Convolutional Neural Network Techniques for Brain Tumor Classification (from 2015 to 2022): Review, Challenges, and Future Perspectives

Abstract

Convolutional neural networks (CNNs) constitute a widely used deep learning approach that has frequently been applied to the problem of brain tumor diagnosis. Such techniques still face some critical challenges in moving towards clinic application. The main objective of this work is to present a comprehensive review of studies using CNN architectures to classify brain tumors using MR images with the aim of identifying useful strategies for and possible impediments in the development of this technology. Relevant articles were identified using a predefined, systematic procedure. For each article, data were extracted regarding training data, target problems, the network architecture, validation methods, and the reported quantitative performance criteria. The clinical relevance of the studies was then evaluated to identify limitations by considering the merits of convolutional neural networks and the remaining challenges that need to be solved to promote the clinical application and development of CNN algorithms. Finally, possible directions for future research are discussed for researchers in the biomedical and machine learning communities. A total of 83 studies were identified and reviewed. They differed in terms of the precise classification problem targeted and the strategies used to construct and train the chosen CNN. Consequently, the reported performance varied widely, with accuracies of 91.63-100% in differentiating meningiomas, gliomas, and pituitary tumors (26 articles) and of 60.0-99.46% in distinguishing low-grade from high-grade gliomas (13 articles). The review provides a survey of the state of the art in CNN-based deep learning methods for brain tumor classification. Many networks demonstrated good performance, and it is not evident that any specific methodological choice greatly outperforms the alternatives, especially given the inconsistencies in the reporting of validation methods, performance metrics, and training data encountered. Few studies have focused on clinical usability.

Keywords: brain tumor classification; clinical application; clinical effectiveness; computer-aided diagnosis; convolutional neural network; deep learning; magnetic resonance imaging.

Figure 1

The PRISMA flowchart of this review. n: number of articles.

Figure 2

The basic workflow of a typical CNN-based brain tumor classification study with four high-level steps: Step 1. Input Image: 2D or 3D Brain MR samples are fed into the classification model; Step 2. Preprocessing: several preprocessing techniques are used to remove the skull, normalize the images, resize the images, and augment the number of training examples; Step 3. CNN Classification: the preprocessed dataset is propagated into the CNN model and is involved in training, validation, and testing processes; Step 4. Performance Evaluation: evaluation of the classification performance of a CNN algorithm with accuracy, specificity, F₁ score, area under the curve, and sensitivity metrics.

Figure 3

Data augmentation: (a) original image; (b) 18° rotation. When rotating by an arbitrary number of degrees (non-modulo 90), rotation will result in the image being padded in each corner. Then, a crop is taken from the center of the newly rotated image to retain the largest crop possible while maintaining the image’s aspect ratio; (c) left–right flipping; (d) top–bottom flipping; (e) scaling by 1.5 times; (f) cropping by center cropping to the size 150 × 150; (g) random brightness enhancement; (h) random contrast enhancement.

Figure 4

Number of articles published from 2015 to 2022.

Figure 5

Usage of preprocessing techniques from 2017 to 2022.

Figure 6

Usage of state-of-the-art CNN models from 2015 and 2022.

Figure 7

Classification accuracy by publication year.

Figure 8

Classification accuracy by classification task.

Figure 9

Classification accuracy by number of patients.

Figure 10

Classification accuracy by number of images.

Figure 11

Classification accuracy by CNN architecture.

Figure 12

Classification accuracy by number of preprocessing operations.

Figure 13

Classification accuracy by number of data augmentation operations.

Figure 14

Histogram (left scale) and cumulative distribution (right scale) of factors not fully reported or considered in the studies reported in Table 4.