Clinical Decision Support Framework for Segmentation and Classification of Brain Tumor MRIs Using a U-Net and DCNN Cascaded Learning Algorithm

Abstract

Brain tumors (BTs) are an uncommon but fatal kind of cancer. Therefore, the development of computer-aided diagnosis (CAD) systems for classifying brain tumors in magnetic resonance imaging (MRI) has been the subject of many research papers so far. However, research in this sector is still in its early stage. The ultimate goal of this research is to develop a lightweight effective implementation of the U-Net deep network for use in performing exact real-time segmentation. Moreover, a simplified deep convolutional neural network (DCNN) architecture for the BT classification is presented for automatic feature extraction and classification of the segmented regions of interest (ROIs). Five convolutional layers, rectified linear unit, normalization, and max-pooling layers make up the DCNN's proposed simplified architecture. The introduced method was verified on multimodal brain tumor segmentation (BRATS 2015) datasets. Our experimental results on BRATS 2015 acquired Dice similarity coefficient (DSC) scores, sensitivity, and classification accuracy of 88.8%, 89.4%, and 88.6% for high-grade gliomas. When it comes to segmenting BRATS 2015 BT images, the performance of our proposed CAD framework is on par with existing state-of-the-art methods. However, the accuracy achieved in this study for the classification of BT images has improved upon the accuracy reported in prior studies. Image classification accuracy for BRATS 2015 BT has been improved from 88% to 88.6%.

Keywords: CAD system; CNN; U-Net; brain tumor; classification; clinical decision; segmentation.

Figure 1

The proposed U-Net and CNN cascaded framework for the segmentation and classification of brain tumor MRIs. (a) The U-Net layers for the segmentation of input images. (b) The CNN layers for the classification of the segmented ROIs.

Figure 2

Rendering of T1, T1c, T2, and FLAIR at a resolution of 32 × 32 × 32.

Figure 3

Rendering of T1, T1c, T2, and FLAIR at a resolution of 64 × 64 × 64.

Figure 4

T1, T1c, T2, and FLAIR rendered at a resolution of 128 × 128 × 128.

Figure 5

Segmentation of high-grade gliomas (HGG) from four MRI sequences (FLAIR, T1, T1C, and T2), respectively, using the U-Net model. Left column: input image; middle column: ground truth images; right column: segmented images.

Figure 5

Segmentation of high-grade gliomas (HGG) from four MRI sequences (FLAIR, T1, T1C, and T2), respectively, using the U-Net model. Left column: input image; middle column: ground truth images; right column: segmented images.

Figure 6

Segmentation of low-grade gliomas (LGG) from four MRI sequences (FLAIR, T1, T1C, and T2), respectively, using the U-Net model. Left column: input image; middle column: ground truth images; right column: segmented images.

Figure 6

Segmentation of low-grade gliomas (LGG) from four MRI sequences (FLAIR, T1, T1C, and T2), respectively, using the U-Net model. Left column: input image; middle column: ground truth images; right column: segmented images.