Multimodal Brain Image Analysis and Survival Prediction Using Neuromorphic Attention-Based Neural Networks

Han, Il Song

doi:10.1007/978-3-030-72084-1_18

Cited by 6 publications

(12 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As shown in Fig. 7, the authors [29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46] employed the following pre-processing methods to account for intensity inhomogeneity throughout the dataset.…”

Section: Pre-processingmentioning

confidence: 99%

“…Han et al [46] used a neuromorphic inspired pre-processing technique for tumour segmentation. It followed the concept of attention-based learner combined with Convolutional Neural Network (CNN) to generate saliency maps which aided in better image analysis.…”

Section: Neuromorphic Map Generationmentioning

confidence: 99%

“…Han et al [46] developed the neuromorphic saliency map to improve image analysis. An attention and saliency map was created to enhance tumour segmentation by simulating the visual brain and applying neuromorphic pre-processing.…”

Section: Feature Extraction and Selectionmentioning

confidence: 99%

“…Han et al [46] created a feedforward neural network (FFNN) with four inputs for OS prediction. The FFNN topology took three survival time outputs in three categories: short, medium, and long.…”

Section: Deep Learning Modelsmentioning

confidence: 99%

“…The accuracy of the categorization of patients as long, short, and midsurvivors was initially tested [37,45]. Moreover, a pair-wise error analysis was performed for the regression model [40] between the predicted and actual OS (in days), with the use [33] RFR with five-fold cross-validation Miron et al [34] Extra trees with a depth of seven levels Chato et al [69] Simple NN Anand et al [32] RFR Parmar et al [30] RFC with multi-fold cross-validation Soltaninejad et al [40] RFC with 50 trees having a depth of 10, fivefold cross-validation Pang et al [43] RFR with fivefold cross-validation Ali et al [41] RFR with grid search Russo et al [45] GLM with Tweedie distribution Akbar et al [42] MobileNetV1/MobileNetV2 with Droput 0.1 Marti Asenjo et al [49] DT (ensembled by RUSBoost method), SVM with quadratic kernel function, ensembled of regression trees (Matlab ML models) Carmo et al [31] MultiATTUNet Han et al [46] FFNN Pei et al [39] RFR with grid search Suter et al [78] ARD, AdaBoost, DT, Extra Tree(s), Gaussian processes, linear, MLP, Nearest Neighbors, passive-aggressive, radius neighbours, RANSAC, RF, stochastic gradient descent, SVR and Theil-Sen regression Zhao et al [38] 3D ResNet50 Patel et al [35] Cox proportional hazards model Dai et al [36] Linear regression McKinley et al [37] Linear regression model and the RF classification model Zhang et al [95] Linear classification, linear regression models of metrics like mean square error (MSE), median square error (median SE), standard deviation of square errors (std SE), and Spearman correlation coefficient (Spearman R) [45]. The standard evaluation framework for tumour survival prediction were based on these metrics: Accuracy: The prediction performance was evaluated using the classification accuracy (i.e., the number of adequately categorised cases) as indicated in Eq.…”

Section: Hd95mentioning

confidence: 99%

See 4 more Smart Citations

State-of-the-art techniques using pre-operative brain MRI scans for survival prediction of glioblastoma multiforme patients and future research directions

Kaur

Rana

Arora

2022

Clin Transl Imaging

View full text Add to dashboard Cite

Objective Glioblastoma multiforme (GBM) is a grade IV brain tumour with very low life expectancy. Physicians and oncologists urgently require automated techniques in clinics for brain tumour segmentation (BTS) and survival prediction (SP) of GBM patients to perform precise surgery followed by chemotherapy treatment. Methods This study aims at examining the recent methodologies developed using automated learning and radiomics to automate the process of SP. Automated techniques use pre-operative raw magnetic resonance imaging (MRI) scans and clinical data related to GBM patients. All SP methods submitted for the multimodal brain tumour segmentation (BraTS) challenge are examined to extract the generic workflow for SP. ResultsThe maximum accuracies achieved by 21 state-of-the-art different SP techniques reviewed in this study are 65.5 and 61.7% using the validation and testing subsets of the BraTS dataset, respectively. The comparisons based on segmentation architectures, SP models, training parameters and hardware configurations have been made. ConclusionThe limited accuracies achieved in the literature led us to review the various automated methodologies and evaluation metrics to find out the research gaps and other findings related to the survival prognosis of GBM patients so that these accuracies can be improved in future. Finally, the paper provides the most promising future research directions to improve the performance of automated SP techniques and increase their clinical relevance.

show abstract

Section: Pre-processingmentioning

confidence: 99%

Section: Neuromorphic Map Generationmentioning

confidence: 99%

Section: Feature Extraction and Selectionmentioning

confidence: 99%

Section: Deep Learning Modelsmentioning

confidence: 99%

Section: Hd95mentioning

confidence: 99%

See 3 more Smart Citations

State-of-the-art techniques using pre-operative brain MRI scans for survival prediction of glioblastoma multiforme patients and future research directions

Kaur

Rana

Arora

2022

Clin Transl Imaging

View full text Add to dashboard Cite

show abstract

Deep learning and machine learning‐based early survival predictions of glioblastoma patients using pre‐operative three‐dimensional brain magnetic resonance imaging modalities

Kaur

Rana

Arora

2022

Int J Imaging Syst Tech

View full text Add to dashboard Cite

To propose and implement an automated machine learning (ML) based methodology to predict the overall survival of glioblastoma multiforme (GBM) patients. In the proposed methodology, we used deep learning (DL) based 3D U‐shaped Convolutional Neural Network inspired encoder‐decoder architecture to segment the brain tumor. Further, feature extraction was performed on these segmented and raw magnetic resonance imaging (MRI) scans using a pre‐trained 2D residual neural network. The dimension‐reduced principal components were integrated with clinical data and the handcrafted features of tumor subregions to compare the performance of regression‐based automated ML techniques. Through the proposed methodology, we achieved the mean squared error (MSE) of 87 067.328, median squared error of 30 915.66, and a SpearmanR correlation of 0.326 for survival prediction (SP) with the validation set of Multimodal Brain Tumor Segmentation 2020 dataset. These results made the MSE far better than the existing automated techniques for the same patients. Automated SP of GBM patients is a crucial topic with its relevance in clinical use. The results proved that DL‐based feature extraction using 2D pre‐trained networks is better than many heavily trained 3D and 2D prediction models from scratch. The ensembled approach has produced better results than single models. The most crucial feature affecting GBM patients' survival is the patient's age, as per the feature importance plots presented in this work. The most critical MRI modality for SP of GBM patients is the T2 fluid attenuated inversion recovery, as evident from the feature importance plots.

show abstract

Extracting Radiomic features from pre-operative and segmented MRI scans improved survival prognosis of glioblastoma Multiforme patients through machine learning: a retrospective study

Kaur

Rana²,

Arora³

2022

Multimed Tools Appl

View full text Add to dashboard Cite

Multimodal Brain Image Analysis and Survival Prediction Using Neuromorphic Attention-Based Neural Networks

Cited by 6 publications

References 19 publications

State-of-the-art techniques using pre-operative brain MRI scans for survival prediction of glioblastoma multiforme patients and future research directions

State-of-the-art techniques using pre-operative brain MRI scans for survival prediction of glioblastoma multiforme patients and future research directions

Deep learning and machine learning‐based early survival predictions of glioblastoma patients using pre‐operative three‐dimensional brain magnetic resonance imaging modalities

Extracting Radiomic features from pre-operative and segmented MRI scans improved survival prognosis of glioblastoma Multiforme patients through machine learning: a retrospective study

Contact Info

Product

Resources

About