Deep learning-integrated MRI brain tumor analysis: feature extraction, segmentation, and Survival Prediction using Replicator and volumetric networks.

in Scientific reports by Deependra Rastogi, Prashant Johri, Massimo Donelli, Seifedine Kadry, Arfat Ahmad Khan, Giuseppe Espa, Paola Feraco, Jungeun Kim

TLDR

  • A deep learning approach was developed to segment brain tumors and predict survival rates in patients with gliomas, with promising results on the BRATS2020 benchmarks dataset.
  • The model's accuracy and reliability improved significantly through the use of 2D volumetric convolutional neural networks and a majority rule.
  • This study has significant implications for the diagnosis and treatment of gliomas and has the potential to improve patient outcomes.

Abstract

The most prevalent form of malignant tumors that originate in the brain are known as gliomas. In order to diagnose, treat, and identify risk factors, it is crucial to have precise and resilient segmentation of the tumors, along with an estimation of the patients' overall survival rate. Therefore, we have introduced a deep learning approach that employs a combination of MRI scans to accurately segment brain tumors and predict survival in patients with gliomas. To ensure strong and reliable tumor segmentation, we employ 2D volumetric convolution neural network architectures that utilize a majority rule. This method helps to significantly decrease model bias and improve performance. Additionally, in order to predict survival rates, we extract radiomic features from the tumor regions that have been segmented, and then use a Deep Learning Inspired 3D replicator neural network to identify the most effective features. The model presented in this study was successful in segmenting brain tumors and predicting the outcome of enhancing tumor and real enhancing tumor. The model was evaluated using the BRATS2020 benchmarks dataset, and the obtained results are quite satisfactory and promising.

Overview

  • The study aims to develop a deep learning approach for accurate segmentation of brain tumors and predicting survival rates in patients with gliomas.
  • The methodology employs 2D volumetric convolutional neural networks and a majority rule to ensure strong and reliable tumor segmentation.
  • To predict survival rates, the study extracts radiomic features from the tumor regions and uses a Deep Learning Inspired 3D replicator neural network to identify the most effective features.

Comparative Analysis & Findings

  • The model was successful in segmenting brain tumors, especially enhancing tumors.
  • The study used the BRATS2020 benchmarks dataset to evaluate the model, and the obtained results were satisfactory and promising.
  • The model's performance improved significantly due to the use of 2D volumetric convolutional neural networks and the majority rule, which reduced model bias and improved performance.

Implications and Future Directions

  • The study's findings have significant implications for the diagnosis, treatment, and identification of risk factors for gliomas.
  • Future research can focus on combining this approach with other modalities, such as PET scans, to improve the accuracy of tumor segmentation and survival rate prediction.
  • The study's results also highlight the potential benefits of using radiomic features and 3D replicator neural networks for predicting survival rates in cancer patients.
Read Full Article