Related Experiment Video
Updated: Jan 17, 2026
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Non-invasive Prediction of Meningioma Tumor Grade by Quantification of Shape-based Radiomics Features and Surface
Martin Diebold1,2,3, Lucas Becker4, Theo Demerath4
1Institute of Neuropathology, Medical Center-University of Freiburg, Faculty of Medicine, University of Freiburg, 79106, Freiburg, Germany.
Purpose:
Prognosis and therapeutic approaches to meningiomas are determined by their proliferative activity and tumor grade. The early identification of aggressive phenotypes is hence essential to improving outcomes. Here, we investigated the potential of shape-based radiomics features and surface regularity (SR) as non-invasive biomarkers of atypical (grade 2) and anaplastic (grade 3) meningiomas.
Methods:
In this retrospective single-center study, we assessed individuals with treatment-naive meningiomas in a comprehensive analysis of MR imaging with histopathological grading and quantification of mitotic activity. Radiomics (sphericity, elongation, flatness) and SR measures were calculated on segmented contrast-enhancing tumor components and tested for association with WHO grade (n = 62 grade 1, n = 71 grade 2, and n = 19 grade 3) and proliferation.
Results:
All tumor grades were significantly different in characteristics of sphericity and SR with decreasing values at higher tumor grades as an expression of a more irregular shape. Additionally, radiomics features of tumor elongation and flatness discriminated grade 3 cases from grade 1 (p = 0.0008 and p = 0.0003) and grade 2 meningiomas (p = 0.0008 and p = 0.0013). Mitotic activity exhibited a significant negative correlation with the four imaging markers assessed (p < 0.003).
Conclusion:
Shape-based radiomics features of sphericity and SR can serve as preoperative grading biomarkers of meningiomas in routine contrast-enhanced T1w MRI.
