A single MRI scan contains sufficient imaging information for accurate prediction of meningioma growth risk.
Authors
Affiliations (9)
Affiliations (9)
- Auckland Bioengineering Institute, The University of Auckland, Auckland 1010, New Zealand. Electronic address: [email protected].
- Department of Neurosurgery, Auckland City and Starship Hospitals, Auckland 1023, New Zealand.
- Auckland Bioengineering Institute, The University of Auckland, Auckland 1010, New Zealand.
- Auckland Bioengineering Institute, The University of Auckland, Auckland 1010, New Zealand; Department of Engineering Science and Biomedical Engineering, The University of Auckland, Auckland 1010, New Zealand.
- Brisbane Clinical Neuroscience Centre, South Brisbane, QLD 4101, Australia.
- Centre for Brain Research, The University of Auckland, Auckland 1023, New Zealand; Mātai Medical Imaging Institute, Gisborne 4010, New Zealand; Department of Anatomy and Medical Imaging, The University of Auckland, Auckland 1023, New Zealand.
- Centre for Brain Research, The University of Auckland, Auckland 1023, New Zealand; Departments of Pharmacology and Clinical Pharmacology, The University of Auckland, Auckland 1023, New Zealand.
- Centre for Brain Research, The University of Auckland, Auckland 1023, New Zealand; Department of Anatomy and Medical Imaging, The University of Auckland, Auckland 1023, New Zealand.
- Auckland Bioengineering Institute, The University of Auckland, Auckland 1010, New Zealand; Centre for Brain Research, The University of Auckland, Auckland 1023, New Zealand. Electronic address: [email protected].
Abstract
Neurosurgical strategies for monitoring meningiomas and evaluating their growth risk largely rely on serial imaging or invasive sampling, practices that place considerable burdens on both patients and clinical resources. In this study, we present a novel framework for predicting meningioma growth risk using only a single contrast-enhanced MRI scan. Our approach compares a custom-trained fully convolutional neural network encoder and PyRadiomics features at both tumor- and whole-image scale, capturing tumor-specific and peritumoral image features, evaluated using conventional machine learning classifiers. The study cohort includes 192 patients with single meningiomas, categorized as growing, stable, or shrinking, based on volumetric assessments by expert neurosurgeons. Classifiers trained on encoder-derived features achieved the highest F1-scores of 0.97 ± 0.01, demonstrating strong predictive performance particularly when edema content was included. Ensemble learning on encoder- and PyRadiomics-extracted features did not improve accuracy compared to the individual approaches. Prediction performance varied across scanner vendor, field strength, tumor location, and volume quintiles, with 3 T scanners yielding superior results, notably higher accuracy for smaller tumors under 1.73 cm<sup>3</sup>, and comparatively reduced performance in foramen magnum and intraventricular regions. This represents an important advance with clear clinical relevance, as smaller tumors are more difficult to classify regarding future growth. Our findings establish the feasibility of predicting meningioma growth risk from a single MRI scan, offering a non-invasive approach for early risk stratification and personalized surveillance strategies. By reducing reliance on serial imaging, the approach has the potential to support informed clinical decisions while improving resource allocation and ensuring timely intervention.