Related Experiment Video
Updated: Jan 12, 2026
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
7.4K
Improving radiomics-based differentiation of supratentorial malignant brain tumors preoperatively with
Zeyu Ma1, Chaoli Zhang2, Yang Guo3
1Department of Neurosurgery, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, China.
Summary
This study developed a multiparametric radiomic model using MRI and diffusion-weighted imaging (DWI) to accurately classify high-grade gliomas (HGG), brain metastases (BM), and primary central nervous system lymphomas (PCNSL). The model significantly improved diagnostic performance for BM and PCNSL.
Area of Science:
- Radiology
- Medical Imaging
- Oncology
Background:
- Supratentorial brain tumors, including high-grade glioma (HGG), brain metastases (BM), and primary lymphomas of the central nervous system (PCNSL), present diagnostic challenges.
- Accurate classification is crucial for appropriate treatment planning and patient management.
Purpose of the Study:
- To develop and validate a three-class multiparametric radiomic model for classifying HGG, BM, and PCNSL.
- To assess the added diagnostic value of diffusion-weighted imaging (DWI)-based radiomic features.
Main Methods:
- A multiparametric radiomic model was constructed using 31 features from conventional MRI (T1WI, T1c, T2WI, FLAIR) and apparent diffusion coefficient (ADC) maps.
- A conventional radiomic model using 28 features from conventional MRI was also developed.
- Both models were trained and validated on internal and external datasets using a three-class random forest method.
Main Results:
- The multiparametric radiomic model achieved high diagnostic performance, with AUCs of 0.978 (HGG), 0.918 (BM), and 0.914 (PCNSL) in the internal validation set.
- In the external validation set, AUCs were 0.916 (HGG), 0.898 (BM), and 0.883 (PCNSL).
- The multiparametric model significantly outperformed the conventional model in classifying BM and PCNSL.
Conclusions:
- Diffusion-weighted imaging (DWI)-based radiomic features provide incremental value beyond conventional MRI.
- The multiparametric radiomic model effectively improves the discrimination of common supratentorial malignant brain tumors.
- This approach enhances diagnostic accuracy for HGG, BM, and PCNSL.
Keywords:
Brain metastasesDiffusion-weighted imagingGliomaPrimary lymphoma of the central nervous systemRadiomics
