Medical Imaging Research Articles

Multi-modal fusion deep transfer learning for accurate brain tumor classification using magnetic resonance imaging images

Early identification and treatment of brain tumors depend critically on accurate classification. Accurate brain tumor classification in medical imaging is essential for clinical decisions and individualized treatment plans. This paper introduces a novel method for classifying brain tumors called multimodal fusion deep transfer learning (MMFDTL) using original, contoured, and annotated magnetic resonance imaging (MRI) images to showcase its capabilities. The MMFDTL can capture complex tumor features frequently missed in analyzing individual modalities. The MMFDTL model employs three deep learning models for extracting features VGG16, Inception V3, and ResNet 50. The accuracy rate improves when combined with decision based multimodal fusion. It produces impressive outcomes of sensitivity 92.96%, specificity 98.54%, precision 93.60%, accuracy 98.80%, F1-score 93.26%, and kappa 91.86%. This research can improve medical imaging and brain tumor analysis through its multi modal fusion approach. It could give healthcare practitioners vital insights for personalized treatment plans and informed decision making.

A Fast Alpha-Tree Algorithm for Extreme Dynamic Range Pixel Dissimilarities.

The α-tree algorithm is a useful hierarchical representation technique which facilitates comprehension of images such as remote sensing and medical images. Most α-tree algorithms make use of priority queues to process image edges in a correct order, but because traditional priority queues are inefficient in α-tree algorithms using extreme-dynamic-range pixel dissimilarities, they run slower compared with other related algorithms such as component tree. In this paper, we propose a novel hierarchical heap priority queue algorithm that can process α-tree edges much more efficiently than other state-of-the-art priority queues. Experimental results using 48-bit Sentinel-2 A remotely sensed images and randomly generated images have shown that the proposed hierarchical heap priority queue improved the timings of the flooding α-tree algorithm by replacing the heap priority queue with the proposed queue: 1.68 times in 4-N and 2.41 times in 8-N on Sentinel-2 A images, and 2.56 times and 4.43 times on randomly generated images.

FSformer: A combined frequency separation network and transformer for LDCT denoising

Low-dose computed tomography (LDCT) has been widely concerned in the field of medical imaging because of its low radiation hazard to humans. However, under low-dose radiation scenarios, a large amount of noise/artifacts are present in the reconstructed image, which reduces the clarity of the image and is not conducive to diagnosis. To improve the LDCT image quality, we proposed a combined frequency separation network and Transformer (FSformer) for LDCT denoising. Firstly, FSformer decomposes the LDCT images into low-frequency images and multi-layer high-frequency images by frequency separation blocks. Then, the low-frequency components are fused with the high-frequency components of different layers to remove the noise in the high-frequency components with the help of the potential texture of low-frequency parts. Next, the estimated noise images can be obtained by using Transformer stage in the frequency aggregation denoising block. Finally, they are fed into the reconstruction prediction block to obtain improved quality images. In addition, a compound loss function with frequency loss and Charbonnier loss is used to guide the training of the network. The performance of FSformer has been validated and evaluated on AAPM Mayo dataset, real Piglet dataset and clinical dataset. Compared with previous representative models in different architectures, FSformer achieves the optimal metrics with PSNR of 33.7714 dB and SSIM of 0.9254 on Mayo dataset, the testing time is 1.825 seconds. The experimental results show that FSformer is a state-of-the-art (SOTA) model with noise/artifact suppression and texture/organization preservation. Moreover, the model has certain robustness and can effectively improve LDCT image quality.

Abdominal Injuries Detection

Abstract: Abdominal trauma refers to physical injuries or damage to the abdominal region, which includes the area between the casket and pelvis. This can involve colorful structures within the tummy, similar as the organs (e.g., liver, spleen, bowel, kidney, and extravasation). Traumatic injury affects individualities across all age groups and stands as one of the primary causes of mortality encyclopedically. The World Health Organization reports that nearly 5 million individualities succumb to traumatic injuries annually. The main ideal of an abdominal injuries’ discovery design is to ameliorate the early and accurate identification of abdominal injuries performing from trauma. This design generally involves the development and perpetration of colorful tools, ways, and protocols to achieve the objects like perfecting the delicacy of CT for detecting abdominal trauma and prognosticate the probability of specific traumatic injuries to abdominal organs. Diagnosing abdominal trauma clinically through physical examination, patient symptoms, or laboratory tests can frequently be grueling. thus, timely opinion using medical imaging plays a pivotal part in patient care. Our design will identify whether the case is suffering from abdominal trauma or not and help the cases by informing when immediate surgery is needed. Abdominal trauma discovery generally involves analysis of CT images to prognosticate the probability of specific traumatic injuries to abdominal organs, similar as the bowel, extravasation (generally refers to bleeding from blood vessel), liver.

Self-supervised learning for medical image analysis: a comprehensive review

Self-supervised learning for medical image analysis: a comprehensive review

The Emerging Role and Clinical Applications of Morphomics in Diagnostic Imaging.

Analytic morphomics refers to the accurate measurement of specific biological markers of human body composition in diagnostic medical imaging. The increasing prevalence of disease processes that alter body composition including obesity, cachexia, and sarcopenia has generated interest in specific targeted measurement of these metrics to possibly prevent or reduce negative health outcomes. Typical morphomic measurements include the area and density of muscle, bone, vascular calcification, visceral fat, and subcutaneous fat on a specific validated axial level in the patient's cross-sectional diagnostic imaging. A distinct advantage of these measurements is that they can be made retrospectively and opportunistically with pre-existing datasets. We provide a narrative review of the current state of art in morphomics, but also consider some potential future directions for this exciting field. Imaging based quantitative assessment of body composition has enormous potential across the breadth and scope of modern clinical practice. From risk stratification to treatment planning, and outcome assessment, all can be enhanced with the use of analytic morphomics. Moreover, it is likely that many new opportunities for personalized medicine will emerge as the field evolves. As radiologists, embracing analytic morphomics will enable us to contribute added value in the care of every patient.

Pure Vision Transformer (CT-ViT) with Noise2Neighbors Interpolation for Low-Dose CT Image Denoising.

Convolutional neural networks (CNN) have been used for a wide variety of deep learning applications, especially in computer vision. For medical image processing, researchers have identified certain challenges associated with CNNs. These challenges encompass the generation of less informative features, limitations in capturing both high and low-frequency information within feature maps, and the computational cost incurred when enhancing receptive fields by deepening the network. Transformers have emerged as an approach aiming to address and overcome these specific limitations of CNNs in the context of medical image analysis. Preservation of all spatial details of medical images is necessary to ensure accurate patient diagnosis. Hence, this research introduced the use of a pure Vision Transformer (ViT) for a denoising artificial neural network for medical image processing specifically for low-dose computed tomography (LDCT) image denoising. The proposed model follows a U-Net framework that contains ViT modules with the integration of Noise2Neighbor (N2N) interpolation operation. Five different datasets containing LDCT and normal-dose CT (NDCT) image pairs were used to carry out this experiment. To test the efficacy of the proposed model, this experiment includes comparisons between the quantitative and visual results among CNN-based (BM3D, RED-CNN, DRL-E-MP), hybrid CNN-ViT-based (TED-Net), and the proposed pure ViT-based denoising model. The findings of this study showed that there is about 15-20% increase in SSIM and PSNR when using self-attention transformers than using the typical pure CNN. Visual results also showed improvements especially when it comes to showing fine structural details of CT images.

An Automated Multi-scale Feature Fusion Network for Spine Fracture Segmentation Using Computed Tomography Images.

Spine fractures represent a critical health concern with far-reaching implications for patient care and clinical decision-making. Accurate segmentation of spine fractures from medical images is a crucial task due to its location, shape, type, and severity. Addressing these challenges often requires the use of advanced machine learning and deep learning techniques. In this research, a novel multi-scale feature fusion deep learning model is proposed for the automated spine fracture segmentation using Computed Tomography (CT) to these challenges. The proposed model consists of six modules; Feature Fusion Module (FFM), Squeeze and Excitation (SEM), Atrous Spatial Pyramid Pooling (ASPP), Residual Convolution Block Attention Module (RCBAM), Residual Border Refinement Attention Block (RBRAB), and Local Position Residual Attention Block (LPRAB). These modules are used to apply multi-scale feature fusion, spatial feature extraction, channel-wise feature improvement, segmentation border results border refinement, and positional focus on the region of interest. After that, a decoder network is used to predict the fractured spine. The experimental results show that the proposed approach achieves better accuracy results in solving the above challenges and also performs well compared to the existing segmentation methods.

[Perspective] Delta Radiomics — Potential Role in Head Neck Cancer

INTRODUCTION: Delta radiomics is a tool used to assess the response of oncologic patients undergoing immunotherapy. It extracts high-dimensional quantitative features from medical images, providing information about cancer's phenotype, genotype, and tumoral microenvironment. This analysis could help avoid invasive procedures and help choose the most suitable therapeutic in multiple therapeutic options. METHOD: Radiomics has gained interest as an imaging biomarker for predicting response to various immunotherapies. Delta radiomics assesses feature variations from one time point to another based on subsequent images, offering higher value for treatment-outcome prediction or patient stratification into risk categories. CONCLUSION: It has potential benefits for clinical endpoints in oncology, such as differential diagnosis, prognosis, treatment response prediction, and evaluation of side effects. Further research with prospective and multicentre studies is needed for clinical validation of delta radiomics approaches. STATEMENT OF CLINICAL SIGNIFICANCE: In head and neck oncology, delta radiomics can be used to enhance the precision of diagnosis, assess tumor response, forecast normal tissue toxicity, predict clinical outcome, and pinpoint characteristics for treatment modification. Patients' quality of life may be enhanced by it. It can support post-treatment surveillance. Additionally, it can support the delivery of individualized care based on a patient's reaction to medication and radiation.

Improvement of Basal Ganglia Detectability in Brain Single Photon Emission Computerized Tomography by Wavelet Transformation in Image Processing Domain: A XCAT Phantom Study

Purpose: Noise in brain Single Photon Emission Computed Tomography (SPECT) images limits an early diagnosis of Parkinson's Disease (PD). To overcome the limitation, as an image processing approach, wavelet transformation was used to denoising the images also with a segmentation method to differentiate the basal ganglia in brain SPECT.
 Materials and Methods: The brain scans of the human 4D Extended Cardiac Torso (XCAT) phantom through the Simulating Medical Imaging Nuclear Detectors (SIMIND) simulated SPECT system were imported to the MATLAB toolkit for image processing. The reconstructed brain images by iterative reconstruction were de-noised through 9 methods of wavelet transformation at different levels, and then six segmentation methods were applied to differentiate the caudate and putamen. The Dice coefficient, Specificity, and Sensitivity evaluation criteria were calculated based on the adaptive thresholding of the selected images from segmentation. A ground truth image was manually marked by a clinical nuclear medicine specialist.
 Results: The dice coefficient was obtained in a range from 0.3979 to 0.6299, as well as the specificity criterion from 0.7682 to 0.8168 and the sensitivity from 0.9049 to 0.9871.
 The results from adaptive threshold segmentation and the evaluation criteria showed that the best levels of the nucleuses detectability were provided by level 7 of Biorthogonal, levels 4 and 7 of Coiflet, level 6 of Daubechies, level 5 of Haar, level 6 of Morlet and level 6 of Symlet methods.
 Conclusion: Parkinson’s disease may be diagnosed in the early stage by an image processing approach to improve the quality of brain SPECT images.

DiffSwinTr: A diffusion model using 3D Swin Transformer for brain tumor segmentation

AbstractAutomatic medical image segmentation has shown great potential in recent years. Howerver, magnetic resonance images (MRI) usually have the characteristics of noise and artifacts, existing methods cannot accurately segment the boundaries. In addition, most existing algorithms are unable to effectively capture the global dependencies to offset the local inductive bias. In this work, we present a novel denoising diffusion model based on 3D Swin Transformer for brain tumor segmentation, called DiffSwinTr. Furthermore, a conditional encoder module is designed to extract multi‐scale features and enhance the ability of local feature perception. We extensively evaluate our model on three BraTS datasets. The proposed DiffSwinTr can achieve average Dice Scores of 79.91%, 83.07%, and 85.38%, as well as the average Hausdorff Distances (95%) of 3.361, 3.334, and 2.975mm, respectively. The experimental results illuminate that DiffSwinTr outperforms the state‐of‐the‐art segmentation methods. Besides, our proposed model is robust to segmenting images with noise and artifacts.

Development and Validation of a Deep Learning Model for Histopathological Slide Analysis in Lung Cancer Diagnosis

Lung cancer is the leading cause of cancer-related deaths worldwide. Two of the crucial factors contributing to these fatalities are delayed diagnosis and suboptimal prognosis. The rapid advancement of deep learning (DL) approaches provides a significant opportunity for medical imaging techniques to play a pivotal role in the early detection of lung tumors and subsequent monitoring during treatment. This study presents a DL-based model for efficient lung cancer detection using whole-slide images. Our methodology combines convolutional neural networks (CNNs) and separable CNNs with residual blocks, thereby improving classification performance. Our model improves accuracy (96% to 98%) and robustness in distinguishing between cancerous and non-cancerous lung cell images in less than 10 s. Moreover, the model’s overall performance surpassed that of active pathologists, with an accuracy of 100% vs. 79%. There was a significant linear correlation between pathologists’ accuracy and years of experience (r Pearson = 0.71, 95% CI 0.14 to 0.93, p = 0.022). We conclude that this model enhances the accuracy of cancer detection and can be used to train junior pathologists.

Toward “super-scintillation” with nanomaterials and nanophotonics

Abstract Following the discovery of X-rays, scintillators are commonly used as high-energy radiation sensors in diagnostic medical imaging, high-energy physics, astrophysics, environmental radiation monitoring, and security inspections. Conventional scintillators face intrinsic limitations including a low extraction efficiency of scintillated light and a low emission rate, leading to efficiencies that are less than 10 % for commercial scintillators. Overcoming these limitations will require new materials including scintillating nanomaterials (“nanoscintillators”), as well as new photonic approaches that increase the efficiency of the scintillation process, increase the emission rate of materials, and control the directivity of the scintillated light. In this perspective, we describe emerging nanoscintillating materials and three nanophotonic platforms: (i) plasmonic nanoresonators, (ii) photonic crystals, and (iii) high-Q metasurfaces that could enable high performance scintillators. We further discuss how a combination of nanoscintillators and photonic structures can yield a “super scintillator” enabling ultimate spatio-temporal resolution while enabling a significant boost in the extracted scintillation emission.

6. Persistent spinal pain syndrome type 2.

Persistent Spinal Pain Syndrome (PSPS) refers to chronic axial pain and/or extremity pain. Two subtypes have been defined: PSPS-type 1 is chronic pain without previous spinal surgery and PSPS-type 2 is chronic pain, persisting after spine surgery, and is formerly known as Failed Back Surgery Syndrome (FBSS) or post-laminectomy syndrome. The etiology of PSPS-type 2 can be gleaned using elements from the patient history, physical examination, and additional medical imaging. Origins of persistent pain following spinal surgery may be categorized into an inappropriate procedure (eg a lumbar fusion at an incorrect level or for sacroiliac joint [SIJ] pain); technical failure (eg operation at non-affected levels, retained disk fragment, pseudoarthrosis), biomechanical sequelae of surgery (eg adjacent segment disease or SIJ pain after a fusion to the sacrum, muscle wasting, spinal instability); and complications (eg battered root syndrome, excessive epidural fibrosis, and arachnoiditis), or undetermined. The literature on the diagnosis and treatment of PSPS-type 2 was retrieved and summarized. There is low-quality evidence for the efficacy of conservative treatments including exercise, rehabilitation, manipulation, and behavioral therapy, and very limited evidence for the pharmacological treatment of PSPS-type 2. Interventional treatments such as pulsed radiofrequency (PRF) of the dorsal root ganglia, epidural adhesiolysis, and spinal endoscopy (epiduroscopy) might be beneficial in patients with PSPS-type 2. Spinal cord stimulation (SCS) has been shown to be an effective treatment for chronic, intractable neuropathic limb pain, and possibly well-selected candidates with axial pain. The diagnosis of PSPS-type 2 is based on patient history, clinical examination, and medical imaging. Low-quality evidence exists for conservative interventions. Pulsed radiofrequency, adhesiolysis and SCS have a higher level of evidence with a high safety margin and should be considered as interventional treatment options when conservative treatment fails.

Analysis of linear feedback shift registers and chaos-based techniques for image encryption

ABSTRACT Image encryption is not new. Several different encryption algorithms have been used. Some of the algorithms are based on random sequences generated using chao systems and DNA sequencing and others are based on compression. Still, others are based on Linear Feedback Shift Registers. This paper focuses on research solutions for securing digital images including medical images. Medical images are important and are generally regarded as sensitive in nature. In this research paper, higher order linear feedback shift registers are developed using combination of lower order LFSRs by suitable clocking and the ciphers generated are utilized to encrypt digital images. Our analysis involves using suitable metrics to compare the chaos-based and LFSR-based stream ciphers for image encryption.

Target Image Processing based on Super-resolution Reconstruction and Machine Learning Algorithm

This article proposes a target image processing method based on super-resolution reconstruction and machine learning algorithms, which solves the low-resolution problem in medical images during imaging. This method uses nonlocal autoregressive learning based on a medical image super-resolution reconstruction method. The autoregressive model is introduced into the sparse representation-based medical image super-resolution reconstruction model by utilizing medical image data inherent nonlocal similarity characteristics. At the same time, a clustering algorithm is used to obtain a classification dictionary, improving experimental efficiency. The experimental results show that ten randomly selected CT/MR images are used as test images, and each image's peak signal-to-noise ratio and structural similarity values are calculated separately. Compared with other methods, the method proposed in this paper is significantly better and can achieve ideal results, with the highest value being 31.49. This method demonstrates the feasibility of using super-resolution reconstruction and machine learning algorithms in medical image resolution.  

Adaptive Feature Medical Segmentation Network: an adaptable deep learning paradigm for high-performance 3D brain lesion segmentation in medical imaging

IntroductionIn neurological diagnostics, accurate detection and segmentation of brain lesions is crucial. Identifying these lesions is challenging due to its complex morphology, especially when using traditional methods. Conventional methods are either computationally demanding with a marginal impact/enhancement or sacrifice fine details for computational efficiency. Therefore, balancing performance and precision in compute-intensive medical imaging remains a hot research topic.MethodsWe introduce a novel encoder-decoder network architecture named the Adaptive Feature Medical Segmentation Network (AFMS-Net) with two encoder variants: the Single Adaptive Encoder Block (SAEB) and the Dual Adaptive Encoder Block (DAEB). A squeeze-and-excite mechanism is employed in SAEB to identify significant data while disregarding peripheral details. This approach is best suited for scenarios requiring quick and efficient segmentation, with an emphasis on identifying key lesion areas. In contrast, the DAEB utilizes an advanced channel spatial attention strategy for fine-grained delineation and multiple-class classifications. Additionally, both architectures incorporate a Segmentation Path (SegPath) module between the encoder and decoder, refining segmentation, enhancing feature extraction, and improving model performance and stability.ResultsAFMS-Net demonstrates exceptional performance across several notable datasets, including BRATs 2021, ATLAS 2021, and ISLES 2022. Its design aims to construct a lightweight architecture capable of handling complex segmentation challenges with high precision.DiscussionThe proposed AFMS-Net addresses the critical balance issue between performance and computational efficiency in the segmentation of brain lesions. By introducing two tailored encoder variants, the network adapts to varying requirements of speed and feature. This approach not only advances the state-of-the-art in lesion segmentation but also provides a scalable framework for future research in medical image processing.

Can surgical computer vision benefit from large-scale visual foundation models?

We investigate whether foundation models pretrained on diverse visual data could be beneficial to surgical computer vision. We use instrument and uterus segmentation in mini-invasive procedures as benchmarks. We propose multiple supervised, unsupervised and few-shot supervised adaptations of foundation models, including two novel adaptation methods. We use DINOv1, DINOv2, DINOv2 with registers, and SAM backbones, with the ART-Net surgical instrument and the SurgAI3.8K uterus segmentation datasets. We investigate five approaches: DINO unsupervised, few-shot learning with a linear decoder, supervised learning with the proposed DINO-UNet adaptation, DPT with DINO encoder, and unsupervised learning with the proposed SAM adaptation. We evaluate 17 models for instrument segmentation and 7 models for uterus segmentation and compare to existing ad hoc models for the tasks at hand. We show that the linear decoder can be learned with few shots. The unsupervised and linear decoder methods obtain slightly subpar results but could be considered useful in data scarcity settings. The unsupervised SAM model produces finer edges but has inconsistent outputs. However, DPT and DINO-UNet obtain strikingly good results, defining a new state of the art by outperforming the previous-best by 5.6 and 4.1 pp for instrument and 4.4 and 1.5 pp for uterus segmentation. Both methods obtain semantic and spatial precision, accurately segmenting intricate details. Our results show the huge potential of using DINO and SAM for surgical computer vision, indicating a promising role for visual foundation models in medical image analysis, particularly in scenarios with limited or complex data.

False-negative factors of percutaneous transluminal clamp biopsy for suspected malignant biliary stricture: 194 cases analyzed from a single center

ObjectiveTo study the predictive factors of false negatives in the diagnosis of biliary stricture (BS) by percutaneous transluminal clamp biopsy (PTCB).MethodFrom January 2016 to January 2021, 194 patients with a high suspicion of malignant tumors due to BS underwent PTCB during biliary drainage at our department. The final diagnosis was confirmed by postoperative pathology, other tissue or cell evidence, or medical imaging follow-up. Univariate and multivariate regression analyses were performed on the pathological results, summarizing the independent risk factors for false-negative value (FNV) to help further clinical diagnosis and treatment.ResultsOf the 194 cases, 176 and 18 cases were finally diagnosed as malignant and benign BS, respectively, compared to 144 and 50 cases by PTCB, including 32 false-negative cases. The sensitivity, specificity, false-positive value, and FNV of PTCB were 81.8%, 100%, 0%, and 18.2%, respectively. Multivariate analysis showed that non-cholangiocarcinoma BS was an independent risk factor for FNV of PTCB (odds ratio 7.5 (95% CI 1.74–32.6), p < 0.01).ConclusionPTCB is an effective minimally invasive interventional technique for BS diagnosis. Non-cholangiocarcinoma BS is an independent risk factor for FNV.Critical relevance statementIdentifying factors that are predictive of false-negative results by percutaneous transluminal clamp biopsy in the setting of biliary stricture may have a guiding effect on clinical practice.Key points• Factors predictive of false negatives in the diagnosis of biliary stricture etiology by PTCB may aid in the interpretation of results.• Non-cholangiocarcinoma BS is an independent risk factor for FNV on PTCB.• PTCB is an effective minimally invasive interventional technique for BS diagnosis.Graphical

Optimised ResNet50 for Multi-Class Classification of Brain Tumors

Categorizing brain cancers, including glioma, meningioma, and pituitary tumors, based on magnetic resonance imaging (MRI) images presents a significant challenge. Deep learning and machine learning techniques have shown promise in enhancing image categorization. To address this challenge, we leverage the power of the optimized ResNet50 model. Our approach involves classifying medical images using Convolutional Neural Network (CNN) features, which are then compared with the ResNet50 model. The primary goal is to detect brain tumors at an early stage using an advanced deep-learning model. We utilize an accessible dataset from Figshare, containing MRI images of the three distinct categories of brain tumors. Existing brain tumor models face limitations in handling multi-class problems and early-stage diagnosis. Therefore, we propose a fully automated approach employing Convolutional Neural Networks (CNN) to extract diverse properties from brain MRI scans. This method aims to provide accurate tumor diagnosis, even with a high number of classes and limited information in MRI data. Our proposed model involves the creation of identification blocks within a four-layered primary architecture, followed by testing and assessment of the interconnected layers. The results demonstrate that our model outperforms existing methods, achieving an impressive overall classification accuracy of 99.03%,