Embedding MRI information into MRSI data source extraction improves brain tumour delineation in animal models

Ortega‐Martorell, Sandra; Candiota, Ana Paula; Thomson, Ryan; Riley, Patrick; Julià‐Sapé, Margarida; Olier, Iván

doi:10.1371/journal.pone.0220809

Cited by 4 publications

(4 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Para el desarrollo de nuestra investigación se optará por esta última opción, para lo cual se cuenta con las bases de datos generada a partir de los estudios previos realizados por (Ortega-Martorell et al, 2019) en los años 2012, 2013 y 2016, reseñados en la bibliografía de esta propuesta.…”

Section: Alcance Y Limitacionesunclassified

Brain Tumor Segmentation and Classification using hybrid Deep CNN with LuNet Classifier

Balamurugan¹,

Gnanamanoharan

2022

Preprint

View full text Add to dashboard Cite

The early diagnosis of brain tumors is critical to enhancing patient survival and prospects. The magnetic resonance imaging (MRI) brain tumor images must be physically analyzed in this work. As a result, computerized approaches for more precise tumor diagnostics are required. However, evaluating shape, volume, borders, Tumor detection, size, segmentation, and classification remains challenging. This work proposes a hybrid Deep Convolutional Neural Network (DCNN) classifier using an enhanced LuNet classifier algorithm in this proposed work. The primary intention of this work is to determine the area of the tumor site and classify brain tumors as benign or malignant. Initially, we split the data using the extended LuNet algorithm. Grey-level co-occurrence matrix (GLCM) and VGG16 extracted the features, yielding 13 classification features. For pretreatment, the Laplacian of Gaussian filter (LOG) is used. Overall, the proposed approach tries to increase the performance of non-deep learning classifiers. Traditional classifiers are superior deep learning methods because they require fewer training data sets, have lower computing complexity, lower user costs, and are easier to use by people with less training experience. The proposed algorithm achieves a better accuracy rate of 99.7%. Compared to the existing algorithm, the proposed approach outperforms them.

show abstract

Section: Alcance Y Limitacionesunclassified

Brain Tumor Segmentation and Classification using hybrid Deep CNN with LuNet Classifier

Balamurugan¹,

Gnanamanoharan

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…One reason is that the PROV data model has strong scenario applicability. The other reason is that the PROV data model has relatively complete open-source codes [ 23 , 24 ].…”

Section: Bt and Rm Technologymentioning

confidence: 99%

Implementation of Trusted Traceability Query Using Blockchain and Deep Reinforcement Learning in Resource Management

Jiang

Lei

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

To better track the source of goods and maintain the quality of goods, the present work uses blockchain technology to establish a system for trusted traceability queries and information management. Primarily, the analysis is made on the shortcomings of the traceability system in the field of agricultural products at the present stage; the study is conducted on the application of the traceability system to blockchain technology, and a new model of agricultural product traceability system is established based on the blockchain technology. Then, a study is carried out on the task scheduling problem of resource clusters in cloud computing resource management. The present work expands the task model and uses the deep Q network algorithm in deep reinforcement learning to solve various optimization objectives preset in the task scheduling problem. Next, a resource management algorithm based on a deep Q network is proposed. Finally, the performance of the algorithm is analyzed from the aspects of parameters, structure, and task load. Experiments show that the algorithm is better than Shortest Job First (SJF), T e t r i s ∗ , Packer, and other classic task scheduling algorithms in different optimization objectives. In the traceability system test, the traceability accuracy is 99% for the constructed system in the first group of samples. In the second group, the traceability accuracy reaches 98% for the constructed system. In general, the traceability accuracy of the system proposed here is above 98% in 8 groups of experimental samples, and the traceability accuracy is close for each experimental group. The resource management approach of the traceability system constructed here provides some ideas for the application of reinforcement learning technology in the construction of traceability systems.

show abstract

“…Our groups have been working in the preclinical scenario in order to explore, interpret, analyse and validate the potential of the spectroscopic approaches, especially spectroscopic imaging (MRSI), in the early assessment of therapy responses [7][8][9]. Also, for several years now, machine learning (ML) and deep learning (DL) methods have been consistently shown to be able to help with tasks such as brain tumour detection [10][11][12], diagnosis and grading [13][14][15], classification [16][17][18][19], segmentation of the affected/tumour area [20][21][22] and therapy response prediction to distinguish post-treatment effects and tumour progression [23][24][25][26][27], among others. Hence, we firmly believe that ML and DL methods applied to such metabolomic datasets using whole pattern inputs may help to unravel its potential in a way that would not be possible with a single metabolite or ratio quantitation.…”

Section: Introductionmentioning

confidence: 99%

Tracking Therapy Response in Glioblastoma Using 1D Convolutional Neural Networks

Ortega-Martorell,

Olier,

Hernandez

et al. 2023

Cancers

Self Cite

View full text Add to dashboard Cite

Background: Glioblastoma (GB) is a malignant brain tumour that is challenging to treat, often relapsing even after aggressive therapy. Evaluating therapy response relies on magnetic resonance imaging (MRI) following the Response Assessment in Neuro-Oncology (RANO) criteria. However, early assessment is hindered by phenomena such as pseudoprogression and pseudoresponse. Magnetic resonance spectroscopy (MRS/MRSI) provides metabolomics information but is underutilised due to a lack of familiarity and standardisation. Methods: This study explores the potential of spectroscopic imaging (MRSI) in combination with several machine learning approaches, including one-dimensional convolutional neural networks (1D-CNNs), to improve therapy response assessment. Preclinical GB (GL261-bearing mice) were studied for method optimisation and validation. Results: The proposed 1D-CNN models successfully identify different regions of tumours sampled by MRSI, i.e., normal brain (N), control/unresponsive tumour (T), and tumour responding to treatment (R). Class activation maps using Grad-CAM enabled the study of the key areas relevant to the models, providing model explainability. The generated colour-coded maps showing the N, T and R regions were highly accurate (according to Dice scores) when compared against ground truth and outperformed our previous method. Conclusions: The proposed methodology may provide new and better opportunities for therapy response assessment, potentially providing earlier hints of tumour relapsing stages.

show abstract

Embedding MRI information into MRSI data source extraction improves brain tumour delineation in animal models

Cited by 4 publications

References 41 publications

Brain Tumor Segmentation and Classification using hybrid Deep CNN with LuNet Classifier

Brain Tumor Segmentation and Classification using hybrid Deep CNN with LuNet Classifier

Implementation of Trusted Traceability Query Using Blockchain and Deep Reinforcement Learning in Resource Management

Tracking Therapy Response in Glioblastoma Using 1D Convolutional Neural Networks

Contact Info

Product

Resources

About