Search Results (21)

Adoptive cell therapies have transformed the treatment landscape for hematologic malignancies. Yet, translation to solid tumors remains constrained by antigen heterogeneity, an immunosuppressive tumor microenvironment (TME), and poor persistence of conventional CAR-T cells. In response, innate immune cell platforms, particularly chimeric antigen receptor–engineered natural killer (CAR-NK) cells and chimeric antigen receptor–macrophages (CAR-MΦ), have emerged as promising alternatives. This review summarizes recent advances in the design and application of CAR-NK and CAR-MΦ therapies for solid tumors. We highlight key innovations, including the use of lineage-specific intracellular signaling domains (e.g., DAP12, 2B4, FcRγ), novel effector constructs (e.g., NKG7-overexpressing CARs, TME-responsive CARs), and scalable induced pluripotent stem cell (iPSC)-derived platforms. Preclinical data support enhanced antitumor activity through mechanisms such as major histocompatibility complex (MHC)-unrestricted cytotoxicity, phagocytosis, trogocytosis, cytokine secretion, and cross-talk with adaptive immunity. Early-phase clinical studies (e.g., CT-0508) demonstrate feasibility and TME remodeling with CAR-MΦ. However, persistent challenges remain, including transient in vivo survival, manufacturing complexity, and risks of off-target inflammation. Emerging combinatorial strategies, such as dual-effector regimens (CAR-NK⁺ CAR-MΦ), cytokine-modulated cross-support, and bispecific or logic-gated CARs, may overcome these barriers and provide more durable, tumor-selective responses. Taken together, CAR-NK and CAR-MΦ platforms are poised to expand the reach of engineered cell therapy into the solid tumor domain. Full article

The HepACAGA (Hepatic Arteriography and C-arm CT-Guided Ablation) technique, which integrates C-arm CT guidance with transcatheter C-arm CT hepatic arteriography (C-arm CTHA), significantly improves liver tumor ablation outcomes by enhancing tumor visualization, navigation, and the intraprocedural assessment of ablation margins. The two key advantages of using C-arm CT over conventional CT for image guidance are firstly that the entire procedure can be performed in the angiography suite, eliminating the need for patient transfer between the angiography suite (catheterization) and CT-room (ablation), and secondly, that integrated C-arm needle guidance software can greatly reduce the difficulty of needle placement. Beyond these advantages, the HepACAGA technique offers additional benefits across four domains: (1) the direct conversion of ablation to intra-arterial liver-directed therapies (e.g., radioembolization or chemoembolization) upon the intraprocedural detection of disease progression; (2) the direct combination of ablation with intra-arterial treatments or portal vein embolization in one session; (3) the enhanced ablation effect through heat sink effect reduction with adjunct bland embolization or balloon occlusion; and (4) the immediate hemorrhage control through direct embolization. This pictorial essay demonstrates the advantages of combining C-arm CT guidance with real-time C-arm CTHA in the percutaneous thermal ablation of liver tumors, with clinical cases illustrating each of the aforementioned four key domains. Full article

Italian ryegrass (Lolium multiflorum Lam.) is a persistent weed species that poses significant management challenges in key agricultural crops such as wheat, corn, cotton, and soybean. This study investigated the prevalence of resistance to ACCase inhibitor herbicides, specifically diclofop and pinoxaden, among field-collected Italian ryegrass populations. The survey revealed widespread resistance to diclofop and emerging cross-resistance to pinoxaden. To elucidate the physiological mechanism of ACCase herbicide resistance, we investigated mutations in the carboxyl-transferase (CT) domain of the ACCase enzyme, a critical region for herbicide sensitivity. Using dCAPS assays and CT domain sequencing, several known resistance-conferring mutations were detected in diclofop survivors, including I1781L, W2027C, I2041N, D2078G, and C2088R. Additionally, other mutations such as L1701M, E1874A, N1878H, G1946E/Q, V1992D, and E2039D were identified. To understand the functional role of these mutations in herbicide resistance, homology modeling was performed using AutoDock Vina for selected mutation combinations. The computational analysis revealed that all mutations and their combinations resulted in reduced binding affinity with diclofop and pinoxaden compared to the wild-type ACCase CT domain. Computational binding energy predictions indicated that the G1946E mutation and the L1701M + I1781L + E1874A + N1878H combination exhibited the lowest affinities for diclofop and pinoxaden, respectively. This study provides valuable insights into the molecular basis of ACCase inhibitor resistance in Italian ryegrass. However, further research is needed to validate the functional significance of each new substitution and its combinations in conferring herbicide resistance. Full article

Deep learning (DL) algorithms used for DOTATATE PET lesion detection typically require large, well-annotated training datasets. These are difficult to obtain due to low incidence of gastroenteropancreatic neuroendocrine tumors (GEP-NETs) and the high cost of manual annotation. Furthermore, networks trained and tested with data acquired from site specific PET/CT instrumentation, acquisition and processing protocols have reduced performance when tested with offsite data. This lack of generalizability requires even larger, more diverse training datasets. The objective of this study is to investigate the feasibility of improving DL algorithm performance by better matching the background noise in training datasets to higher noise, out-of-domain testing datasets. ⁶⁸Ga-DOTATATE PET/CT datasets were obtained from two scanners: Scanner1, a state-of-the-art digital PET/CT (GE DMI PET/CT; n = 83 subjects), and Scanner2, an older-generation analog PET/CT (GE STE; n = 123 subjects). Set1, the data set from Scanner1, was reconstructed with standard clinical parameters (5 min; Q.Clear) and list-mode reconstructions (VPFXS 2, 3, 4, and 5-min). Set2, data from Scanner2 representing out-of-domain clinical scans, used standard iterative reconstruction (5 min; OSEM). A deep neural network was trained with each dataset: Network1 for Scanner1 and Network2 for Scanner2. DL performance (Network1) was tested with out-of-domain test data (Set2). To evaluate the effect of training sample size, we tested DL model performance using a fraction (25%, 50% and 75%) of Set1 for training. Scanner1, list-mode 2-min reconstructed data demonstrated the most similar noise level compared that of Set2, resulting in the best performance (F₁ = 0.713). This was not significantly different compared to the highest performance, upper-bound limit using in-domain training for Network2 (F₁ = 0.755; p-value = 0.103). Regarding sample size, the F1 score significantly increased from 25% training data (F₁ = 0.478) to 100% training data (F₁ = 0.713; p < 0.001). List-mode data from modern PET scanners can be reconstructed to better match the noise properties of older scanners. Using existing data and their associated annotations dramatically reduces the cost and effort in generating these datasets and significantly improves the performance of existing DL algorithms. List-mode reconstructions can provide an efficient, low-cost method to improve DL algorithm generalizability. Full article

The combined effects of temperature (from −80 °C to +80 °C) and 100 kV X-ray exposure (up to 108 kGy(SiO₂)) on the physical properties of Brillouin scattering and losses in three differently doped silica-based optical fibers, with varying dopant type and concentration (4 wt%(Ge), 10 wt%(Ge) and 1 wt%(F)), are experimentally studied in this work. The dependencies of Brillouin Frequency Shifts (BFS), Radiation-Induced Attenuation (RIA) levels, Brillouin gain attenuation, Brillouin frequency temperature (C_T) and strain (C_ε) sensitivity coefficients are studied under X-rays in a wide temperature range [−80 °C; +80 °C]. Brillouin sensing capabilities are investigated using a Brillouin Optical Time Domain Analyzer (BOTDA), and several properties are reported: (i) similar behavior of the Brillouin gain amplitude decrease with the increase in the RIA; (ii) the F-doped and heavily Ge-doped fibers do not exhibit a temperature dependence under radiation for their responses in Brillouin gain losses. Increasing Ge dopant concentration also reduces the irradiation temperature effect on RIA. In addition, Radiation-Induced Brillouin Frequency Shift (RI-BFS) manifests a slightly different behavior for lower temperatures than RIA, presenting an opportunity for a comprehensive understanding of RI-BFS origins. Related temperature and strain sensors are designed for harsh environments over an extended irradiation temperature range, which is useful for a wide range of applications. Full article

One of the early manifestations of systemic atherosclerosis, which leads to blood circulation issues, is the enhanced arterial light reflex (EALR). Fundus images are commonly used for regular screening purposes to intervene and assess the severity of systemic atherosclerosis in a timely manner. However, there is a lack of automated methods that can meet the demands of large-scale population screening. Therefore, this study introduces a novel cross-scale transformer-based multi-instance learning method, named MIL-CT, for the detection of early arterial lesions (e.g., EALR) in fundus images. MIL-CT utilizes the cross-scale vision transformer to extract retinal features in a multi-granularity perceptual domain. It incorporates a multi-head cross-scale attention fusion module to enhance global perceptual capability and feature representation. By integrating information from different scales and minimizing information loss, the method significantly improves the performance of the EALR detection task. Furthermore, a multi-instance learning module is implemented to enable the model to better comprehend local details and features in fundus images, facilitating the classification of patch tokens related to retinal lesions. To effectively learn the features associated with retinal lesions, we utilize weights pre-trained on a large fundus image Kaggle dataset. Our validation and comparison experiments conducted on our collected EALR dataset demonstrate the effectiveness of the MIL-CT method in reducing generalization errors while maintaining efficient attention to retinal vascular details. Moreover, the method surpasses existing models in EALR detection, achieving an accuracy, precision, sensitivity, specificity, and F1 score of 97.62%, 97.63%, 97.05%, 96.48%, and 97.62%, respectively. These results exhibit the significant enhancement in diagnostic accuracy of fundus images brought about by the MIL-CT method. Thus, it holds potential for various applications, particularly in the early screening of cardiovascular diseases such as hypertension and atherosclerosis. Full article

The human fovea is a specialized pit structure in the central retina. Foveal hypoplasia is a condition where the foveal pit does not fully develop, and it is associated with poor vision. Autosomal dominant isolated foveal hypoplasia (FVH1) is a rare condition of foveal hypoplasia (FH) that lacks any other ocular manifestations. FVH1 is associated with hypomorphic mutations in the PAX6 gene that encodes a sequence-specific DNA-binding transcription factor for morphogenesis and evolution of the eye. We report our findings in 17 patients with PAX6 mutations associated with FVH1 or FH with aniridia and corneal opacities. Patients with three mutations, p.V78E, p.V83F and p.R128H, in the C-terminal subdomain of the paired domain (CTS) consistently have severe FH. Luciferase assays for a single reporter containing a representative PAX6 binding site indicated that the transcriptional activities of these mutations were significantly reduced, comparable to that of the truncation mutation of p.G65Rfs*5. Patients with p.P20S in the N-terminal subdomain of the paired domain, and a patient with p.N365K in the proline-serine-threonine-rich domain (PSTD) had mild FH. A patient with p.Q255L in the homeodomain had severe FH. The P20S and Q255L mutants did not affect the transcriptional activity. Mutant N365K has a retained DNA-binding activity but a reduced transcriptional activity, due to a low PSTD transactivation. These findings demonstrated that mutations associated with FVH1 underlie a functional divergence between DNA-binding ability and transcriptional activity. We conclude that a wide range of mutations in the PAX6 gene is not limited to the CST region and are responsible for FVH1. Full article

Medical image analysis plays an important role in clinical diagnosis. In this paper, we examine the recent Segment Anything Model (SAM) on medical images, and report both quantitative and qualitative zero-shot segmentation results on nine medical image segmentation benchmarks, covering various imaging modalities, such as optical coherence tomography (OCT), magnetic resonance imaging (MRI), and computed tomography (CT), as well as different applications including dermatology, ophthalmology, and radiology. Those benchmarks are representative and commonly used in model development. Our experimental results indicate that while SAM presents remarkable segmentation performance on images from the general domain, its zero-shot segmentation ability remains restricted for out-of-distribution images, e.g., medical images. In addition, SAM exhibits inconsistent zero-shot segmentation performance across different unseen medical domains. For certain structured targets, e.g., blood vessels, the zero-shot segmentation of SAM completely failed. In contrast, a simple fine-tuning of it with a small amount of data could lead to remarkable improvement of the segmentation quality, showing the great potential and feasibility of using fine-tuned SAM to achieve accurate medical image segmentation for a precision diagnostics. Our study indicates the versatility of generalist vision foundation models on medical imaging, and their great potential to achieve desired performance through fine-turning and eventually address the challenges associated with accessing large and diverse medical datasets in support of clinical diagnostics. Full article

A variety of medical computer vision applications analyze 2D slices of computed tomography (CT) scans, whereas axial slices from the body trunk region are usually identified based on their relative position to the spine. A limitation of such systems is that either the correct slices must be extracted manually or labels of the vertebrae are required for each CT scan to develop an automated extraction system. In this paper, we propose an unsupervised domain adaptation (UDA) approach for vertebrae detection and identification based on a novel Domain Sanity Loss (DSL) function. With UDA the model’s knowledge learned on a publicly available (source) data set can be transferred to the target domain without using target labels, where the target domain is defined by the specific setup (CT modality, study protocols, applied pre- and processing) at the point of use (e.g., a specific clinic with its specific CT study protocols). With our approach, a model is trained on the source and target data set in parallel. The model optimizes a supervised loss for labeled samples from the source domain and the DSL loss function based on domain-specific “sanity checks” for samples from the unlabeled target domain. Without using labels from the target domain, we are able to identify vertebra centroids with an accuracy of $72.8$%. By adding only ten target labels during training the accuracy increases to $89.2$%, which is on par with the current state-of-the-art for full supervised learning, while using about 20 times less labels. Thus, our model can be used to extract 2D slices from 3D CT scans on arbitrary data sets fully automatically without requiring an extensive labeling effort, contributing to the clinical adoption of medical imaging by hospitals. Full article

Duplication and deletion of the peripheral myelin protein 22 (PMP22) gene cause Charcot-Marie-Tooth disease type 1A (CMT1A) and hereditary neuropathy with liability to pressure palsies (HNPP), respectively, while point mutations or small insertions and deletions (indels) usually cause CMT type 1E (CMT1E) or HNPP. This study was performed to identify PMP22 mutations and to analyze the genotype–phenotype correlation in Korean CMT families. By the application of whole-exome sequencing (WES) and targeted gene panel sequencing (TS), we identified 14 pathogenic or likely pathogenic PMP22 mutations in 21 families out of 850 CMT families who were negative for 17p12 (PMP22) duplication. Most mutations were located in the well-conserved transmembrane domains. Of these, eight mutations were not reported in other populations. High frequencies of de novo mutations were observed, and the mutation sites of c.68C>G and c.215C>T were suggested as the mutational hotspots. Affected individuals showed an early onset-severe phenotype and late onset-mild phenotype, and more than 40% of the CMT1E patients showed hearing loss. Physical and electrophysiological symptoms of the CMT1E patients were more severely damaged than those of CMT1A while similar to CMT1B caused by MPZ mutations. Our results will be useful for the reference data of Korean CMT1E and the molecular diagnosis of CMT1 with or without hearing loss. Full article

Chlamydia trachomatis (Ct) is the most common bacterial sexual transmitted pathogen, yet a vaccine is not currently available. Here, we used the immunogenic bacteriophage MS2 virus-like particle (VLP) technology to engineer vaccines against the Ct major outer membrane protein variable domain 4 (MOMP-VD4), which contains a conserved neutralizing epitope (TTLNPTIAG). A previously described monoclonal antibody to the MOMP-VD4 (E4 mAb) is capable of neutralizing all urogenital Ct serovars and binds this core epitope, as well as several non-contiguous amino acids. This suggests that this core epitope may require conformational context in order to elicit neutralizing antibodies to Ct. In order to identify immunogens that could elicit neutralizing antibodies to the TTLNPTIAG epitope, we used two approaches. First, we used affinity selection with a bacteriophage MS2-VLP library displaying random peptides in a constrained, surface-exposed loop to identify potential E4 mAb mimotopes. After four rounds of affinity selection, we identified a VLP-displayed peptide (HMVGSTKWTN) that could bind to the E4 mAb and elicited serum IgG that bound weakly to Ct elementary bodies by ELISA. Second, two versions of the core conserved TTLNPTIAG epitope (TTLNPTIAG and TTLNPTIAGA) were recombinantly expressed on the coat protein of the MS2 VLP in a constrained, surface-exposed loop. Mouse immune sera IgG bound to Ct elementary bodies by ELISA. Immunization with these MS2 VLPs provided protection from vaginal Chlamydia infection in a murine challenge model. These data suggest that short peptide epitopes targeting the MOMP-VD4 could be appropriate for Ct vaccine design when displayed on an immunogenic bacteriophage VLP vaccine platform. Full article

Connexin (Cx43)-formed channels have been linked to cardiac arrhythmias and diseases of the heart associated with myocardial tissue loss and fibrosis. These pathologies include ischemic heart disease, ischemia-reperfusion injury, heart failure, hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, and Duchenne muscular dystrophy. A number of Cx43 mimetic peptides have been reported as therapeutic candidates for targeting disease processes linked to Cx43, including some that have advanced to clinical testing in humans. These peptides include Cx43 sequences based on the extracellular loop domains (e.g., Gap26, Gap 27, and Peptide5), cytoplasmic-loop domain (Gap19 and L2), and cytoplasmic carboxyl-terminal domain (e.g., JM2, Cx43tat, CycliCX, and the alphaCT family of peptides) of this transmembrane protein. Additionally, RYYN peptides binding to the Cx43 carboxyl-terminus have been described. In this review, we survey preclinical and clinical data available on short mimetic peptides based on, or directly targeting, Cx43, with focus on their potential for treating heart disease. We also discuss problems that have caused reluctance within the pharmaceutical industry to translate peptidic therapeutics to the clinic, even when supporting preclinical data is strong. These issues include those associated with the administration, stability in vivo, and tissue penetration of peptide-based therapeutics. Finally, we discuss novel drug delivery technologies including nanoparticles, exosomes, and other nanovesicular carriers that could transform the clinical and commercial viability of Cx43-targeting peptides in treatment of heart disease, stroke, cancer, and other indications requiring oral or parenteral administration. Some of these newly emerging approaches to drug delivery may provide a path to overcoming pitfalls associated with the drugging of peptide therapeutics. Full article

Training medical image analysis models traditionally requires large amounts of expertly annotated imaging data which is time-consuming and expensive to obtain. One solution is to automatically extract scan-level labels from radiology reports. Previously, we showed that, by extending BERT with a per-label attention mechanism, we can train a single model to perform automatic extraction of many labels in parallel. However, if we rely on pure data-driven learning, the model sometimes fails to learn critical features or learns the correct answer via simplistic heuristics (e.g., that “likely” indicates positivity), and thus fails to generalise to rarer cases which have not been learned or where the heuristics break down (e.g., “likely represents prominent VR space or lacunar infarct” which indicates uncertainty over two differential diagnoses). In this work, we propose template creation for data synthesis, which enables us to inject expert knowledge about unseen entities from medical ontologies, and to teach the model rules on how to label difficult cases, by producing relevant training examples. Using this technique alongside domain-specific pre-training for our underlying BERT architecture i.e., PubMedBERT, we improve F1 micro from 0.903 to 0.939 and F1 macro from 0.512 to 0.737 on an independent test set for 33 labels in head CT reports for stroke patients. Our methodology offers a practical way to combine domain knowledge with machine learning for text classification tasks. Full article

Medical image segmentation has gained greater attention over the past decade, especially in the field of image-guided surgery. Here, robust, accurate and fast segmentation tools are important for planning and navigation. In this work, we explore the Convolutional Neural Network (CNN) based approaches for multi-dataset segmentation from CT examinations. We hypothesize that selection of certain parameters in the network architecture design critically influence the segmentation results. We have employed two different CNN architectures, 3D-UNet and VGG-16, given that both networks are well accepted in the medical domain for segmentation tasks. In order to understand the efficiency of different parameter choices, we have adopted two different approaches. The first one combines different weight initialization schemes with different activation functions, whereas the second approach combines different weight initialization methods with a set of loss functions and optimizers. For evaluation, the 3D-UNet was trained with the Medical Segmentation Decathlon dataset and VGG-16 using LiTS data. The quality assessment done using eight quantitative metrics enhances the probability of using our proposed strategies for enhancing the segmentation results. Following a systematic approach in the evaluation of the results, we propose a few strategies that can be adopted for obtaining good segmentation results. Both of the architectures used in this work were selected on the basis of general acceptance in segmentation tasks for medical images based on their promising results compared to other state-of-the art networks. The highest Dice score obtained in 3D-UNet for the liver, pancreas and cardiac data was 0.897, 0.691 and 0.892. In the case of VGG-16, it was solely developed to work with liver data and delivered a Dice score of 0.921. From all the experiments conducted, we observed that two of the combinations with Xavier weight initialization (also known as Glorot), Adam optimiser, Cross Entropy loss ($Gl{o}_{CE}^{Adam}$) and LeCun weight initialization, cross entropy loss and Adam optimiser $Le{c}_{CE}^{Adam}$ worked best for most of the metrics in a 3D-UNet setting, while Xavier together with cross entropy loss and Tanh activation function ($Gl{o}_{CE}^{tanh}$) worked best for the VGG-16 network. Here, the parameter combinations are proposed on the basis of their contributions in obtaining optimal outcomes in segmentation evaluations. Moreover, we discuss that the preliminary evaluation results show that these parameters could later on be used for gaining more insights into model convergence and optimal solutions.The results from the quality assessment metrics and the statistical analysis validate our conclusions and we propose that the presented work can be used as a guide in choosing parameters for the best possible segmentation results for future works. Full article

Pathogenic microorganisms produce various virulence factors, e.g., enzymes, cytotoxins, effectors, which trigger development of pathologies in infectious diseases. Cholera toxin (CT) produced by O1 and O139 serotypes of Vibrio cholerae (V. cholerae) is a major cytotoxin causing severe diarrhea. Cholix cytotoxin (Cholix) was identified as a novel eukaryotic elongation factor 2 (eEF2) adenosine-diphosphate (ADP)-ribosyltransferase produced mainly in non-O1/non-O139 V. cholerae. The function and role of Cholix in infectious disease caused by V. cholerae remain unknown. The crystal structure of Cholix is similar to Pseudomonas exotoxin A (PEA) which is composed of an N-terminal receptor-recognition domain and a C-terminal ADP-ribosyltransferase domain. The endocytosed Cholix catalyzes ADP-ribosylation of eEF2 in host cells and inhibits protein synthesis, resulting in cell death. In a mouse model, Cholix caused lethality with severe liver damage. In this review, we describe the mechanism underlying Cholix-induced cytotoxicity. Cholix-induced apoptosis was regulated by mitogen-activated protein kinase (MAPK) and protein kinase C (PKC) signaling pathways, which dramatically enhanced tumor necrosis factor-α (TNF-α) production in human liver, as well as the amount of epithelial-like HepG2 cancer cells. In contrast, Cholix induced apoptosis in hepatocytes through a mitochondrial-dependent pathway, which was not stimulated by TNF-α. These findings suggest that sensitivity to Cholix depends on the target cell. A substantial amount of information on PEA is provided in order to compare/contrast this well-characterized mono-ADP-ribosyltransferase (mART) with Cholix. Full article