An Attentive Multi-Modal CNN for Brain Tumor Radiogenomic Classification

Qu, Ruyi; Xiao, Zhifeng

doi:10.3390/info13030124

Cited by 14 publications

(7 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Convolutional neural networks and recurrent neural networks (RNN) are different types of artificial neural networks that can perform representational learning on imaging data and provide different hierarchical feature representations at each network layer ( 28 ). It is precisely the stacking employment of multiple network layers with non-linear activation functions that make the feature representation complex and diverse.…”

Section: Methodsmentioning

confidence: 99%

Deep-learning and conventional radiomics to predict IDH genotyping status based on magnetic resonance imaging data in adult diffuse glioma

Zhang,

Fan,

Zhang

et al. 2023

Front. Oncol.

View full text Add to dashboard Cite

ObjectivesIn adult diffuse glioma, preoperative detection of isocitrate dehydrogenase (IDH) status helps clinicians develop surgical strategies and evaluate patient prognosis. Here, we aim to identify an optimal machine-learning model for prediction of IDH genotyping by combining deep-learning (DL) signatures and conventional radiomics (CR) features as model predictors.MethodsIn this study, a total of 486 patients with adult diffuse gliomas were retrospectively collected from our medical center (n=268) and the public database (TCGA, n=218). All included patients were randomly divided into the training and validation sets by using nested 10-fold cross-validation. A total of 6,736 CR features were extracted from four MRI modalities in each patient, namely T1WI, T1CE, T2WI, and FLAIR. The LASSO algorithm was performed for CR feature selection. In each MRI modality, we applied a CNN+LSTM–based neural network to extract DL features and integrate these features into a DL signature after the fully connected layer with sigmoid activation. Eight classic machine-learning models were analyzed and compared in terms of their prediction performance and stability in IDH genotyping by combining the LASSO–selected CR features and integrated DL signatures as model predictors. In the validation sets, the prediction performance was evaluated by using accuracy and the area under the curve (AUC) of the receiver operating characteristics, while the model stability was analyzed by using the relative standard deviation of the AUC (RSDAUC). Subgroup analyses of DL signatures and CR features were also individually conducted to explore their independent prediction values.ResultsLogistic regression (LR) achieved favorable prediction performance (AUC: 0.920 ± 0.043, accuracy: 0.843 ± 0.044), whereas support vector machine with the linear kernel (l-SVM) displayed low prediction performance (AUC: 0.812 ± 0.052, accuracy: 0.821 ± 0.050). With regard to stability, LR also showed high robustness against data perturbation (RSDAUC: 4.7%). Subgroup analyses showed that DL signatures outperformed CR features (DL, AUC: 0.915 ± 0.054, accuracy: 0.835 ± 0.061, RSDAUC: 5.9%; CR, AUC: 0.830 ± 0.066, accuracy: 0.771 ± 0.051, RSDAUC: 8.0%), while DL and DL+CR achieved similar prediction results.ConclusionIn IDH genotyping, LR is a promising machine-learning classification model. Compared with CR features, DL signatures exhibit markedly superior prediction values and discriminative capability.

show abstract

Section: Methodsmentioning

confidence: 99%

Deep-learning and conventional radiomics to predict IDH genotyping status based on magnetic resonance imaging data in adult diffuse glioma

Zhang,

Fan,

Zhang

et al. 2023

Front. Oncol.

View full text Add to dashboard Cite

show abstract

“…By utilizing this ensemble model, the authors achieved improved performance in segmenting brain tumors. In the domain of brain tumor radiogenomic classification, Qu and Xiao [26] proposed an attention-based multimodal CNN. The model integrates various techniques including multimodal feature aggregation, separable embedding, model-wise shortcut connections, and a lightweight attention mechanism, all designed to enhance overall performance.…”

Section: Related Workmentioning

confidence: 99%

Brain Tumor Detection Using 3D-UNet Segmentation Features and Hybrid Machine Learning Model

Mallampati,

Ishaq,

Rustam

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Machine learning has significantly improved disease diagnosis, enhancing the efficiency and accuracy of the healthcare system. One critical area where it proves beneficial is diagnosing brain tumors, a life-threatening disease, where early and accurate predictions can save lives. This study focuses on deploying a machine learning-based approach for brain tumor detection, utilizing Magnetic Resonance Imaging (MRI) features. We train the proposed model using 3D-UNet and 2D-UNet segmentation features extracted from MRI, encompassing shape, statistics, gray level size zone matrix, gray level dependence matrix, gray level co-occurrence matrix, and gray level run length matrix values. To improve performance, we propose a hybrid model that combines the strengths of two machine learning models, K-nearest neighbor (KNN) and gradient boosting classifier (GBC), using soft voting criteria. We combine them because, in cases where KNN exhibits poor performance for certain data points, GBC demonstrates significant performance, and vice versa, where GBC shows poor results, KNN performs significantly better. With 2D-UNet segmentation features, the model achieves a 64% accuracy. By training it on 3D-UNet segmentation features, we achieve a significant accuracy of 71% which surpasses existing state-of-the-art models that utilize 3D-UNet segmentation features.

show abstract

“…This work focuses on densnet121 and ensemble methods to classify brain tumors. [1]. In this work, a novel approach of categorizing brain tumour pictures from several modalities is put forward.…”

Section: Introductionmentioning

confidence: 99%

Brain Tumor Radiogenomic Classification Using Deep Learning

Manasa¹,

Jeevitha²,

Chandana³

et al. 2023

Int. J. Res. Publ. Rev.

View full text Add to dashboard Cite

A brain tumor is an abnormal cell formation within the mass of your brain. The cancer is a term used to represent malignant tumours. Many different types of brain tumours exist. Brain tumors may also be benign (noncancerous) or malignant (cancerous). Imaging techniques like Computed tomography (CT) or Magnetic Resonance Imaging (MRI) are commonly used to find brain cancerous areas. Other techniques for finding brain tumors include lumbar puncture, cerebral angiography, positron emission tomography, and molecular testing. A unique attentive deep-learning-based approach for recognizing malignant tumors is proposed in this research. Less than 10% of people with glioblastoma (GBM), survive for less than five years since it grows quickly. The model built predicts the genetic subtype of glioblastoma by detecting the presence of MGMT (O( 6)-methyl guanine-DNA methyl transferase) promoter methylation. Using MRI images, MGMT methylation prediction models were created. MRI scans will be used to see if MGMT activation in gliomas may be identified, we thoroughly investigated contemporary CNN architectures. By using these models, brain surgeries count can be decreased. The using a better model which has high level of accuracywill be used to detect brain tumour and doctors will be able to use it. By providing appropriate medical treatment, it has the potential to save many lives.

show abstract

An Attentive Multi-Modal CNN for Brain Tumor Radiogenomic Classification

Cited by 14 publications

References 37 publications

Deep-learning and conventional radiomics to predict IDH genotyping status based on magnetic resonance imaging data in adult diffuse glioma

Deep-learning and conventional radiomics to predict IDH genotyping status based on magnetic resonance imaging data in adult diffuse glioma

Brain Tumor Detection Using 3D-UNet Segmentation Features and Hybrid Machine Learning Model

Brain Tumor Radiogenomic Classification Using Deep Learning

Contact Info

Product

Resources

About