A Deep Learning Model Collaborates with an Expert Radiologist to Classify Brain Tumors from MR Images

Abstract

The brain, which consists of nerve cells called neurons, is the center of the nervous system. The rapid and abnormal growth of nerve cells by interacting with each other is called a brain tumor. Undiagnosed or delayed diagnosis of brain tumors lead to death. Although it depends on experience, manually diagnosing and classifying brain tumors is challenging for physicians. Artificial intelligence-based computer systems can help doctors detect brain tumors using the developments in hardware technology and the amount of data increasing daily. This study proposes a deep learning-based system to classify brain MRI images as tumorous or normal using the pre-trained EfficientNet-B0 model. Our radiologist validated a public dataset containing 3000 brain MRI images. The dataset is divided into 70% train, 20% validation, and 10% test. In the test phase after the training, the pre-trained EfficientNet-B0 model achieved high performance with 99.33% accuracy, 99.33% sensitivity, and 99.33% F1 score. In addition, in the evaluation of the test images, the heat maps obtained by the Grad-CAM method were examined by our radiology specialist. The result of evaluations shows that the pre-trained EfficientNet-B0 deep model chooses the right focus areas in its predictions and can be used for clinical tumor detection due to its explainable structure.

Keywords

• Brain tumor
• MRI
• deep learning
• EfficientNet
• Grad-CAM

References

1. Perkins A, Liu G. Primary brain tumors in adults: diagnosis and treatment. American family physician, 2016; 93(3): 211-217.
2. Villanueva-Meyer J. E, Mabray M. C, Cha S. Current clinical brain tumor imaging. Neurosurgery, 2017; 81(3): 397-415.
3. Zhang Y, Wang S, Wu H, Hu K, Ji S. Brain Tumors Classification for MR images based on Attention Guided Deep Learning Model. 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), 2021; 3233-3236.
4. Chartrand G, Cheng P. M, Vorontsov E, Drozdzal M, Turcotte S, Pal C. J, Tang A. Deep learning: a primer for radiologists. Radiographics, 2017; 37(7): 2113-2131.
5. Arsalan M, Owais M, Mahmood T, Choi J, Park K. R. Artificial intelligence-based diagnosis of cardiac and related diseases. Journal of Clinical Medicine, 2020; 9(3): 871.
6. Biswas M, Kuppili V, Saba L, Edla D. R, Suri H. S, Cuadrado-Godia, E, Suri J. S. State-of-the-art review on deep learning in medical imaging. Front Biosci (Landmark Ed), 2019; 24: 392-426.
7. Khan, H. A, Jue W, Mushtaq M, Mushtaq M. U. Brain tumor classification in MRI image using convolutional neural network. Math. Biosci. Eng, 2020; 17(5): 6203-6216.
8. Singh V, Sharma S, Goel S, Lamba S, Garg N. Brain Tumor Prediction by Binary Classification Using VGG‐16. Smart and Sustainable Intelligent Systems, 2021; 127-138.

9. Pundir A, Kumar R. Brain Tumor Classification in MRI Images Using Transfer Learning. Machine Learning for Intelligent Multimedia Analytics, 2021; 307-319.
10. A. Hamada, Br35h: Brain Tumor Detection 2020, version 5, 2020, https://www.kaggle.com/ahmedhamada0/brain-tumordetection.
11. Tan M, Le Q. Efficientnet: Rethinking model scaling for convolutional neural networks. International conference on machine learning, 2019; 6105-6114.
12. Deng J, Dong W, Socher R, Li L. J, Li K, Fei-Fei L. Imagenet: A large-scale hierarchical image database. Computer vision and pattern recognition, 2009; 248-255.
13. Selvaraju R. R, Cogswell M, Das A, Vedantam R, Parikh D, Batra D. Grad-cam: Visual explanations from deep networks via gradient-based localization. International conference on computer vision, 2017; 618-626.