Related Experiment Video
Updated: Jul 3, 2025
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
Evaluating Autoencoders for Dimensionality Reduction of MRI-derived Radiomics and Classification of Malignant Brain
Mikayla L Biggs1, Yaohua Wang2, Neetu Soni3
1Foundation Medicine Cambridge, Massachusetts, USA.
Abstract:
Malignant brain tumors including parenchymal metastatic (MET) lesions, glioblastomas (GBM), and lymphomas (LYM) account for 29.7% of brain cancers. However, the characterization of these tumors from MRI imaging is difficult due to the similarity of their radiologically observed image features. Radiomics is the extraction of quantitative imaging features to characterize tumor intensity, shape, and texture. Applying machine learning over radiomic features could aid diagnostics by improving the classification of these common brain tumors. However, since the number of radiomic features is typically larger than the number of patients in the study, dimensionality reduction is needed to balance feature dimensionality and model complexity. Autoencoders are a form of unsupervised representation learning that can be used for dimensionality reduction. It is similar to PCA but uses a more complex and non-linear model to learn a compact latent space. In this work, we examine the effectiveness of autoencoders for dimensionality reduction on the radiomic feature space of multiparametric MRI images and the classification of malignant brain tumors: GBM, LYM, and MET. We further aim to address the class imbalances imposed by the rarity of lymphomas by examining different approaches to increase overall predictive performance through multiclass decomposition strategies.
Insights
This study uses autoencoders for dimensionality reduction in brain tumor radiomics, improving the classification of malignant brain tumors like glioblastomas (GBM), lymphomas (LYM), and metastatic (MET) lesions.
Area of Science:
- Medical imaging analysis
- Machine learning in oncology
- Radiomics and artificial intelligence
Background:
- Malignant brain tumors, including metastatic (MET) lesions, glioblastomas (GBM), and lymphomas (LYM), represent a significant portion of brain cancers.
- Characterizing these tumors via MRI is challenging due to similar imaging features.
- Radiomics offers quantitative feature extraction for tumor characterization, but high dimensionality requires reduction for machine learning applications.
Purpose of the Study:
- To evaluate autoencoders for dimensionality reduction in the radiomic feature space of multiparametric MRI.
- To improve the classification accuracy of malignant brain tumors (GBM, LYM, MET).
- To address class imbalance issues, particularly for rare lymphomas, using multiclass decomposition strategies.
Main Methods:
- Extraction of radiomic features from multiparametric MRI data.
- Application of autoencoders for unsupervised dimensionality reduction of radiomic features.
- Development and evaluation of machine learning models for classifying GBM, LYM, and MET tumors.
- Implementation of multiclass decomposition strategies to handle data imbalance.
Main Results:
- Autoencoders effectively reduced the dimensionality of radiomic features.
- The proposed approach demonstrated improved classification performance for malignant brain tumors.
- Multiclass decomposition strategies showed potential in enhancing predictive accuracy, especially for rare classes like lymphomas.
Conclusions:
- Autoencoders are a viable tool for dimensionality reduction in brain tumor radiomics.
- This method enhances the machine learning-based classification of common malignant brain tumors.
- Addressing class imbalance is crucial for robust diagnostic models in neuro-oncology.
