Multimodal wavelet embedding representation for data combination (MaWERiC): integrating magnetic resonance imaging and spectroscopy for prostate cancer detection

Recently, both Magnetic Resonance (MR) Imaging (MRI) and Spectroscopy (MRS) have emerged as promising tools for detection of prostate cancer (CaP). However, due to the inherent dimensionality differences in MR imaging and spectral information, quantitative integration of T2 weighted MRI (T2w MRI) and MRS for improved CaP detection has been a major challenge. In this paper, we present a novel computerized decision support system called multimodal wavelet embedding representation for data combination (MaWERiC) that employs, (i) wavelet theory to extract 171 Haar wavelet features from MRS and 54 Gabor features from T2w MRI, (ii) dimensionality reduction to individually project wavelet features from MRS and T2w MRI into a common reduced Eigen vector space, and (iii), a random forest classifier for automated prostate cancer detection on a per voxel basis from combined 1.5 T in vivo MRI and MRS. A total of 36 1.5 T endorectal in vivo T2w MRI and MRS patient studies were evaluated per voxel by MaWERiC using a three‐fold cross validation approach over 25 iterations. Ground truth for evaluation of results was obtained by an expert radiologist annotations of prostate cancer on a per voxel basis who compared each MRI section with corresponding ex vivo wholemount histology sections with the disease extent mapped out on histology. Results suggest that MaWERiC based MRS T2w meta‐classifier (mean AUC, μ = 0.89 ± 0.02) significantly outperformed (i) a T2w MRI (using wavelet texture features) classifier (μ = 0.55 ± 0.02), (ii) a MRS (using metabolite ratios) classifier (μ = 0.77 ± 0.03), (iii) a decision fusion classifier obtained by combining individual T2w MRI and MRS classifier outputs (μ = 0.85 ± 0.03), and (iv) a data combination method involving a combination of metabolic MRS and MR signal intensity features (μ = 0.66 ± 0.02). Copyright © 2011 John Wiley & Sons, Ltd.

[1]  H. Hotelling.  Analysis of a complex of statistical variables into principal components. , 1933 .

[2]  P. Dettmar,et al.  Limited value of endorectal magnetic resonance imaging and transrectal ultrasonography in the staging of clinically localized prostate cancer , 2001, BJU international.

[3]  Mario Marchand,et al.  Feature Selection with Conjunctions of Decision Stumps and Learning from Microarray Data , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  J. Tenenbaum,et al.  A global geometric framework for nonlinear dimensionality reduction. , 2000, Science.

[5]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[6]  C. Floyd,et al.  Optimized approach to decision fusion of heterogeneous data for breast cancer diagnosis. , 2006, Medical physics.

[7]  Robert M. Haralick,et al.  Textural Features for Image Classification , 1973, IEEE Trans. Syst. Man Cybern..

[8]  C. Kim,et al.  Localization of Prostate Cancer Using 3T MRI: Comparison of T2-Weighted and Dynamic Contrast-Enhanced Imaging , 2006, Journal of computer assisted tomography.

[9]  Shekhar Verma,et al.  Feature selection using Haar wavelet power spectrum , 2006, BMC Bioinformatics.

[10]  Pierre Comon,et al.  Independent component analysis, A new concept? , 1994, Signal Process..

[11]  M. Kattan,et al.  Prediction of organ-confined prostate cancer: incremental value of MR imaging and MR spectroscopic imaging to staging nomograms. , 2006, Radiology.

[12]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[13]  G. De Meerleer,et al.  A qualitative approach to combined magnetic resonance imaging and spectroscopy in the diagnosis of prostate cancer. , 2010, European journal of radiology.

[14]  J. Ross Quinlan,et al.  C4.5: Programs for Machine Learning , 1992 .

[15]  J. Kurhanewicz,et al.  Improved water and lipid suppression for 3D PRESS CSI using rf band selective inversion with gradient dephasing (basing) , 1997, Magnetic resonance in medicine.

[16]  S. Pizer,et al.  The Image Processing Handbook , 1994 .

[17]  Dimitris N. Metaxas,et al.  Automated detection of prostatic adenocarcinoma from high-resolution ex vivo MRI , 2005, IEEE Transactions on Medical Imaging.

[18]  M S Cohen,et al.  Rapid and effective correction of RF inhomogeneity for high field magnetic resonance imaging , 2000, Human brain mapping.

[19]  A. Jemal,et al.  Cancer Statistics, 2010 , 2010, CA: a cancer journal for clinicians.

[20]  A. Wunderlich,et al.  Multimodal Cranial Neuronavigation: Direct Integration of Functional Magnetic Resonance Imaging and Positron Emission Tomography Data: Technical Note , 2001, Neurosurgery.

[21]  Thomas G. Dietterich.  Multiple Classifier Systems , 2000, Lecture Notes in Computer Science.

[22]  Arend Heerschap,et al.  Combination of feature‐reduced MR spectroscopic and MR imaging data for improved brain tumor classification , 2005, NMR in biomedicine.

[23]  R. Polikar,et al.  Ensemble based systems in decision making , 2006, IEEE Circuits and Systems Magazine.

[24]  David G. Stork,et al.  Pattern Classification (2nd ed.) , 1999 .

[25]  B Michael Kelm,et al.  Automated estimation of tumor probability in prostate magnetic resonance spectroscopic imaging: Pattern recognition vs quantification , 2007, Magnetic resonance in medicine.

[26]  Anant Madabhushi,et al.  A hierarchical spectral clustering and nonlinear dimensionality reduction scheme for detection of prostate cancer from magnetic resonance spectroscopy (MRS). , 2009, Medical physics.

[27]  J. Suykens,et al.  Classification of brain tumours using short echo time 1H MR spectra. , 2004, Journal of magnetic resonance.

[28]  Dirk Van den Poel,et al.  FACULTEIT ECONOMIE , 2007 .

[29]  Thomas Hambrock,et al.  Computer-assisted analysis of peripheral zone prostate lesions using T2-weighted and dynamic contrast enhanced T1-weighted MRI , 2010, Physics in medicine and biology.

[30]  William Wells,et al.  Detection of prostate cancer by integration of line-scan diffusion, T2-mapping and T2-weighted magnetic resonance imaging; a multichannel statistical classifier. , 2003, Medical physics.

[31]  M. Sdika,et al.  Time‐domain semi‐parametric estimation based on a metabolite basis set , 2005, NMR in biomedicine.

[32]  C. Amling,et al.  Extensive repeat transrectal ultrasound guided prostate biopsy in patients with previous benign sextant biopsies. , 2000, The Journal of urology.

[33]  M. Zweig,et al.  Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. , 1993, Clinical chemistry.

[34]  P. Carroll,et al.  Three-dimensional H-1 MR spectroscopic imaging of the in situ human prostate with high (0.24-0.7-cm3) spatial resolution. , 1996, Radiology.

[35]  Harshinder Singh,et al.  Application of the Random Forest Method in Studies of Local Lymph Node Assay Based Skin Sensitization Data. , 2005 .

[36]  Xin Liu,et al.  Prostate Cancer Segmentation With Simultaneous Estimation of Markov Random Field Parameters and Class , 2009, IEEE Transactions on Medical Imaging.

[37]  K. Berberidis,et al.  A computer-aided system for the detection of prostate cancer based on magnetic resonance image analysis , 2008, 2008 3rd International Symposium on Communications, Control and Signal Processing.

[38]  P. Carroll,et al.  Prostate cancer: localization with three-dimensional proton MR spectroscopic imaging--clinicopathologic study. , 1999, Radiology.

[39]  Ronald R. Coifman,et al.  Entropy-based algorithms for best basis selection , 1992, IEEE Trans. Inf. Theory.

[40]  E. Stoner,et al.  Intra‐ and interobserver variability of MRI prostate volume measurements , 1997, The Prostate.

[41]  David G. Stork,et al.  Pattern classification, 2nd Edition , 2000 .

[42]  Torsten Rohlfing,et al.  Information Fusion in Biomedical Image Analysis: Combination of Data vs. Combination of Interpretations , 2005, IPMI.

[43]  Yun Zhang,et al.  Comparison and improvement of wavelet‐based image fusion , 2008 .

[44]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[45]  Nicu Sebe,et al.  Wavelet based texture classification , 2000, Proceedings 15th International Conference on Pattern Recognition. ICPR-2000.

[46]  Matilde Santos Peñas,et al.  Classification of Biomedical Signals Using a Haar 4 Wavelet Transform and a Hamming Neural Network , 2007, IWINAC.

[47]  R E Lenkinski,et al.  Current role of MR imaging in the staging of adenocarcinoma of the prostate. , 1993, Radiology.

[48]  José R. Álvarez,et al.  Nature Inspired Problem-Solving Methods in Knowledge Engineering, Second International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2007, La Manga del Mar Menor, Spain, June 18-21, 2007, Proceedings, Part II , 2007, IWINAC.

[49]  S. Cerutti,et al.  A wavelet packets decomposition algorithm for quantification of in vivo (1)H-MRS parameters. , 2002, Medical engineering & physics.

[50]  P. Carroll,et al.  Prostate depiction at endorectal MR spectroscopic imaging: investigation of a standardized evaluation system. , 2004, Radiology.

[51]  D. Gabor,et al.  Theory of communication. Part 1: The analysis of information , 1946 .

[52]  H. Huisman,et al.  Prostate cancer localization with dynamic contrast-enhanced MR imaging and proton MR spectroscopic imaging. , 2006, Radiology.

[53]  Jan Luts,et al.  Effect of feature extraction for brain tumor classification based on short echo time 1H MR spectra , 2008, Magnetic resonance in medicine.

[54]  P. Choyke,et al.  Prostate cancer: value of multiparametric MR imaging at 3 T for detection--histopathologic correlation. , 2010, Radiology.

[55]  Ramón Díaz-Uriarte,et al.  Gene selection and classification of microarray data using random forest , 2006, BMC Bioinformatics.

[56]  P. Baraldi,et al.  Multimodal MRI in the characterization of glial neoplasms: the combined role of single-voxel MR spectroscopy, diffusion imaging and echo-planar perfusion imaging , 2007, Neuroradiology.

[57]  Zengqi Sun,et al.  Wavelet Denoise on MRS Data Based on ICA and PCA , 2005, ISNN.

[58]  Abbas Z. Kouzani,et al.  Random forest based lung nodule classification aided by clustering , 2010, Comput. Medical Imaging Graph..

[59]  J Kurhanewicz,et al.  Prostate cancer: prediction of extracapsular extension with endorectal MR imaging and three-dimensional proton MR spectroscopic imaging. , 1999, Radiology.

[60]  J. Kurhanewicz,et al.  Combined magnetic resonance imaging and spectroscopic imaging approach to molecular imaging of prostate cancer , 2002, Journal of magnetic resonance imaging : JMRI.

[61]  S T Roweis,et al.  Nonlinear dimensionality reduction by locally linear embedding. , 2000, Science.

[62]  David G. Stork,et al.  Pattern Classification , 1973 .

[63]  Stéphane Mallat,et al.  A Theory for Multiresolution Signal Decomposition: The Wavelet Representation , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[64]  Pramod K. Varshney,et al.  An adaptive multimodal biometric management algorithm , 2005, IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews).

[65]  Zhuowen Tu,et al.  Probabilistic boosting-tree: learning discriminative models for classification, recognition, and clustering , 2005, Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1.

[66]  Nello Cristianini,et al.  A statistical framework for genomic data fusion , 2004, Bioinform..

[67]  Sabine Van Huffel,et al.  Nosologic imaging of the brain: segmentation and classification using MRI and MRSI , 2009, NMR in biomedicine.

[68]  Dennis Gabor,et al.  Theory of communication , 1946 .

[69]  A. Madabhushi,et al.  Integrated diagnostics: a conceptual framework with examples , 2010, Clinical chemistry and laboratory medicine.

[70]  Yoav Freund,et al.  A decision-theoretic generalization of on-line learning and an application to boosting , 1995, EuroCOLT.

[71]  Dana H. Ballard,et al.  Computer Vision , 1982 .

[72]  P. Luyten,et al.  Accurate quantification of in vivo 31P NMR signals using the variable projection method and prior knowledge , 1988, Magnetic resonance in medicine.

[73]  R. Lenkinski,et al.  Elastic registration of multimodal prostate MRI and histology via multiattribute combined mutual information. , 2011, Medical physics.

[74]  P. Carroll,et al.  Multiparametric magnetic resonance imaging in prostate cancer: present and future , 2008, Current opinion in urology.

[75]  Gustavo Carneiro,et al.  Detection and Measurement of Fetal Anatomies from Ultrasound Images using a Constrained Probabilistic Boosting Tree , 2008, IEEE Transactions on Medical Imaging.

[76]  M. van der Graaf,et al.  In vivo proton MR spectroscopy reveals altered metabolite content in malignant prostate tissue. , 1997, Anticancer research.

[77]  John C. Russ,et al.  The Image Processing Handbook , 2016, Microscopy and Microanalysis.

[78]  Vanhamme,et al.  Improved method for accurate and efficient quantification of MRS data with use of prior knowledge , 1997, Journal of magnetic resonance.

[79]  M. Kattan,et al.  The utility of magnetic resonance imaging and spectroscopy for predicting insignificant prostate cancer: an initial analysis , 2007, BJU international.

[80]  E. Squillaci,et al.  MR spectroscopy of prostate cancer. Initial clinical experience. , 2005, Journal of experimental & clinical cancer research : CR.

[81]  J Kurhanewicz,et al.  Magnetic resonance imaging and spectroscopic imaging of prostate cancer. , 2003, Cancer investigation.

[82]  Jun Zheng,et al.  Wavelet based feature reduction method for effective classification of hyperspectral data , 2003, Proceedings ITCC 2003. International Conference on Information Technology: Coding and Computing.

[83]  J. Kurhanewicz,et al.  Very selective suppression pulses for clinical MRSI studies of brain and prostate cancer , 2000, Magnetic resonance in medicine.