Deep learning-based detection of pediatric brain tumor presence on MRI.
The study reports cross-validated development of a CNN using T1-weighted MRI from 285 pediatric patients to distinguish tumor from nontumor images, with the best 40-epoch configuration achieving 99% precision, 100% recall, and a 99% F1-score.
Open original publication →What the AI sees
The study reports cross-validated development of a CNN using T1-weighted MRI from 285 pediatric patients to distinguish tumor from nontumor images, with the best 40-epoch configuration achieving 99% precision, 100% recall, and a 99% F1-score.
Research significance
Evidence: the CNN showed high tumor-detection performance within the supplied pediatric MRI dataset; inference: after rigorous external and prospective validation, such a model might support earlier diagnosis, reduce missed tumors, or improve imaging workflows, but no treatment-selection or patient-outcome benefit is demonstrated here.
Source abstract
INTRODUCTION: Pediatric cancer imposes a significant global burden, with ∼400 000 new cases annually. Magnetic resonance imaging (MRI) is a fundamental technique for timely diagnosis; however, interpretation can be subjective. In this context, convolutional neural networks (CNNs) have emerged as a promising tool for early and accurate detection of brain tumors in pediatric patients. OBJECTIVE: To develop and validate a CNN model for tumor detection tasks (tumor vs nontumor) of pediatric brain MRI images using T1-weighted imaging. METHODOLOGY: T1-weighted brain MRI images from 285 pediatric patients were included (140 with confirmed tumors and 145 controls). A CNN architecture with convolutional layers, max-pooling, and dropout regularization was implemented. The model was trained and evaluated using cross-validation, employing the following metrics: accuracy, precision, recall, and F1-score. Model performance was compared across training configurations of 10-50 and 100 epochs to determine the optimal setup. RESULTS: The model trained for 40 epochs achieved the best overall performance, with a precision of 99%, a recall of 100%, and an F1-score of 99%. These metrics demonstrate excellent tumor detection while eliminating false negatives, a crucial aspect in pediatric oncology. CONCLUSION: The CNN demonstrates strong potential as a diagnostic aid for detection brain tumors using pediatric MRI images. Its integration into clinical environments could contribute to more objective interpretation, optimize diagnostic workflows, and provide imaging follow-up before and after treatment.