Morphometric Analyses of Macrophages.

Cell morphology and motility drive the cellular capabilities to interact with the environment. For example, microglia, the longest known tissue-resident macrophages, show a highly branched process tree with which they continuously scan their environment. Computational image analysis allows to quantify morphology and/or motility from images of tissue-resident macrophages. Here, I describe a step-by-step protocol for analyzing the morphology (and motility) of macrophages with our recently described, freely available software MotiQ, which provides a broad band of parameters and thereby serves as a versatile tool for studies of morphology and motility.

Automated Cell Counting of Macrophages In Situ.

Using the open-source image analysis software CellProfiler to automatically quantify antibody-stained or fluorescently labeled macrophages in situ provides accurate and reproducible cell counts. It is a vastly enhanced alternative method to both manual cell counting and estimation of cell marker expression based on fluorescence intensity. Quantification of tissue-resident macrophages acquired on widefield or confocal microscopes can be batch processed using our pipeline to produce data within minutes.

Investigating the slice thickness effect on noise and diagnostic content of single-source multi-slice computerized axial tomography.

High-quality and detailed CT scan images are crucial for accurate diagnosis. Factors such as image noise and slice thickness affect image quality. This study aimed to determine the optimal slice thickness that minimized image noise while maintaining sufficient diagnostic information using the single-source computed tomography head protocol. Single-source CT images were examined using the Linux Operating system Ge Revolution 64-slice CT scanner, and a combination of statical analysis and DICOM CT image analysis was employed. The single-source energy head CT protocol was used to investigate the effect of slice thickness on noise and visibility in images. Different values of slice thickness 0.625, 1.25, 2.5, 3.75, 5, 7.5, and 10 were prepared, and then quantitative analysis was performed. Thinner slice thickness decreased image noise, increased visibility, and improved detection. Therefore, the balance between changing the thickness of the slice with the diagnostic content and image noise must be considered. Maximum slice thickness enhances CT image detail and structure despite more noise. Based on the results, a slice thickness of 1.25mm was identified as the optimal choice for reducing image noise and achieving better and more accurate detection using the single-source computed tomography head protocol. The study revealed that image noise tends to increase with greater slice thickness according to the Linux operating system. These findings can serve as a valuable guide for quality control methods in CT centers, emphasizing the need to determine the appropriate slice thickness to ensure an accurate diagnosis.

Invited Session II: Myopia and myopia control: Early functional changes in the myopic retina compromise emmetropization.

There is abundant evidence that emmetropization is controlled by visual experience, and that the retina is able to extract the information necessary to fine-tune axial eye growth during development. Emmetropization represents a closed-loop feedback system that uses defocus as error signal. It involves two pathways, one stimulating eye growth and the other restraining it. Both are different at several levels (1) different genes (2) different biochemical pathways and pharmacological interventions (3) different modes of retinal image processing. Knowing all this, the question arises why myopia does not limit itself and why undercorrection does not inhibit eye growth as expected from experiments in animal models. We found that only the emmetropic human retina can generate the growth-inhibiting signals when the focal plane is in front of the retina while the myopic retina has largely lost this ability. The functional deficit concerns retinal image processing, not the biochemical signaling cascades to choroid and sclera. Most recently, we found that the emmetropic human retina uses chromatic differences in focus to determine the sign of defocus. Again, we found that the myopic retina has lost this ability. The questions are now: (1) why and when occur the changes in the myopic retina that make myopia an open loop system, (2) what is the biological sense of this functional loss at a time when vision is otherwise normal (with correction) and (3) what are the underlying retinal circuits that seem to "give up"?

Award Lecture: 2021 Tillyer Award Lecture: The Early Visual Encoding: Computations and Consequences.

The visual system takes sensory measurements of the light incident at the eyes and uses these to make perceptual inferences about external world. The sensory measurements do not preserve all of the information available in the light signal. One approach to understanding the implications of the initial stages of visual processing is ideal observer analysis, which evaluates the information available to support psychophysical discriminations at various stages of the early visual representation. We are interested in extending this type of analysis to take into account the statistical structure of natural images. To do so, we developed an open-source computational model of the initial visual encoding, ISETBio (isetbio.org). ISETBio incorporates specification of visual displays, retinal image formation through the eye's optics, spatio-spectral sampling by the retinal cone mosaic, Poisson noise in the cone photopigment excitations, transduction of excitations to photocurrent, and fixational eye movements. In this talk, I will introduce ISETBio and illustrate a set of insights it enables about visual processing by reviewing a number of computational examples. These examples will include ways that combining ISETBio with Bayesian image-reconstruction methods helps us understand how the interaction of the visual encoding and the statistical structure of natural images shapes visual performance.

Deep Learning-Based Versus Iterative Image Reconstruction for Unenhanced Brain CT: A Quantitative Comparison of Image Quality.

This exploratory retrospective study aims to quantitatively compare the image quality of unenhanced brain computed tomography (CT) reconstructed with an iterative (AIDR-3D) and a deep learning-based (AiCE) reconstruction algorithm. After a preliminary phantom study, AIDR-3D and AiCE reconstructions (0.5 mm thickness) of 100 consecutive brain CTs acquired in the emergency setting on the same 320-detector row CT scanner were retrospectively analyzed, calculating image noise reduction attributable to the AiCE algorithm, artifact indexes in the posterior cranial fossa, and contrast-to-noise ratios (CNRs) at the cortical and thalamic levels. In the phantom study, the spatial resolution of the two datasets proved to be comparable; conversely, AIDR-3D reconstructions showed a broader noise pattern. In the human study, median image noise was lower with AiCE compared to AIDR-3D (4.7 vs. 5.3, p < 0.001, median 19.6% noise reduction), whereas AIDR-3D yielded a lower artifact index than AiCE (7.5 vs. 8.4, p < 0.001). AiCE also showed higher median CNRs at the cortical (2.5 vs. 1.8, p < 0.001) and thalamic levels (2.8 vs. 1.7, p < 0.001). These results highlight how image quality improvements granted by deep learning-based (AiCE) and iterative (AIDR-3D) image reconstruction algorithms vary according to different brain areas.

Quantitative evaluation of fiber network structure-property relationships in bacterial cellulose hydrogels.

The present study attempts to elucidate the network structure-property relationships of bacterial cellulose (BC) hydrogels comprising cellulose nanofibrils with favorable mechanical properties. To achieve this, it is necessary to establish a method based on quantitative evaluation of nanofibril network structure, rather than a simple application of classical polymer network theory. BC hydrogels with various network structures related to their mechanical properties were prepared from seven bacterial strains. The crosslink densities of the gels were determined quantitatively by a combination of fluorescence microscopy and image analysis. The tensile tests showed that the stress-strain curves of BC hydrogels exhibited strain hardening according to the power law for strain, and the power exponent had a linear relationship with the crosslink density. This result provides insight into the structure-property relationships of BC hydrogels, which could be used to inform quality control, process optimization, and high-throughput property prediction during manufacture.

Spatial transformation of multi-omics data unlocks novel insights into cancer biology.

The application of next-generation sequencing (NGS) has transformed cancer research. As costs have decreased, NGS has increasingly been applied to generate multiple layers of molecular data from the same samples, covering genomics, transcriptomics, and methylomics. Integrating these types of multi-omics data in a combined analysis is now becoming a common issue with no obvious solution, often handled on an ad hoc basis, with multi-omics data arriving in a tabular format and analyzed using computationally intensive statistical methods. These methods particularly ignore the spatial orientation of the genome and often apply stringent p-value corrections that likely result in the loss of true positive associations. Here, we present GENIUS (GEnome traNsformatIon and spatial representation of mUltiomicS data), a framework for integrating multi-omics data using deep learning models developed for advanced image analysis. The GENIUS framework is able to transform multi-omics data into images with genes displayed as spatially connected pixels and successfully extract relevant information with respect to the desired output. We demonstrate the utility of GENIUS by applying the framework to multi-omics datasets from the Cancer Genome Atlas. Our results are focused on predicting the development of metastatic cancer from primary tumors, and demonstrate how through model inference, we are able to extract the genes which are driving the model prediction and are likely associated with metastatic disease progression. We anticipate our framework to be a starting point and strong proof of concept for multi-omics data transformation and analysis without the need for statistical correction.

2DeteCT - A large 2D expandable, trainable, experimental Computed Tomography dataset for machine learning.

Recent research in computational imaging largely focuses on developing machine learning (ML) techniques for image reconstruction, which requires large-scale training datasets consisting of measurement data and ground-truth images. However, suitable experimental datasets for X-ray Computed Tomography (CT) are scarce, and methods are often developed and evaluated only on simulated data. We fill this gap by providing the community with a versatile, open 2D fan-beam CT dataset suitable for developing ML techniques for a range of image reconstruction tasks. To acquire it, we designed a sophisticated, semi-automatic scan procedure that utilizes a highly-flexible laboratory X-ray CT setup. A diverse mix of samples with high natural variability in shape and density was scanned slice-by-slice (5,000 slices in total) with high angular and spatial resolution and three different beam characteristics: A high-fidelity, a low-dose and a beam-hardening-inflicted mode. In addition, 750 out-of-distribution slices were scanned with sample and beam variations to accommodate robustness and segmentation tasks. We provide raw projection data, reference reconstructions and segmentations based on an open-source data processing pipeline.

As artificial intelligence goes multimodal, medical applications multiply.

Machines don't have eyes, but you wouldn't know that if you followed the progression of deep learning models for accurate interpretation of medical images, such as x-rays, computed tomography (CT) and magnetic resonance imaging (MRI) scans, pathology slides, and retinal photos. Over the past several years, there has been a torrent of studies that have consistently demonstrated how powerful "machine eyes" can be, not only compared with medical experts but also for detecting features in medical images that are not readily discernable by humans. For example, a retinal scan is rich with information that people can't see, but machines can, providing a gateway to multiple aspects of human physiology, including blood pressure; glucose control; risk of Parkinson's, Alzheimer's, kidney, and hepatobiliary diseases; and the likelihood of heart attacks and strokes. As a cardiologist, I would not have envisioned that machine interpretation of an electrocardiogram would provide information about the individual's age, sex, anemia, risk of developing diabetes or arrhythmias, heart function and valve disease, kidney, or thyroid conditions. Likewise, applying deep learning to a pathology slide of tumor tissue can also provide insight about the site of origin, driver mutations, structural genomic variants, and prognosis. Although these machine vision capabilities for medical image interpretation may seem impressive, they foreshadow what is potentially far more expansive terrain for artificial intelligence (AI) to transform medicine. The big shift ahead is the ability to transcend narrow, unimodal tasks, confined to images, and broaden machine capabilities to include text and speech, encompassing all input modes, setting the foundation for multimodal AI.

Validation of deep learning-based CT image reconstruction for treatment planning.

Deep learning-based CT image reconstruction (DLR) is a state-of-the-art method for obtaining CT images. This study aimed to evaluate the usefulness of DLR in radiotherapy. Data were acquired using a large-bore CT system and an electron density phantom for radiotherapy. We compared the CT values, image noise, and CT value-to-electron density conversion table of DLR and hybrid iterative reconstruction (H-IR) for various doses. Further, we evaluated three DLR reconstruction strength patterns (Mild, Standard, and Strong). The variations of CT values of DLR and H-IR were large at low doses, and the difference in average CT values was insignificant with less than 10 HU at doses of 100 mAs and above. DLR showed less change in CT values and smaller image noise relative to H-IR. The noise-reduction effect was particularly large in the low-dose region. The difference in image noise between DLR Mild and Standard/Strong was large, suggesting the usefulness of reconstruction intensities higher than Mild. DLR showed stable CT values and low image noise for various materials, even at low doses; particularly for Standard or Strong, the reduction in image noise was significant. These findings indicate the usefulness of DLR in treatment planning using large-bore CT systems.

Nuclear expression of Ku70/80 is associated with CHEK2 germline mutations in breast cancer.

Ku70/80 protein inhibitors reduce the repair of DNA double-strand breaks via the Ku70/80 pathway, so they can be used to treat cancers with Ku70/80 overexpression. Since the association of Ku70/80 with germline CHEK2 mutations in breast cancer is unknown, in this study we evaluated the expression of Ku70/80 in breast cancers with germline CHEK2 mutations. Immunohistochemistry with a Ku70/80 antibody on tissue microarrays from 225 CHEK2-associated breast cancers was used and automatically assessed with computerized image analysis. We report that the vast majority of breast cancers expressed high level of nuclear Ku70/80 and a small percentage of tumors (3.5%) were negative for Ku70/80 expression. There was a significant difference between the nuclear Ku70/80 expression in CHEK2-associated vs. CHEK2-non-associated breast cancers in all tumors (p = 0.009), and in the estrogen receptor (ER) positive subgroup of breast cancers (p = 0.03). This study is the first reporting an association of Ku70/80 expression with CHEK2 germline mutations in breast cancer. The results suggest that evaluation of Ku70/80 expression in breast cancer may improve the selection of breast cancer patients for Ku70/80 inhibitor therapy, and point to CHEK2-associated breast cancer and a subset of ER-positive breast cancer as potential suitable targets for such therapy.

Improved breast ultrasound tumor classification using dual-input CNN with GAP-guided attention loss.

Ultrasonography is a widely used medical imaging technique for detecting breast cancer. While manual diagnostic methods are subject to variability and time-consuming, computer-aided diagnostic (CAD) methods have proven to be more efficient. However, current CAD approaches neglect the impact of noise and artifacts on the accuracy of image analysis. To enhance the precision of breast ultrasound image analysis for identifying tissues, organs and lesions, we propose a novel approach for improved tumor classification through a dual-input model and global average pooling (GAP)-guided attention loss function. Our approach leverages a convolutional neural network with transformer architecture and modifies the single-input model for dual-input. This technique employs a fusion module and GAP operation-guided attention loss function simultaneously to supervise the extraction of effective features from the target region and mitigate the effect of information loss or redundancy on misclassification. Our proposed method has three key features: (i) ResNet and MobileViT are combined to enhance local and global information extraction. In addition, a dual-input channel is designed to include both attention images and original breast ultrasound images, mitigating the impact of noise and artifacts in ultrasound images. (ii) A fusion module and GAP operation-guided attention loss function are proposed to improve the fusion of dual-channel feature information, as well as supervise and constrain the weight of the attention mechanism on the fused focus region. (iii) Using the collected uterine fibroid ultrasound dataset to train ResNet18 and load the pre-trained weights, our experiments on the BUSI and BUSC public datasets demonstrate that the proposed method outperforms some state-of-the-art methods. The code will be publicly released at https://github.com/425877/Improved-Breast-Ultrasound-Tumor-Classification.

pomBseen: An automated pipeline for analysis of fission yeast images.

Fission yeast is a model organism widely used for studies of eukaryotic cell biology. As such, it is subject to bright-field and fluorescent microscopy. Manual analysis of such data can be laborious and subjective. Therefore, we have developed pomBseen, an image analysis pipeline for the quantitation of fission yeast micrographs containing a bright-field channel and up to two fluorescent channels. It accepts a wide range of image formats and produces a table with the size and total and nuclear fluorescent intensities of the cells in the image. Benchmarking of the pipeline against manually annotated datasets demonstrates that it reliably segments cells and acquires their image parameters. Written in MATLAB, pomBseen is also available as a standalone application.

The implementation of artificial intelligence to the low-cost metaphase finder.

Biological dosimetry is used to estimate one's dose by biological phenomena. The most popular and 'gold standard' phenomenon is the appearance of dicentric chromosomes in metaphases. The metaphase finder is a tool for biological dosimetry that finds metaphase cells on glass slides. It consists of an automated microscope, auto-focus system, X-Y stage, camera and computer. It does the image analysis of the microscopic images of the glass slides and displays the positions of metaphase cells. In this paper, the author reported that the prototype of AI-implemented metaphase finder was combined with the microscope system by file sharing and image transfer program, and that the metaphase finder system's accuracy was compared with previous non-AI system, using the same samples.

Development of high-throughput systems for biodosimetry.

Biomarkers for ionising radiation exposure have great utility in scenarios where there has been a potential exposure and physical dosimetry is missing or in dispute, such as for occupational and accidental exposures. Biomarkers that respond as a function of dose are particularly useful as biodosemeters to determine the dose of radiation to which an individual has been exposed. These dose measurements can also be used in medical scenarios to track doses from medical exposures and even have the potential to identify an individual's response to radiation exposure that could help tailor treatments. The measurement of biomarkers of exposure in medicine and for accidents, where a larger number of samples would be required, is limited by the throughput of analysis (i.e. the number of samples that could be processed and analysed), particularly for microscope-based methods, which tend to be labour-intensive. Rapid analysis in an emergency scenario, such as a large-scale accident, would provide dose estimates to medical practitioners, allowing timely administration of the appropriate medical countermeasures to help mitigate the effects of radiation exposure. In order to improve sample throughput for biomarker analysis, much effort has been devoted to automating the process from sample preparation through automated image analysis. This paper will focus mainly on biological endpoints traditionally analysed by microscopy, specifically dicentric chromosomes, micronuclei and gamma-H2AX. These endpoints provide examples where sample throughput has been improved through automated image acquisition, analysis of images acquired by microscopy, as well as methods that have been developed for analysis using imaging flow cytometry.

Sensitivity of standardised radiomics algorithms to mask generation across different software platforms.

The field of radiomics continues to converge on a standardised approach to image processing and feature extraction. Conventional radiomics requires a segmentation. Certain features can be sensitive to small contour variations. The industry standard for medical image communication stores contours as coordinate points that must be converted to a binary mask before image processing can take place. This study investigates the impact that the process of converting contours to mask can have on radiomic features calculation. To this end we used a popular open dataset for radiomics standardisation and we compared the impact of masks generated by importing the dataset into 4 medical imaging software. We interfaced our previously standardised radiomics platform with these software using their published application programming interface to access image volume, masks and other data needed to calculate features. Additionally, we used super-sampling strategies to systematically evaluate the impact of contour data pre processing methods on radiomic features calculation. Finally, we evaluated the effect that using different mask generation approaches could have on patient clustering in a multi-center radiomics study. The study shows that even when working on the same dataset, mask and feature discrepancy occurs depending on the contour to mask conversion technique implemented in various medical imaging software. We show that this also affects patient clustering and potentially radiomic-based modelling in multi-centre studies where a mix of mask generation software is used. We provide recommendations to negate this issue and facilitate reproducible and reliable radiomics.

CTA-UNet: CNN-transformer architecture UNet for dental CBCT images segmentation.

In view of the limitations of current deep learning models in segmenting dental cone-beam computed tomography (CBCT) images, specifically dealing with complex root morphological features, fuzzy boundaries between tooth roots and alveolar bone, and the need for costly annotation of dental CBCT images. We collected dental CBCT data from 200 patients and annotated 45 of them for network training, and proposed a CNN-Transformer Architecture UNet network, which combines the advantages of CNN and Transformer. The CNN component effectively extracts local features, while the Transformer captures long-range dependencies. Multiple spatial attention modules were included to enhance the network's ability to extract and represent spatial information. Additionally, we introduced a novel Masked image modeling method to pre-train the CNN and Transformer modules simultaneously, mitigating limitations due to a smaller amount of labeled training data. Experimental results demonstrate that the proposed method achieved superior performance (DSC of 87.12%, IoU of 78.90%, HD95 of 0.525 mm, ASSD of 0.199 mm), and provides a more efficient and effective approach to automatically and accurately segment dental CBCT images, has real-world applicability in orthodontics and dental implants.

MCSF-Net: a multi-scale channel spatial fusion network for real-time polyp segmentation.

Colorectal cancer is a globally prevalent cancer type that necessitates prompt screening. Colonoscopy is the established diagnostic technique for identifying colorectal polyps. However, missed polyp rates remain a concern. Early detection of polyps, while still precancerous, is vital for minimizing cancer-related mortality and economic impact. In the clinical setting, precise segmentation of polyps from colonoscopy images can provide valuable diagnostic and surgical information. Recent advances in computer-aided diagnostic systems, specifically those based on deep learning techniques, have shown promise in improving the detection rates of missed polyps, and thereby assisting gastroenterologists in improving polyp identification. In the present investigation, we introduce MCSF-Net, a real-time automatic segmentation framework that utilizes a multi-scale channel space fusion network. The proposed architecture leverages a multi-scale fusion module in conjunction with spatial and channel attention mechanisms to effectively amalgamate high-dimensional multi-scale features. Additionally, a feature complementation module is employed to extract boundary cues from low-dimensional features, facilitating enhanced representation of low-level features while keeping computational complexity to a minimum. Furthermore, we incorporate shape blocks to facilitate better model supervision for precise identification of boundary features of polyps. Our extensive evaluation of the proposed MCSF-Net on five publicly available benchmark datasets reveals that it outperforms several existing state-of-the-art approaches with respect to different evaluation metrics. The proposed approach runs at an impressive ∼45 FPS, demonstrating notable advantages in terms of scalability and real-time segmentation.

Ceusia-Breast: computer-aided diagnosis with contrast enhanced ultrasound image analysis for breast lesions.

BACKGROUND: In recent years, contrast-enhanced ultrasonography (CEUS) has been used for various applications in breast diagnosis. The superiority of CEUS over conventional B-mode imaging in the ultrasound diagnosis of the breast lesions in clinical practice has been widely confirmed. On the other hand, there have been many proposals for computer-aided diagnosis of breast lesions on B-mode ultrasound images, but few for CEUS. We propose a semi-automatic classification method based on machine learning in CEUS of breast lesions. METHODS: The proposed method extracts spatial and temporal features from CEUS videos and breast tumors are classified as benign or malignant using linear support vector machines (SVM) with combination of selected optimal features. In the proposed method, tumor regions are extracted using the guidance information specified by the examiners, then morphological and texture features of tumor regions obtained from B-mode and CEUS images and TIC features obtained from CEUS video are extracted. Then, our method uses SVM classifiers to classify breast tumors as benign or malignant. During SVM training, many features are prepared, and useful features are selected. We name our proposed method "Ceucia-Breast" (Contrast Enhanced UltraSound Image Analysis for BREAST lesions). RESULTS: The experimental results on 119 subjects show that the area under the receiver operating curve, accuracy, precision, and recall are 0.893, 0.816, 0.841 and 0.920, respectively. The classification performance is improved by our method over conventional methods using only B-mode images. In addition, we confirm that the selected features are consistent with the CEUS guidelines for breast tumor diagnosis. Furthermore, we conduct an experiment on the operator dependency of specifying guidance information and find that the intra-operator and inter-operator kappa coefficients are 1.0 and 0.798, respectively. CONCLUSION: The experimental results show a significant improvement in classification performance compared to conventional classification methods using only B-mode images. We also confirm that the selected features are related to the findings that are considered important in clinical practice. Furthermore, we verify the intra- and inter-examiner correlation in the guidance input for region extraction and confirm that both correlations are in strong agreement.