2019
Cleland, Theo; Mainprize, J. G.; Alonzo-Proulx, Olivier; Martel, Anne L.; Harvey, J. A.; Yaffe, Martin J.
Use of convolutional neural networks to predict risk of masking by mammographic density. Proceedings Article
In: SPIE Medical Imaging 2019: Computer Aided Diagnosis, pp. 10950–69, SPIE, San Diego, CA, 2019.
Abstract | Links | BibTeX | Tags:
@inproceedings{Cleland2019,
title = {Use of convolutional neural networks to predict risk of masking by mammographic density.},
author = {Theo Cleland and J. G. Mainprize and Olivier Alonzo-Proulx and Anne L. Martel and J. A. Harvey and Martin J. Yaffe},
doi = {10.1117/12.2513063},
year = {2019},
date = {2019-01-01},
booktitle = {SPIE Medical Imaging 2019: Computer Aided Diagnosis},
pages = {10950–69},
publisher = {SPIE},
address = {San Diego, CA},
abstract = {Sensitivity of screening mammography is reduced by increased mammographic density (MD). MD can obscure or “mask” developing lesions making them harder to detect. Predicting masking risk may be an effective tool for a stratified screening program where selected women can receive alternative screening modalities that are less susceptible to masking. Here, we investigate whether the use of artificial intelligence can accurately predict the masking risk and compare its performance to conventional BI-RADS density classification. The analysis was based on 214 subjects comprised of 147 women with a screen-detected (SD) or “non-masked” cancer and 67 that developed a non-screen detected (NSD) or “masked” cancer within 2 years following a negative screen. Prior to analysis, mammograms were pre-processed into quantitative MD maps using an in-house algorithm. A convolutional neural network (CNN) consisting of a VGG-16 base with a basic classifier was used. The CNN was pre-trained on 5,865 mammograms to classify by BI-RADS density category. The CNN was then trained in a seven cross-fold approach to classify masking status. Using BI-RADS density as a masking risk predictor has an AUC of 0.64 [0.57-0.71 95CI]. The CNN masking risk (CNN-mask) yields an AUC of 0.76 [0.70-0.82]. Combining the CNN-mask with our previous hand-crafted masking risk predictor, the AUC improved to 0.78 [0.70-0.84]. The AUC improved to 0.82 [0.70-0.87] when analysis was restricted to 1 year NSD cancers. This work suggests that using a CNN masking risk predictor can be used to guide a stratified screening program to overcome the limitations of screening mammography in dense breasts.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Sensitivity of screening mammography is reduced by increased mammographic density (MD). MD can obscure or “mask” developing lesions making them harder to detect. Predicting masking risk may be an effective tool for a stratified screening program where selected women can receive alternative screening modalities that are less susceptible to masking. Here, we investigate whether the use of artificial intelligence can accurately predict the masking risk and compare its performance to conventional BI-RADS density classification. The analysis was based on 214 subjects comprised of 147 women with a screen-detected (SD) or “non-masked” cancer and 67 that developed a non-screen detected (NSD) or “masked” cancer within 2 years following a negative screen. Prior to analysis, mammograms were pre-processed into quantitative MD maps using an in-house algorithm. A convolutional neural network (CNN) consisting of a VGG-16 base with a basic classifier was used. The CNN was pre-trained on 5,865 mammograms to classify by BI-RADS density category. The CNN was then trained in a seven cross-fold approach to classify masking status. Using BI-RADS density as a masking risk predictor has an AUC of 0.64 [0.57-0.71 95CI]. The CNN masking risk (CNN-mask) yields an AUC of 0.76 [0.70-0.82]. Combining the CNN-mask with our previous hand-crafted masking risk predictor, the AUC improved to 0.78 [0.70-0.84]. The AUC improved to 0.82 [0.70-0.87] when analysis was restricted to 1 year NSD cancers. This work suggests that using a CNN masking risk predictor can be used to guide a stratified screening program to overcome the limitations of screening mammography in dense breasts.
2018
Akbar, Shazia; Martel, Anne L.
Cluster-Based Learning from Weakly Labeled Bags in Digital Pathology Workshop
Machine Learning for Health (ML4H), NeurIPS 2018, Montreal, Canada, 2018.
Abstract | Links | BibTeX | Tags: digital pathology
@workshop{Akbar2018c,
title = {Cluster-Based Learning from Weakly Labeled Bags in Digital Pathology},
author = {Shazia Akbar and Anne L. Martel },
url = {https://arxiv.org/abs/1812.00884},
year = {2018},
date = {2018-12-08},
booktitle = {Machine Learning for Health (ML4H), NeurIPS 2018},
address = {Montreal, Canada},
abstract = {To alleviate the burden of gathering detailed expert annotations when training deep neural networks, we propose a weakly supervised learning approach to recognize metastases in microscopic images of breast lymph nodes. We describe an alternative training loss which clusters weakly labeled bags in latent space to inform relevance of patch-instances during training of a convolutional neural network. We evaluate our method on the Camelyon dataset which contains high-resolution digital slides of breast lymph nodes, where labels are provided at the image-level and only subsets of patches are made available during training.},
keywords = {digital pathology},
pubstate = {published},
tppubtype = {workshop}
}
To alleviate the burden of gathering detailed expert annotations when training deep neural networks, we propose a weakly supervised learning approach to recognize metastases in microscopic images of breast lymph nodes. We describe an alternative training loss which clusters weakly labeled bags in latent space to inform relevance of patch-instances during training of a convolutional neural network. We evaluate our method on the Camelyon dataset which contains high-resolution digital slides of breast lymph nodes, where labels are provided at the image-level and only subsets of patches are made available during training.
Kuling, Grey; Fashandi, Homa; Lu, YingLi; Wu, Hongbo; Martel, Anne L.
Breast Volume and Fibroglandular Tissue Segmentation in MRI using a Deep Learning Unet Workshop
ISMRM Workshop on Breast MRI: Advancing the State of the Art, 2018.
Links | BibTeX | Tags: _breast_segmentation, Breast MRI, Breast-CAD
@workshop{Kuling2018,
title = {Breast Volume and Fibroglandular Tissue Segmentation in MRI using a Deep Learning Unet},
author = {Grey Kuling and Homa Fashandi and YingLi Lu and Hongbo Wu and Anne L. Martel},
url = {http://martellab.com/wp-content/uploads/2019/09/GCK_ISMRMAbstract_DLSegmatation_072018-3.pdf},
year = {2018},
date = {2018-09-10},
urldate = {2018-09-10},
booktitle = {ISMRM Workshop on Breast MRI: Advancing the State of the Art},
keywords = {_breast_segmentation, Breast MRI, Breast-CAD},
pubstate = {published},
tppubtype = {workshop}
}
Akbar, Shazia; Peikari, Mohammad; Salama, Sherine; Nofech-Mozes, Sharon; Martel, Anne L.
Determining tumor cellularity in digital slides using ResNet Proceedings Article
In: SPIE Medical Imaging, Houston, Texas, 2018.
Abstract | Links | BibTeX | Tags: digital pathology
@inproceedings{Akbar2018a,
title = {Determining tumor cellularity in digital slides using ResNet},
author = {Shazia Akbar and Mohammad Peikari and Sherine Salama and Sharon Nofech-Mozes and Anne L. Martel},
url = {https://www.spiedigitallibrary.org/conference-proceedings-of-spie/10581/105810U/Determining-tumor-cellularity-in-digital-slides-using-ResNet/10.1117/12.2292813.full},
year = {2018},
date = {2018-01-01},
booktitle = {SPIE Medical Imaging},
volume = {10581},
address = {Houston, Texas},
abstract = {The residual cancer burden index is a powerful prognostic factor which is used to measure neoadjuvant therapy response in invasive breast cancers. Tumor cellularity is one component of the residual cancer burden index and is currently measured manually through eyeballing. As such it is subject to inter- and intra-variability and is currently restricted to discrete values. We propose a method for automatically determining tumor cellularity in digital slides using deep learning techniques. We train a series of ResNet architectures to output both discrete and continuous values and compare our outcomes with scores acquired manually by an expert pathologist. Our configurations were validated on a dataset of image patches extracted from digital slides, each containing various degrees of tumor cellularity. Results showed that, in the case of discrete values, our models were able to distinguish between regions-of-interest containing tumor and healthy cells with over 97% test accuracy rates. Overall, we achieved 76% accuracy over four predefined tumor cellularity classes (no tumor/tumor; low, medium and high tumor cellularity). When computing tumor cellularity scores on a continuous scale, ResNet showed good correlations with manually-identified scores, showing potential for computing reproducible scores consistent with expert opinion using deep learning techniques.},
keywords = {digital pathology},
pubstate = {published},
tppubtype = {inproceedings}
}
The residual cancer burden index is a powerful prognostic factor which is used to measure neoadjuvant therapy response in invasive breast cancers. Tumor cellularity is one component of the residual cancer burden index and is currently measured manually through eyeballing. As such it is subject to inter- and intra-variability and is currently restricted to discrete values. We propose a method for automatically determining tumor cellularity in digital slides using deep learning techniques. We train a series of ResNet architectures to output both discrete and continuous values and compare our outcomes with scores acquired manually by an expert pathologist. Our configurations were validated on a dataset of image patches extracted from digital slides, each containing various degrees of tumor cellularity. Results showed that, in the case of discrete values, our models were able to distinguish between regions-of-interest containing tumor and healthy cells with over 97% test accuracy rates. Overall, we achieved 76% accuracy over four predefined tumor cellularity classes (no tumor/tumor; low, medium and high tumor cellularity). When computing tumor cellularity scores on a continuous scale, ResNet showed good correlations with manually-identified scores, showing potential for computing reproducible scores consistent with expert opinion using deep learning techniques.
Akbar, Shazia; Peikari, Mohammad; Salama, Sherine; Nofech-Mozes, Sharon; Martel, Anne L.
The transition module: A method for preventing overfitting in convolutional neural networks Journal Article
In: Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, pp. 1-6, 2018.
Abstract | Links | BibTeX | Tags:
@article{Akbar2018b,
title = {The transition module: A method for preventing overfitting in convolutional neural networks},
author = {Shazia Akbar and Mohammad Peikari and Sherine Salama and Sharon Nofech-Mozes and Anne L. Martel},
url = {https://www.tandfonline.com/doi/abs/10.1080/21681163.2018.1427148?tab=permissions&scroll=top},
doi = {10.1080/21681163.2018.1427148},
year = {2018},
date = {2018-01-01},
journal = {Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization},
pages = {1-6},
abstract = {Digital pathology has advanced substantially over the last decade with the adoption of slide scanners in pathology labs. The use of digital slides to analyse diseases at the microscopic level is both cost-effective and efficient. Identifying complex tumor patterns in digital slides is a challenging problem but holds significant importance for tumor burden assessment, grading and many other pathological assessments in cancer research. The use of convolutional neural networks (CNNs) to analyze such complex images has been well adopted in digital pathology. However in recent years, the architecture of CNNs have altered with the introduction of inception modules which have shown great promise for classification tasks. In this paper, we propose a modified "transition'' module which encouraging generalization in a deep learning framework with few training samples. In the transition module, filters of varying sizes are used to encourage class-specific filters at multiple spatial resolutions followed by global average pooling. We demonstrate the performance of the transition module in AlexNet and ZFNet, for classifying breast tumors in two independent datasets of scanned histology sections; the inclusion of the transition module in these CNNs improved performance.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Digital pathology has advanced substantially over the last decade with the adoption of slide scanners in pathology labs. The use of digital slides to analyse diseases at the microscopic level is both cost-effective and efficient. Identifying complex tumor patterns in digital slides is a challenging problem but holds significant importance for tumor burden assessment, grading and many other pathological assessments in cancer research. The use of convolutional neural networks (CNNs) to analyze such complex images has been well adopted in digital pathology. However in recent years, the architecture of CNNs have altered with the introduction of inception modules which have shown great promise for classification tasks. In this paper, we propose a modified "transition'' module which encouraging generalization in a deep learning framework with few training samples. In the transition module, filters of varying sizes are used to encourage class-specific filters at multiple spatial resolutions followed by global average pooling. We demonstrate the performance of the transition module in AlexNet and ZFNet, for classifying breast tumors in two independent datasets of scanned histology sections; the inclusion of the transition module in these CNNs improved performance.
Levman, Jacob; Takahashi, Emi; Forgeron, Cynthia; MacDonald, Patrick; Stewart, Natalie; Lim, Ashley; Martel, Anne
A Sorting Statistic with Application in Neurological Magnetic Resonance Imaging of Autism Journal Article
In: Journal of Healthcare Engineering, vol. 2018, pp. 1–6, 2018, ISSN: 2040-2295.
Abstract | Links | BibTeX | Tags:
@article{Levman2018,
title = {A Sorting Statistic with Application in Neurological Magnetic Resonance Imaging of Autism},
author = {Jacob Levman and Emi Takahashi and Cynthia Forgeron and Patrick MacDonald and Natalie Stewart and Ashley Lim and Anne Martel},
url = {https://www.hindawi.com/journals/jhe/2018/8039075/},
doi = {10.1155/2018/8039075},
issn = {2040-2295},
year = {2018},
date = {2018-01-01},
journal = {Journal of Healthcare Engineering},
volume = {2018},
pages = {1–6},
abstract = {Effect size refers to the assessment of the extent of differences between two groups of samples on a single measurement. Assessing effect size in medical research is typically accomplished with Cohen's d statistic. Cohen's d statistic assumes that average values are good estimators of the position of a distribution of numbers and also assumes Gaussian (or bell-shaped) underlying data distributions. In this paper, we present an alternative evaluative statistic that can quantify differences between two data distributions in a manner that is similar to traditional effect size calculations; however, the proposed approach avoids making assumptions regarding the shape of the underlying data distribution. The proposed sorting statistic is compared with Cohen's d statistic and is demonstrated to be capable of identifying feature measurements of potential interest for which Cohen's d statistic implies the measurement would be of little use. This proposed sorting statistic has been evaluated on a large clinical autism dataset from Boston Children's Hospital , Harvard Medical School , demonstrating that it can potentially play a constructive role in future healthcare technologies.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Effect size refers to the assessment of the extent of differences between two groups of samples on a single measurement. Assessing effect size in medical research is typically accomplished with Cohen's d statistic. Cohen's d statistic assumes that average values are good estimators of the position of a distribution of numbers and also assumes Gaussian (or bell-shaped) underlying data distributions. In this paper, we present an alternative evaluative statistic that can quantify differences between two data distributions in a manner that is similar to traditional effect size calculations; however, the proposed approach avoids making assumptions regarding the shape of the underlying data distribution. The proposed sorting statistic is compared with Cohen's d statistic and is demonstrated to be capable of identifying feature measurements of potential interest for which Cohen's d statistic implies the measurement would be of little use. This proposed sorting statistic has been evaluated on a large clinical autism dataset from Boston Children's Hospital , Harvard Medical School , demonstrating that it can potentially play a constructive role in future healthcare technologies.
2017
Akbar, Shazia; Peikari, Mohammad; Salama, Sherine; Nofech-Mozes, Sharon; Martel, Anne L.
Transitioning between Convolutional and Fully Connected Layers in Neural Networks Proceedings Article
In: 3rd workshop on Deep Learning in Medical Image Analysis (DLMIA), MICCAI 2017, 2017.
Abstract | Links | BibTeX | Tags: CBCF-DCIS
@inproceedings{Akbar2017a,
title = {Transitioning between Convolutional and Fully Connected Layers in Neural Networks},
author = {Shazia Akbar and Mohammad Peikari and Sherine Salama and Sharon Nofech-Mozes and Anne L. Martel},
url = {https://arxiv.org/abs/1707.05743},
year = {2017},
date = {2017-00-00},
booktitle = {3rd workshop on Deep Learning in Medical Image Analysis (DLMIA), MICCAI 2017},
abstract = {Digital pathology has advanced substantially over the last decade however tumor localization continues to be a challenging problem due to highly complex patterns and textures in the underlying tissue bed. The use of convolutional neural networks (CNNs) to analyze such complex images has been well adopted in digital pathology. However in recent years, the architecture of CNNs have altered with the introduction of inception modules which have shown great promise for classification tasks. In this paper, we propose a modified "transition" module which learns global average pooling layers from filters of varying sizes to encourage class-specific filters at multiple spatial resolutions. We demonstrate the performance of the transition module in AlexNet and ZFNet, for classifying breast tumors in two independent datasets of scanned histology sections, of which the transition module was superior.},
keywords = {CBCF-DCIS},
pubstate = {published},
tppubtype = {inproceedings}
}
Digital pathology has advanced substantially over the last decade however tumor localization continues to be a challenging problem due to highly complex patterns and textures in the underlying tissue bed. The use of convolutional neural networks (CNNs) to analyze such complex images has been well adopted in digital pathology. However in recent years, the architecture of CNNs have altered with the introduction of inception modules which have shown great promise for classification tasks. In this paper, we propose a modified "transition" module which learns global average pooling layers from filters of varying sizes to encourage class-specific filters at multiple spatial resolutions. We demonstrate the performance of the transition module in AlexNet and ZFNet, for classifying breast tumors in two independent datasets of scanned histology sections, of which the transition module was superior.
2016
Wu, Hongbo
Automatic Computer Aided Diagnosis of Breast Cancer in Dynamic Contrast Enhanced Magnetic Resonance Images Masters Thesis
University of Toronto, Department Medical Biophysics , 2016.
Abstract | Links | BibTeX | Tags: Breast-CAD
@mastersthesis{Wu2016,
title = {Automatic Computer Aided Diagnosis of Breast Cancer in Dynamic Contrast Enhanced Magnetic Resonance Images},
author = {Wu, Hongbo},
url = {http://hdl.handle.net/1807/76226},
year = {2016},
date = {2016-11-01},
address = {Department Medical Biophysics },
school = {University of Toronto},
abstract = {Automated Computer Aided Diagnosis (CADx) systems have the potential to improve the diagnostic accuracy of radiologists. Most CADx algorithms use features generated from outlined regions to differentiate between benign and malignant lesions. Manually outlining these regions for the purpose of analysis is not viable and therefore an automated segmentation method is essential. Our proposed method uses a trained deep Artificial Neural Network (ANN) to classify overlapping tiles in breast Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI) images as lesion or non-lesion. The classified tiles are then grouped into regions. Additional morphological, kinetic and textural features are computed for each detected region. A cascaded Random Forests Classifier (RFC) classifies the regions as malignant or benign. Our method was tested on a dataset containing 71 malignant, 140 benign, and 316 normal studies. Free-response Receiver Operating Characteristic (FROC) analysis of our method shows 94.4% sensitivity at 0.12 false positive detections per normal study.},
keywords = {Breast-CAD},
pubstate = {published},
tppubtype = {mastersthesis}
}
Automated Computer Aided Diagnosis (CADx) systems have the potential to improve the diagnostic accuracy of radiologists. Most CADx algorithms use features generated from outlined regions to differentiate between benign and malignant lesions. Manually outlining these regions for the purpose of analysis is not viable and therefore an automated segmentation method is essential. Our proposed method uses a trained deep Artificial Neural Network (ANN) to classify overlapping tiles in breast Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI) images as lesion or non-lesion. The classified tiles are then grouped into regions. Additional morphological, kinetic and textural features are computed for each detected region. A cascaded Random Forests Classifier (RFC) classifies the regions as malignant or benign. Our method was tested on a dataset containing 71 malignant, 140 benign, and 316 normal studies. Free-response Receiver Operating Characteristic (FROC) analysis of our method shows 94.4% sensitivity at 0.12 false positive detections per normal study.
Rushin Shojaii, Anne L Martel
Optimized SIFTFlow for registration of whole-mount histology to reference optical images Journal Article
In: Journal of Medical Imaging, vol. 3, no. 4, pp. 047501-047501, 2016.
Abstract | Links | BibTeX | Tags: digital pathology, registration
@article{Shojaii2016,
title = {Optimized SIFTFlow for registration of whole-mount histology to reference optical images},
author = {Rushin Shojaii, Anne L Martel},
url = {http://medicalimaging.spiedigitallibrary.org/article.aspx?articleid=2571703},
doi = {10.1117/1.JMI.3.4.047501},
year = {2016},
date = {2016-10-19},
journal = {Journal of Medical Imaging},
volume = {3},
number = {4},
pages = {047501-047501},
abstract = {The registration of two-dimensional histology images to reference images from other modalities is an important preprocessing step in the reconstruction of three-dimensional histology volumes. This is a challenging problem because of the differences in the appearances of histology images and other modalities, and the presence of large nonrigid deformations which occur during slide preparation. This paper shows the feasibility of using densely sampled scale-invariant feature transform (SIFT) features and a SIFTFlow deformable registration algorithm for coregistering whole-mount histology images with blockface optical images. We present a method for jointly optimizing the regularization parameters used by the SIFTFlow objective function and use it to determine the most appropriate values for the registration of breast lumpectomy specimens. We demonstrate that tuning the regularization parameters results in significant improvements in accuracy and we also show that SIFTFlow outperforms a previously described edge-based registration method. The accuracy of the histology images to blockface images registration using the optimized SIFTFlow method was assessed using an independent test set of images from five different lumpectomy specimens and the mean registration error was 0.32±0.22 mm0.32±0.22 mm.},
keywords = {digital pathology, registration},
pubstate = {published},
tppubtype = {article}
}
The registration of two-dimensional histology images to reference images from other modalities is an important preprocessing step in the reconstruction of three-dimensional histology volumes. This is a challenging problem because of the differences in the appearances of histology images and other modalities, and the presence of large nonrigid deformations which occur during slide preparation. This paper shows the feasibility of using densely sampled scale-invariant feature transform (SIFT) features and a SIFTFlow deformable registration algorithm for coregistering whole-mount histology images with blockface optical images. We present a method for jointly optimizing the regularization parameters used by the SIFTFlow objective function and use it to determine the most appropriate values for the registration of breast lumpectomy specimens. We demonstrate that tuning the regularization parameters results in significant improvements in accuracy and we also show that SIFTFlow outperforms a previously described edge-based registration method. The accuracy of the histology images to blockface images registration using the optimized SIFTFlow method was assessed using an independent test set of images from five different lumpectomy specimens and the mean registration error was 0.32±0.22 mm0.32±0.22 mm.
Peikari, Mohammad; Martel, Anne L.
Automatic cell detection and segmentation from H and E stained pathology slides using colorspace decorrelation stretching Proceedings Article
In: Gurcan, Metin N.; Madabhushi, Anant (Ed.): Medical Imaging 2016: Digital Pathology, pp. 979114-1: 979114-6 , SPIE SPIE, 2016.
Abstract | Links | BibTeX | Tags: digital pathology
@inproceedings{Peikari2016,
title = {Automatic cell detection and segmentation from H and E stained pathology slides using colorspace decorrelation stretching},
author = {Mohammad Peikari and Anne L. Martel},
editor = {Metin N. Gurcan and Anant Madabhushi},
url = {http://proceedings.spiedigitallibrary.org/proceeding.aspx?articleid=2506827},
doi = {10.1117/12.2216507},
year = {2016},
date = {2016-03-23},
booktitle = {Medical Imaging 2016: Digital Pathology},
volume = {9791},
number = {2},
pages = {979114-1: 979114-6 },
publisher = {SPIE},
organization = {SPIE},
abstract = {Purpose: Automatic cell segmentation plays an important role in reliable diagnosis and prognosis of patients. Most of the state-of-the-art cell detection and segmentation techniques focus on complicated methods to subtract foreground cells from the background. In this study, we introduce a preprocessing method which leads to a better detection and segmentation results compared to a well-known state-of-the-art work. Method: We transform the original red-green-blue (RGB) space into a new space defined by the top eigenvectors of the RGB space. Stretching is done by manipulating the contrast of each pixel value to equalize the color variances. New pixel values are then inverse transformed to the original RGB space. This altered RGB image is then used to segment cells. Result: The validation of our method with a well-known state-of-the-art technique revealed a statistically significant improvement on an identical validation set. We achieved a mean F1-score of 0.901. Conclusion: Preprocessing steps to decorrelate colorspaces may improve cell segmentation performances},
keywords = {digital pathology},
pubstate = {published},
tppubtype = {inproceedings}
}
Purpose: Automatic cell segmentation plays an important role in reliable diagnosis and prognosis of patients. Most of the state-of-the-art cell detection and segmentation techniques focus on complicated methods to subtract foreground cells from the background. In this study, we introduce a preprocessing method which leads to a better detection and segmentation results compared to a well-known state-of-the-art work. Method: We transform the original red-green-blue (RGB) space into a new space defined by the top eigenvectors of the RGB space. Stretching is done by manipulating the contrast of each pixel value to equalize the color variances. New pixel values are then inverse transformed to the original RGB space. This altered RGB image is then used to segment cells. Result: The validation of our method with a well-known state-of-the-art technique revealed a statistically significant improvement on an identical validation set. We achieved a mean F1-score of 0.901. Conclusion: Preprocessing steps to decorrelate colorspaces may improve cell segmentation performances
2011
Lausch, Anthony
Nonrigid Registration of Dynamic Contrast-enhanced MRI Data using Motion Informed Intensity Corrections Masters Thesis
University of Toronto, Medical Biophysics, 2011.
Abstract | Links | BibTeX | Tags: registration
@mastersthesis{Lausch2011,
title = {Nonrigid Registration of Dynamic Contrast-enhanced MRI Data using Motion Informed Intensity Corrections},
author = {Lausch, Anthony},
url = {http://hdl.handle.net/1807/31294},
year = {2011},
date = {2011-12-13},
address = {Medical Biophysics},
school = {University of Toronto},
abstract = {Effective early detection and monitoring of patient response to cancer therapy is important for improved patient outcomes, avoiding unnecessary procedures and their associated toxicities, as well as the development of new therapies. Dynamic contrast-enhanced magnetic resonance imaging shows promise as a way to evaluate tumour vasculature and assess the efficacy of new anti-angiogenic drugs. However, unavoidable patient motion can decrease the accuracy of subsequent analyses rendering the data unusable. Motion correction algorithms are challenging to develop for contrast-enhanced data since intensity changes due to contrast-enhancement and patient motion must somehow be differentiated from one another. A novel method is presented that employs a motion-informed intensity correction in order to facilitate the registration of contrast enhanced data. The intensity correction simulates the presence or absence of contrast agent in the image volumes to be registered in an attempt to emulate the level of contrast-enhancement present in a single reference image volume.},
keywords = {registration},
pubstate = {published},
tppubtype = {mastersthesis}
}
Effective early detection and monitoring of patient response to cancer therapy is important for improved patient outcomes, avoiding unnecessary procedures and their associated toxicities, as well as the development of new therapies. Dynamic contrast-enhanced magnetic resonance imaging shows promise as a way to evaluate tumour vasculature and assess the efficacy of new anti-angiogenic drugs. However, unavoidable patient motion can decrease the accuracy of subsequent analyses rendering the data unusable. Motion correction algorithms are challenging to develop for contrast-enhanced data since intensity changes due to contrast-enhancement and patient motion must somehow be differentiated from one another. A novel method is presented that employs a motion-informed intensity correction in order to facilitate the registration of contrast enhanced data. The intensity correction simulates the presence or absence of contrast agent in the image volumes to be registered in an attempt to emulate the level of contrast-enhancement present in a single reference image volume.
Gallego, Cristina
Automatic 3D Segmentation of the Breast in MRI Masters Thesis
University of Toronto, Medical Biophysics, 2011.
Abstract | Links | BibTeX | Tags: Breast-CAD
@mastersthesis{Gallego2011,
title = {Automatic 3D Segmentation of the Breast in MRI},
author = {Gallego, Cristina},
url = {http://hdl.handle.net/1807/30619},
year = {2011},
date = {2011-12-08},
address = {Medical Biophysics},
school = {University of Toronto},
abstract = {Breast cancer is currently the most common diagnosed cancer among women and a significant cause of death. Breast density is considered a significant risk factor and an important biomarker influencing the later risk of breast cancer. Therefore, ongoing epidemiological studies using MRI are evaluating quantitatively breast density in young women. One of the challenges is segmenting the breast in order to calculate total breast volume and exclude non-breast surrounding tissues. This thesis describes an automatic 3D breast volume segmentation based on 3D local edge detection using phase congruency and Poisson surface reconstruction to extract the total breast volume in 3D. The boundary localization framework is integrated on a subsequent atlas-based segmentation using a Laplacian framework. The 3D segmentation achieves breast-air and breast-chest wall boundary localization errors with a median of 1.36 mm and 2.68 mm respectively when tested on 409 MRI datasets.
},
keywords = {Breast-CAD},
pubstate = {published},
tppubtype = {mastersthesis}
}
Breast cancer is currently the most common diagnosed cancer among women and a significant cause of death. Breast density is considered a significant risk factor and an important biomarker influencing the later risk of breast cancer. Therefore, ongoing epidemiological studies using MRI are evaluating quantitatively breast density in young women. One of the challenges is segmenting the breast in order to calculate total breast volume and exclude non-breast surrounding tissues. This thesis describes an automatic 3D breast volume segmentation based on 3D local edge detection using phase congruency and Poisson surface reconstruction to extract the total breast volume in 3D. The boundary localization framework is integrated on a subsequent atlas-based segmentation using a Laplacian framework. The 3D segmentation achieves breast-air and breast-chest wall boundary localization errors with a median of 1.36 mm and 2.68 mm respectively when tested on 409 MRI datasets.
2010
Levman, Jacob
University of Toronto, 2010.
Abstract | Links | BibTeX | Tags: Breast-CAD
@phdthesis{Levman2010,
title = {Pattern Recognition Applied to the Computer-aided Detection and Diagnosis of Breast Cancer from Dynamic Contrast-enhanced Magnetic Resonance Breast Images},
author = {Jacob Levman },
url = {http://hdl.handle.net/1807/24361},
year = {2010},
date = {2010-04-21},
urldate = {2010-04-21},
address = {Medical Biophysics},
school = {University of Toronto},
abstract = {The goal of this research is to improve the breast cancer screening process based on magnetic resonance imaging (MRI). In a typical MRI breast examination, a radiologist is responsible for visually examining the MR images acquired during the examination and identifying suspect tissues for biopsy. It is known that if multiple radiologists independently analyze the same examinations and we biopsy any lesion that any of our radiologists flagged as suspicious then the overall screening process becomes more sensitive but less specific. Unfortunately cost factors prohibit the use of multiple radiologists for the screening of every breast MR examination. It is thought that instead of having a second expert human radiologist to examine each set of images, that the act of second reading of the examination can be performed by a computer-aided detection and diagnosis system. The research presented in this thesis is focused on the development of a computer-aided detection and diagnosis system for breast cancer screening from dynamic contrast-enhanced magnetic resonance imaging examinations. This thesis presents new computational techniques in supervised learning, unsupervised learning and classifier visualization. The techniques have been applied to breast MR lesion data and have been shown to outperform existing methods yielding a computer aided detection and diagnosis system with a sensitivity of 89% and a specificity of 70%.},
keywords = {Breast-CAD},
pubstate = {published},
tppubtype = {phdthesis}
}
The goal of this research is to improve the breast cancer screening process based on magnetic resonance imaging (MRI). In a typical MRI breast examination, a radiologist is responsible for visually examining the MR images acquired during the examination and identifying suspect tissues for biopsy. It is known that if multiple radiologists independently analyze the same examinations and we biopsy any lesion that any of our radiologists flagged as suspicious then the overall screening process becomes more sensitive but less specific. Unfortunately cost factors prohibit the use of multiple radiologists for the screening of every breast MR examination. It is thought that instead of having a second expert human radiologist to examine each set of images, that the act of second reading of the examination can be performed by a computer-aided detection and diagnosis system. The research presented in this thesis is focused on the development of a computer-aided detection and diagnosis system for breast cancer screening from dynamic contrast-enhanced magnetic resonance imaging examinations. This thesis presents new computational techniques in supervised learning, unsupervised learning and classifier visualization. The techniques have been applied to breast MR lesion data and have been shown to outperform existing methods yielding a computer aided detection and diagnosis system with a sensitivity of 89% and a specificity of 70%.
2008
Levman, Jacob; Leung, Tony; Causer, Petrina; Plewes, Don; Martel, Anne L.
Classification of dynamic contrast-enhanced magnetic resonance breast lesions by support vector machines Journal Article
In: IEEE Transactions on Medical Imaging , vol. 27, no. 5, pp. 688 - 696, 2008.
Abstract | Links | BibTeX | Tags: Breast-CAD
@article{Levman2008,
title = {Classification of dynamic contrast-enhanced magnetic resonance breast lesions by support vector machines},
author = {Jacob Levman and Tony Leung and Petrina Causer and Don Plewes and Anne L. Martel},
url = {https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2891012/},
doi = {10.1109/TMI.2008.916959},
year = {2008},
date = {2008-05-01},
journal = {IEEE Transactions on Medical Imaging },
volume = {27},
number = {5},
pages = {688 - 696},
abstract = {Early detection of breast cancer is one of the most important factors in determining prognosis for women with malignant tumors. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) has been shown to be the most sensitive modality for screening high-risk women. Computer-aided diagnosis (CAD) systems have the potential to assist radiologists in the early detection of cancer. A key component of the development of such a CAD system will be the selection of an appropriate classification function responsible for separating malignant and benign lesions. The purpose of this study is to evaluate the effects of variations in temporal feature vectors and kernel functions on the separation of malignant and benign DCE-MRI breast lesions by support vector machines (SVMs). We also propose and demonstrate a classifier visualization and evaluation technique. We show that SVMs provide an effective and flexible framework from which to base CAD techniques for breast MRI, and that the proposed classifier visualization technique has potential as a mechanism for the evaluation of classification solutions.},
keywords = {Breast-CAD},
pubstate = {published},
tppubtype = {article}
}
Early detection of breast cancer is one of the most important factors in determining prognosis for women with malignant tumors. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) has been shown to be the most sensitive modality for screening high-risk women. Computer-aided diagnosis (CAD) systems have the potential to assist radiologists in the early detection of cancer. A key component of the development of such a CAD system will be the selection of an appropriate classification function responsible for separating malignant and benign lesions. The purpose of this study is to evaluate the effects of variations in temporal feature vectors and kernel functions on the separation of malignant and benign DCE-MRI breast lesions by support vector machines (SVMs). We also propose and demonstrate a classifier visualization and evaluation technique. We show that SVMs provide an effective and flexible framework from which to base CAD techniques for breast MRI, and that the proposed classifier visualization technique has potential as a mechanism for the evaluation of classification solutions.