Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Brain tumor classification based on long echo proton MRS signals.

L Lukas1, A Devos, J A K Suykens

  • 1SCD-SISTA, Department of Electrical Engineering, Katholieke Universiteit Leuven, Kasteelpark Arenberg 10, 3001 Heverlee (Leuven), Belgium.

Artificial Intelligence in Medicine
|June 9, 2004
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Eligibility assessment for yttrium-90 radioembolisation in patients with recurrent glioblastoma: A single-centre cohort study based on the FRONTIER clinical trial criteria (NCT05303467).

Radiologia·2026
Same author

Linear Accelerator Mask-Based Radiosurgery for Trigeminal Neuralgia: Prospective Evaluation of 50 Patients.

World neurosurgery·2026
Same author

Circular over-the-wire pulsed field ablation for atrial fibrillation: differences in outcomes between conscious sedation or general anesthesia.

Journal of interventional cardiac electrophysiology : an international journal of arrhythmias and pacing·2025
Same author

Electrical reconnection after pulsed field ablation with a circular over-the-wire catheter in atrial fibrillation.

Europace : European pacing, arrhythmias, and cardiac electrophysiology : journal of the working groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology of the European Society of Cardiology·2025
Same author

Reduction of RF Heating Near Bilateral Deep Brain Stimulation Leads Using Two-Channel RF Shimming at 3T.

NMR in biomedicine·2025
Same author

The utility of dynamic contrast-enhanced intranodal magnetic resonance lymphangiography (MRL) in the investigation of primary lymphatic anomalies.

Clinical radiology·2024

Proton magnetic resonance spectroscopy (1H MRS) effectively classifies brain tumors. Kernel-based methods like SVM and LS-SVM show high performance without dimensionality reduction, matching LDA for tumor discrimination.

Area of Science:

  • Neuroimaging
  • Medical Diagnostics
  • Machine Learning in Medicine

Background:

  • Growing interest in brain tumor classification using proton magnetic resonance spectroscopy (1H MRS).
  • Acquisition of significant long echo 1H MRS data for brain tumor classification by four EU INTERPRET project research centers.

Purpose of the Study:

  • To objectively compare linear and non-linear classification techniques for discriminating four brain tumor types.
  • Evaluate the performance of Linear Discriminant Analysis (LDA), Support Vector Machines (SVM), and Least Squares SVM (LS-SVM) with linear and RBF kernels.

Main Methods:

  • Comparison of LDA, linear SVM, linear LS-SVM, and non-linear LS-SVM (RBF kernel).
  • Inclusion of optimal input variable selection and parameter estimation.
  • Performance evaluation using 200 stratified random train/test samplings.

Related Experiment Videos

  • Receiver Operating Characteristic (ROC) curve analysis for binary classification and accuracy for multiclass classification.

    • Main Results:

    • Automated binary classifiers achieved an Area Under the ROC Curve (AUC) > 0.9 for most tumor types, except glioblastomas versus metastases.
    • No statistically significant performance difference was found between LDA and kernel-based methods (SVM, LS-SVM).
    • Kernel-based methods demonstrated high performance without requiring dimensionality reduction.

    Conclusions:

    • Long echo 1H MRS data, analyzed with various classifiers, shows strong potential for brain tumor classification.
    • Kernel-based methods (SVM, LS-SVM) offer advantages in handling high-dimensional data for brain tumor discrimination.
    • While LDA performed comparably, SVM and LS-SVM provide a robust approach without the need for feature selection or dimensionality reduction.