Identification Of Tumors Research Articles

Potential diagnostic, prognostic, and predictive biomarkers of gastric cancer.

Gastric cancer (GC), a malignant epithelial tumor, is the fourth leading cause of cancer-related death worldwide. Therapeutic strategies for GC, despite the biggest challenges, can significantly improve survival rates through early detection and effective screening methods. To provide brief information on the necessity of multiple specific diagnostic, prognostic, and predictive markers for GC. This review was conducted using a variety of search engines, including PubMed Central, Scopus, Web of Science, Google Scholar, and others. Some potential biomarkers that provide essential information include circulating tumor cells (CTCs), DNA methylation, claudin 18.2, fibroblast growth factor receptor 2 (FGFR2), long noncoding RNAs (lncRNAs), cell-free DNA (cfDNA), microRNAs, and serum pepsinogens. Multiple tumor markers are essential for screening, tumor identification, staging, prognostic assessment, and monitoring recurrence after therapy due to the absence of a single tumor indicator for diagnosing, prognosticating, and predicting GC.

SCG2 and CPE may be novel markers for the identification of pancreatic neuroendocrine tumors and solid pseudopapillary neoplasms.

Distinguishing pancreatic neuroendocrine tumors (pNETs) from solid pseudopapillary neoplasms (SPNs) is challenging, primarily due to their overlapping pathological characteristics. To address this, our study aims to identify and validate novel biomarkers that effectively differentiate between these two conditions. We focus on the exploration of new immunohistochemical markers to enhance this distinction. In this study, we analyzed genetic variations in pNETs and SPNs using the GSE43795 dataset from the Gene Expression Omnibus (GEO) database. Our approach was to identify genes with higher expression in pNETs compared to SPNs and normal pancreatic tissues. We conducted enrichment analyses to understand the functions of these genes. Furthermore, protein-protein interaction (PPI) network analysis was utilized to identify key genes associated with pNETs. Our sample consisted of 163 pancreatic tumor specimens, comprising 78 pNETs and 85 SPNs. We also collected clinicopathological data and used immunohistochemistry to measure the expression levels of these key genes. The enrichment analysis revealed that genes overexpressed in pNETs were mainly involved in signal release, vesicle transport, and ion pathway activation, playing significant roles in endocrine processes like insulin secretion, dopamine synapses, and circadian rhythm regulation. The PPI analysis identified secretogranin II (SCG2), carboxypeptidase E (CPE), and chromogranin A (CgA, CHGA) as key markers for differentiating pNETs from SPNs. Immunohistochemical validation of these markers demonstrated high sensitivity (SCG2: 98.7%, CPE: 97.4%) and specificity (100%), indicating their superior discriminative power compared to traditional markers like CgA, β-catenin, lymphoid enhancer-binding factor 1 (LEF1), and vimentin. Our study indicates that SCG2 and CPE are effective, novel immunohistochemical biomarkers for differentiating pNETs from SPNs.

Comparative Analysis of SVM and CNN Techniques for Brain Tumor Detection

A brain tumor is the most common disease on earth and it is harmful to people. Tumors are the uncontrolled growth of cells and tissues in the human brain called a tumor. The image is acquired using CT scans and Magnetic Resonance Images. The identification of tumors at an early stage is critical and challenging for researchers. A patient comes to the hospital when he starts suffering from pain, headache, omission etc and at that time, if he has a tumor, To recognize the tumor early stage it is very different to identify whether it is benign (non-cancerous) or malignant (cancerous), many techniques or methods are available for detection of tumor here we apply SVM algorithm and CNN on brain Magnetic Resonance Images for classification of a benign or malignant tumor. Here, we propose a system based on the new concept of simple tumor detection that uses feature extraction techniques, segmentation algorithm and classification. To identify similar patients who have or do not have a brain tumor, as well as to ascertain the type of tumor they have and their tumor sizes. By comparing both SVM & CNN which technique is more beneficial and which one is better in both? The performance of SVM classifiers is measured in terms of training effectiveness and classification accuracy. With 95% accuracy, it manages the process of brain tumor categorization in MRI scans. The efficacy of training and classification accuracy of the CNN classifier is compared (96.33%). Both methods get high accuracy but as compared to SVM, CNN provides more accuracy and consumes less time for execution.

A Review on: Identification of Brain Tumor in MRI Images by the Use of SVM and Fuzzy C-Means Combination

A Review on: Identification of Brain Tumor in MRI Images by the Use of SVM and Fuzzy C-Means Combination

A COMPUTER-AIDED BRAIN GLIOMA IDENTIFICATION USING MRI BASED TEXTURE ANALYSIS

Brain gliomas are deadly tumors that are discovered after a stressful, lengthy, and difficult procedure. Radiology is a broad and varied field that detects brain tumors, but interpreting radiological pictures of the brain needs advanced training, experience, and subject-matter expertise. The endeavor is challenging due to the wide variation in brain tumor tissues among individuals and similar cases within normal tissues. Magnetic resonance imaging (MRI)-based computer-aided biomedical image processing solves the challenges associated with brain tumor localization and identification while simultaneously addressing the shortage of qualified radiologists. This study proposes a computer-aided brain glioma identification (CABGD) model for brain glioma identification using the texture analysis of brain MRIs. The proposed model makes use of machine learning classifiers and brain MRI data. There were 200 MRIs of both normal and glioma brains in the experimental dataset. Firstly, the MRI dataset was pre-processed to crop the MRIs, size equalization, and gray-level conversion. Next, noises were removed by applying filters. Two ROIs of the sizes (10 ×10) were taken on each MRI after the tumor region was segmented. After extracting COM texture features from each ROI, thirty optimal features were obtained through a compound supervised feature selection, blend of Fisher (FSHR), probabilistic model of error (POE), average correlation (AVC), and mutual information (MUI). The optimal feature brain MRI dataset was classified into glioma and normal brain by applying machine learning classifiers named Bayas Net (BN), logistic model tree (LMT), and partial decision tree (PART); the classification results of the NB, LMT, and PART classifiers were 79.75%, 82.75, and 85%, respectively.

MRI-Based Brain Tumor Classification Using a Dilated Parallel Deep Convolutional Neural Network

Brain tumors are frequently classified with high accuracy using convolutional neural networks (CNNs) to better comprehend the spatial connections among pixels in complex pictures. Due to their tiny receptive fields, the majority of deep convolutional neural network (DCNN)-based techniques overfit and are unable to extract global context information from more significant regions. While dilated convolution retains data resolution at the output layer and increases the receptive field without adding computation, stacking several dilated convolutions has the drawback of producing a grid effect. This research suggests a dilated parallel deep convolutional neural network (PDCNN) architecture that preserves a wide receptive field in order to handle gridding artifacts and extract both coarse and fine features from the images. This article applies multiple preprocessing strategies to the input MRI images used to train the model. By contrasting various dilation rates, the global path uses a low dilation rate (2,1,1), while the local path uses a high dilation rate (4,2,1) for decremental even numbers to tackle gridding artifacts and to extract both coarse and fine features from the two parallel paths. Using three different types of MRI datasets, the suggested dilated PDCNN with the average ensemble method performs best. The accuracy achieved for the multiclass Kaggle dataset-III, Figshare dataset-II, and binary tumor identification dataset-I is 98.35%, 98.13%, and 98.67%, respectively. In comparison to state-of-the-art techniques, the suggested structure improves results by extracting both fine and coarse features, making it efficient.

Brain Tumor Detection and Classification Using Adjusted InceptionV3, AlexNet, VGG16, VGG19 with ResNet50-152 CNN Model

INTRODUCTION: Brain tumors have become a major global health concern, characterized by the abnormal growth of brain cells that can negatively affect surrounding tissues. These cells can either be malignant (cancerous) or benign (non-cancerous), with their impact varying based on their location, size and type. OBJECTIVE: Early detection and classification of brain tumors are challenging due to their complex and variable structural makeup. Accurate early diagnosis is crucial to minimize mortality rates. METHOD: To address this challenge, researchers proposed an optimized model based on Convolutional Neural Networks (CNNs) with transfer learning, utilizing architectures like Inception-V3, AlexNet, VGG16, and VGG19. This study evaluates the performance of these adjusted CNN models for brain tumor identification and classification using MRI data. The TCGA-LGG and The TCIA, two well-known open-source datasets, were employed to assess the model's performance. The optimized CNN architecture leveraged pre-trained weights from large image datasets through transfer learning. RESULTS: The refined ResNet50-152 model demonstrated impressive performance metrics: for the non-tumor class, it achieved a precision of 0.98, recall of 0.95, F1 score of 0.93, and accuracy of 0.94; for the tumor class, it achieved a precision of 0.87, recall of 0.92, F1 score of 0.88, and accuracy of 0.96. CONCLUSION: These results indicate that the refined CNN model significantly improves accuracy in classifying brain tumors from MRI scans, showcasing its potential for enhancing early diagnosis and treatment planning.

HGG-47. SPATIAL SINGLE-CELL PROTEOMIC ANALYSIS MAPS THE DISTINCT PEDIATRIC-TYPE DIFFUSE HIGH-GRADE GLIOMA TUMOR ENTITIES AND DEFINES THEIR TUMOR AND IMMUNE MICROENVIRONMENTS

Abstract Paediatric-type diffuse high-grade gliomas (pDHGG) are a group of molecularly and clinically defined tumor entities of the central nervous system. The investigation of the spatial organization of pDHGG tissues may provide new insights to decipher their heterogeneous nature and identify vulnerabilities in their aggressive behaviour. Here we exploit Imaging Mass Cytometry (IMC), a spatial biology approach enabling multiplex tissue phenotyping at single-cell resolution. We used 16 total formalin-fixed paraffin-embedded (FFPE) tissues from 4 patients including 1 Diffuse hemispheric glioma H3G34-mutant (DHG-H3G34), 1 Diffuse pediatric-type high-grade glioma H3-wildtype and IDH-wildtype (DHGG-WT) and 2 Diffuse midline glioma H3K27-altered (DMG-H3K27, 1 H3.3K27M and 1 H3.1K27M). Classical pathology assessment and DNA methylation profile were also performed. For IMC staining, 2 metal-tag custom-designed antibody panels (tumour and immune) of 31 antibodies were used. The panels include antibodies for differentiated and stem-cell markers, extracellular matrices, immune markers and H3K27M and H3.3G34R variants. An IMC analysis pipeline was developed: images are extracted in tiff format from the proprietary MCD files produced by Hyperion instrument, filtered to remove hot pixels via top-hat and OpenCV in Python, segmented using steinbock and analysed using the IMC Data Analysis toolbox in R, using docker. A total of 209 mm2 of tissue were analysed. IMC analysis showed that DHG-H3G34 displays a higher heterogeneity in terms of cell phenotypes compared to the other tumor entities. Moreover, immune infiltrate of T, B cells, NK and monocytes was observed in the DHG-H3G34 and DHGG-WT tissues, while the DMG-H3K27 had low/no immune infiltrate with the H3.3K27M showing higher PDL-1 expression. Analysis of longitudinal and multi-region samples highlighted trajectories of tumor evolution and identification of extracellular matrix niches in tumor core versus the infiltrative front. Our study provides novel insights into pDHGG complex biology by investigating their tissue architecture, spatial organization and cell-cell interactions.

The Extra-Extended Translabyrinthine Approach for Resection of Large Acoustic Neuroma: 2-Dimensional Operative Video.

The extended translabyrinthine approach to acoustic neuroma (AN) was created to allow improved visualization and access to larger tumors.1,2 The dural opening, however, remained confined to the presigmoid space. Other authors have introduced modifications to increase the dura exposure around the internal auditory canal (IAC).3-5 The extra-extended translabyrinthine approach was conceptualized by the senior author (CC) to maximize AN exposure and early cranial nerve identification. The tentorial peeling was added to allow extradural mobilization of the temporal lobe.6 This allows further safe bone removal around the IAC and petrous apex and consistent opening of the facial canal at IAC fundus. This modification creates 280-to-360-degree dura exposure at the IAC. The dural opening extends to the petrous apex superiorly and the prepontine arachnoid cistern inferiorly and includes resection of a tentorium dural flap created by the tentorial peeling.6 This exposure allows for near circumferential exposure of the tumor and early identification of the glossopharyngeal nerve in the cochlear aqueduct area, the trigeminal nerve at the porus trigeminal, and the facial nerve (FN) at IAC fundus. In addition, this ample exposure permits identification of the FN trajectory in the tumor capsule before any tumor dissection. We present a detailed video of extra-extended translabyrinthine approach technique in a patient with a large left AN (Hannover classification T4B).7 This video does not involve any human research projects not requiring Institutional Review Board/ethic committee approval. The patient consented to the procedure and to the publication of his image. Complete resection was obtained. The FN function was House-Brackman I/VI.

Adaptive Detection and Classification of Brain Tumour Images Based on Photoacoustic Imaging

A new imaging technique called photoacoustic imaging (PAI) combines the advantages of ultrasound imaging and optical absorption to provide structural and functional details of tissues. It has broad application prospects in the accurate diagnosis and treatment monitoring of brain tumours. However, the existing photoacoustic image classification algorithms cannot effectively distinguish benign tumours from malignant tumours. To address this problem, the YoLov8-MedSAM model is proposed in this research to provide precise and adaptable brain tumour identification and detection segmentation. Additionally, it employs convolutional neural networks (CNNs) to classify and identify tumours in order to distinguish between benign and malignant variations in PAI. The experimental results show that the method proposed in this study not only effectively detects and segments brain tumours of various shapes and sizes but also increases the accuracy of brain tumour classification to 97.02%. The method provides richer and more valuable diagnostic information to the clinic and effectively optimizes the diagnosis and treatment strategy of brain tumours.

Enhancing brain tumor detection in MRI with a rotation invariant Vision Transformer.

The Rotation Invariant Vision Transformer (RViT) is a novel deep learning model tailored for brain tumor classification using MRI scans. RViT incorporates rotated patch embeddings to enhance the accuracy of brain tumor identification. Evaluation on the Brain Tumor MRI Dataset from Kaggle demonstrates RViT's superior performance with sensitivity (1.0), specificity (0.975), F1-score (0.984), Matthew's Correlation Coefficient (MCC) (0.972), and an overall accuracy of 0.986. RViT outperforms the standard Vision Transformer model and several existing techniques, highlighting its efficacy in medical imaging. The study confirms that integrating rotational patch embeddings improves the model's capability to handle diverse orientations, a common challenge in tumor imaging. The specialized architecture and rotational invariance approach of RViT have the potential to enhance current methodologies for brain tumor detection and extend to other complex imaging tasks.

CNN-Based Segmentation and Detection of Brain Tumors MRI Images: A Review

In order to provide effective therapy and enhance patient outcomes, it is essential to diagnose brain tumors as early as possible and with high accuracy. Due to the fact that it provides very specific anatomical information, magnetic resonance imaging (MRI) is an extremely important tool for diagnosing brain tumors. The manual segmentation and identification of brain tumors from MRI images, on the other hand, would take a significant amount of time and are prone to human mistake. The use of convolutional neural networks, often known as CNNs, has become more popular as a resource for automating certain activities. The purpose of this review article is to investigate the current developments in CNN-based methods for the segmentation and identification of brain tumors in magnetic resonance imaging (MRI) images. We describe the most significant difficulties that are connected with the analysis of brain tumors using MRI, investigate the different CNN designs that are used for this purpose, and evaluate the performance metrics of these architectures. The purpose of this study is to offer a complete overview of the present state-of-the-art in CNN-based brain tumor analysis of MRI data, emphasizing both the promise and limits of this method.

An exceptional presentation of pituitary apoplexy in thyrotoxicosis

Pituitary apoplexy (PA) is a rare and potentially life-threatening condition characterized by severe headaches, vomiting, visual disturbances, ophthalmoplegia, altered mental status, and possible pan-hypopituitarism. A macroadenoma-induced pituitary haemorrhage is the primary cause. Various factors such as systemic hypertension, altered intracranial pressure, head trauma, anticoagulation, and pregnancy can trigger PA. A 35-year-old non-smoking male presented with headaches and an enlarged thyroid gland. Initial imaging revealed a haemorrhage in the sellar area, prompting a clinical diagnosis of pituitary apoplexy with T3 thyrotoxicosis. Despite the absence of visual impairment, conservative management was chosen, including anti-hypertensive therapy and follow-up MRI after three months. PA diagnosis can be challenging without prior identification of a pituitary tumour. Imaging modalities like MRI are crucial for diagnosis, and radiological follow-up is recommended. The relationship between PA and hyperthyroidism remains unclear, necessitating further investigation in patients presenting with PA symptoms.

A Metadata-Driven Approach to Malignant Tumor Identification in Dermoscopy Images based on Genetic-Inspired Crow Search Optimized Convolutional Multilayer Perceptron Network

A Metadata-Driven Approach to Malignant Tumor Identification in Dermoscopy Images based on Genetic-Inspired Crow Search Optimized Convolutional Multilayer Perceptron Network

18F-AlF-NOTA-octreotide PET/CT in the localization of tumor-induced osteomalacia: case series and literature review.

This study explores tumor-induced osteomalacia (TIO) through a case series and literature review, assessing the diagnostic potential of 18F-AlF-NOTA-octreotide (18F-OC) positron emission tomography/computed tomography (PET/CT). We analyzed TIO patients who underwent 18F-OC PET/CT. Parameters such as tumor dimension, the maximum standardized uptake value (SUVmax), the mean standardized uptake value (SUVmean) and metabolic tumor volume (MTV) were meticulously assessed. Clinical features and imaging characteristics pertinent to TIO were reviewed. 6 patients with clinical suspicion of TIO exhibited hypophosphatemia (0.25 to 0.64 mmol/L), elevated alkaline phosphatase (ALP) levels (142 to 506 U/L), and increased parathyroid hormone (PTH) levels (92.9 to 281.7 pg/mL). Of these patients, two underwent FGF-23 testing, with results of 3185.00 pg/ml and 17.56 pg/ml, respectively. Conventional imaging modalities depicted widespread osteoporosis, with several cases demonstrating fractures indicative of osteomalacic and associated pathological fractures. Subsequent 18F-OC PET/CT facilitated the accurate localization of causative tumors, with histopathological examination confirming the diagnosis of phosphaturic mesenchymal tumor (PMT). The interval from initial clinical presentation to definitive TIO diagnosis spanned approximately 2.5 years (range: 1 - 4 years), with tumors varying in size (maximum diameter: 7.8 to 40.0mm), SUVmax (5.47 to 25.69), SUVmean (3.43 to 7.26), and MTV (1.27 to 18.59 cm3). The implementation of whole-body 18F-OC PET/CT imaging emerges as a critical tool in the identification of occult tumors causing TIO. Future investigations incorporating a broader cohort are imperative to further delineate the diagnostic and therapeutic implications of 18F-OC PET/CT in managing TIO.

Prevalence of neurotrophic tropomyosin receptor kinase (NTRK) fusion gene positivity in patients with solid tumors in Japan.

Members of the neurotrophic tropomyosin receptor kinase (NTRK) gene family, NTRK1, NTRK2, and NTRK3 encode TRK receptor tyrosine kinases. Intra- or inter-chromosomal gene rearrangements produce NTRK gene fusions encoding fusion proteins which are oncogenic drivers in various solid tumors. This study investigated the prevalence of NTRK fusion genes and identified fusion partners in Japanese patients with solid tumors recorded in the Center for Cancer Genomics and Advanced Therapeutics database of comprehensive genomic profiling test. In the analysis population (n = 46,621), NTRK fusion genes were detected in 91 patients (0.20%). The rate was higher in pediatric cases (<18 years; 1.69%) than in adults (0.16%). NTRK gene fusions were identified in 21 different solid tumor types involving 38 different partner genes including 22 (57.9%) previously unreported NTRK gene fusions. The highest frequency of NTRK gene fusions was head and neck cancer (1.31%) and thyroid cancer (1.31%), followed by soft tissue sarcoma (STS; 0.91%). A total of 97 NTRK fusion gene partners were analyzed involving mainly NTRK1 (49.5%) or NTRK3 (44.2%) gene fusions. The only fusion gene detected in head and neck cancer was ETV6::NTRK3 (n = 22); in STS, ETV6::NTRK3 (n = 7) and LMNA::NTRK1 (n = 5) were common. Statistically significant mutual exclusivity of NTRK fusions with alterations was confirmed in TP53, KRAS, and APC. NTRK gene fusion was detected from 11 STS cases: seven unclassified sarcoma, three sarcoma NOS, and one Ewing sarcoma. NTRK gene fusion identification in solid tumors enables accurate diagnosis and potential TRK inhibitor therapy.

Diagnosis of the Initial Stage of Hepatocellular Carcinoma: A Review.

Hepatocellular carcinoma (HCC) is the third most common cause of cancer-related deaths worldwide. There may be more than a million instances of hepatocellular carcinoma by 2025, making it a persistent concern for global health. The most common form of hepatocellular carcinoma accounts for more than 90% of cases. There is no known cure for hepatocellular carcinoma, which is usually detected late in life. Unlike most other common malignancies, such as lung, prostate, and breast cancers, where mortality rates are declining, rates of death are rising by around 2-3% every year. It is extremely difficult to diagnose hepatocellular carcinoma in its early stages. Alpha-fetoprotein serology studies and ultrasonography (US) monitoring were historically the primary methods for early detection of hepatocellular cancer. However, the sensitivity or specificity of ultrasonography/alpha-fetoprotein (US/AFP) is not high enough to detect hepatocellular carcinoma in its early stages. Alpha-fetoprotein, or AFP, is an amino acid that is normally produced by the liver or yolk sac of an embryonic baby. In adults, AFP levels are typically modest. Adults with high levels of AFP have been associated with several illnesses, the most well-known of which are certain types of cancer. It is still possible to diagnose hepatocellular carcinoma early because of current technological advancements. We address the advancements in the diagnosis of hepatocellular carcinoma in this article, with a focus on new imaging techniques and diagnostic markers for early-stage tumor identification.

FCM and CBAC based Brain Tumor Identification and Segmentation

A brain tumor are an abnormal growth of cells within the brain, forming a mass that can be either cancerous (malignant) or non-cancerous (benign). Despite their differences, both types of tumors can pose serious health risks. As these tumors grow, they can increase intracranial pressure, leading to potential brain damage. This increased pressure can result in various symptoms such as headaches, seizures, vision problems, and changes in cognitive function. The potential for life-threatening consequences makes early detection and treatment crucial. The objective of the research is to develop a system or algorithm capable of accurately identifying the presence of brain tumors within medical imaging data (CT or MRI scans) and subsequently segmenting the tumor regions from the surrounding healthy brain tissue. This research aims at building an automated multi stage reliable system for classifying MRI images as tumor or non-tumor images. However, the research aims to diagnose brain tumor by extracting the tumor region accurately. The main contribution of this work is to automatically segment the tumor region from the MRI brain images, using Fuzzy C-Means (FCM) Clustering and the Content-Based Active Contour (CBAC) method. The CBAC method helps to resolve the issues of saddle points and broken edges in the extracted tumor region.

Prediction of Abnormal Growth of Breast Cancer Cell Using Support Vector Machine

Abstract: Breast cancer is a serious health concern that affects millions of women worldwide. Early detection of breast cancer is crucial for effective treatment and improved survival rates. Artificial intelligence (AI) has shown great promise in the field of medical imaging and has been increasingly used for breast cancer detection. Cancer is one of the most dangerous diseases to humans, and yet no permanent cure has been developed for it. Breast cancer is one of the most common cancer types. According to the National Breast Cancer foundation, in 2020 alone, more than 276,000 new cases of invasive breast cancer and more than 48,000 non-invasive cases were diagnosed in the US. To put these figures in perspective, 64% of these cases are diagnosed early in the disease’s cycle, giving patients a 99% chance of survival. In this paper we develop a system for diagnosis, prognosis and prediction of breast cancer using AI. This will assist the doctors in diagnosis of the disease. In this research work, we systematically reviewed previous work done on detection and treatment of breast cancer using genetic sequencing or histopathological imaging with the help of deep learning and machine learning We use of AI for breast cancer detection, including studies that have used mammography, ultrasound, and magnetic resonance imaging (MRI) as imaging modalities. AI can also improve the accuracy and efficiency of breast cancer screening and diagnosis, as well as the challenges that must be addressed to ensure the successful integration of AI into clinical practice. However, further research and development are needed to optimize AI algorithms and ensure their safety and effectiveness in clinical practice. Breast cancer detection has emerged as a critical area of research, leveraging the power of Artificial Intelligence (AI) to enhance diagnostic accuracy and efficiency. Utilizing AI technologies, particularly deep learning, enables the development of robust and precise models for early breast cancer detection. These models analyze mammographic images, identifying subtle patterns indicative of malignancies that may escape human eye detection. Transfer learning, employing pre-trained neural network architectures, has demonstrated significant success in leveraging existing knowledge for improved performance. The integration of convolutional neural networks (CNNs) and image processing techniques empowers AI models to discern subtle abnormalities, aiding in the timely identification of potential breast tumors. Additionally, AI-driven diagnostic tools can streamline radiologists' workflow, providing rapid and reliable assessments, ultimately contributing to early intervention and improved patient outcomes. As these technologies continue to evolve, the synergy between AI and breast cancer detection promises to revolutionize screening methods, fostering a paradigm shift towards more accurate, efficient, and accessible diagnostic practices in the realm of breast health.

A landmark-supervised registration framework for multi-phase CT images with cross-distillation

Objective. Multi-phase computed tomography (CT) has become a leading modality for identifying hepatic tumors. Nevertheless, the presence of misalignment in the images of different phases poses a challenge in accurately identifying and analyzing the patient’s anatomy. Conventional registration methods typically concentrate on either intensity-based features or landmark-based features in isolation, so imposing limitations on the accuracy of the registration process. Method. We establish a nonrigid cycle-registration network that leverages semi-supervised learning techniques, wherein a point distance term based on Euclidean distance between registered landmark points is introduced into the loss function. Additionally, a cross-distillation strategy is proposed in network training to further improve registration performance which incorporates response-based knowledge concerning the distances between feature points. Results. We conducted experiments using multi-centered liver CT datasets to evaluate the performance of the proposed method. The results demonstrate that our method outperforms baseline methods in terms of target registration error. Additionally, Dice scores of the warped tumor masks were calculated. Our method consistently achieved the highest scores among all the comparing methods. Specifically, it achieved scores of 82.9% and 82.5% in the hepatocellular carcinoma and the intrahepatic cholangiocarcinoma dataset, respectively. Significance. The superior registration performance indicates its potential to serve as an important tool in hepatic tumor identification and analysis.