Change detection of medical images using dictionary learning techniques and principal component analysis.

Introduction Advances in medical imaging technology have made it possible routinely to acquire high-resolution, three-dimensional images of human anatomy and function using a variety of imaging modalities. Today the number of acquired 2D images per exam varies from 150 images for screening up to 700 to 3000 for the diagnosis of complex cases. This large amount of images per case together with the growing importance of medical imaging in clinical practice, have continuously increased the workload of the radiologist, which explains the need for computer-assisted medical image analysis. Furthermore there is a quest for objective, quantitative information from medical images. In radiotherapy for instance, outlines of the irradiation volume and the neighboring organs at risk are delineated in 3D images and used to calculate a suitable radiation therapy. In neurology, degenerative diseases of the brain, such as multiple sclerosis, Alzheimer's, or schizophrenia, are studied by measuring brain shape and tissue changes in MR images. In cardiology, the health condition of the heart is assessed by studying the dynamics, the perfusion, and tissue characteristics of the heart muscle as revealed by MR or ultrasound images, and so forth. Traditionally, medical images are interpreted by visual inspection of the 2D images displayed slice by slice. Such radiological protocol is necessarily subjective , as it is based on the perception by a human observer and is usually restricted to mere qualitative statements and judgments. Moreover, the traditional 2D display of 3D images allows immediate inspection of anatomical structures in the two dimensions of the image plane only, whereas the third dimension has to be reconstructed mentally by the radiologist by looking at adjacent image slices.

A.P. Dhawan, H.K. Huang, D.-S. Kim, eds. World Scientific Publishing; 2008, 868 pages, $228.00. At first glance at the cover and size of Principles and Advanced Methods in Medical Imaging and Image Analysis , a passerby might think that it is the latest Harry Potter novel; however, the gadgets and

Tourette syndrome (TS) is a neurological disorder characterized by involuntary and repetitive movements known as tics. A retrospective analysis of magnetic resonance imaging (MRI) scans from 39 children and adolescents with TS was performed and subsequently compared with MRI scans from 834 neurotypical controls. The purpose of this study was to identify any differences in the regions of motor circuitry in TS to further our understanding of their disturbances in motor control (i.e., motor tics). Measures of volume, cortical thickness, surface area, and surface curvature for specific motor regions were derived from each MRI scan. The results revealed increased surface curvature in the opercular part of the inferior frontal gyrus and the triangular part of the inferior frontal gyrus in the TS group compared with the neurotypical control group. These novel findings offer some of the first evidence for surface curvature differences in motor circuitry regions in TS, which may be associated with known motor and vocal tics.

With technological advances and increasing demand for finer resolution images, tomographic medical imaging can be a huge computational problem, consequently, the processing time for image construction can be prohibitively large and impractical for real-time applications. The main bottlenecks are retrieving data and solving the inverse problem to attain medical images. Proper Orthogonal Decomposition (POD) is a model-reduction technique that can be used to compress a large set of images into an orthonormal basis whose elements can accurately generate any images in the original collection with the fewest possible modes. Applicability of POD on tomographic images is not possible without the linearity of the inverse problem. Due to its structure, the first few POD elements contain all the dominant features of the entire image collection. Thus, instead of performing the inverse Radon transform on all medical tomographic images, one needs to process only the primary POD modes once and reuse them to construct all tomographic images, rendering its computational savings. In this article, we improve the POD method further by implementing its hybrid version. Namely, the computation of the covariance matrix and associated eigenvalues and eigenvectors can be expensive for large-size images. This process can be sped by working with the down-sampled data and use the resulting coefficients to reconstruct the full-resolution images. Image reconstruction of fish eggs in a test tube will be presented. Errors and computational savings will also be discussed.

BACKGROUND Brain metastases, secondary brain tumors originating from cancers elsewhere in the body, can present with focal neurologic deficits without systemic symptoms and are usually evaluated using magnetic resonance imaging (MRI). Multimodal imaging methods are usually used to determine the primary cancer source. Radiomic analysis involves extracting quantitative features from medical images using data-characterization algorithms. Recent advances in deep learning and radiomics, which quantitatively analyze imaging data, offer the potential for improved prediction of primary tumor origin through MR imaging analysis. Using longitudinal MRI data, we developed and validated a mixed-effects neural network model for predicting the primary origin of brain metastases. METHODS We used the open-source dataset from 5 different institutions made available by Ocaña-Tienda et al. (2023) to build the model. Radiomic feature selection was performed, followed by building a novel quantitative radiomic approach that uses mixed-effects deep learning to analyze longitudinal radiomic data of brain metastases from MRI scans. Model performance was evaluated using receiver operating characteristic curve (ROC) analysis. RESULTS Our mixed-effects model demonstrated high accuracy (90.5%), precision (92.7%), recall (87.8%), and F1 score (90.2%), with an area under the ROC (AUC-ROC) curve of 0.944 in identifying the primary tumor site. The nested mixed-effects model was superior to the fixed-effects model with an AUC-ROC of 0.719. CONCLUSIONS Our mixed-effects neural network model accurately predicts the primary origin of brain metastases with high internal and external validity. Using advanced radiomic feature selection and deep learning techniques. Our model achieved high-performance metrics, thus providing a promising tool for enhancing diagnostic capability and tailoring patient-specific management strategies. Further research could explore adding more data from other institutions and studying the effect of integrating this model in clinical workflows to optimize patient care.

Magnetic resonance imaging (MRI) scans in the elderly commonly show white matter findings that may raise concerns. We sought to document incidence, manifestations, and predictors of worsening white matter grade on serial imaging. The Cardiovascular Health Study is a population-based, longitudinal study of 5888 people aged 65 years and older, of whom 1919 have had extensive initial and follow-up evaluations, including 2 MRI scans separated by 5 years. Scans were read without clinical information in standard side-by-side fashion to determine worsening white matter grade. Worsening was evident in 538 participants (28%), mostly (85%) by 1 grade. Although similar at initial scan, participants with worsening white matter grade, compared with those without, experienced greater decline on modified Mini-Mental State examination and Digit-Symbol Substitution test (both P< or =0.001) after controlling for potential confounding factors, including occurrence of transient ischemic attack or stroke between scans. Independent predictors of worsening white matter grade included cigarette smoking before initial scan and infarct on initial scan. Otherwise, predictors differed according to white matter grade on initial scan. For low initial grade, increased age, increased diastolic blood pressure, increased high-density lipoprotein cholesterol, and decreased low-density lipoprotein cholesterol were associated with increased risk of worsening. For high initial grade, any cardiovascular disease and low ankle-arm index were associated with decreased risk of worsening, whereas use of diuretics and statins were associated with increased risk. Worsening white matter grade on serial MRI scans in elderly is common, is associated with cognitive decline, and has complex relations with cardiovascular risk factors.

The cumulative effects and clinical safety of repeat magnetic resonance imaging on an MRI-conditional pacemaker system at 1.5 tesla

Corpus callosum atrophy and neuropsychological outcome following carbon monoxide poisoning

Segmenting brain tumors in magnetic resonance imaging (MRI) scans has been a challenging issue with medical imaging, because of the great diversity in tumors regarding their sizes, shapes, and textures. Traditional 3D convolutional neural networks (3D-CNNs) have been shown to be very good at extracting local features, but bad at capturing relationships across large distances. To overcome this drawback, the current study offers a hybrid deep-learning architecture, which is the 3D Dual-Path Transformer Residual Attention U-Net (3D DPT-RAUNet) to realize an accurate segmentation of the brain-tumor. The architecture proposed will combine two orthogonal modules: 3D Residual Attention U-Net branch that will focus on fine-grained learning of local features, and Transformer-based Dual-Path Encoder which will focus on learning global spatial cues in the entire MRI volume. Tumor-relevant regions are selectively highlighted by means of the attention mechanism, and gradient propagation stabilized by residual connections is achieved by the model to achieve better convergence. Experiments with the BraTS2023 dataset using substantial preprocessing include bias field correction, intensity normalisation, and patch extraction show that the 3D DPT - RAUNet obtains a mean Dice similarity coefficient of 0.957 which is significantly better than current 3D U-Net and Transformer-based models and is at a low cost of about 18 - percent. The findings highlight the superior performance, strength, and clinical practicability of the hybrid architecture in automating the analysis of brain-tumors.

Organ-on-a-chip (OOC) is a new type of biochip technology. Various types of OOC systems have been developed rapidly in the past decade and found important applications in drug screening and precision medicine. However, due to the complexity in the structure of both the chip-body itself and the engineered-tissue inside, the imaging and analysis of OOC have still been a big challenge for biomedical researchers. Considering that medical imaging is moving towards higher spatial and temporal resolution and has more applications in tissue engineering, this paper aims to review medical imaging methods, including CT, micro-CT, MRI, small animal MRI, and OCT, and introduces the application of 3D printing in tissue engineering and OOC in which medical imaging plays an important role. The achievements of medical imaging assisted tissue engineering are reviewed, and the potential applications of medical imaging in organoids and OOC are discussed. Moreover, artificial intelligence - especially deep learning - has demonstrated its excellence in the analysis of medical imaging; we will also present the application of artificial intelligence in the image analysis of 3D tissues, especially for organoids developed in novel OOC systems.

Magnetic resonance imaging (MRI) scans are known to suffer from a variety of acquisition artifacts as well as equipment-based variations that impact image appearance and segmentation performance. It is still unclear whether a direct relationship exists between magnetic resonance (MR) image quality metrics (IQMs) (e.g., signal-to-noise, contrast-to-noise) and segmentation accuracy. Deep learning (DL) approaches have shown significant promise for automated segmentation of brain tumors on MRI but depend on the quality of input training images. We sought to evaluate the relationship between IQMs of input training images and DL-based brain tumor segmentation accuracy toward developing more generalizable models for multi-institutional data. We trained a 3D DenseNet model on the BraTS 2020 cohorts for segmentation of tumor subregions enhancing tumor (ET), peritumoral edematous, and necrotic and non-ET on MRI; with performance quantified via a 5-fold cross-validated Dice coefficient. MRI scans were evaluated through the open-source quality control tool MRQy, to yield 13 IQMs per scan. The Pearson correlation coefficient was computed between whole tumor (WT) dice values and IQM measures in the training cohorts to identify quality measures most correlated with segmentation performance. Each selected IQM was used to group MRI scans as "better" quality (BQ) or "worse" quality (WQ), via relative thresholding. Segmentation performance was re-evaluated for the DenseNet model when (i) training on BQ MRI images with validation on WQ images, as well as (ii) training on WQ images, and validation on BQ images. Trends were further validated on independent test sets derived from the BraTS 2021 training cohorts. For this study, multimodal MRI scans from the BraTS 2020 training cohorts were used to train the segmentation model and validated on independent test sets derived from the BraTS 2021 cohort. Among the selected IQMs, models trained on BQ images based on inhomogeneity measurements (coefficient of variance, coefficient of joint variation, coefficient of variation of the foreground patch) and the models trained on WQ images based on noise measurement peak signal-to-noise ratio (SNR) yielded significantly improved tumor segmentation accuracy compared to their inverse models. Our results suggest that a significant correlation may exist between specific MR IQMs and DenseNet-based brain tumor segmentation performance. The selection of MRI scans for model training based on IQMs may yield more accurate and generalizable models in unseen validation.

Summary Introduction : There has been a rapid development of deep learning (DL) models for medical imaging. However, DL requires a large labeled dataset for training the models. Getting large-scale labeled data remains a challenge, and multi-center datasets suffer from heterogeneity due to patient diversity and varying imaging protocols. Domain adaptation (DA) has been developed to transfer the knowledge from a labeled data domain to a related but unlabeled domain in either image space or feature space. DA is a type of transfer learning (TL) that can improve the performance of models when applied to multiple different datasets. Objective : In this survey, we review the state-of-the-art DL-based DA methods for medical imaging. We aim to summarize recent advances, highlighting the motivation, challenges, and opportunities, and to discuss promising directions for future work in DA for medical imaging. Methods : We surveyed peer-reviewed publications from leading biomedical journals and conferences between 2017-2020, that reported the use of DA in medical imaging applications, grouping them by methodology, image modality, and learning scenarios. Results : We mainly focused on pathology and radiology as application areas. Among various DA approaches, we discussed domain transformation (DT) and latent feature-space transformation (LFST). We highlighted the role of unsupervised DA in image segmentation and described opportunities for future development. Conclusion : DA has emerged as a promising solution to deal with the lack of annotated training data. Using adversarial techniques, unsupervised DA has achieved good performance, especially for segmentation tasks. Opportunities include domain transferability, multi-modal DA, and applications that benefit from synthetic data.

Memetic improved cuckoo search algorithm for automatic B-spline border approximation of cutaneous melanoma from macroscopic medical images

Medical imaging and information management systems require transmission of medical images over the Internet. Many image watermarking techniques have been proposed in recent years to ensure the integrity and authenticity of medical images transferred over insecure networks. In this work, we propose a new medical image watermarking technique to detect tampered regions on medical images with finer accuracy by authenticating 4×4 blocks and without restricting region of interest (ROI) size. The proposed method can mark a 4×4 pixel block if it has even one tampered pixel, while similar methods (which have no ROI size restriction) mark 8×8, 16×16, and 40×40 pixel blocks. Modified difference expansion (MDE) and least significant bit (LSB) embedding techniques are used together first in the literature by the method to embed authentication bits into corresponding blocks. The method uses a small 4×4 window to mark the tampered region. Experimental results show that the proposed method detects tampered regions on medical images with high accuracy and can be used by all medical image modalities. The results also indicate that the method has finer accuracy and no ROI size restriction compared to similar works reported in the literature.

Factors predicting local tumor control after gamma knife stereotactic radiosurgery for benign intracranial meningiomas