Journal Article
. 2012 Apr;39(4).
doi: 10.1118/1.3682173.

Comparative analysis of nonlinear dimensionality reduction techniques for breast MRI segmentation

Alireza Akhbardeh 1 Michael A Jacobs  
  • PMID: 22482648
  •     30 References
  •     8 citations


Purpose: Visualization of anatomical structures using radiological imaging methods is an important tool in medicine to differentiate normal from pathological tissue and can generate large amounts of data for a radiologist to read. Integrating these large data sets is difficult and time-consuming. A new approach uses both supervised and unsupervised advanced machine learning techniques to visualize and segment radiological data. This study describes the application of a novel hybrid scheme, based on combining wavelet transform and nonlinear dimensionality reduction (NLDR) methods, to breast magnetic resonance imaging (MRI) data using three well-established NLDR techniques, namely, ISOMAP, local linear embedding (LLE), and diffusion maps (DfM), to perform a comparative performance analysis.

Methods: Twenty-five breast lesion subjects were scanned using a 3T scanner. MRI sequences used were T1-weighted, T2-weighted, diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE) imaging. The hybrid scheme consisted of two steps: preprocessing and postprocessing of the data. The preprocessing step was applied for B(1) inhomogeneity correction, image registration, and wavelet-based image compression to match and denoise the data. In the postprocessing step, MRI parameters were considered data dimensions and the NLDR-based hybrid approach was applied to integrate the MRI parameters into a single image, termed the embedded image. This was achieved by mapping all pixel intensities from the higher dimension to a lower dimensional (embedded) space. For validation, the authors compared the hybrid NLDR with linear methods of principal component analysis (PCA) and multidimensional scaling (MDS) using synthetic data. For the clinical application, the authors used breast MRI data, comparison was performed using the postcontrast DCE MRI image and evaluating the congruence of the segmented lesions.

Results: The NLDR-based hybrid approach was able to define and segment both synthetic and clinical data. In the synthetic data, the authors demonstrated the performance of the NLDR method compared with conventional linear DR methods. The NLDR approach enabled successful segmentation of the structures, whereas, in most cases, PCA and MDS failed. The NLDR approach was able to segment different breast tissue types with a high accuracy and the embedded image of the breast MRI data demonstrated fuzzy boundaries between the different types of breast tissue, i.e., fatty, glandular, and tissue with lesions (>86%).

Conclusions: The proposed hybrid NLDR methods were able to segment clinical breast data with a high accuracy and construct an embedded image that visualized the contribution of different radiological parameters.

Incremental nonlinear dimensionality reduction by manifold learning.
Martin H C Law, Anil K Jain.
IEEE Trans Pattern Anal Mach Intell, 2006 Mar 11; 28(3). PMID: 16526424
Cognition. The manifold ways of perception.
H S Seung, D D Lee.
Science, 2001 Feb 24; 290(5500). PMID: 11188725
Method to correct intensity inhomogeneity in MR images for atherosclerosis characterization.
Olivier Salvado, Claudia Hillenbrand, Shaoxiang Zhang, David L Wilson.
IEEE Trans Med Imaging, 2006 May 13; 25(5). PMID: 16689259
3-T dynamic contrast-enhanced MRI of the breast: pharmacokinetic parameters versus conventional kinetic curve analysis.
Riham H El Khouli, Katarzyna J Macura, +2 authors, David A Bluemke.
AJR Am J Roentgenol, 2011 Nov 24; 197(6). PMID: 22109308    Free PMC article.
Multiparametric magnetic resonance imaging of breast cancer.
Michael A Jacobs.
J Am Coll Radiol, 2009 Jun 30; 6(7). PMID: 19560072
Riemannian manifold learning.
Tong Lin, Hongbin Zha.
IEEE Trans Pattern Anal Mach Intell, 2008 Mar 29; 30(5). PMID: 18369250
Multispectral analysis of magnetic resonance images.
M W Vannier, R L Butterfield, +3 authors, M Gado.
Radiology, 1985 Jan 01; 154(1). PMID: 3964938
Perception-based visualization of manifold-valued medical images using distance-preserving dimensionality reduction.
Ghassan Hamarneh, Chris McIntosh, Mark S Drew.
IEEE Trans Med Imaging, 2011 Feb 08; 30(7). PMID: 21296705
Monitoring of neoadjuvant chemotherapy using multiparametric, ²³Na sodium MR, and multimodality (PET/CT/MRI) imaging in locally advanced breast cancer.
Michael A Jacobs, Ronald Ouwerkerk, +6 authors, Vered Stearns.
Breast Cancer Res Treat, 2011 Apr 02; 128(1). PMID: 21455671    Free PMC article.
A hierarchical spectral clustering and nonlinear dimensionality reduction scheme for detection of prostate cancer from magnetic resonance spectroscopy (MRS).
Pallavi Tiwari, Mark Rosen, Anant Madabhushi.
Med Phys, 2009 Oct 09; 36(9). PMID: 19810465    Free PMC article.
MRI techniques to monitor MS evolution: the present and the future.
Massimo Filippi, Robert I Grossman.
Neurology, 2002 Apr 24; 58(8). PMID: 11971079
A global geometric framework for nonlinear dimensionality reduction.
J B Tenenbaum, V de Silva, J C Langford.
Science, 2000 Dec 23; 290(5500). PMID: 11125149
Highly Cited.
Multi-parametric neuroimaging reproducibility: a 3-T resource study.
Bennett A Landman, Alan J Huang, +14 authors, Peter C M van Zijl.
Neuroimage, 2010 Nov 26; 54(4). PMID: 21094686    Free PMC article.
Highly Cited.
Diffusion-weighted imaging improves the diagnostic accuracy of conventional 3.0-T breast MR imaging.
Riham H Ei Khouli, Michael A Jacobs, +4 authors, David A Bluemke.
Radiology, 2010 Jun 25; 256(1). PMID: 20574085    Free PMC article.
Highly Cited.
Differentiation of clinically benign and malignant breast lesions using diffusion-weighted imaging.
Yong Guo, You-Quan Cai, +5 authors, Jia-Hong Gao.
J Magn Reson Imaging, 2002 Aug 31; 16(2). PMID: 12203765
Highly Cited.
Geometric diffusions as a tool for harmonic analysis and structure definition of data: diffusion maps.
R R Coifman, S Lafon, +4 authors, S W Zucker.
Proc Natl Acad Sci U S A, 2005 May 19; 102(21). PMID: 15899970    Free PMC article.
Highly Cited.
Magnetic resonance imaging of the breast.
Riham H El Khouli, Michael A Jacobs, David A Bluemke.
Semin Roentgenol, 2008 Sep 09; 43(4). PMID: 18774031
Overview on research and development of computer-aided diagnostic schemes.
Kunio Doi.
Semin Ultrasound CT MR, 2004 Nov 24; 25(5). PMID: 15559123
Nonlinear dimensionality reduction by locally linear embedding.
S T Roweis, L K Saul.
Science, 2000 Dec 23; 290(5500). PMID: 11125150
Highly Cited.
Tumor invasion margin on the Riemannian space of brain fibers.
Dana Cobzas, Parisa Mosayebi, Albert Murtha, Martin Jagersand.
Med Image Comput Comput Assist Interv, 2009 Jan 01; 12(Pt 2). PMID: 20426153
Elastic registration in the presence of intensity variations.
Senthil Periaswamy, Hany Farid.
IEEE Trans Med Imaging, 2003 Aug 09; 22(7). PMID: 12906240
Physical radiotherapy treatment planning based on functional PET/CT data.
Daniela Thorwarth, Xavier Geets, Marta Paiusco.
Radiother Oncol, 2010 Aug 03; 96(3). PMID: 20673689
Multiparametric magnetic resonance imaging, spectroscopy and multinuclear (²³Na) imaging monitoring of preoperative chemotherapy for locally advanced breast cancer.
Michael A Jacobs, Vered Stearns, +9 authors, Ronald Ouwerkerk.
Acad Radiol, 2010 Sep 25; 17(12). PMID: 20863721    Free PMC article.
Small lesions evaluation based on unsupervised cluster analysis of signal-intensity time courses in dynamic breast MRI.
A Meyer-Baese, T Schlossbauer, O Lange, A Wismueller.
Int J Biomed Imaging, 2009 Jan 01; 2009. PMID: 20379361    Free PMC article.
On analyzing diffusion tensor images by identifying manifold structure using isomaps.
Ragini Verma, Parmeshwar Khurd, Christos Davatzikos.
IEEE Trans Med Imaging, 2007 Aug 08; 26(6). PMID: 17679328
Combined use of MRI and PET to monitor response and assess residual disease for locally advanced breast cancer treated with neoadjuvant chemotherapy.
Xiaoming Chen, Mark O Moore, +5 authors, Robert B Livingston.
Acad Radiol, 2004 Nov 09; 11(10). PMID: 15530804
Unsupervised segmentation of multiparameter MRI in experimental cerebral ischemia with comparison to T2, diffusion, and ADC MRI parameters and histopathological validation.
M A Jacobs, R A Knight, +5 authors, M Chopp.
J Magn Reson Imaging, 2000 Apr 15; 11(4). PMID: 10767072
Functional magnetic resonance imaging in prostate cancer.
Michael Seitz, Amita Shukla-Dave, +6 authors, Anno Graser.
Eur Urol, 2009 Feb 03; 55(4). PMID: 19185981
Model-Free Visualization of Suspicious Lesions in Breast MRI Based on Supervised and Unsupervised Learning.
Thorsten Twellmann, Anke Meyer-Baese, +2 authors, Tim W Nattkemper.
Eng Appl Artif Intell, 2009 Mar 04; 21(2). PMID: 19255616    Free PMC article.
Multiparametric and multinuclear magnetic resonance imaging of human breast cancer: current applications.
Michael A Jacobs, Ronald Ouwerkerk, +8 authors, David A Bluemke.
Technol Cancer Res Treat, 2004 Nov 25; 3(6). PMID: 15560711
Strain-encoded breast MRI in phantom and ex vivo specimens with histological validation: preliminary results.
Ahmed A Haruoni, Jakir Hossain, +4 authors, Michael A Jacobs.
Med Phys, 2012 Dec 13; 39(12). PMID: 23231318    Free PMC article.
Locally linear embedding (LLE) for MRI based Alzheimer's disease classification.
Xin Liu, Duygu Tosun, +2 authors, Alzheimer's Disease Neuroimaging Initiative.
Neuroimage, 2013 Jun 25; 83. PMID: 23792982    Free PMC article.
Whole Body MRI at 3T with Quantitative Diffusion Weighted Imaging and Contrast-Enhanced Sequences for the Characterization of Peripheral Lesions in Patients with Neurofibromatosis Type 2 and Schwannomatosis.
Laura M Fayad, Jaishri Blakeley, +2 authors, Michael A Jacobs.
ISRN Radiol, 2013 Jan 01; 2013. PMID: 24967287    Free PMC article.
Automatic computation of left ventricular volume changes over a cardiac cycle from echocardiography images by nonlinear dimensionality reduction.
Zahra Alizadeh Sani, Ahmad Shalbaf, Hamid Behnam, Reza Shalbaf.
J Digit Imaging, 2014 Jul 26; 28(1). PMID: 25059548    Free PMC article.
Echocardiography without electrocardiogram using nonlinear dimensionality reduction methods.
Ahmad Shalbaf, Zahra AlizadehSani, Hamid Behnam.
J Med Ultrason (2001), 2015 Nov 19; 42(2). PMID: 26576567
Radiomics: a new application from established techniques.
Vishwa Parekh, Michael A Jacobs.
Expert Rev Precis Med Drug Dev, 2017 Jan 04; 1(2). PMID: 28042608    Free PMC article.
Highly Cited.
Visualizing Alzheimer's disease progression in low dimensional manifolds.
Kangwon Seo, Rong Pan, +3 authors, Alzheimer's Disease Neuroimaging Initiative.
Heliyon, 2019 Aug 14; 5(8). PMID: 31406946    Free PMC article.
Integrated Multiparametric Radiomics and Informatics System for Characterizing Breast Tumor Characteristics with the OncotypeDX Gene Assay.
Michael A Jacobs, Christopher B Umbricht, +5 authors, Antonio C Wolff.
Cancers (Basel), 2020 Oct 01; 12(10). PMID: 32992569    Free PMC article.