Model-free visualization of suspicious lesions in breast MRI based on supervised and unsupervised learning

Twellmann, Thorsten; Meyer‐Baese, Anke; Lange, Oliver; Foo, Simon Y.; Nattkemper, Tim Wilhelm

doi:10.1016/j.engappai.2007.04.005

Cited by 26 publications

(20 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…13,15,[19][20][21][22][23][24][25][26] Automated CADe approaches usually exploit the fact that malignant lesions typically have different signal intensity kinetic profiles on DCE-MRI compared to normal parenchyma. [13][14][15][16][17][18]29,30 Some of these methods have been shown to have a detection accuracy comparable to manual detection. [13][14][15][16][17][18]29,30 However, a CADx system assumes that the lesion detection has been solved either manually or via CADe, and it is typically comprised of two modules: (1) a quantitative feature extractor and (2) a classifier that employs the attributes extracted from the lesion to discriminate lesion classes.…”

Section: Previous Work and Motivationmentioning

confidence: 99%

“…[13][14][15][16][17][18]29,30 Some of these methods have been shown to have a detection accuracy comparable to manual detection. [13][14][15][16][17][18]29,30 However, a CADx system assumes that the lesion detection has been solved either manually or via CADe, and it is typically comprised of two modules: (1) a quantitative feature extractor and (2) a classifier that employs the attributes extracted from the lesion to discriminate lesion classes. Several different CADx classifiers for DCE-MRI have been proposed including linear discriminant analysis (LDA), 17 artificial neural networks, 6,15,19,20 and support vector machine (SVM) classifiers.…”

Section: Previous Work and Motivationmentioning

confidence: 99%

“…11,[13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] Several computer-based image analysis systems have been proposed [13][14][15][16][17][18][19][20][21][22][23][24][25][26]29,30 with the aim of reducing interobserver variability on breast DCE-MRI. CAD approaches for breast MRI are typically either for automatically (a) detecting a lesion (computer-aided detection (CADe)) [13][14][15][16][17][18]29,30 or (b) classifying a lesion as benign or malignant (computer-aided diagnosis (CADx)). 13,15,[19][20][21][22][23][24][25][26] Automated CADe approaches usually exploit the fact that malignant lesions typically have different signal intensity kinetic profiles on DCE-MRI compared to normal parenchyma.…”

Section: Previous Work and Motivationmentioning

confidence: 99%

See 2 more Smart Citations

Textural Kinetics: A Novel Dynamic Contrast-Enhanced (DCE)-MRI Feature for Breast Lesion Classification

et al. 2010

View full text Add to dashboard Cite

Dynamic contrast-enhanced (DCE)-magnetic resonance imaging (MRI) of the breast has emerged as an adjunct imaging tool to conventional X-ray mammography due to its high detection sensitivity. Despite the increasing use of breast DCE-MRI, specificity in distinguishing malignant from benign breast lesions is low, and interobserver variability in lesion classification is high. The novel contribution of this paper is in the definition of a new DCE-MRI descriptor that we call textural kinetics, which attempts to capture spatiotemporal changes in breast lesion texture in order to distinguish malignant from benign lesions. We qualitatively and quantitatively demonstrated on 41 breast DCE-MRI studies that textural kinetic features outperform signal intensity kinetics and lesion morphology features in distinguishing benign from malignant lesions. A probabilistic boosting tree (PBT) classifier in conjunction with textural kinetic descriptors yielded an accuracy of 90%, sensitivity of 95%, specificity of 82%, and an area under the curve (AUC) of 0.92. Graph embedding, used for qualitative visualization of a low-dimensional representation of the data, showed the best separation between benign and malignant lesions when using textural kinetic features. The PBT classifier results and trends were also corroborated via a support vector machine classifier which showed that textural kinetic features outperformed the morphological, static texture, and signal intensity kinetics descriptors. When textural kinetic attributes were combined with morphologic descriptors, the resulting PBT classifier yielded 89% accuracy, 99% sensitivity, 76% specificity, and an AUC of 0.91.

show abstract

Section: Previous Work and Motivationmentioning

confidence: 99%

Section: Previous Work and Motivationmentioning

confidence: 99%

Section: Previous Work and Motivationmentioning

confidence: 99%

See 1 more Smart Citation

Textural Kinetics: A Novel Dynamic Contrast-Enhanced (DCE)-MRI Feature for Breast Lesion Classification

et al. 2010

View full text Add to dashboard Cite

show abstract

“…Supervised learning techniques have been applied to breast MRI data in multiple studies [6,[11][12][13]. Previous work in the scientific literature on assessing the vascular heterogeneity of breast cancer from MRI examinations has involved the use of co-occurrence matrix texture analysis [14,15].…”

Section: Introductionmentioning

confidence: 99%

A Vector Machine Formulation with Application to the Computer-Aided Diagnosis of Breast Cancer from DCE-MRI Screening Examinations

2013

View full text Add to dashboard Cite

This study investigates the use of a proposed vector machine formulation with application to dynamic contrastenhanced magnetic resonance imaging examinations in the context of the computer-aided diagnosis of breast cancer. This paper describes a method for generating feature measurements that characterize a lesion's vascular heterogeneity as well as a supervised learning formulation that represents an improvement over the conventional support vector machine in this application. Spatially varying signal-intensity measures were extracted from the examinations using principal components analysis and the machine learning technique known as the support vector machine (SVM) was used to classify the results. An alternative vector machine formulation was found to improve on the results produced by the established SVM in randomized bootstrap validation trials, yielding a receiveroperating characteristic curve area of 0.82 which represents a statistically significant improvement over the SVM technique in this application.

show abstract

“…The supervised pattern recognition refers to the techniques with the priori knowledge about the category membership of samples used for the classification (Berrueta et al 2007;Twellmanna et al 2008). The identification model is developed on a training set of samples within categories.…”

mentioning

confidence: 99%

Eggshell crack detection based on acoustic impulse response and supervised pattern recognition

Hao¹,

Zhao²,

Chen³

et al. 2009

Czech J. Food Sci.

View full text Add to dashboard Cite

Lin H., Zhao J., Chen Q., Cai J., Zhou P. (2009): Eggshell crack detection based on acoustic impulse response and supervised pattern recognition. Czech J. Food Sci., 27: 393-402.A system based on acoustic resonance was developed for eggshell crack detection. It was achieved by the analysis of the measured frequency response of eggshell excited with a light mechanism. The response signal was processed by recursive least squares adaptive filter, which resulted in the signal-to-noise ratio of the acoustic impulse response reing remarkably enhanced. Five features variables were exacted from the response frequency signals. To develop a robust discrimination model, three pattern recognition algorithms (i.e. K-nearest neighbours, artificial neural network, and support vector machine) were examined comparatively in this work. Some parameters of the model were optimised by cross-validation in the building model. The experimental results showed that the performance of the support vector machine model is the best in comparison to k-nearest neighbours and artificial neural network models. The optimal support vector machine model was obtained with the identification rates of 95.1% in the calibration set, and 97.1% in the prediction set, respectively. Based on the results, it was concluded that the acoustic resonance system combined with the supervised pattern recognition has a significant potential for the cracked eggs detection.

show abstract

Model-free visualization of suspicious lesions in breast MRI based on supervised and unsupervised learning

Cited by 26 publications

References 25 publications

Textural Kinetics: A Novel Dynamic Contrast-Enhanced (DCE)-MRI Feature for Breast Lesion Classification

Textural Kinetics: A Novel Dynamic Contrast-Enhanced (DCE)-MRI Feature for Breast Lesion Classification

A Vector Machine Formulation with Application to the Computer-Aided Diagnosis of Breast Cancer from DCE-MRI Screening Examinations

Eggshell crack detection based on acoustic impulse response and supervised pattern recognition

Contact Info

Product

Resources

About