Integrating Text and Image Analysis: Exploring GPT-4V's Capabilities in Advanced Radiological Applications Across Subspecialties.

This study demonstrates that GPT-4V outperforms GPT-4 across radiology subspecialties in analyzing 207 cases with 1312 images from the Radiological Society of North America Case Collection.

The antagonism mechanism of astilbin against cadmium-induced injury in chicken lungs via Treg/Th1 balance signaling pathway.

The purpose of this study was to investigate the effect of Treg/Th1 imbalance in cadmium-induced lung injury and the potential protective effect of astilbin against cadmium-induced lung injury in chicken. Cadmium exposure significantly decreased T-AOC and GSH-Px levels and SOD activity in the chicken lung tissues. In contrast, it significantly increased the MDA and NO levels. These results indicate that cadmium triggers oxidative stress in lungs. Histopathological analysis revealed that cadmium exposure further induced infiltration of lymphocytes in the chicken lungs, indicating that cadmium causes pulmonary damage. Further analysis revealed that cadmium decreased the expression of IL-4 and IL-10 but increased those of IL-17, Foxp3, TNF-α, and TGF-ß, indicating that the exposure of cadmium induced the imbalance of Treg/Th1. Moreover, cadmium adversely affected chicken lung function by activating the NF-kB pathway and inducing expression of genes downstream to these pathways (COX-2, iNOS), associated with inflammatory injury in the lung tissue. Astilbin reduced cadmium-induced oxidative stress and inflammation in the lungs by increasing antioxidant enzyme activities and restoring Treg/Th1 balance. In conclusion, our results suggest that astilbin treatment alleviated the effects of cadmium-mediated lung injury in chickens by restoring the Treg/Th1 balance.

New insights into starch, lipid, and protein interactions - Colon microbiota fermentation.

Starch, lipids, and proteins are essential biological macromolecules that play a crucial role in providing energy and nutrition to our bodies. Interactions between these macromolecules have been shown to impact starch digestibility. Understanding and controlling starch digestibility is a key area of research. Investigating the mechanisms behind the interactions of these three components and their influence on starch digestibility is of significant practical importance. Moreover, these interactions can result in the formation of resistant starch, which can be fermented by gut microbiota in the colon, leading to various health benefits. While current research has predominantly focused on the digestive properties of starch in the small intestine, there is a notable gap in understanding the colonic microbial fermentation phase of resistant starch. The benefits of fermentation of resistant starch in the colon may outweigh its glucose-lowering effect in the small intestine. Thus, it is crucial to study the fermentation behavior of resistant starch in the colon. This paper investigates the impact of interactions among starch, lipids, and proteins on starch digestion, with a specific focus on the fermentation phase of indigestible carbohydrates in the colon. Furthermore, valuable insights are offered for guiding future research endeavors.

Phosphorylated SHMT2 Regulates Oncogenesis Through m6A Modification in Lung Adenocarcinoma.

Targeting cancer-specific metabolic processes is a promising therapeutic strategy. Here, this work uses a compound library that directly inhibits metabolic enzymes to screen the potential metabolic targets in lung adenocarcinoma (LUAD). SHIN1, the specific inhibitor of serine hydroxymethyltransferase 1/2 (SHMT1/2), has a highly specific inhibitory effect on LUAD cells, and this effect depends mainly on the overexpression of SHMT2. This work clarifies that mitogen-activated protein kinase 1 (MAPK1)-mediated phosphorylation at Ser90 is the key mechanism underlying SHMT2 upregulation in LUAD and that this phosphorylation stabilizes SHMT2 by reducing STIP1 homology and U-box containing protein 1 (STUB1)-mediated ubiquitination and degradation. SHMT2-Ser90 dephosphorylation decreases S-adenosylmethionine levels in LUAD cells, resulting in reduced N6-methyladenosine (m6A) levels in global RNAs without affecting total protein or DNA methylation. Methylated RNA immunoprecipitation sequencing (MeRIP-Seq) and RNA sequencing (RNA-Seq) analyses further demonstrate that SHMT2-Ser90 dephosphorylation accelerates the RNA degradation of oncogenic genes by reducing m6A modification, leading to the inhibition of tumorigenesis. Overall, this study elucidates a new regulatory mechanism of SHMT2 during oncogenesis and provides a theoretical basis for targeting SHMT2 as a therapeutic target in LUAD.

Comparative Analysis of Multimodal Large Language Model Performance on Clinical Vignette Questions.

This study compares 2 large language models and their performance vs that of competing open-source models.

Study on numerical simulation and mechanical properties of anchor cable with C-shaped tube subjected to shearing.

To examine the disparity in deformation behavior and mechanical qualities between anchor cables with C-shaped tubes and regular anchor cables under shear conditions. The double-sided shear tests of free-section anchor cables and anchor cables with C-shaped tubes were conducted utilizing the indoor large-scale double-shear test equipment with varying pretension loads. The indoor double-shear tests indicate that the inclusion of the C-shaped tube alters the stress distribution of the anchor cables inside the anchor cables with C-shaped tubes and mitigates the impact of stress concentration. Moreover, it facilitates the transition of the anchor cable's failure mode from a mix of tensile and shear breaking to mainly tensile breakage. In addition, ABAQUS finite element analysis software was used to establish a double shear test model of the anchor cable with C-shaped tube to accurately simulate the interaction and stress distribution among the anchor cable, C-shaped tube, and concrete block in the double shear test. The findings of the simulation results reveal that the numerical model can adequately depict the evolution of the stress distribution in the prestressed anchored structure and the damage of the concrete block with increasing shear displacement. The relational equation for the yield state of the anchor cable with C-shaped tube under combined tensile and shear loads is found by integrating the experimental and simulation data, the static beam theory, and the concept of minimal potential energy.

Comparison of Locking Plate Alone and Locking Plate Combined with 3D Printed Polymethylmethacrylate Augmentation in Treating Proximal Humerus Fractures in the Elderly.

BACKGROUND: Locking plates are widely used in open reduction internal fixation (ORIF) for proximal humeral fracture (PHF). However, the optimal surgical treatment of unstable, displaced PHF in elderly patients remains controversial. This study aimed to compare the radiological and clinical outcomes of surgical treatment of PHF in the elderly with locking plate (LP) alone and locking plate combined with 3D printed polymethylmethacrylate (PMMA) prosthesis augmentation (LP-PA). METHODS: From May 2015 to April 2021, a total of 97 patients aged ≥ 60 years with acute unstable PHF who underwent osteosynthesis with either LP (46 patients) or LP-PA (51 patients) were retrospectively analyzed. For the LP-PA group, a customized proximal humeral prosthesis made of PMMA cement was intra-operatively fabricated by a three-dimensional (3D) printed prototype mold for the humeral medial support. Radiological outcomes were analyzed by measuring the value of neck-shaft angle (NSA) and humeral head height (HHH). The clinical outcomes were evaluated using Constant-Murley Score (CMS), Disabilities of the Arm Shoulder and Hand (DASH) score, American Shoulder and Elbow Surgeons (ASES) score, and the shoulder range of motion (ROM). Pain was measured using a visual analogue scale (VAS). RESULTS: At the one-year follow-up, all fractures healed radiologically and clinically. The mean changes of NSA and HHH over the follow-up period were markedly smaller in the LP-PA group (3.8 ± 0.9° and 1.7 ± 0.3 mm) than those in the LP group (9.7 ± 2.1° and 3.2 ± 0.6 mm, both P < 0.0001). The LP-PA group also presented lower DASH score (17.1 ± 3.6), higher ASES score (89.5 ± 11.2) and better ROM in forward elevation (142 ± 26°) and external rotation (59 ± 11°) compared to the LP group (28.9 ± 4.8 for DASH score, P < 0.0001; 82.3 ± 9.0 for ASES score, P < 0.001; 129 ± 21° for forward elevation, P = 0.008; and 52 ± 9° for external rotation, P = 0.001). There was no significant difference in overall complication rate between the two groups, although the complication rate of screw perforation was higher in the LP-PA group (P = 0.172). CONCLUSIONS: For PHF in elderly patients, the combination of LP fixation and PMMA prosthesis augmentation effectively improved humeral head support and reduction maintenance, providing satisfactory outcomes both radiologically and clinically. This technique also reduced the incidence of screw perforation associated with plate fixation alone, making it a reasonable option to ensure satisfactory clinical outcomes.

Transplantation of Neural Stem Cells Loaded in an IGF-1 Bioactive Supramolecular Nanofiber Hydrogel for the Effective Treatment of Spinal Cord Injury.

Spinal cord injury (SCI) leads to massive cell death, disruption, and demyelination of axons, resulting in permanent motor and sensory dysfunctions. Stem cell transplantation is a promising therapy for SCI. However, owing to the poor microenvironment that develops following SCI, the bioactivities of these grafted stem cells are limited. Cell implantation combined with biomaterial therapies is widely studied for the development of tissue engineering technology. Herein, an insulin-like growth factor-1 (IGF-1)-bioactive supramolecular nanofiber hydrogel (IGF-1 gel) is synthesized that can activate IGF-1 downstream signaling, prevent the apoptosis of neural stem cells (NSCs), improve their proliferation, and induce their differentiation into neurons and oligodendrocytes. Moreover, implantation of NSCs carried out with IGF-1 gels promotes neurite outgrowth and myelin sheath regeneration at lesion sites following SCI. In addition, IGF-1 gels can enrich extracellular vesicles (EVs) derived from NSCs or from nerve cells differentiated from these NSCs via miRNAs related to axonal regeneration and remyelination, even in an inflammatory environment. These EVs are taken up by autologous endogenous NSCs and regulate their differentiation. This study provides adequate evidence that combined treatment with NSCs and IGF-1 gels is a potential therapeutic strategy for treating SCI.

Large language models streamline automated machine learning for clinical studies.

A knowledge gap persists between machine learning (ML) developers (e.g., data scientists) and practitioners (e.g., clinicians), hampering the full utilization of ML for clinical data analysis. We investigated the potential of the ChatGPT Advanced Data Analysis (ADA), an extension of GPT-4, to bridge this gap and perform ML analyses efficiently. Real-world clinical datasets and study details from large trials across various medical specialties were presented to ChatGPT ADA without specific guidance. ChatGPT ADA autonomously developed state-of-the-art ML models based on the original study's training data to predict clinical outcomes such as cancer development, cancer progression, disease complications, or biomarkers such as pathogenic gene sequences. Following the re-implementation and optimization of the published models, the head-to-head comparison of the ChatGPT ADA-crafted ML models and their respective manually crafted counterparts revealed no significant differences in traditional performance metrics (p ≥ 0.072). Strikingly, the ChatGPT ADA-crafted ML models often outperformed their counterparts. In conclusion, ChatGPT ADA offers a promising avenue to democratize ML in medicine by simplifying complex data analyses, yet should enhance, not replace, specialized training and resources, to promote broader applications in medical research and practice.

Mechanism and role of mitophagy in the development of severe infection.

Mitochondria produce adenosine triphosphate and potentially contribute to proinflammatory responses and cell death. Mitophagy, as a conservative phenomenon, scavenges waste mitochondria and their components in the cell. Recent studies suggest that severe infections develop alongside mitochondrial dysfunction and mitophagy abnormalities. Restoring mitophagy protects against excessive inflammation and multiple organ failure in sepsis. Here, we review the normal mitophagy process, its interaction with invading microorganisms and the immune system, and summarize the mechanism of mitophagy dysfunction during severe infection. We highlight critical role of normal mitophagy in preventing severe infection.

Open Distal Femur Fractures Treated with Bone Cement Intramedullary Support Combined with Locked Plate Fixation.

Objective: The objective of this study was to assess the short-term clinical efficacy of the short-term clinical efficacy of bone cement intramedullary support combined with locked plate fixation in the treatment of such fractures. Methods: A retrospective study including 21 patients was reviewed at an urban level one trauma center. There were 17 males and 4 females, with a mean age of 33.9 years. Gustilo grade was II (12 cases), III-A (6 cases), III-B (2 cases), and III-C (1 case). Two fractures were AO-OTA type 33A3, 9 cases were type 33C2, and 10 cases were type 33C3. After the first stage debridement and temporary external fixation, all patients received bone cement intramedullary support combined with locked plate fixation through an anterolateral incision at the second stage.. The perioperative complications, need for bone graft, alignment, and radiographic union were recorded. At 1-year follow-up, the range of knee motion was recorded, and functional results were evaluated by the Hospital for Special Surgery (HSS) knee score. Results: All 21 patients were followed up for 12-36 months, with an average of 18.7 months. 1 case had superficial wound infection, and 2 cases had partial skin edge necrosis of the original open wound. After symptomatic dressing changes, they all healed well. 4 cases had autogenous bone grafting. 18 patients (85.7%) achieved radiographic union, with a mean union time of 6.2 months. Two patients underwent secondary operation 9 months after surgery due to nonunion and finally united after autologous bone grafting. One patient developed a deep infection 8 months after surgery and was successfully treated with Masquelet technique. Finally, bone union was achieved 7 months after surgery. The alignment was good in 17 patients (81.0%). No deep infection or hardware failure occurred during 1-year follow-up. The average range of knee extension and flexion was 5.2 ° and 106.8 °, respectively. The HSS score averaged 83.6. Conclusions: Bone cement intramedullary support combined with locked plate fixation was an effective treatment modality of open distal femur fractures with high union rate, low complication, adequate alignment and satisfactory functional outcomes.

Encrypted federated learning for secure decentralized collaboration in cancer image analysis.

Artificial intelligence (AI) has a multitude of applications in cancer research and oncology. However, the training of AI systems is impeded by the limited availability of large datasets due to data protection requirements and other regulatory obstacles. Federated and swarm learning represent possible solutions to this problem by collaboratively training AI models while avoiding data transfer. However, in these decentralized methods, weight updates are still transferred to the aggregation server for merging the models. This leaves the possibility for a breach of data privacy, for example by model inversion or membership inference attacks by untrusted servers. Somewhat-homomorphically-encrypted federated learning (SHEFL) is a solution to this problem because only encrypted weights are transferred, and model updates are performed in the encrypted space. Here, we demonstrate the first successful implementation of SHEFL in a range of clinically relevant tasks in cancer image analysis on multicentric datasets in radiology and histopathology. We show that SHEFL enables the training of AI models which outperform locally trained models and perform on par with models which are centrally trained. In the future, SHEFL can enable multiple institutions to co-train AI models without forsaking data governance and without ever transmitting any decryptable data to untrusted servers.

Schiff base fluorescent sensor with aggregation induced emission characteristics for the sensitive and specific Fe3+ detection.

An aggregation induced emission based compound ((E)-4-((2-hydroxy-5-methoxybenzylidene)amino)benzoic acid) was synthesized through facile Schiff base condensation and characterized by various spectral techniques. The as-prepared compound represented a typical aggregation induced emission behavior in aqueous solution and exploited as a turn-off fluorescent sensor for Fe3+ detection in THF-H2O system (3:7, v/v) with high sensitivity and selectivity. The mechanism of the fluorescence quenching was intensively studied, which was attributed to both dynamic quenching and inner filter effect. The fluorescence probe displayed a highly broad dynamic response range (0.5-500 µM) for selective detection of Fe3+ with a limit of detection of 0.079 µM. The proposed method was successfully employed for detection and quantification of Fe3+ in human urine samples and proved to have potential for practical applications in biological field.

Multimodal Deep Learning for Integrating Chest Radiographs and Clinical Parameters: A Case for Transformers.

Background Clinicians consider both imaging and nonimaging data when diagnosing diseases; however, current machine learning approaches primarily consider data from a single modality. Purpose To develop a neural network architecture capable of integrating multimodal patient data and compare its performance to models incorporating a single modality for diagnosing up to 25 pathologic conditions. Materials and Methods In this retrospective study, imaging and nonimaging patient data were extracted from the Medical Information Mart for Intensive Care (MIMIC) database and an internal database comprised of chest radiographs and clinical parameters inpatients in the intensive care unit (ICU) (January 2008 to December 2020). The MIMIC and internal data sets were each split into training (n = 33 893, n = 28 809), validation (n = 740, n = 7203), and test (n = 1909, n = 9004) sets. A novel transformer-based neural network architecture was trained to diagnose up to 25 conditions using nonimaging data alone, imaging data alone, or multimodal data. Diagnostic performance was assessed using area under the receiver operating characteristic curve (AUC) analysis. Results The MIMIC and internal data sets included 36 542 patients (mean age, 63 years ± 17 [SD]; 20 567 male patients) and 45 016 patients (mean age, 66 years ± 16; 27 577 male patients), respectively. The multimodal model showed improved diagnostic performance for all pathologic conditions. For the MIMIC data set, the mean AUC was 0.77 (95% CI: 0.77, 0.78) when both chest radiographs and clinical parameters were used, compared with 0.70 (95% CI: 0.69, 0.71; P < .001) for only chest radiographs and 0.72 (95% CI: 0.72, 0.73; P < .001) for only clinical parameters. These findings were confirmed on the internal data set. Conclusion A model trained on imaging and nonimaging data outperformed models trained on only one type of data for diagnosing multiple diseases in patients in an ICU setting. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Kitamura and Topol in this issue.

SUCLG2 Regulates Mitochondrial Dysfunction through Succinylation in Lung Adenocarcinoma.

Mitochondrial dysfunction and abnormal energy metabolism are major features of cancer. However, the mechanisms underlying mitochondrial dysfunction during cancer progression are far from being clarified. Here, it is demonstrated that the expression level of succinyl-coenzyme A (CoA) synthetase GDP-forming subunit ß (SUCLG2) can affect the overall succinylation of lung adenocarcinoma (LUAD) cells. Succinylome analysis shows that the deletion of SUCLG2 can upregulate the succinylation level of mitochondrial proteins and inhibits the function of key metabolic enzymes by reducing either enzymatic activity or protein stability, thus dampening mitochondrial function in LUAD cells. Interestingly, SUCLG2 itself is also succinylated on Lys93, and this succinylation enhances its protein stability, leading to the upregulation of SUCLG2 and promoting the proliferation and tumorigenesis of LUAD cells. Sirtuin 5 (SIRT5) desuccinylates SUCLG2 on Lys93, followed by tripartite motif-containing protein 21 (TRIM21)-mediated ubiquitination through K63-linkage and degradation in the lysosome. The findings reveal a new role for SUCLG2 in mitochondrial dysfunction and clarify the mechanism of the succinylation-mediated protein homeostasis of SUCLG2 in LUAD, thus providing a theoretical basis for developing anti-cancer drugs targeting SUCLG2.

Management of segmental defects post open distal femur fracture using a titanium cage combined with the Masquelet technique A single-centre report of 23 cases.

INTRODUCTION: The segmental bone defects post open distal femur fracture presents a reconstructive challenge, which often requires extreme solutions. The present study reviewed a new treatment strategy which used a cylindrical titanium mesh cage as an adjunct to the Masquelet technique. METHODS: We retrospectively reviewed a consecutive series of 23 patients treated for segmental bone defects post open distal femur fracture using a titanium mesh cage combined with the Masquelet technique under a 2-staged protocol in our institution from 2017 to 2021. The study group consisted of 13 men and 10 women with an average age of 44.1 years. The surgical debridement was performed with antibiotic polymethylmethacrylate (PMMA) cement spacer implanted into the bone defect combined with cement-wrapped plate stabilization, or antibiotic beads with vacuum sealing drainage (VSD) to cover the wound. The second stage of the Masquelet technique for bone defect repair began at least 4-6 weeks after the first stage, once all signs of possible infection were eliminated. After the cement spacer was removed, the definitive reconstruction was completed with exchange to a cylindrical titanium mesh cage filled with cancellous autograft within the induced membrane. The bone defect with cage was stabilized with a distal femoral Less Invasive Stabilization System (LISS). The radiological and clinical records of the enrolled patients were retrospectively analyzed. RESULTS: The mean follow-up was 38.6 months. The average number of operations before the second stage was 1.3. The mean interval between the two stages was 12.7 weeks. The average length of the defect measured 8.3 cm (ranging from 6.1 to 12.4 cm). All the defects filled with autograft within the cage achieved bony union, with a mean healing time of 8.4 months. At the latest follow-up, the mean knee extension measured 6.2° (ranging from 0° to 20°), and the mean flexion measured 101.8° (ranging from 60° to 120°). Complications included two instances of superficial stitch abscess, which eventually healed. CONCLUSIONS: The use of a titanium cage implanted into an induced membrane in a 2-staged Masquelet protocol could achieve satisfactory clinical outcomes in cases of segmental defects following open distal femur fractures.

Quality analysis and metabolomic profiling of the effects of exogenous abscisic acid on rabbiteye blueberry.

Blueberry is a characteristic berry fruit shrub of the genus Vaccinium in the Rhododendron family. The fruit is rich in anthocyanins and has a variety of nutritional and health functions. This study aimed to systematically study the effect of exogenous abscisic acid (ABA) application on ripening and metabolites in blueberry fruits. Blueberry fruit ripening was divided into six stages for further analysis. In this study, nontarget metabolomics was performed to demonstrate the effect on metabolite levels. The results showed that 1000 mg/L ABA significantly promoted fruit ripening and increased anthocyanin content. Moreover, exogenous ABA treatment can affect endogenous ABA levels and improve its antioxidant capacity. Important metabolites of the flavonoid pathway were detected, and the results showed that anthocyanin synthesis increased, and some other bioactive metabolite levels decreased. After comprehensive assessments, we believe that 1000 mg/L exogenous ABA application will have positive impacts on blueberry fruit quality and economic benefits.

STUB1-mediated ubiquitination regulates the stability of GLUD1 in lung adenocarcinoma.

The dysregulation of glutamine metabolism provides survival advantages for tumors by supplementing tricarboxylic acid cycle. Glutamate dehydrogenase 1 (GLUD1) is one of the key enzymes in glutamine catabolism. Here, we found that enhanced protein stability was the key factor for the upregulation of GLUD1 in lung adenocarcinoma. We discovered that GLUD1 showed a high protein expression in lung adenocarcinoma cells or tissues. We elucidated that STIP1 homology and U-box-containing protein 1 (STUB1) was the key E3 ligase responsible for ubiquitin-mediated proteasomal degradation of GLUD1. We further showed that lysine 503 (K503) was the main ubiquitination site of GLUD1, inhibiting the ubiquitination at this site promoted the proliferation and tumor growth of lung adenocarcinoma cells. Taken together, this study clarifies the molecular mechanism of GLUD1 in maintaining protein homeostasis in lung adenocarcinoma, which provides a theoretical basis for the development of anti-cancer drugs targeting GLUD1.

Medical transformer for multimodal survival prediction in intensive care: integration of imaging and non-imaging data.

When clinicians assess the prognosis of patients in intensive care, they take imaging and non-imaging data into account. In contrast, many traditional machine learning models rely on only one of these modalities, limiting their potential in medical applications. This work proposes and evaluates a transformer-based neural network as a novel AI architecture that integrates multimodal patient data, i.e., imaging data (chest radiographs) and non-imaging data (clinical data). We evaluate the performance of our model in a retrospective study with 6,125 patients in intensive care. We show that the combined model (area under the receiver operating characteristic curve [AUROC] of 0.863) is superior to the radiographs-only model (AUROC = 0.811, p < 0.001) and the clinical data-only model (AUROC = 0.785, p < 0.001) when tasked with predicting in-hospital survival per patient. Furthermore, we demonstrate that our proposed model is robust in cases where not all (clinical) data points are available.

A multimodal comparison of latent denoising diffusion probabilistic models and generative adversarial networks for medical image synthesis.

Although generative adversarial networks (GANs) can produce large datasets, their limited diversity and fidelity have been recently addressed by denoising diffusion probabilistic models, which have demonstrated superiority in natural image synthesis. In this study, we introduce Medfusion, a conditional latent DDPM designed for medical image generation, and evaluate its performance against GANs, which currently represent the state-of-the-art. Medfusion was trained and compared with StyleGAN-3 using fundoscopy images from the AIROGS dataset, radiographs from the CheXpert dataset, and histopathology images from the CRCDX dataset. Based on previous studies, Progressively Growing GAN (ProGAN) and Conditional GAN (cGAN) were used as additional baselines on the CheXpert and CRCDX datasets, respectively. Medfusion exceeded GANs in terms of diversity (recall), achieving better scores of 0.40 compared to 0.19 in the AIROGS dataset, 0.41 compared to 0.02 (cGAN) and 0.24 (StyleGAN-3) in the CRMDX dataset, and 0.32 compared to 0.17 (ProGAN) and 0.08 (StyleGAN-3) in the CheXpert dataset. Furthermore, Medfusion exhibited equal or higher fidelity (precision) across all three datasets. Our study shows that Medfusion constitutes a promising alternative to GAN-based models for generating high-quality medical images, leading to improved diversity and less artifacts in the generated images.