s2MRI-ADNet: an interpretable deep learning framework integrating Euclidean-graph representations of Alzheimer's disease solely from structural MRI.

OBJECTIVE: To establish a multi-dimensional representation solely on structural MRI (sMRI) for early diagnosis of AD. METHODS: A total of 3377 participants' sMRI from four independent databases were retrospectively identified to construct an interpretable deep learning model that integrated multi-dimensional representations of AD solely on sMRI (called s2MRI-ADNet) by a dual-channel learning strategy of gray matter volume (GMV) from Euclidean space and the regional radiomics similarity network (R2SN) from graph space. Specifically, the GMV feature map learning channel (called GMV-Channel) was to take into consideration spatial information of both long-range spatial relations and detailed localization information, while the node feature and connectivity strength learning channel (called NFCS-Channel) was to characterize the graph-structured R2SN network by a separable learning strategy. RESULTS: The s2MRI-ADNet achieved a superior classification accuracy of 92.1% and 91.4% under intra-database and inter-database cross-validation. The GMV-Channel and NFCS-Channel captured complementary group-discriminative brain regions, revealing a complementary interpretation of the multi-dimensional representation of brain structure in Euclidean and graph spaces respectively. Besides, the generalizable and reproducible interpretation of the multi-dimensional representation in capturing complementary group-discriminative brain regions revealed a significant correlation between the four independent databases (p < 0.05). Significant associations (p < 0.05) between attention scores and brain abnormality, between classification scores and clinical measure of cognitive ability, CSF biomarker, metabolism, and genetic risk score also provided solid neurobiological interpretation. CONCLUSION: The s2MRI-ADNet solely on sMRI could leverage the complementary multi-dimensional representations of AD in Euclidean and graph spaces, and achieved superior performance in the early diagnosis of AD, facilitating its potential in both clinical translation and popularization.

Jointly constrained group sparse connectivity representation improves early diagnosis of Alzheimer's disease on routinely acquired T1-weighted imaging-based brain network.

Background: Radiomics-based morphological brain networks (radMBN) constructed from routinely acquired structural MRI (sMRI) data have gained attention in Alzheimer's disease (AD). However, the radMBN suffers from limited characterization of AD because sMRI only characterizes anatomical changes and is not a direct measure of neuronal pathology or brain activity. Purpose: To establish a group sparse representation of the radMBN under a joint constraint of group-level white matter fiber connectivity and individual-level sMRI regional similarity (JCGS-radMBN). Methods: Two publicly available datasets were adopted, including 120 subjects from ADNI with both T1-weighted image (T1WI) and diffusion MRI (dMRI) for JCGS-radMBN construction, 818 subjects from ADNI and 200 subjects solely with T1WI from AIBL for validation in early AD diagnosis. Specifically, the JCGS-radMBN was conducted by jointly estimating non-zero connections among subjects, with the regularization term constrained by group-level white matter fiber connectivity and individual-level sMRI regional similarity. Then, a triplet graph convolutional network was adopted for early AD diagnosis. The discriminative brain connections were identified using a two-sample t-test, and the neurobiological interpretation was validated by correlating the discriminative brain connections with cognitive scores. Results: The JCGS-radMBN exhibited superior classification performance over five brain network construction methods. For the typical NC vs. AD classification, the JCGS-radMBN increased by 1-30% in accuracy over the alternatives on ADNI and AIBL. The discriminative brain connections exhibited a strong connectivity to hippocampus, parahippocampal gyrus, and basal ganglia, and had significant correlation with MMSE scores. Conclusion: The proposed JCGS-radMBN facilitated the AD characterization of brain network established on routinely acquired imaging modality of sMRI. Supplementary Information: The online version of this article (10.1007/s13755-023-00269-0) contains supplementary material, which is available to authorized users.

Cross2SynNet: cross-device-cross-modal synthesis of routine brain MRI sequences from CT with brain lesion.

OBJECTIVES: CT and MR are often needed to determine the location and extent of brain lesions collectively to improve diagnosis. However, patients with acute brain diseases cannot complete the MRI examination within a short time. The aim of the study is to devise a cross-device and cross-modal medical image synthesis (MIS) method Cross2SynNet for synthesizing routine brain MRI sequences of T1WI, T2WI, FLAIR, and DWI from CT with stroke and brain tumors. MATERIALS AND METHODS: For the retrospective study, the participants covered four different diseases of cerebral ischemic stroke (CIS-cohort), cerebral hemorrhage (CH-cohort), meningioma (M-cohort), glioma (G-cohort). The MIS model Cross2SynNet was established on the basic architecture of conditional generative adversarial network (CGAN), of which, the fully convolutional Transformer (FCT) module was adopted into generator to capture the short- and long-range dependencies between healthy and pathological tissues, and the edge loss function was to minimize the difference in gradient magnitude between synthetic image and ground truth. Three metrics of mean square error (MSE), peak signal-to-noise ratio (PSNR), and structure similarity index measure (SSIM) were used for evaluation. RESULTS: A total of 230 participants (mean patient age, 59.77 years ± 13.63 [standard deviation]; 163 men [71%] and 67 women [29%]) were included, including CIS-cohort (95 participants between Dec 2019 and Feb 2022), CH-cohort (69 participants between Jan 2020 and Dec 2021), M-cohort (40 participants between Sep 2018 and Dec 2021), and G-cohort (26 participants between Sep 2019 and Dec 2021). The Cross2SynNet achieved averaged values of MSE = 0.008, PSNR = 21.728, and SSIM = 0.758 when synthesizing MRIs from CT, outperforming the CycleGAN, pix2pix, RegGAN, Pix2PixHD, and ResViT. The Cross2SynNet could synthesize the brain lesion on pseudo DWI even if the CT image did not exhibit clear signal in the acute ischemic stroke patients. CONCLUSIONS: Cross2SynNet could achieve routine brain MRI synthesis of T1WI, T2WI, FLAIR, and DWI from CT with promising performance given the brain lesion of stroke and brain tumor.

Circ-KATNAL1 Knockdown Reduces Neuronal Apoptosis and Alleviates Spinal Cord Injury Through the miR-98-5p/PRDM5 Regulatory Axis.

Spinal cord injury (SCI) is a common disease of the central nervous system. circRNAs play a crucial role in neurological disease. The purpose of this study was to investigate the role of circ-KATNAL1 in SCI and its regulatory mechanism. T9-L10 spinal segment of Sprague Dawley rats was compressed or contused after T10 laminectomy to establish the SCI rat model. Then, rats were divided into SCI group, si-NC group, si-circ-KATNAL1 group, si-circ-KATNAL1 + antagomir NC group, si-circ-KATNAL1 + miR-98-5p antagomir group, si-circ-KATNAL1 + oe-NC group, and si-circ-KATNAL1 + oe-PRDM5 group, with 6 rats in each group. There was another sham operation group that received no treatment. Basso, Beattie, and Bresnahan (BBB) scores were used to evaluate the neural function of rats. In addition to that, the pathological changes of spinal cord tissue, neuronal apoptosis, and inflammatory responses were correspondingly observed and analyzed. Quantitative measurements of circ-KATNAL1, miR-98-5p, and PRDM5 levels were conducted, as well as analyses of their interrelationship. Circ-KATNAL1 was up-regulated in the spinal cord tissue of SCI rats, and circ-KATNAL1 knockdown could improve neural function, alleviate pathological changes of spinal cord tissue, and inhibit neuronal apoptosis and inflammatory responses in SCI rats. For miR-98-5p, circ-KATNAL1 was an upstream factor, while PRDM5 was a downstream actor. miR-98-5p deficiency or PRDM5 restoration impaired the remission effect of circ-KATNAL1 knockdown on SCI. Circ-KATNAL1 knockdown reduces neuronal apoptosis and alleviates SCI through miR-98-5p/PRDM5 regulatory axis.

Relations of heart-type and brain-type fatty acid-binding proteins with postoperative cognitive dysfunction in elderly patients undergoing spinal surgery.

OBJECTIVE: The aim of this study is to analyze the relations of heart-type fatty acid-binding protein (H-FABP) and brain-type fatty acid-binding protein (B-FABP) with postoperative cognitive dysfunction (POCD) in elderly patients undergoing spinal surgery. METHODS: One hundred and twenty-five patients who underwent spinal surgery were enrolled in this study. According to whether patients had POCD within 5 days after surgery, the participants were divided into POCD group and non-POCD group. Before surgery and 6 h after surgery, the serum H-FABP and B-FABP contents were detected. RESULTS: There were 33 (26.4%) patients in POCD group, and 92 (73.60%) patients in non-POCD group. After surgery, the serum H-FABP and B-FABP contents in POCD group were significantly higher than those before surgery, respectively (p<0.05), and those in non-POCD group were significantly lower than those before surgery, respectively (p<0.05). After surgery, the serum H-FABP and B-FABP contents in POCD group were significantly higher than those in non-POCD group, respectively (p<0.05). CONCLUSION: The serum H-FABP and B-FABP contents are positively related to the occurrence of POCD in elderly patients undergoing spinal surgery.

[Three dimensional finite element analysis of optimal distribution model of fillers in vertebroplasty].

OBJECTIVE: To establish a three-dimensional finite element model of osteoporosis and to study the stiffness recovery of injured vertebrae and stress analysis of adjacent vertebrae after percutaneous vertebroplasty under different perfusion and distribution conditions by simulating fluid flow into the vertebral body. METHODS: A male healthy volunteer was selected. CT scans were performed from T11 to L2. Mimics 15.0 and ABAQUS 6.11 software were used to extract CT images. The vertebral model of osteoporotic fracture was established. The flow physical field and conduction and diffusion physical field were coupled to simulate the process and parts of the injection of bone cement into the vertebral fracture model. The amount of bone cement injected into the vertebral fracture model was 2 ml, 4 ml, 6 ml respectively. The diffusion range of bone cement was simulated on the image, and the post injection model of bone cement was obtained. Vertical downward, forward and backward pressure of 300 N were applied on the surface of the model to simulate vertebral movement. The stress changes of upper and lower vertebrae and diseased vertebrae under different conditions were calculated. RESULTS: (1) The VonMises stress of T12 inferior endplate was the largest in the three states before and after fracture.(2) The VonMises stress of the intervertebral disc and each endplate after fracture was significantly higher than before fracture. When percutaneous vertebroplasty was applied, as the amount of bone cement injection increases, the VonMises stress of the adjacent vertebral endplates increases. In the diseased vertebrae, as the amount of bone cement increases, the VonMises stress of the vertebral body endplate showed a downward trend. CONCLUSION: Reliable biomechanical model of lumbar vertebral fracture can be established by using CT scanning data through software simulation. Vertebral fracture and vertebroplasty will cause biomechanical changes of adjacent vertebral bodies. With the increase of bone cement injection, the influence of biomechanical changes will increase significantly. Neighbouring vertebral fractures are more likely. For this experiment, percutaneous vertebroplasty has a suitable amount of cement injection of 4 ml.

Three-dimensional finite element analysis of optimal distribution model of vertebroplasty.

BACKGROUND: Establishment of a three-dimensional (3D) finite element model of osteoporosis, the simulation fluid was used to enter the vertebral body to study the stiffness recovery of injured vertebral body under different perfusion and distribution conditions, and the stress analysis of adjacent vertebral body after percutaneous vertebroplasty (PVP) was carried out. METHODS: A healthy male volunteer was selected. Computed tomography (CT) scanning was performed from T11 to L2. MIMICS 15.0 and ABAQUS 6.11 software was used to extract CT findings, and a vertebral model of osteoporotic fracture was established. The flow physical field and conduction and diffusion physical field were coupled to simulate the process and parts of the bone cement injection into the vertebral fracture model. The quantities of bone cement injected into the vertebral fracture model were 2, 4, and 6 mL, respectively. The diffusion range of bone cement was simulated on the simulated image, and the postinjection model of bone cement was obtained. For the simulation of vertebral movement, vertical downward, forward, and backward pressure of 300 N was applied on the model's surface. The stress changes in the upper and lower vertebrae and diseased vertebrae were calculated under different conditions. RESULTS: It was revealed that the von Mises stress in the endplate under T12 was the highest in the three different states before and after fracture. The von Mises stress in the intervertebral discs and endplates was significantly higher after fracture than before fracture. When PVP was applied, the von Mises stress in adjacent endplates was increased with the increase of cement injection, while the von Mises stress was decreased in the adjacent endplates with cement injection compared with diseased vertebrae. CONCLUSIONS: A reliable biomechanical model of lumbar vertebral fracture can be established through numerical simulation of CT scanning data. Vertebral fracture and vertebroplasty may cause biomechanical changes in adjacent vertebrae. The influence of biomechanical changes may notably increase along with the amount of bone cement injected. In this study, PVP revealed 4 mL to be the optimal amount for cement injection.

Identifying hub genes and potential mechanisms associated with senescence in human annulus cells by gene expression profiling and bioinformatics analysis.

The aim of the present study was to reveal the potential hub genes and regulatory mechanisms associated with senescence in human annulus cells by analyzing microarray data using bioinformatics. The gene expression dataset GSE17077, of senescent and non­senescent annulus cells obtained from patients with disc degenerative diseases (DDD), was downloaded from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) were identified. Functional and pathway annotations were performed using Gene Ontology and Kyoto Encyclopedia of Genes and Genomes terms, respectively. Web­based Gene Set Analysis Toolkit and Chip Enrichment Analysis were used to identify key transcription factors (TFs). A protein­protein interaction (PPI) network was constructed to analyze the hub genes associated with senescence in DDD. A total of 667 DEGs were screened, including 368 up­ and 299 down­regulated genes. These DEGs were enriched in phosphorylation, regulation of apoptosis and regulation of programmed cell death. In addition, DEGs were involved in axon guidance, natural killer cell­mediated cytotoxicity, purine metabolism and the mitogen­activated protein kinase (MAPK) signaling pathway. The TFs activator protein 1 (AP1), specificity protein 1 and aryl hydrocarbon receptor may serve regulatory roles in gene expression in senescent cells. Certain key target genes of TFs, including heat shock protein 90 (HSP90) and C­X­C motif chemokine 5 (CXCL5), within the DEGs were revealed to have a high connectivity degree by PPI analysis. The results of the present study indicated that the MAPK­regulated AP1 pathway may contribute to senescence­associated disc degeneration. The DEGs, including HSP90 and CXCL5, with a high degree of connectivity may be potential targets for future investigations into molecular biomarkers.

IL-1ß promotes ADAMTS enzyme-mediated aggrecan degradation through NF-κB in human intervertebral disc.

BACKGROUND: The purpose of this study is to investigate IL-1ß regulation of a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS-4 and ADAMTS-5) expression through nuclear factor kappa B (NF-κB) in human nucleus pulposus (NP) cells. METHODS: qRT-PCR and Western blot were used to measure ADAMTS expression. Transfections and gene silencing were used to determine the role of NF-κB on cytokine-mediated ADAMTS expression and its role in aggrecan degradation. RESULTS: IL-1ß increased ADAMTS expression in NP cells. Treatment with NF-κB inhibitors abolished the inductive effect of the cytokines on ADAMTS expression. Silencing of p65 confirmed their role in IL-1ß-dependent ADAMTS-4 and ADAMTS-5 expression and aggrecan degradation. CONCLUSIONS: By controlling the activation of NF-κB signaling, IL-1ß modulates the expression of ADAMTS in NP cells. To our knowledge, this is the first study that shows the contribution of both ADAMTS-4 and ADAMTS-5 to aggrecan degradation in human NP cells.