Tumor Segmentation Research Articles

A Radiograph Dataset for the Classification, Localization, and Segmentation of Primary Bone Tumors

A Radiograph Dataset for the Classification, Localization, and Segmentation of Primary Bone Tumors

Synthetic [18F]FET-PET Generation and Hotspot Prediction Via Preoperative MRI Fusion of Glioma Lacking Radiographic High-Grade Characteristics

Abstract Background Limited amino acid availability for Positron Emission Tomography (PET) imaging hinders therapeutic decision-making for gliomas without typical high-grade imaging features. To address this gap, we evaluated a generative artificial intelligence (AI) approach for creating synthetic [18F]FET-PET and predicting high [18F]FET-uptake from magnetic resonance imaging (MRI). Methods We trained a deep learning (DL)-based model to segment tumors in MRI and extracted radiomic features using the pyradiomics package and classification with a Random Forest classifier. To generate [18F]FET-PET images, we employed a Generative Adversarial Network (GAN) framework and utilized a split-input fusion module for processing different MRI sequences through feature extraction, concatenation, and self-attention. Results We included MR and PET images from 215 studies for the hotspot classification and 211 studies for the synthetic PET generation task. The top-performing radiomic features achieved 80% accuracy for hotspot prediction. From the synthetic [18F]FET-PET, 85% were classified as clinically useful by senior physicians. Peak Signal-to-Noise Ratio analysis indicated high signal fidelity with a peak at 40 dB, while Structural Similarity Index values showed structural congruence. Root Mean Square Error analysis demonstrated lower values below 5.6. Most Visual Information Fidelity scores ranged between 0.6 and 0.7. This indicates that synthetic PET images retain the essential information required for clinical assessment and diagnosis. Conclusion For the first time, we demonstrate that predicting high [18F]FET-uptake and generating synthetic PET images from preoperative MRI in LGG and HGG is feasible. Advanced MRI modalities and other generative AI models will be used to improve the algorithm further in future studies.

Mapping Pituitary Neuroendocrine Tumors: An Annotated MRI Dataset Profiling Tumor and Carotid Characteristics

Pituitary neuroendocrine tumors remain one of the most common intracranial tumors. While radiomic research related to pituitary tumors is progressing, public data sets for external validation remain scarce. We introduce an open dataset comprising high-resolution T1 contrast-enhanced MR scans of 136 patients with pituitary tumors, annotated for tumor segmentation and accompanied by clinical, radiological and pathological metadata. This diverse dataset captures variations in tumor size, location, and pathological activity, essential for understanding this complex condition. Expert annotations of both the tumor and adjacent carotid arteries ensure precise delineation, facilitating the development of automated segmentation algorithms. Our initiative addresses the need for standardized data in pituitary oncology, fosters collaboration and innovation, and enables the development and benchmarking of workflows that utilize pituitary radiomics for treatment planning and outcome prediction.

Direct Prediction of 48 Month Survival Status in Patients with Uveal Melanoma Using Deep Learning and Digital Cytopathology Images

Background: Uveal melanoma (UM) is the most common primary intraocular malignancy in adults. The median overall survival time for patients who develop metastasis is approximately one year. In this study, we aim to leverage deep learning (DL) techniques to analyze digital cytopathology images and directly predict the 48 month survival status on a patient level. Methods: Fine-needle aspiration biopsy (FNAB) of the tumor was performed in each patient diagnosed with UM. The cell aspirate was smeared on a glass slide and stained with H&E. Each slide then underwent whole-slide scanning. Within each whole-slide image, regions of interest (ROIs) with UM cells were automatically extracted. Each ROI was converted into super pixels, and the super pixels were automatically detected, segmented and annotated as “tumor cell” or “background” using DL. Cell-level features were extracted from the segmented tumor cells. The cell-level features were aggregated into slide-level features which were learned by a fully connected layer in an artificial neural network, and the patient survival status was predicted directly from the slide-level features. The data were partitioned at the patient level (78% training and 22% testing). Our DL model was trained to perform the binary prediction of yes-versus-no survival by Month 48. The ground truth for patient survival was established via a retrospective chart review. Results: A total of 74 patients were included in this study (43% female; mean age at the time of diagnosis: 61.8 ± 11.6 years), and 207,260 unique ROIs were generated for model training and testing. By Month 48 after diagnosis, 18 patients (24%) died from UM metastasis. Our hold-out test set contained 16 patients, where 6 patients had passed away and 10 patients were alive at Month 48. When using a sensitivity threshold of 80% in predicting UM-specific death by Month 48, our model achieved an overall accuracy of 75%. Within the subgroup of patients who died by Month 48, our model achieved a prediction accuracy of 83%. Of note, one patient in our test set was a clinical surprise, namely death by Month 48 despite having a GEP class 1A tumor, which typically portends a good prognosis. Our model correctly predicted this clinical surprise as well. Conclusions: Our DL model was able to predict the Month 48 survival status directly from digital cytopathology images obtained from FNABs of UM tumors with reasonably robust performance. This approach, if validated prospectively, could serve as an alternative survival prediction tool for patients with UM to whom GEP is not available.

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

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.

Deep learning MRI models for the differential diagnosis of tumefactive demyelination versus IDH-wildtype glioblastoma.

Diagnosis of tumefactive demyelination can be challenging. The diagnosis of indeterminate brain lesions on MRI often requires tissue confirmation via brain biopsy. Noninvasive methods for accurate diagnosis of tumor and non-tumor etiologies allows for tailored therapy, optimal tumor control, and a reduced risk of iatrogenic morbidity and mortality. Tumefactive demyelination has imaging features that mimic isocitrate dehydrogenase-wildtype glioblastoma (IDHwt GBM). We hypothesized that deep learning applied to postcontrast T1-weighted (T1C) and T2-weighted (T2) MRI images can discriminate tumefactive demyelination from IDHwt GBM. Patients with tumefactive demyelination (n=144) and IDHwt GBM (n=455) were identified by clinical registries. A 3D DenseNet121 architecture was used to develop models to differentiate tumefactive demyelination and IDHwt GBM using both T1C and T2 MRI images, as well as only T1C and only T2 images. A three-stage design was used: (i) model development and internal validation via five-fold cross validation using a sex-, age-, and MRI technology-matched set of tumefactive demyelination and IDHwt GBM, (ii) validation of model specificity on independent IDHwt GBM, and (iii) prospective validation on tumefactive demyelination and IDHwt GBM. Stratified AUCs were used to evaluate model performance stratified by sex, age at diagnosis, MRI scanner strength, and MRI acquisition. The deep learning model developed using both T1C and T2 images had a prospective validation area under the receiver operator characteristic curve (AUC) of 88% (95% CI: 0.82 - 0.95). In the prospective validation stage, a model score threshold of 0.28 resulted in 91% sensitivity of correctly classifying tumefactive demyelination and 80% specificity (correctly classifying IDHwt GBM). Stratified AUCs demonstrated that model performance may be improved if thresholds were chosen stratified by age and MRI acquisition. MRI images can provide the basis for applying deep learning models to aid in the differential diagnosis of brain lesions. Further validation is needed to evaluate how well the model generalizes across institutions, patient populations, and technology, and to evaluate optimal thresholds for classification. Next steps also should incorporate additional tumor etiologies such as CNS lymphoma and brain metastases. AUC = area under the receiver operator characteristic curve; CNS = central nervous system; CNSIDD = central nervous system inflammatory demyelinating disease; FeTS = federated tumor segmentation; GBM = glioblastoma; IDHwt = isocitrate dehydrogenase wildtype; IHC = immunohistochemistry; MOGAD = myelin oligodendrocyte glycoprotein antibody associated disorder; MS = multiple sclerosis; NMOSD = neuromyelitis optica spectrum disorder; wt = wildtype.

SVMVGGNet-16: A Novel Machine and Deep Learning Based Approaches for Lung Cancer Detection Using Combined SVM and VGGNet-16.

Lung cancer remains a leading cause of cancer-related mortality worldwide, necessitating early and accurate detection methods. Our study aims to enhance lung cancer detection by integrating VGGNet-16 form of Convolutional Neural Networks (CNNs) and Support Vector Machines (SVM) into a hybrid model (SVMVGGNet-16), leveraging the strengths of both models for high accuracy and reliability in classifying lung cancer types in different 4 classes such as adenocarcinoma (ADC), large cell carcinoma (LCC), Normal, and squamous cell carcinoma (SCC). Using the LIDC-IDRI dataset, we pre-processed images with a median filter and histogram equalization, segmented lung tumors through thresholding and edge detection, and extracted geometric features such as area, perimeter, eccentricity, compactness, and circularity. VGGNet-16 and SVM employed for feature extraction and classification, respectively. Performance matrices were evaluated using accuracy, AUC, recall, precision, and F1-score. Both VGGNet-16 and SVM underwent comparative analysis during the training, validation, and testing phases. The SVMVGGNet-16 model outperformed both, with a training accuracy (97.22%), AUC (99.42%), recall (94.22%), precision (95.28%), and F1- score (94.68%). In testing, our SVMVGGNet-16 model maintained high accuracy (96.72%), with an AUC (96.87%), recall (84.67%), precision (87.40%), and F1-score (85.73%). Our experimental results demonstrate the potential of SVMVGGNet-16 in improving diagnostic performance, leading to earlier detection and better treatment outcomes. Future work includes refining the model, expanding datasets, conducting clinical trials, and integrating the system into clinical practice to ensure practical usability.

Whole-body tumor segmentation from FDG-PET/CT: Leveraging a segmentation prior from tissue-wise projections.

Background: Accurate tumor detection and quantification are important for optimized therapy planning and evaluation. Total tumor burden is also an appealing biomarker for clinical trials. Manual examination and annotation of oncologic PET/CT is labor-intensive and demands a high level of expertise. One significant challenge is the risk for human error, leading to potential omission of especially small tumors and tumors with low FDG uptake. Purpose: In this study, we introduced an automated framework with segmentation prior, from a tissue-wise multi-channel multi-angled based approach, to enhance tumor segmentation in whole-body FDG-PET/CT. Method: The proposed framework utilized a segmentation prior generated from tumor segmentations in tissue-wise multi-channel projections of the standardized uptake value (SUV) from PET. Projections were created from various angles and the tissues were identified based on their CT Hounsfield values. The resulting segmentation masks were subsequently backprojected into a unified 3D volume for creation of the segmentation prior. Finally, the segmentation prior was provided as an additional input channel along with the CT and SUV images to three variants of 3D segmentation networks (3D UNet, dynUNet, nnUNet) to enhance the overall tumor segmentation performance. All the methods were independently evaluated using 5-fold cross-validation on the autoPET dataset and subsequently tested on the U-CAN dataset. Results: Combining the segmentation prior with the original SUV and CT images improved overall tumor segmentation performance significantly compared to a baseline network. The increase in Dice coefficient for lymphoma, lung cancer, and melanoma across different segmentation networks were: 3D UNet ( , , ), dynUNet ( , , ), and nnUNet ( , , ), respectively; *, p-value < 0.05; ns, non-significance. Conclusion: The increased segmentation accuracy could be attributed to the segmentation prior generated from tissue-wise SUV projections, revealing information from various tissues that was useful for segmentation of tumors. The results from this study highlight the potential of the proposed method as a valuable future tool for time-efficient quantification of tumor burden in oncologic FDG-PET/CT.

Medical Federated Model with Mixture of Personalized and Shared Components.

Although data-driven methods usually have noticeable performance on disease diagnosis and treatment, they are suspected of leakage of privacy due to collecting data for model training. Recently, federated learning provides a secure and trustable alternative to collaboratively train model without any exchange of medical data among multiple institutes. Therefore, it has draw much attention due to its natural merit on privacy protection. However, when heterogenous medical data exists between different hospitals, federated learning usually has to face with degradation of performance. In the paper, we propose a new personalized framework of federated learning to handle the problem. It successfully yields personalized models based on awareness of similarity between local data, and achieves better tradeoff between generalization and personalization than existing methods. After that, we further design a differentially sparse regularizer to improve communication efficiency during procedure of model training. Additionally, we propose an effective method to reduce the computational cost, which improves computation efficiency significantly. Furthermore, we collect five real medical datasets, including two public medical image datasets and three private multi-center clinical diagnosis datasets, and evaluate its performance by conducting nodule classification, tumor segmentation, and clinical risk prediction tasks. Comparing with 14 existing related methods, the proposed method successfully achieves the best model performance, and meanwhile up to 60% improvement of communication efficiency. Source code is public, and can be accessed at: https://github.com/ ApplicationTechnologyOfMedicalBigData/pFedNet-code.

Detection and segmentation of meningioma tumors using improved cloud empowered visual geometry group (cloud-ivgg) deep learning structure

Detection and segmentation of meningioma brain tumor is a complex process due to its similar textural pattern with other tumors. In this paper Meningioma Tumor Detection System (MTDS) approach is proposed to detect and classify the meningioma brain images from the healthy brain images. The training work flow of the proposed MTDS approach consists of Spatial Gabor Transform (SGT), feature computations and deep learning structure. The features are computed from the meningioma brain image dataset images and the normal brain image dataset images and these features are fed into the classification architecture. In this paper, the proposed CLOUD-IVGG architecture is derived from the existing Cloud empowered Visual Geometry Group (VGG) architecture to improve the detection rate of the proposed system and to decrease the computational time complexity. The testing work flow of the proposed system is also consist of SGT, feature computation and the CLOUD-IVGG architecture to produce the classification result of the source brain images into either normal or meningioma. Further, the tumor regions in this meningioma image have been located using the Morphological segmentation algorithm. In this research work, two independent resource brain imaging datasets has been involved to estimate and validate the performance efficiency of the proposed MTDS. The datasets are Kaggle Brain Imaging (KBI) and BRATS Imaging 2020 (BI20). The performance efficiency has been analyzed with respect to detection rate, precision, recall and Jaccard index

EF-VPT-Net: Enhanced Feature-Based Vision Patch Transformer Network for Accurate Brain Tumor Segmentation in Magnetic Resonance Imaging

EF-VPT-Net: Enhanced Feature-Based Vision Patch Transformer Network for Accurate Brain Tumor Segmentation in Magnetic Resonance Imaging

Federated learning for brain tumor segmentation and classification : A statistical approach

Enhancing the chances of survival for cancer patients requires early brain tumour identification. This study suggests a federated learning-based method to deal with classification problems caused by overfitting. With the use of the BRATS 2019 and 2020 datasets, the method successfully divides and categorises brain tumours. Data preprocessing involves min-max normalization, followed by CNN segmentation from MRI images. Extracted features undergo shape, texture, and Resnet-50-based feature extraction. These features are then inputted into an LSTM classifier, utilizing Federated Learning for improved classification. Findings indicate that the suggested FL utilizing the CNN-LSTM model outperforms other techniques like U-net and ensemble models, achieving high accuracy: 99.59% on BRATS 2019 and 99.61% on BRATS 2020 datasets.

Adaptive Fusion and Edge‐Oriented Enhancement for Brain Tumor Segmentation With Missing Modalities

Adaptive Fusion and Edge‐Oriented Enhancement for Brain Tumor Segmentation With Missing Modalities

Hierarchical Transformer with Multi-Scale Parallel Aggregation for Breast Tumor Segmentation

Hierarchical Transformer with Multi-Scale Parallel Aggregation for Breast Tumor Segmentation

Customized U-Net Model based Brain Tumor Segmentation in MRI Images and Ensemble based Tumor Classification Systems

Customized U-Net Model based Brain Tumor Segmentation in MRI Images and Ensemble based Tumor Classification Systems

Deep learning-based brain tumor segmentation: A comparison of U-Net and segNet algorithms

Brain tumors are among the diseases that pose a serious health concern worldwide and can lead to fatal outcomes if left untreated. The segmentation of brain tumors is a critical step for the accurate diagnosis of the disease and effective management of the treatment process. This study was conducted to examine the success rates of deep learning-based U-Net and SegNet algorithms in brain tumor segmentation. MRI brain images and black and white masks belonging to these images were used in the study. Image processing techniques, including histogram equalization, edge detection, noise reduction, contrast enhancement, and Gaussian blurring, were applied. These image processing steps improved the quality of the MRI images, contributing to more accurate segmentation results. As a result of the segmentation operations performed with U-Net and SegNet algorithms, the U-Net algorithm achieved an accuracy rate of 96%, while the SegNet algorithm’s accuracy rate was measured at 94%. The study determined that the U-Net algorithm provided a higher success rate and was more effective in brain tumor segmentation. In particular, the contribution of image processing steps to segmentation success was observed.

Classification of Brain Tumor based on Machine Learning Algorithms: A Review

Brain tumor classification using machine learning algorithms is pivotal for medical diagnostics, particularly in magnetic resonance imaging (MRI) analysis. This review provides a comprehensive overview of recent advancements in brain tumor classification methodologies, emphasizing preprocessing, feature extraction, and classification phases. Preprocessing steps involve noise reduction and intensity normalization, while feature extraction encompasses texture analysis and deep learning-based methods. Machine learning algorithms such as support vector machines (SVM) and convolutional neural networks (CNNs) are utilized for accurate classification based on extracted features. Recent literature highlights the significance of diverse datasets, hyperparameter tuning, and segmentation techniques in improving diagnostic accuracy. Noteworthy methodologies include deep learning models for glioma grading and novel optimization techniques for tumor segmentation. The review underscores interdisciplinary collaboration between medical professionals and computational scientists to refine existing methodologies and overcome challenges. A systematic search of academic databases, including PubMed, IEEE Xplore, ACM Digital Library, ScienceDirect, Springer and Google Scholar, was conducted. Continued evolution in brain tumor classification promises enhanced diagnostic accuracy and personalized treatment strategies, ultimately improving patient outcomes in neuro-oncology.

A Deep Learning Algorithm Improvement for Brain Tumour Segmentation Using Kernel-Based CNN and M-SVM

A Deep Learning Algorithm Improvement for Brain Tumour Segmentation Using Kernel-Based CNN and M-SVM

Comparative Analysis of Edge Detection Operators Using a Threshold Estimation Approach on Medical Noisy Images with Different Complexities

The manuscript conducts a comparative analysis to assess the impact of noise on medical images using a proposed threshold value estimation approach. It applies an innovative method for edge detection on images of varying complexity, considering different noise types and concentrations of noise. Five edges are evaluated on images with low, medium, and high detail levels. This study focuses on medical images from three distinct datasets: retinal images, brain tumor segmentation, and lung segmentation from CT scans. The importance of noise analysis is heightened in medical imaging, as noise can significantly obscure the critical features and potentially lead to misdiagnoses. Images are categorized based on the complexity, providing a multidimensional view of noise’s effect on edge detection. The algorithm utilized the grid search (GS) method and random search with nine values (RS9). The results demonstrate the effectiveness of the proposed approach, especially when using the Canny operator, across diverse noise types and intensities. Laplace operators are most affected by noise, yet significant improvements are observed with the new approach, particularly when using the grid search method. The obtained results are compared with the most popular techniques for edge detection using deep learning like AlexNet, ResNet, VGGNet, MobileNetv2, and Inceptionv3. The paper presents the results via graphs and edge images, along with a detailed analysis of each operator’s performance with noisy images using the proposed approach.

Advanced Hybrid Brain Tumor Segmentation in MRI: Elephant Herding Optimization Combined with Entropy-Guided Fuzzy Clustering

Accurate and early detection of brain tumors is essential for improving clinical outcomes and guiding effective treatment planning. Traditional segmentation techniques in MRI often struggle with challenges such as noise, intensity variations, and complex tumor morphologies, which can hinder their effectiveness in critical healthcare scenarios. This study proposes an innovative hybrid methodology that integrates advanced metaheuristic optimization and entropy-based fuzzy clustering to enhance segmentation precision in brain tumor detection. This method combines the nature-inspired Elephant Herding Optimization (EHO) algorithm with Entropy-Driven Fuzzy C-Means (EnFCM) clustering, offering significant improvements over conventional methods. EHO is utilized to optimize the clustering process, enhancing the algorithm’s ability to delineate tumor boundaries, while entropy-based fuzzy clustering accounts for intensity inhomogeneity and diverse tumor characteristics, promoting more consistent and reliable segmentation results. This approach was evaluated using the BraTS challenge dataset, a benchmark in the field of brain tumor segmentation. The results demonstrate marked improvements across several performance metrics, including Dice similarity, mean squared error (MSE), peak signal-to-noise ratio (PSNR), and the Tanimoto coefficient (TC), underscoring this method’s robustness and segmentation accuracy. By managing image noise and reducing computational demands, the EHO-EnFCM approach not only captures intricate tumor structures but also facilitates efficient image processing, making it suitable for real-time clinical applications. Overall, the findings reveal the potential of this hybrid approach to advance MRI-based tumor detection, offering a promising tool that enhances both accuracy and computational efficiency for medical imaging and diagnosis.