Medical Imaging Research Articles

Computed Tomography- Based Maxillary Sinus Morphometry: A Tool for Gender Differentiation in Libyan Population

Introduction: In forensic contexts, gender identification is a multifaceted process that involves examining various anatomical characteristics to determine an individual's gender. Recent advancements in medical imaging techniques, such as computed tomography (CT) scans have provided researchers and healthcare professionals with more detailed insights into anatomical differences between genders such as the dimensions of paranasal sinuses. Aim: This study aimed to compare the maxillary sinus (MS) dimensions between genders in a Libyan sample using CT scans. Materials and methods: One hundred consecutive Libyan patients, comprising 50 women and 50 men, were included in the study. The Width, height, and length were determined based on CT images. Additionally, the perimeter, area, and volume of the MS were calculated mathematically. The collected data analyzed to compare the MS parameters between genders in both sides. Statistical significance was accepted at p<0.05. Results: Statistically we found that the length, width and length of the MS were significantly different between male & female in both sides, while there was no significant difference in case of sinus area and volume. The accuracy level for correctly classifying gender based on width of right MS was 91.8%, demonstrating strong predictive power of this parameter. The result showed also that 94 % of males and 90% of females were sexed correctly with an overall accuracy of 92%. Conclusion: This study highlights the importance of understanding sinus morphology in forensic anthropology for gender determination.

Artificial Intelligence and Machine Learning in Hepatocellular Carcinoma Screening, Diagnosis and Treatment - A Comprehensive Systematic Review

Background: Medical image analysis plays a crucial role in the screening, monitoring, diagnosis, and prognosis of diseases. Hepatocellular Carcinoma (HCC) often remain asymptomatic in their early stages. Timely identification and intervention are crucial to prevent the progression to decompensated liver diseases and advanced-stage HCC, minimizing morbidity and mortality. Methodology: This study examined twenty relevant research articles that met the inclusion criteria. Utilizing the PRISMA criteria, a systematic search was conducted on PubMed/MEDLINE and Google Scholar for studies on HCC screening employing Artificial Intelligence (AI) and Machine Learning (ML). The search focused on the keywords "artificial intelligence" and "hepatocellular carcinoma," with inclusion criteria specifying studies in the English language published in and after 2020. Exclusions were made for histology, animal research, and investigations conducted before 2020. Titles and abstracts were thoroughly reviewed, and any discrepancies were discussed. Results: The comprehensive review reveals the transformative impact of AI and ML on HCC screening, diagnosis, and therapy. ML models demonstrated effectiveness in early HCC diagnosis, distinguishing hepatic lesions, predicting treatment responses, and assessing recurrence risks across various techniques. Integration of mass spectrometry-based technologies, advanced imaging, and real-world data significantly improved diagnostic accuracy and clinical decision support. AI models exhibited diagnostic expertise with potential applications in therapy suggestions, personalized surveillance, and prognostic evaluations. However, further validation and seamless integration into clinical practice are essential for realizing their full potential. Conclusion: This systematic study underscores the progress made in the application of AI and ML in HCC screening, diagnosis, and treatment. Numerous studies highlight the capability of AI and ML systems to enhance hepatocellular carcinoma diagnosis, predict outcomes, and optimize therapy. The enduring and versatile nature of these technologies points towards a revolutionary future in personalized and efficient HCC management.

High-Resolution X-ray Image from Copper-Based Perovskite Hybrid Polymer.

Cs3Cu2I5 nanocrystals (NCs) are considered to be promising materials due to their high photoluminescence efficiency, lack of lead toxicity, and X-ray responsiveness. However, during the crystallization process, NCs are prone to agglomeration and exhibit uneven size distribution, resulting in several light scattering that severely affect their imaging resolution. Herein, we successfully developed a high-resolution scintillator film by growing copper-based perovskite NCs within a hybrid polymer matrix. By leveraging the ingenious integration of polyvinylidene fluoride (PVDF) and polymethyl methacrylate (PMMA), the size and distribution uniformity of Cs3Cu2I5 NCs can be effectively controlled. Consequently, a high spatial resolution of 14.3 lp mm-1 and a low detection limit of 105 nGy s-1 are achieved, and the scintillator film has excellent flexibility and stability. These results highlight the promising application of Cs3Cu2I5 scintillator films in low-cost, flexible, and high-performance medical imaging.

Hi-gMISnet: generalized medical image segmentation using DWT based multilayer fusion and dual mode attention into high resolution pGAN

Objective. Automatic medical image segmentation is crucial for accurately isolating target tissue areas in the image from background tissues, facilitating precise diagnoses and procedures. While the proliferation of publicly available clinical datasets led to the development of deep learning-based medical image segmentation methods, a generalized, accurate, robust, and reliable approach across diverse imaging modalities remains elusive. Approach. This paper proposes a novel high-resolution parallel generative adversarial network (pGAN)-based generalized deep learning method for automatic segmentation of medical images from diverse imaging modalities. The proposed method showcases better performance and generalizability by incorporating novel components such as partial hybrid transfer learning, discrete wavelet transform (DWT)-based multilayer and multiresolution feature fusion in the encoder, and a dual mode attention gate in the decoder of the multi-resolution U-Net-based GAN. With multi-objective adversarial training loss functions including a unique reciprocal loss for enforcing cooperative learning in pGANs, it further enhances the robustness and accuracy of the segmentation map. Main results. Experimental evaluations conducted on nine diverse publicly available medical image segmentation datasets, including PhysioNet ICH, BUSI, CVC-ClinicDB, MoNuSeg, GLAS, ISIC-2018, DRIVE, Montgomery, and PROMISE12, demonstrate the proposed method’s superior performance. The proposed method achieves mean F1 scores of 79.53%, 88.68%, 82.50%, 93.25%, 90.40%, 94.19%, 81.65%, 98.48%, and 90.79%, respectively, on the above datasets, surpass state-of-the-art segmentation methods. Furthermore, our proposed method demonstrates robust multi-domain segmentation capabilities, exhibiting consistent and reliable performance. The assessment of the model’s proficiency in accurately identifying small details indicates that the high-resolution generalized medical image segmentation network (Hi-gMISnet) is more precise in segmenting even when the target area is very small. Significance. The proposed method provides robust and reliable segmentation performance on medical images, and thus it has the potential to be used in a clinical setting for the diagnosis of patients.

Boosting wisdom of the crowd for medical image annotation using training performance and task features.

A crucial bottleneck in medical artificial intelligence (AI) is high-quality labeled medical datasets. In this paper, we test a large variety of wisdom of the crowd algorithms to label medical images that were initially classified by individuals recruited through an app-based platform. Individuals classified skin lesions from the International Skin Lesion Challenge 2018 into 7 different categories. There was a large dispersion in the geographical location, experience, training, and performance of the recruited individuals. We tested several wisdom of the crowd algorithms of varying complexity from a simple unweighted average to more complex Bayesian models that account for individual patterns of errors. Using a switchboard analysis, we observe that the best-performing algorithms rely on selecting top performers, weighting decisions by training accuracy, and take into account the task environment. These algorithms far exceed expert performance. We conclude by discussing the implications of these approaches for the development of medical AI.

Application of CNN Model on Medical Images (Chest X-Rays)

Chest diseases encompass a range of conditions affecting the lungs, including infections caused by bacteria, viruses, or other microorganisms. These diseases can vary in severity, posing a potential threat to vulnerable individuals such as young children, older adults, and those with weakened immune systems. Detecting chest diseases involves employing diverse methods, such as physical examination, chest x-rays, blood tests, and sputum culture. During a physical examination, healthcare professionals utilize a stethoscope to listen for abnormal breathing sounds or crackling noises in the chest. They also assess factors like fever and respiratory rate. Chest x-rays provide a visual representation of the lungs, enabling the identification of areas displaying inflammation or fluid accumulation. Sputum culture involves collecting a sample of mucus expelled from the lungs through coughing and testing it for the presence of microorganisms. Key Words: Deep Learning, CNN, Chest X-Ray Images.

Duodenal-Jejunal Stenosis as a Manifestation of Eosinophilic Gastroenteritis: A Case Report

Introduction: Eosinophilic gastroenteritis (EGE) is a rare disorder characterized by eosinophilic infiltration of the stomach and small intestine. Its clinical manifestations depend on the extent, location, and depth of the inflammatory infiltrate. The pathogenesis remains unclear, but it may be related to food allergy or other hypersensitivity reactions in some patients. Untreated cases can cause intestinal stricture or perforation, making timely diagnosis and intervention essential. Clinical case: A 13-year-old male adolescent was hospitalized due to symptoms suggesting intestinal obstruction or alterations in intestinal motility. Medical imaging revealed significant gastric distension and delayed gastric emptying. An endoscopic study showed a “snake-skin” pattern, duodenitis due to food stasis, a 2 cm ulcer, and stenosis in the duodenum. The biopsy confirmed EGE with significant irregular eosinophilic infiltration of the lamina propria. A panendoscopy was performed after steroid treatment without being able to assess the site of stenosis and perform pneumatic dilation. After completing the immunomodulatory management, a contrast-enhanced abdominal tomography was performed showed a significant decrease in the intestinal lumen, therefore an exploratory laparotomy with a lateral to lateral jejune-jejunal anastomosis was performed. Discussion: EGE is a rare inflammatory disorder characterized by eosinophilic infiltration of the stomach and small intestine. Symptoms are based on the affected gastrointestinal tract segment. In this case, our patient presented intestinal obstruction symptoms with endoscopic biopsies and a histopathological study confirmed eosinophil infiltration. Treatment typically involves identifying potential triggers and corticosteroids to reduce inflammation and symptoms. However, in this case, a decrease in the intestinal lumen was observed, therefore an elective surgical resection was performed. Conclusion: Eosinophilic gastroenteritis is a rare disorder that requires strong clinical suspicion to be diagnosed, which is why its knowledge is important. In this case, gastrointestinal symptoms, radiological and endoscopic images, in addition to the histopathological report, were key tools for integrating the diagnosis. Although management with steroids and immunomodulators was not successful, surgical management became an appropriate option as definitive treatment.

Deep Learning Analysis of 3D Volume of Brain Anomaly Using MRI

In this study, we present a comprehensive investigation into the efficacy of two state-of-the-art deep learning architectures, 3D UNet and VNet, for brain tumor detection and segmentation in volumetric medical images.

Production of carbon-11 for PET preclinical imaging using a high-repetition rate laser-driven proton source

Most advanced medical imaging techniques, such as positron-emission tomography (PET), require tracers that are produced in conventional particle accelerators. This paper focuses on the evaluation of a potential alternative technology based on laser-driven ion acceleration for the production of radioisotopes for PET imaging. We report for the first time the use of a high-repetition rate, ultra-intense laser system for the production of carbon-11 in multi-shot operation. Proton bunches with energies up to 10–14 MeV were systematically accelerated in long series at pulse rates between 0.1 and 1 Hz using a PW-class laser. These protons were used to activate a boron target via the 11\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{11}$$\\end{document}B(p,n)11\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{11}$$\\end{document}C nuclear reaction. A peak activity of 234 kBq was obtained in multi-shot operation with laser pulses with an energy of 25 J. Significant carbon-11 production was also achieved for lower pulse energies. The experimental carbon-11 activities measured in this work are comparable to the levels required for preclinical PET, which would be feasible by operating at the repetition rate of current state-of-the-art technology (10 Hz). The scalability of next-generation laser-driven accelerators in terms of this parameter for sustained operation over time could increase these overall levels into the clinical PET range.

Early detection of glaucoma: feature visualization with a deep convolutional network

ABSTRACT Computer-aided diagnosis of ocular disorders like glaucoma can be effectively performed with retinal fundus images. Since the advent of machine learning and later deep learning techniques in medical image analysis, research focusing on automated and early detection of glaucoma has gained prominence. In this paper, we show how a deep Convolutional Neural Network model can be used in visualising the features related to Glaucoma that are consistent with (a subset of) those used by ophthalmologists for diagnosis. However, since the doctors do not necessarily depend entirely on these features but back them up with other important measurements such as visual field testing (HVF), any deep neural network such as CNN may not necessarily be expected to yield 100% accuracy. Further, other important parameters such as Intraocular Pressure (IOP), which may be built into a deep learning model, may not necessarily correlate well with the presence of Glaucoma. In this work, a deep CNN model has been trained and tested on a large number of high-quality fundus images, both normal and glaucomatous. We compare the results with transfer learning models such as VGG16, ResNet50, and MobileNetV2. On an average, we obtained an accuracy of 93.75% in identifying glaucoma, focusing only on the features of the fundus image.

Enhancing EfficientNetv2 with global and efficient channel attention mechanisms for accurate MRI-Based brain tumor classification

AbstractThe early and accurate diagnosis of brain tumors is critical for effective treatment planning, with Magnetic Resonance Imaging (MRI) serving as a key tool in the non-invasive examination of such conditions. Despite the advancements in Computer-Aided Diagnosis (CADx) systems powered by deep learning, the challenge of accurately classifying brain tumors from MRI scans persists due to the high variability of tumor appearances and the subtlety of early-stage manifestations. This work introduces a novel adaptation of the EfficientNetv2 architecture, enhanced with Global Attention Mechanism (GAM) and Efficient Channel Attention (ECA), aimed at overcoming these hurdles. This enhancement not only amplifies the model’s ability to focus on salient features within complex MRI images but also significantly improves the classification accuracy of brain tumors. Our approach distinguishes itself by meticulously integrating attention mechanisms that systematically enhance feature extraction, thereby achieving superior performance in detecting a broad spectrum of brain tumors. Demonstrated through extensive experiments on a large public dataset, our model achieves an exceptional high-test accuracy of 99.76%, setting a new benchmark in MRI-based brain tumor classification. Moreover, the incorporation of Grad-CAM visualization techniques sheds light on the model’s decision-making process, offering transparent and interpretable insights that are invaluable for clinical assessment. By addressing the limitations inherent in previous models, this study not only advances the field of medical imaging analysis but also highlights the pivotal role of attention mechanisms in enhancing the interpretability and accuracy of deep learning models for brain tumor diagnosis. This research sets the stage for advanced CADx systems, enhancing patient care and treatment outcomes.

Analytical Study of Medical PET Images by DWT and PCA

The PET medical images need clarification to reach a correct diagnosis of the disease by removing noise from them, and this is done by applying several techniques, including DWT, PCA. DWT is an algorithm that compresses an image and removes noise by dividing the image into several levels, removing the unnecessary (neglected) part of the image. The PCA algorithm compresses the error parameters in the image and then reconstructs them.

THE IMPORTANCE OF ULTRASONOGRAPHY IN THE DETECTION OF CONGENITAL HEART DISEASES IN CHILDREN

Over the last few years, remarkable progress has been made in the field of congenital heart diseases. Improvements considering diagnostic modalities, especially imaging, in surgical and interventional techniques, as well as in postoperative therapy and care, have contributed to a significant reduction in mortality and morbidity. One of the most important applications of medical imaging techniques in children is the detection and treatment of congenital heart anomalies. Objective of this article is to show the importance of ultrasonography in the detection of congenital heart diseases in children. The study was conducted on children with simple and complex congenital heart diseases and was conducted on the Pediatric Clinic, UKCS. The research is descriptive on a representative sample. In our study, 166 children were observed, of which 148 children (77 boys, 71 girls) with simple congenital heart diseases, and 18 children (8 boys, 10 girls) with complex congenital heart diseases. Out of the total number of observed children, 115 had a surgical correction, 97 children with simple congenital heart diseases (45 boys, 52 girls) and 18 children with complex congenital heart diseases (8 boys, 10 girls). The number of children monitored through the Cardiac Counseling Center who didn’t undergo surgical correction was 51, all with simple congenital heart diseases. Out of the total number of observed children who were frequently coming for follow-ups, 28 children had changes on the ECG, and 138 of them had no changes on the ECG, 93 were surgically treated, and 73 of them were on conservative therapy. Based on the results of the research, we conclude that ultrasonography is an important method in the detection and treatment of congenital heart diseases.

PSO-PSP-Net + InceptionV3: An optimized hyper-parameter tuned Computer-Aided Diagnostic model for liver tumor detection using CT scan slices

An automated diagnostic system leads to a supreme requirement in medical image analysis, greatly impacting the death rate due to the high spreading rate of liver tumors. However, the traditional histopathological diagnostic processes cause more disruption in the medical detection framework because of the procedure of manual estimation of CT scans and medical labor-intensive, which lead to the time consumption process. It is a highly challenging task for the segmentation of liver tumor because of the irregular appearance, distinct densities, patchy edges, and tumor sizes. To overcome this issue, an automated computer-aided diagnosis framework has been utilized to diagnose the liver cancer quickly and accurately with the less workload of expert radiologists.Presently, Deep learning-based techniques have been applied to identify the tumor regions and assist radiologists in the initial diagnosis and scrutinizing the severity level. The proposed work presents a modified pyramid scene parsing network (PSP-Net) model based on InceptionV3 deep network architecture with a particle swarm optimization algorithm, which increases the efficiency and effectiveness of the liver diagnostic model. Three publicly accessible benchmark datasets have been engaged with the pre-processing task accomplished by image normalization and data augmentation techniques. Moreover, a hyper-parameter tuned improved PSO-PSP-Net algorithm has analyzed the optimal set of gathered features. In the experimental evaluation, the proposed model outperformed than grey wolf optimization (GWO-PSP-Net), PSO-U-Net, GWO-U-Net, opposition based spotted hyena optimization (O-SHO-CRNN), and grey wolf-class topper optimization (GW-CTO) + U-Net by 8.42 %, 4.9 %, 9.54 %, 5.12 %, and 1.75 %, respectively in terms of accuracy for LiTS dataset. The developed PSO-PSP-Net model attained better accuracy than GWO-PSP-Net, PSO-U-Net, GWO-U-Net, and Residual Atrous U-Net (RA-U-Net) for the segmentation of Type-I and Type-II liver tumors by 6.88 %, 8.44 %, 13.75 %, 5.99 % and 6.12 %, 9.74 %, 17.3 %, and 5.11 %, respectively for 3DIRCAD-b1 dataset. The obtained accuracy of the developed model is 14.09 %, 7.54 %, 16.36 %, 2.64 %, and 14.9 %, 8.86 %, 17.76 %, and 4.87 % maximized than GWO-PSP-Net, PSO-U-Net, GWO-U-Net, and RA-U-Net, respectively, for the segmentation of the infected arteries and bones in the 3DIRCAD-b1 dataset. For the CHAOS dataset, the improved PSO-PSP-Net model outperformed GWO-PSP-Net, PSO-U-Net, and GWO-U-Net, respectively, in terms of overall accuracy by 7.44 %, 8.85 %, and 17.2 %, respectively. The investigational analysis validates the proposed PSO-PSP-Net model's effectiveness and performance in obtaining high segmentation accuracy in comparison with other state-of-the-art approaches.

Ultrasound Elastography: Methods, Clinical Applications, and Limitations: A Review Article

Ultrasound is a highly adaptable medical imaging modality that offers several applications and a wide range of uses, both for diagnostic and therapeutic purposes. The principles of sound wave propagation and reflection enable ultrasound imaging to function as a highly secure modality. This technique facilitates the production of real-time visual representations, thereby assisting in the evaluation of various medical conditions such as cardiac, gynecologic, and abdominal diseases, among others. The ultrasound modality encompasses a diverse range of modes and mechanisms that serve to enhance the methodology of pathology and physiology assessment. Doppler imaging and US elastography, in particular, are two such techniques that contribute to this expansion. Elastography-based imaging methods have attracted significant interest in recent years for the non-invasive evaluation of tissue mechanical characteristics. These techniques utilize the changes in soft tissue elasticity in various diseases to generate both qualitative and quantitative data for diagnostic purposes. Specialized imaging techniques collect data by identifying tissue stiffness under mechanical forces such as compression or shear waves. However, in this review paper, we provide a comprehensive examination of the fundamental concepts, underlying physics, and limitations associated with ultrasound elastography. Additionally, we present a concise overview of its present-day clinical utilization and ongoing advancements across many clinical domains.

AI-enhanced integration of genetic and medical imaging data for risk assessment of Type 2 diabetes.

Type 2 diabetes (T2D) presents a formidable global health challenge, highlighted by its escalating prevalence, underscoring the critical need for precision health strategies and early detection initiatives. Leveraging artificial intelligence, particularly eXtreme Gradient Boosting (XGBoost), we devise robust risk assessment models for T2D. Drawing upon comprehensive genetic and medical imaging datasets from 68,911 individuals in the Taiwan Biobank, our models integrate Polygenic Risk Scores (PRS), Multi-image Risk Scores (MRS), and demographic variables, such as age, sex, and T2D family history. Here, we show that our model achieves an Area Under the Receiver Operating Curve (AUC) of 0.94, effectively identifying high-risk T2D subgroups. A streamlined model featuring eight key variables also maintains a high AUC of 0.939. This high accuracy for T2D risk assessment promises to catalyze early detection and preventive strategies. Moreover, we introduce an accessible online risk assessment tool for T2D, facilitating broader applicability and dissemination of our findings.

Monkey Pox Detection using Deep learning

Monkeypox is a rare zoonotic disease caused by the Monkeypox virus, which manifests in humans with symptoms similar to those of smallpox. Early detection and accurate diagnosis are crucial for effective management and containment of outbreaks. This research proposes a novel Monkeypox detection system utilizing machine learning techniques to enhance the speed and accuracy of diagnosis. The proposed system integrates various data sources, including clinical records, medical imaging, and demographic information, to develop a comprehensive dataset for model training. Machine learning algorithms, such as convolutional neural networks (CNNs) for image analysis and ensemble methods for combining diverse data modalities, are employed to identify patterns indicative of Monkeypox infection

Utilizing Artificial Neural Networks for Geometric Bone Model Reconstruction in Mandibular Prognathism Patients

Patient-specific 3D models of the human mandible are finding increasing utility in medical fields such as oral and maxillofacial surgery, orthodontics, dentistry, and forensic sciences. The efficient creation of personalized 3D bone models poses a key challenge in these applications. Existing solutions often rely on 3D statistical models of human bone, offering advantages in rapid bone geometry adaptation and flexibility by capturing a range of anatomical variations, but also a disadvantage in terms of reduced precision in representing specific shapes. Considering this, the proposed parametric model allows for precise manipulation using morphometric parameters acquired from medical images. This paper highlights the significance of employing the parametric model in the creation of a personalized bone model, exemplified through a case study targeting mandibular prognathism reconstruction. A personalized model is described as 3D point cloud determined through the utilization of series of parametric functions, determined by the application of geometrical morphometrics, morphology properties, and artificial neural networks in the input dataset of human mandible samples. With 95.05% of the personalized model’s surface area displaying deviations within −1.00–1.00 mm relative to the input polygonal model, and a maximum deviation of 2.52 mm, this research accentuates the benefits of the parametric approach, particularly in the preoperative planning of mandibular deformity surgeries.

TransDiff: medical image segmentation method based on Swin Transformer with diffusion probabilistic model

TransDiff: medical image segmentation method based on Swin Transformer with diffusion probabilistic model

Leveraging Autoencoders for Better Representation Learning

ABSTRACT The effectiveness of a machine learning algorithm depends to a considerable extent on the representation used for the modeling. It is more critical with medical image data like Parkinson’s, where complexity plays a significant role. The paper proposes a methodology that leverages representation learning in Parkinson’s disease data. It uses seven types of autoencoders as learning techniques, identifying the best among them. The method is validated using three additional measures: Firstly, another dataset of Parkinson’s is used to confirm its effectiveness, thereby making the method dataset-agnostic. Secondly, seven machine-learning techniques formulate the problem in a supervised learning setting by taking the representation given by the best autoencoder as the features and the disease stage as the labels. Lastly, it uses three-dimensionality reduction techniques to visualize these latent variables or representations in a lower dimension. The high accuracy demonstrated at the supervised learning level, and the formation of patterns exhibited at the visualization level demonstrate the effectiveness of the proposed methodology.