Tumor Detection Research Articles

Using U-Net models in deep learning for brain tumor detection from MRI scans

Tumor diseases in the nervous system are both dangerous and complex. Magnetic Resonance Imaging (MRI) is crucial for detecting brain disease; however, identifying the presence of tumors from these is time-consuming and requires a professional doctor. Utilizing deep learning for tumor detection in MRI images can reduce waiting times and enhance detection accuracy. We propose a method employing two U-Net models: ResNeXt-50 and EfficientNet architectures, integrated with a Feature Pyramid Network (FPN) for segmenting brain tumor. The models were trained on the BraTS 2021 dataset, consisting of 3,929 MRI scan images with 3,929 corresponding masks, divided into training, testing, and evaluation sets in a 70:15:15 ratio. The results indicate that the hybrid model, which combines EfficientNet and FPN, delivers superior performance, with an average Intersection over Union (IoU) accuracy of 0.90 on the test set compared to 0.50 for ResNeXt-50, and Dice accuracy of 0.92 compared to 0.66 for ResNeXt-50. Furthermore, we developed a web application that implements the EfficientNet with FPN model, facilitating convenient tumor detection from uploaded MRI images for doctors.

Integrating Diffusion-weighted and Perfusion-weighted Imaging for Improved Grading and Characterization of Brain Tumors

Objective: This study aims to evaluate the role of Diffusion-Weighted Imaging (DWI) and Perfusion-Weighted Imaging (PWI) in the detection, characterization, and grading of brain tumors, focusing on their correlation with histopathological findings. Materials and Methods: A total of 40 patients with intracranial mass lesions underwent DWI, PWI, and conventional MRI using a 1.5 Tesla MRI scanner. Apparent Diffusion Coefficient (ADC) values were calculated for tumor core, peritumoral edema, and normal brain tissue. Cerebral Blood Volume (CBV), Cerebral Blood Flow (CBF), and Mean Transit Time (MTT) were analyzed using PWI. Data were statistically analyzed, with a p-value ≤ 0.05 considered significant. Results: The study included 38 neoplastic lesions and 2 tuberculomas. DWI showed distinct ADC values for different tumors, with meningiomas having a mean ADC of 0.84 ± 0.3 × 10⁻³ mm²/s and schwannomas a higher ADC of 2.14 × 10⁻³ mm²/s. PWI revealed elevated relative CBV in high-grade gliomas (mean: 2.1 ± 0.76), while lower rCBV was observed in metastases (mean: 0.5 ± 0.26). Tuberculomas exhibited intermediate perfusion characteristics. Conclusion: DWI and PWI, in conjunction with conventional MRI, enhance the diagnostic accuracy for brain tumors by providing valuable information on tumor cellularity and vascularity. These techniques aid in distinguishing between tumor types, grades, and associated edema, thus improving clinical decision-making and treatment planning.

Integrated miRNA Signatures: Advancing Breast Cancer Diagnosis and Prognosis.

Breast cancer ranks first in incidence and second in deaths worldwide, presenting alarmingly rising mortality rates. Imaging methodologies and/or invasive biopsies are routinely used for screening and detection, although not always with high sensitivity/specificity. MicroRNAs (miRNAs) could serve as diagnostic and prognostic biomarkers for breast cancer. We have designed a computational approach combining transcriptome profiling, survival analyses, and diagnostic power calculations from 1165 patients with breast invasive carcinoma from The Cancer Genome Atlas (TCGA-BRCA). Our strategy yielded two separate miRNA signatures consisting of four up-regulated and five down-regulated miRNAs in breast tumors, with cumulative diagnostic strength of AUC 0.93 and 0.92, respectively. We provide direct evidence regarding the breast cancer-specific expression of both signatures through a multicancer comparison of >7000 biopsies representing 19 solid cancer types, challenging their diagnostic potency beyond any of the current diagnostic methods. Our signatures are functionally implicated in cancer-related processes with statistically significant effects on overall survival and lymph-node invasion in breast cancer patients, which underlie their strong prognostic implication. Collectively, we propose two novel miRNA signatures with significantly elevated diagnostic and prognostic power as a functionally resolved tool for binary and accurate detection of breast cancer and other tumors of the female reproductive system.

Uterine fibroid metastasis as the initial manifestation of lobular breast cancer

The occurrence of breast cancer metastasis in a uterine leiomyoma is exceptional, and its diagnosis prior to detection of the primary tumor is even rarer. We describe at case of a 52-year-old female patient with a large fast growing uterine fibroid that caused pelvic pain and abdominal distention. A total abdominal hysterectomy with bilateral salpingo-oophorectomy was performed. Histopathology of the surgical specimen showed a characteristic pattern of invasive breast lobular carcinoma in the leiomyoma. A magnetic resonance image of the breast was then performed, which reveled primary and bilateral breast cancer. She subsequently underwent neoadjuvant endocrine therapy plus cyclin inhibitors, breast-conserving surgery, and radiotherapy. To date, the patient is alive and asymptomatic after 3 years of follow-up.

Robust brain MRI image classification with SIBOW-SVM

Primary Central Nervous System tumors in the brain are among the most aggressive diseases affecting humans. Early detection and classification of brain tumor types, whether benign or malignant, glial or non-glial, is critical for cancer prevention and treatment, ultimately improving human life expectancy. Magnetic Resonance Imaging (MRI) is the most effective technique for brain tumor detection, generating comprehensive brain scans. However, human examination can be error-prone and inefficient due to the complexity, size, and location variability of brain tumors. Recently, automated classification techniques using machine learning methods, such as Convolutional Neural Networks (CNNs), have demonstrated significantly higher accuracy than manual screening. However, deep learning-based image classification methods, including CNNs, face challenges in estimating class probabilities without proper model calibration (Guo et al., 2017; Minderer et al., 2021). In this paper, we propose a novel brain tumor image classification method called SIBOW-SVM, which integrates the Bag-of-Features model with SIFT feature extraction and weighted Support Vector Machines. This new approach can effectively extract hidden image features, enabling differentiation of various tumor types, provide accurate label predictions, and estimate probabilities of images belonging to each class, offering high-confidence classification decisions. We have also developed scalable and parallelable algorithms to facilitate the practical implementation of SIBOW-SVM for massive image datasets. To benchmark our method, we apply SIBOW-SVM to a public dataset of brain tumor MRI images containing four classes: glioma, meningioma, pituitary, and normal. Our results demonstrate that the new method outperforms state-of-the-art techniques, including CNNs, in terms of uncertainty quantification, classification accuracy, computational efficiency, and data robustness.

Enhanced MRI-based brain tumour classification with a novel Pix2pix generative adversarial network augmentation framework.

The scarcity of medical imaging datasets and privacy concerns pose significant challenges in artificial intelligence-based disease prediction. This poses major concerns to patient confidentiality as there are now tools capable of extracting patient information by merely analysing patient's imaging data. To address this, we propose the use of synthetic data generated by generative adversarial networks as a solution. Our study pioneers the utilisation of a novel Pix2Pix generative adversarial network model, specifically the 'image-to-image translation with conditional adversarial networks,' to generate synthetic datasets for brain tumour classification. We focus on classifying four tumour types: glioma, meningioma, pituitary and healthy. We introduce a novel conditional deep convolutional neural network architecture, developed from convolutional neural network architectures, to process the pre-processed generated synthetic datasets and the original datasets obtained from the Kaggle repository. Our evaluation metrics demonstrate the conditional deep convolutional neural network model's high performance with synthetic images, achieving an accuracy of 86%. Comparative analysis with state-of-the-art models such as Residual Network50, Visual Geometry Group 16, Visual Geometry Group 19 and InceptionV3 highlights the superior performance of our conditional deep convolutional neural network model in brain tumour detection, diagnosis and classification. Our findings underscore the efficacy of our novel Pix2Pix generative adversarial network augmentation technique in creating synthetic datasets for accurate brain tumour classification, offering a promising avenue for improved disease prediction and treatment planning.

Head-to-Head Comparison of [68Ga]Ga-NOTA-RM26 and [18F]FDG PET/CT in Patients with Gastrointestinal Stromal Tumors: A Prospective Study.

Gastrointestinal stromal tumors (GISTs) are the most common stromal tumors in the gastrointestinal tract. This study was designed to evaluate a gastrin-releasing peptide receptor antagonist PET tracer, [68Ga]Ga-NOTA-RM26, and compare it with [18F]FDG PET/CT in the assessment of patients with GISTs. Methods: With institutional review board approval and informed consent, 30 patients with suspected or proven GISTs based on abdominal CT or gastroscopy were recruited. All patients underwent [68Ga]Ga-NOTA-RM26 and [18F]FDG PET/CT scans. Pathology and other patient information were collected. Results: No radiopharmaceutical-related adverse events were observed in the patients. In total, 18 lesions in 16 patients were diagnosed as GIST, 3 patients were diagnosed with schwannoma, and 4 patients were diagnosed with leiomyoma. In 18 GISTs, the mean SUVmax of [68Ga]Ga-NOTA-RM26 PET was significantly higher than that of [18F]FDG PET (17.07 ± 19.57 vs. 2.28 ± 1.65; P < 0.01), and [68Ga]Ga-NOTA-RM26 PET/CT had a higher tumor detection rate than did [18F]FDG PET/CT (88.9% vs. 50%; P < 0.01). The uptake of [68Ga]Ga-NOTA-RM26 in GISTs was significantly higher than that in 2 other benign tumors (leiomyoma or schwannoma) (17.07 ± 19.57 vs. 4.23 ± 1.77; P = 0.014). With the SUVmax cutoff value of 6.0, the sensitivity of 68Ga-NOTA-RM26 PET/CT in diagnosing GISTs is 72% and the specificity is 85.7%. Conclusion: Compared with [18F]FDG PET/CT, [68Ga]Ga-NOTA-RM26 PET/CT is a promising and effective imaging modality for the detection of GISTs.

Correction: Superb microvascular imaging in tumour detection: global trends and research gaps

Correction: Superb microvascular imaging in tumour detection: global trends and research gaps

Head-to-head comparison of 68Ga-FAPI-46 PET/CT, 18F-FDG PET/CT, and contrast-enhanced CT for the detection of various tumors.

FAPI-PET/CT exhibits high tumor uptake and low background accumulation, enabling high-sensitivity tumor detection. We compared the diagnostic performance of 68Ga-FAPI-46 PET/CT plus contrast-enhanced CT (CE-CT), 18F-FDG PET/CT plus CE-CT, and standalone CE-CT in patients with various malignancies. 232 patients underwent 68Ga-FAPI-46 PET/CT,18F-FDG PET/CT, and CE-CT each within 4weeks. Detection rates were assessed by a blinded reader, with ≥ 2weeks between scans of the same patient to avoid recall bias. A sub-analysis of diagnostic performance was performed for 490 histopathologically validated lesions. Detection rates were compared using McNemar's test. Lesion-based detection rates in 68Ga-FAPI-46 PET/CT plus CE-CT, 18F-FDG PET/CT plus CE-CT, and CE-CT alone were 91.2% (1540/1688), 82.5% (1393/1688) and 60.2% (1016/1688). The detection rates were significantly higher for 68Ga-FAPI-46 PET/CT plus CE-CT than for 18F-FDG PET/CT plus CE-CT (p < 0.02 for primary lesions and p < 0.001 for total, abdominopelvic nodal, liver and other visceral lesions) and CE-CT (p < 0.0001 for total, primary, cervicothoracic nodal, abdominopelvic nodal, liver, other visceral, and bone lesions). In the sub-analysis, sensitivity, specificity, positive and negative predictive value, and accuracy were 61.3%, 96.7%, 81.4%, 91.4% and 90.0% for 68Ga-FAPI-46 PET/CT plus CE-CT, 57.0%, 95.7%, 75.7%, 90.5% and 88.4% for 18F-FDG PET/CT plus CE-CT, and 51.6%, 97.2%, 81.4%, 89.6% and 88.6% for CECT, respectively. 68Ga-FAPI-46 PET/CT plus CE-CT demonstrates a higher tumor detection rate than 18F-FDG PET/CT plus CE-CT and CE-CT in a diverse spectrum of malignancies, especially for primary, abdominopelvic nodal, liver, and other visceral lesions. Further studies on which entities draw particular benefit from 68Ga-FAPI-46 PET/CT are warranted to aid appropriate diagnostic workup. A total of N = 232 patients were analyzed. Of these, N = 50 patients were included in a prospective interventional trial (NCT05160051), and N = 175 in a prospective observational trial (NCT04571086) for correlation and clinical follow-up of PET findings; N = 7 patients were analyzed retrospectively.

Self-Supervised Learning for Generic Raman Spectrum Denoising.

Raman and surface-enhanced Raman scattering (SERS) spectroscopy are highly specific and sensitive optical modalities that have been extensively investigated in diverse applications. Noise reduction is demanding in the preprocessing procedure, especially for weak Raman/SERS spectra. Existing denoising methods require manual optimization of parameters, which is time-consuming and laborious and cannot always achieve satisfactory performance. Deep learning has been increasingly applied in Raman/SERS spectral denoising but usually requires massive training data, where the true labels may not exist. Aiming at these challenges, this work presents a generic Raman spectrum denoising algorithm with self-supervised learning for accurate, rapid, and robust noise reduction. A specialized network based on U-Net is established, which first extracts high-level features and then restores key peak profiles of the spectra. A subsampling strategy is proposed to refine the raw Raman spectrum and avoid the underlying biased interference. The effectiveness of the proposed approach has been validated by a broad range of spectral data, exhibiting its strong generalization ability. In the context of photosafe detection of deep-seated tumors, our method achieved signal-to-noise ratio enhancement by over 400%, which resulted in a significant increase in the limit of detection thickness from 10 to 18 cm. Our approach demonstrates superior denoising performance compared to the state-of-the-art denoising methods. The occlusion method further showed that the proposed algorithm automatically focuses on characterized peaks, enhancing the interpretability of our approach explicitly in Raman and SERS spectroscopy.

Application of small-sized magnetically controlled capsule gastroscopy in upper gastrointestinal diseases screening in asymptomatic individuals.

To explore the application of small-sized magnetically controlled capsule gastroscopy (MCCG) in upper gastrointestinal diseases screening in asymptomatic individuals. A retrospective analysis of the clinical data of 2163 asymptomatic individuals who underwent small-sized MCCG at our center from September 2022 to December 2023. The detection of submucosal tumors, polyps and ulcers in the upper gastrointestinal tract, the tolerance and safety of the subjects were statistically analyzed. Suspected submucosal tumors in the upper gastrointestinal tract were detected in 34 (1.57%) of 2136 subjects, with a higher incidence in females and no observed age difference. Polyps were detected in 328 subjects (15.16%), with a higher incidence in females and an increased detection rate with increasing age. Ulcers were detected in 27 subjects (1.25%), with a higher incidence in males and no observed age difference. There was no significant discomfort in all subjects, and no adverse event or capsule retention occurred. Small-sized MCCG can be used for focal lesion screening in the upper gastrointestinal tract and is comfortable and safe, making it a safe and efficient method for examining upper gastrointestinal diseases in the physical examination population.

Near-field metasurface sensor for an early-stage breast cancer detection

In this paper, a novel and highly sensitive metasurface sensor for microwave breast tumor detection is proposed. The proposed sensor array comprises 8 × 8 small-size sensor elements that are capable of reacting to changes in both electric and magnetic fields. This allows the sensor to detect even minor variations in the surrounding medium, resulting in improved sensitivity. Additionally, designing a sensor array improves sensitivity by covering all areas of the breast tissues with multiple small sensor elements. Numerical studies have been conducted to assess the sensor sensitivity using realistic healthy and non-healthy breast models with diagnosed tumors placed at different locations within healthy breast models in the CST simulation environment at varying stand-off distances. An experiment was conducted to validate the sensor’s concept. It involved testing a metasurface sensor with phantoms resembling both healthy female breast tissues and those with a 10 mm tumor. The results from simulations and experiments demonstrate that the metasurface sensor is capable of detecting breast tumors at different distances. For safety compliance, specific absorption rate (SAR) values were obtained through both simulation and experimentation. The simulated SAR values were calculated to be 0.357 W/kg and 0.216 W/kg at 1 g and 10 g, respectively, using 17 dBm for safety. The measured SAR values were 0.101 W/kg and 0.1 W/kg at 1 g and 10 g, respectively.

Detection of brain tumor using Hybridized 3D U-Net model on MRI images

Detection of brain tumor using Hybridized 3D U-Net model on MRI images

YOLO v10-Based Brain Tumor Detection: An Innovative Approach in CT Imaging

YOLO v10-Based Brain Tumor Detection: An Innovative Approach in CT Imaging

Foundational Segmentation Models and Clinical Data Mining Enable Accurate Computer Vision for Lung Cancer.

This study aims to assess the effectiveness of integrating Segment Anything Model (SAM) and its variant MedSAM into the automated mining, object detection, and segmentation (MODS) methodology for developing robust lung cancer detection and segmentation models without post hoc labeling of training images. In a retrospective analysis, 10,000 chest computed tomography scans from patients with lung cancer were mined. Line measurement annotations were converted to bounding boxes, excluding boxes < 1cm or > 7cm. The You Only Look Once object detection architecture was used for teacher-student learning to label unannotated lesions on the training images. Subsequently, a final tumor detection model was trained and employed with SAM and MedSAM for tumor segmentation. Model performance was assessed on a manually annotated test dataset, with additional evaluations conducted on an external lung cancer dataset before and after detection model fine-tuning. Bootstrap resampling was used to calculate 95% confidence intervals. Data mining yielded 10,789 line annotations, resulting in 5403 training boxes. The baseline detection model achieved an internal F1 score of 0.847, improving to 0.860 after self-labeling. Tumor segmentation using the final detection model attained internal Dice similarity coefficients (DSCs) of 0.842 (SAM) and 0.822 (MedSAM). After fine-tuning, external validation showed an F1 of 0.832 and DSCs of 0.802 (SAM) and 0.804 (MedSAM). Integrating foundational segmentation models into the MODS framework results in high-performing lung cancer detection and segmentation models using only mined clinical data. Both SAM and MedSAM hold promise as foundational segmentation models for radiology images.

Mouse Models of HER2+ Human Breast Cancer: A Literature Review

Introduction: HER2+ breast cancer (BC) makes up 15-20% of BC cases, where HER2 overexpression drives BC progression via proliferative signals. Targeted therapies inhibit HER2's activity but face challenges like resistance and metastasis. Thus, HER2+ BC is a developing research area where many preclinical studies are being conducted, and mice are often used due to their genetic and physiological similarities to humans. This study aims to review and compare existing mouse models to determine the best model of HER2+ BC. Methods: A literature search on PubMed in February 2024 using search terms including HER2, neu, neuNT, MMTV, and mammary tumors identified 16 primary research articles that used Neu-overexpressing mice. The papers were categorized under five models: MMTV-Neu, MMTV-NeuNT, FloxNeo-NeuNT, MMTV-NIC, and MMTV-HER2. Results: Among examined studies, Andrechek et al.'s FloxNeo-NeuNT model had the longest average tumor onset at 447 days, while Kuang et al.'s MMTV-NIC had the shortest at 126 days. Ursini-Segal et al.'s MMTV-NIC mice showed 100% mammary tumor incidence. Most models resulted in mammary adenocarcinomas, with lung metastases commonly observed, especially in papers that used MMTV-NIC. Discussion: Human HER2+ BC is commonly adenocarcinoma and often metastasizes to lungs, bones, liver, and brain. The mouse models were also found to be adenocarcinomas, but they primarily showed lung metastasis, possibly due to insufficient tumor detection methods for brain and bone metastasis. Neu copy numbers in Andrechek et al.’s FloxNeo-NeuNT model also align with findings on HER2 copy number amplification in human HER2+ BC. Conclusion: We have found that MMTV-NIC is the optimal mouse model for HER2+ BC research due to its short latency, 100% tumor incidence, and high rate of lung metastasis.

Tumor-induced phosphopenic osteomalacia: modern approaches to diagnostics and treatment

Phosphopenic osteomalacia (PPOM) is a rare variant of paraneoplastic syndrome caused by tumor synthesis of fibroblast growth factor 23 (FGF23). FGF23 secretion leads to a decrease in phosphate reabsorption and calcitriol levels, which leads to the development of severe hypophosphataemia and hypocalcaemia. FGF23 synthesis is predominantly associated with benign mesenchymal tumors, but has also been described in malignant neoplasms. The main clinical manifestations of PPOM are generalized myalgias and myopathy, ostealgia, pathological fractures, etc. The diagnosis of the disease requires a step-by-step investigation using somatostatin receptor-based imaging techniques, as these have the highest sensitivity for the detection of neoplasms causing osteomalacia. Surgical intervention is clearly the treatment of choice. Promising non-surgical methods include treatment with burosumab and somatostatin analogues.

Brain Tumor Detection through Image Fusion Using Cross Guided Filter and Convolutional Neural Network

This Data fusion has become a significant issue in diagnostic imaging, particularly in medical applications like radiation and guided image surgery. Medical image fusion aims to enhance the precision of tumor diagnosis, by preserving the salient information and characteristics of the original images in the fused image. It has been shown that guided filters are capable of maintaining edges well. In this paper, we propose a novel cross-guided filter-based fusion approach for multimodal medical images utilizing convolutional neural networks. The cross-guided filter is used in the proposed algorithm to extract the detailed features from the source images. Convolutional neural networks are used to generate the feature weights of source images derived from the detail layers. The weighted average rule is used to merge the source images based on these weights. We used thirty distinct types of medical images from diverse sources to compare the effectiveness of the proposed strategy to that of existing methods, both numerically and visually. The experimental findings demonstrated that, in terms of both objective evaluation and qualitative image quality, the suggested system performs better than other standard methods already in use. The quantitative results show that compared to existing methods under consideration for comparison, the proposed algorithm improves mutual information by 25%, image entropy by 9.5%, spatial frequency by 21%, standard deviation by 18.1%, structural similarity index by 30%, and edge strength of the fused image by 39%.

Tumor detection in breast cancer pathology patches using a Multi-scale Multi-head Self-attention Ensemble Network on Whole Slide Images

Breast cancer (BC) is the most common type of cancer among women globally and is one of the leading causes of cancer-related deaths among women. In the diagnosis of BC, histopathological assessment is the gold standard, where automated tumor detection technologies play a pivotal role. Utilizing Convolutional Neural Networks (CNNs) for automated analysis of image patches from Whole Slide Images (WSIs) enhances detection accuracy and alleviates the workload of pathologists. However, CNNs often face limitations in handling pathological patches due to a lack of sufficient contextual information and limited feature generation capabilities. To address this, we propose a novel Multi-scale Multi-head Self-attention Ensemble Network (MMSEN), which integrates a multi-scale feature generation module, a convolutional self-attention module, and an adaptive feature integration with an output module, effectively optimizing the performance of classical CNNs. The design of MMSEN optimizes the capture of key information and the comprehensive integration of features in WSIs pathological patches, significantly enhancing the precision of tumor detection. Validation results from a five-fold cross-validation experiment on the PatchCamelyon (PCam) dataset demonstrate that MMSEN achieves a ROC-AUC of 99.01% ± 0.02%, an F1-score of 98.00% ± 0.08%, a Balanced Accuracy (B-Acc) of 98.00% ± 0.08%, and a Matthews Correlation Coefficient (MCC) of 96.00% ± 0.16% (p<0.05). These results demonstrate the effectiveness and potential of MMSEN in detecting tumors from pathological patches in WSIs for BC.

Enhancing Brain Tumor Detection Through Custom Convolutional Neural Networks and Interpretability-Driven Analysis

Brain tumor detection is crucial for effective treatment planning and improved patient outcomes. However, existing methods often face challenges, such as limited interpretability and class imbalance in medical-imaging data. This study presents a novel, custom Convolutional Neural Network (CNN) architecture, specifically designed to address these issues by incorporating interpretability techniques and strategies to mitigate class imbalance. We trained and evaluated four CNN models (proposed CNN, ResNetV2, DenseNet201, and VGG16) using a brain tumor MRI dataset, with oversampling techniques and class weighting employed during training. Our proposed CNN achieved an accuracy of 94.51%, outperforming other models in regard to precision, recall, and F1-Score. Furthermore, interpretability was enhanced through gradient-based attribution methods and saliency maps, providing valuable insights into the model’s decision-making process and fostering collaboration between AI systems and clinicians. This approach contributes a highly accurate and interpretable framework for brain tumor detection, with the potential to significantly enhance diagnostic accuracy and personalized treatment planning in neuro-oncology.