prognostic markers in formalin-fixed paraffin-embedded clinical samples, facilitating measurement of gene expression signatures in routine clinical samples.« less
  • A dynamic and reciprocal flow of information between cells and the extracellular matrix contributes significantly to the regulation of form and function in developing systems. Signals generated by the extracellular matrix do not act in isolation. Instead, they are processed within the context of global signalling hierarchies whose constituent inputs and outputs are constantly modulated by all the factors present in the cell's surrounding microenvironment. This is particularly evident in the mammary gland, where the construction and subsequent destruction of such a hierarchy regulates changes in tissue-specific gene expression, morphogenesis and apoptosis during each developmental cycle of pregnancy, lactation andmore » involution.« less