Automated differentiation of glioblastomas from intracranial metastases using 3T MR spectroscopic and perfusion data

AbstractPurpose   Differentiation of glioblastomas from metastases is clinical important, but may be difficult even for expert observers. To investigate the contribution of machine learning algorithms in the differentiation of glioblastomas multiforme (GB) from metastases, we developed and tested a pattern recognition system based on 3T magnetic resonance (MR) data. Materials and Methods   Single and multi-voxel proton magnetic resonance spectroscopy (1H-MRS) and dynamic susceptibility contrast (DSC) MRI scans were performed on 49 patients with solitary brain tumors (35 glioblastoma multiforme and 14 metastases). Metabolic (NAA/Cr, Cho/Cr, (Lip $$+$$ Lac)/Cr) and perfusion (rCBV) parameters were measured in both intratumoral and peritumoral regions. The statistical significance of these parameters was evaluated. For the classification procedure, three datasets were created to find the optimum combination of parameters that provides maximum differentiation. Three machine learning methods were utilized: Naïve-Bayes, Support Vector Machine (SVM) and $$k$$-nearest neighbor (KNN). The discrimination ability of each classifier was evaluated with quantitative performance metrics. Results   Glioblastoma and metastases were differentiable only in the peritumoral region of these lesions ($$p<0.05$$). SVM achieved the highest overall performance (accuracy 98 %) for both the intratumoral and peritumoral areas. Naïve-Bayes and KNN presented greater variations in performance. The proper selection of datasets plays a very significant role as they are closely correlated to the underlying pathophysiology. Conclusion   The application of pattern recognition techniques using 3T MR-based perfusion and metabolic features may provide incremental diagnostic value in the differentiation of common intraaxial brain tumors, such as glioblastoma versus metastasis.

[1]  I. Tsougos,et al.  Spectroscopic Evaluation of Glioma Grading at 3T: The Combined Role of Short and Long TE , 2012, TheScientificWorldJournal.

[2]  Sabine Van Huffel,et al.  On the Design of a Web-Based Decision Support System for Brain Tumour Diagnosis Using Distributed Agents , 2006, 2006 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology Workshops.

[3]  Sebastian Zander,et al.  A preliminary performance comparison of five machine learning algorithms for practical IP traffic flow classification , 2006, CCRV.

[4]  M P Lichy,et al.  Diagnostic performance of spectroscopic and perfusion MRI for distinction of brain tumors , 2006, Neurology.

[5]  S. Ng,et al.  Primary Cerebral Lymphoma and Glioblastoma Multiforme: Differences in Diffusion Characteristics Evaluated with Diffusion Tensor Imaging , 2008, American Journal of Neuroradiology.

[6]  Matthijs Oudkerk,et al.  Perfusion MR imaging for differentiation of benign and malignant meningiomas , 2008, Neuroradiology.

[7]  A Heerschap,et al.  Discrimination between Metastasis and Glioblastoma Multiforme Based on Morphometric Analysis of MR Images , 2010, American Journal of Neuroradiology.

[8]  Michalis E. Zervakis,et al.  Brain lesion classification using 3T MRS spectra and paired SVM kernels , 2011, Biomed. Signal Process. Control..

[9]  Pat Langley,et al.  Estimating Continuous Distributions in Bayesian Classifiers , 1995, UAI.

[10]  Sarah Jane Delany.  k-Nearest Neighbour Classifiers , 2007 .

[11]  Philip S. Yu,et al.  Top 10 algorithms in data mining , 2007, Knowledge and Information Systems.

[12]  L. G. Davis,et al.  Basic methods in molecular biology , 1986 .

[13]  Christos Davatzikos,et al.  Investigating machine learning techniques for MRI-based classification of brain neoplasms , 2011, International Journal of Computer Assisted Radiology and Surgery.

[14]  Sabine Van Huffel,et al.  Multiproject–multicenter evaluation of automatic brain tumor classification by magnetic resonance spectroscopy , 2008, Magnetic Resonance Materials in Physics, Biology and Medicine.

[15]  Pedro M. Domingos,et al.  On the Optimality of the Simple Bayesian Classifier under Zero-One Loss , 1997, Machine Learning.

[16]  Christos Davatzikos,et al.  Classification of brain tumor type and grade using MRI texture and shape in a machine learning scheme , 2009, Magnetic resonance in medicine.

[17]  Sanjeev Chawla,et al.  Proton Magnetic Resonance Spectroscopy in Differentiating Glioblastomas From Primary Cerebral Lymphomas and Brain Metastases , 2010, Journal of computer assisted tomography.

[18]  F. Howe,et al.  Differentiation of metastases from high‐grade gliomas using short echo time 1H spectroscopy , 2004, Journal of magnetic resonance imaging : JMRI.

[19]  Sabine Van Huffel,et al.  Brain tumor classification based on long echo proton MRS signals , 2004, Artif. Intell. Medicine.

[20]  A. Bjørnerud,et al.  Glioma grading by using histogram analysis of blood volume heterogeneity from MR-derived cerebral blood volume maps. , 2008, Radiology.

[21]  Jie Yang,et al.  Degree prediction of malignancy in brain glioma using support vector machines , 2006, Comput. Biol. Medicine.

[22]  Hairong Qi.  Feature Selection and kNN Fusion in Molecular Classification of Multiple Tumor Types , .

[23]  Leonard E. Trigg,et al.  Technical Note: Naive Bayes for Regression , 2000, Machine Learning.

[24]  G Johnson,et al.  Glial neoplasms: dynamic contrast-enhanced T2*-weighted MR imaging. , 1999, Radiology.

[25]  Mahlon D. Johnson,et al.  MR diffusion tensor and perfusion-weighted imaging in preoperative grading of supratentorial nonenhancing gliomas. , 2011, Neuro-oncology.

[26]  Glyn Johnson,et al.  High-grade gliomas and solitary metastases: differentiation by using perfusion and proton spectroscopic MR imaging. , 2002, Radiology.

[27]  A. W. Simonetti,et al.  The use of multivariate MR imaging intensities versus metabolic data from MR spectroscopic imaging for brain tumour classification. , 2005, Journal of magnetic resonance.

[28]  Linda Zhang,et al.  Pilot study on evaluation of any correlation between MR perfusion (Ktrans) and diffusion (apparent diffusion coefficient) parameters in brain tumors at 3 Tesla , 2012, Cancer imaging : the official publication of the International Cancer Imaging Society.

[29]  R M Weisskoff,et al.  Relative cerebral blood volume maps corrected for contrast agent extravasation significantly correlate with glioma tumor grade, whereas uncorrected maps do not. , 2006, AJNR. American journal of neuroradiology.

[30]  D. Mikulis,et al.  Diagnostic value of peritumoral minimum apparent diffusion coefficient for differentiation of glioblastoma multiforme from solitary metastatic lesions. , 2011, AJR. American journal of roentgenology.

[31]  Elias R Melhem,et al.  Intraaxial brain masses: MR imaging-based diagnostic strategy--initial experience. , 2007, Radiology.

[32]  A. W. Simonetti,et al.  Development of a decision support system for diagnosis and grading of brain tumours using in vivo magnetic resonance single voxel spectra , 2006, NMR in biomedicine.

[33]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[34]  Soonmee Cha,et al.  Neuroimaging in neuro-oncology , 2009, Neurotherapeutics.

[35]  Patricia Svolos,et al.  Differentiation of glioblastoma multiforme from metastatic brain tumor using proton magnetic resonance spectroscopy, diffusion and perfusion metrics at 3 T , 2012, Cancer imaging : the official publication of the International Cancer Imaging Society.

[36]  ChiangIC,et al.  Distinction between high-grade gliomas and solitary metastases using peritumoral 3-T magnetic resonance spectroscopy, diffusion, and perfusion imagings , 2004 .

[37]  Jason Weston,et al.  A user's guide to support vector machines. , 2010, Methods in molecular biology.

[38]  Peter E. Hart,et al.  Nearest neighbor pattern classification , 1967, IEEE Trans. Inf. Theory.

[39]  Sabine Van Huffel,et al.  HealthAgents: distributed multi-agent brain tumor diagnosis and prognosis , 2009, Applied Intelligence.

[40]  J. Kazmierska,et al.  Application of the Naïve Bayesian Classifier to optimize treatment decisions. , 2008, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[41]  Juan Miguel García-Gómez,et al.  Brain Tumor Classification Using Magnetic Resonance Spectroscopy , 2011 .

[42]  V. L. Doyle,et al.  Metabolic profiles of human brain tumors using quantitative in vivo 1H magnetic resonance spectroscopy , 2003, Magnetic resonance in medicine.

[43]  Jerome H. Friedman,et al.  On Bias, Variance, 0/1—Loss, and the Curse-of-Dimensionality , 2004, Data Mining and Knowledge Discovery.

[44]  Donald A. Adjeroh,et al.  Random KNN feature selection - a fast and stable alternative to Random Forests , 2011, BMC Bioinformatics.

[45]  Z. Wu,et al.  In vivo single-voxel proton MR spectroscopy in the differentiation of high-grade gliomas and solitary metastases. , 2004, Clinical radiology.

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

[47]  Leon N. Cooper,et al.  Improving nearest neighbor rule with a simple adaptive distance measure , 2007, Pattern Recognit. Lett..