A Novel 1H-MRS Quantification Approach Based on Spectral Fitting for Lateralization/Localization of Seizure Foci in Patients with Temporal Lobe Epilepsy

  • Neda Mohammadi Quantitative MR Imaging and Spectroscopy Group, Research Center for Molecular and Cellular Imaging, Tehran, Iran. AND Department of Biomedical Engineering and Medical Physics, Tehran University of Medical Sciences, Tehran, Iran.
  • Mohammad Hadi Arabi Quantitative MR Imaging and Spectroscopy Group, Research Center for Molecular and Cellular Imaging, Tehran, Iran. AND Department of Biomedical Engineering and Medical Physics, Tehran University of Medical Sciences, Tehran, Iran.
  • Fatemeh Fadaei Department of Epilepsy, Pars Hospital, Tehran, Iran.
  • Anahita Fathi-Kazerooni Quantitative MR Imaging and Spectroscopy Group, Research Center for Molecular and Cellular Imaging, Tehran, Iran. AND Department of Biomedical Engineering and Medical Physics, Tehran University of Medical Sciences, Tehran, Iran.
  • Jafar Mehvari-Habibabadi Isfahan Neuroscience Research Center, Isfahan University of Medical Sciences, Isfahan, Iran.
  • Mohammad Hossein Harirchian Department of Neurology, School of Medicine, Imam Khomeini Hospital, Tehran University of Medical Sciences, Tehran, Iran.
  • Seyed Sohrab Hashemi-Fesharaki Department of Epilepsy, Pars Hospital, Tehran, Iran.
  • Saeed Sarkar Department of Medical Physics, Faculty of Medicine, Tehran University of Medical Sciences, and Research Center for Science & Technology in Medicine, Imam Khomaini Hospital, Bolvare Keshavarz, Tehran, Iran.
  • Hamidreza Saligheh-Rad Mail Quantitative MR Imaging and Spectroscopy Group, Research Center for Molecular and Cellular Imaging, Tehran, Iran. AND Department of Biomedical Engineering and Medical Physics, Tehran University of Medical Sciences, Tehran, Iran
Keywords:
High dimensional data, Nonlinear dimensionality reduction, Diffusion map, Proton magnetic resonance spec- troscopy (1H-MRS).

Abstract

Purpose: Proton magnetic resonance spectroscopy (1H-MRS) is a non-invasive method with the capability to correctly lateralize the seizure foci in patients with Temporal Lobe Epilepsy (TLE), with the first evidence published in 1993. One major drawback of this modality is that the MRS data is naturally high dimensional. This, along with the time-consuming post-processing and quantification procedures such as spectral fitting, have made MRS impractical for clinical use. Dimension reduction techniques eliminate undesired properties of high-dimensional spaces, suggesting simple and feasible analysis techniques in comparison with quantification procedures.
Methods: In this study, we use two dimension reduction techniques so-called Isomap and Diffusion maps to quantify MRS data obtained from TLE patients for localization seizure foci. Then, we evaluate the results by comparison with obtained ratio of NAA/(Cr+Cho) from the quantification method.
Results: Our results show that the proposed methodology has the ability to localize and/or to lateralize the seizure foci in such patients, while it maintains minimal required amount of computations and time (Sensitivity= 60%, Specificity= 82.81%).
Conclusion: We are hoping that this method broadens new horizons to explore the informative yet complicated MRS modality into Epileptic diagnosis.

Published
2015-03-30
How to Cite
1.
Mohammadi N, Arabi MH, Fadaei F, Fathi-Kazerooni A, Mehvari-Habibabadi J, Harirchian MH, Hashemi-Fesharaki SS, Sarkar S, Saligheh-Rad H. A Novel 1H-MRS Quantification Approach Based on Spectral Fitting for Lateralization/Localization of Seizure Foci in Patients with Temporal Lobe Epilepsy. Frontiers Biomed Technol. 2(1):8-14.
Section
Original Article(s)