The application of large language models in medicine: A scoping review.

This study systematically reviewed the application of large language models (LLMs) in medicine, analyzing 550 selected studies from a vast literature search. LLMs like ChatGPT transformed healthcare by enhancing diagnostics, medical writing, education, and project management. They assisted in drafting medical documents, creating training simulations, and streamlining research processes. Despite their growing utility in assisted diagnosis and improving doctor-patient communication, challenges persisted, including limitations in contextual understanding and the risk of over-reliance. The surge in LLM-related research indicated a focus on medical writing, diagnostics, and patient communication, but highlighted the need for careful integration, considering validation, ethical concerns, and the balance with traditional medical practice. Future research directions suggested a focus on multimodal LLMs, deeper algorithmic understanding, and ensuring responsible, effective use in healthcare.

High triglyceride-glucose (TyG) index is associated with poor prognosis of heart failure with preserved ejection fraction.

BACKGROUND: The impact of insulin resistance on the prognosis of heart failure with preserved ejection fraction (HFpEF) remains unknown. This study aimed to investigate the association between the triglyceride-glucose (TyG) index, an easily calculated marker of insulin resistance, and the long-term prognosis of HFpEF. METHODS: A total of 823 patients with HFpEF were enrolled in the study. The TyG index was determined using the formula ln(fasting triglycerides [mg/dL] × fasting glucose [mg/dL]/2). The primary endpoint was all-cause death. The secondary endpoints were cardiovascular (CV) death and heart failure (HF) rehospitalization. Restricted cubic spline, multivariate Cox proportional hazard models, and competing risk models were used for analyses. RESULTS: During a median follow-up period of 3.16 years, 147 (17.8%) all-cause deaths, 139 (16.8%) CV deaths, and 222 (27.0%) HF rehospitalizations occurred. Restricted cubic spline analysis revealed a J-shaped association between the TyG index and the mortality and rehospitalization rates. In the multivariate Cox proportional hazard models, compared with those in the lowest TyG index tertile, patients in the highest tertile exhibited the greatest susceptibility to all-cause death (HR 1.53, 95% CI 1.19-1.98) and CV death (HR 1.52, 95% CI 1.19-1.96). In the competing risk model, a significant association between the TyG index and HF rehospitalization was observed (HR 1.31, 95% CI, 1.07-1.61). CONCLUSION: A high TyG index is associated with an increased risk of mortality and rehospitalization in patients with HFpEF. The TyG index may serve as a promising prognostic marker for patients with HFpEF.

MSFT-YOLO: Improved YOLOv5 Based on Transformer for Detecting Defects of Steel Surface.

With the development of artificial intelligence technology and the popularity of intelligent production projects, intelligent inspection systems have gradually become a hot topic in the industrial field. As a fundamental problem in the field of computer vision, how to achieve object detection in the industry while taking into account the accuracy and real-time detection is an important challenge in the development of intelligent detection systems. The detection of defects on steel surfaces is an important application of object detection in the industry. Correct and fast detection of surface defects can greatly improve productivity and product quality. To this end, this paper introduces the MSFT-YOLO model, which is improved based on the one-stage detector. The MSFT-YOLO model is proposed for the industrial scenario in which the image background interference is great, the defect category is easily confused, the defect scale changes a great deal, and the detection results of small defects are poor. By adding the TRANS module, which is designed based on Transformer, to the backbone and detection headers, the features can be combined with global information. The fusion of features at different scales by combining multi-scale feature fusion structures enhances the dynamic adjustment of the detector to objects at different scales. To further improve the performance of MSFT-YOLO, we also introduce plenty of effective strategies, such as data augmentation and multi-step training methods. The test results on the NEU-DET dataset show that MSPF-YOLO can achieve real-time detection, and the average detection accuracy of MSFT-YOLO is 75.2, improving about 7% compared to the baseline model (YOLOv5) and 18% compared to Faster R-CNN, which is advantageous and inspiring.

A first-principles study on the electronic property and magnetic anisotropy of ferromagnetic CrOF and CrOCl monolayers.

Two-dimensional ferromagnetic materials with large perpendicular magnetic anisotropy (PMA) hold great potential in realizing low critical switching current, high thermal stability and high density nonvolatile storage in magnetic random-access memories. Our first-principles calculations reveal that CrOF and CrOCl monolayers (MLs) are two-dimensional (2D) ferromagnetic semiconductors with out-of-plane magnetic easy axis, and PMAs of CrOF and CrOCl MLs are mainly contributed by Cr atoms. The magnetic anisotropy of CrOF and CrOCl MLs can be controlled and enhanced by applying biaxial strain. Tensile strain can further enhance PMAs of CrOF and CrOCl MLs by 82.9% and 161.0% higher than those of unstrained systems, respectively. In addition, appropriate compressive strain can switch the magnetic easy axis of CrOF and CrOCl MLs from out-of-plane direction to in-plane direction. The semiconductor natures of CrOF and CrOCl MLs robust against biaxial strain, the band gaps of these systems under biaxial strain are in the range of 1.26 eV to 2.40 eV. By applying biaxial strain, the Curie temperatures of CrOF and CrOCl MLs increase up to 282 K and 163 K, respectively. These tunable properties suggest that CrOF and CrOCl MLs have great application potentials for magnetic data storage.

Biaxial strain modulated the electronic structure of hydrogenated 2D tetragonal silicene.

Silicene-based materials have attracted great attention due to their easier incorporation into silicon-based devices and components. In addition to the reported hydrogenated 2D tetragonal silicene (γ-SiH), we propose two stable atomic configurations of hydrogenated 2D tetragonal silicene (α-SiH and ß-SiH) based on first-principles calculation. The calculated results indicate hydrogenation can effectively open the band gap of 2D tetragonal silicene, α-SiH is a semiconductor with a direct band gap of 2.436 eV whereas ß-SiH is indirect band gap of 2.286 eV. We also find that the electronic band structure of α-SiH, ß-SiH and γ-SiH can be modulated via biaxial strain. By applying biaxial strain in the range of -10% to 12%, the band gap of α-SiH, ß-SiH and γ-SiH can be tuned in a range of 1.732-2.585 eV. Furthermore, direct-indirect or indirect-direct transition can be induced under biaxial strain, showing a high degree of flexibility in electronic band structure. The research not only broadens the diversity of hydrogenated 2D tetragonal silicenes, but also provides more possibilities of their applications in spintronic devices.

Significant Enhancement of MgZnO Metal-Semiconductor-Metal Photodetectors via Coupling with Pt Nanoparticle Surface Plasmons.

We proposed and demonstrated MgZnO metal-semiconductor-metal (MSM) ultraviolet photodetectors (UV) assisted with surface plasmons (SPs) prepared by the radio frequency magnetron sputtering deposition method. After the decoration of their surface with Pt nanoparticles (NPs), the responsivity of all the electrode spacing (3, 5, and 8 µm) photodetectors were enhanced dramatically; to our surprise, comparing with them the responsivity of larger spacing sample, more SPs were gathered which are smaller than others in turn. A physical mechanism focused on SPs and depletion width is given to explain the above results.