Individualized estimation of arterial carbon dioxide partial pressure using machine learning in children receiving mechanical ventilation.

BACKGROUND: Measuring arterial partial pressure of carbon dioxide (PaCO2) is crucial for proper mechanical ventilation, but the current sampling method is invasive. End-tidal carbon dioxide (EtCO2) has been used as a surrogate, which can be measured non-invasively, but its limited accuracy is due to ventilation-perfusion mismatch. This study aimed to develop a non-invasive PaCO2 estimation model using machine learning. METHODS: This retrospective observational study included pediatric patients (< 18 years) admitted to the pediatric intensive care unit of a tertiary children's hospital and received mechanical ventilation between January 2021 and June 2022. Clinical information, including mechanical ventilation parameters and laboratory test results, was used for machine learning. Linear regression, multilayer perceptron, and extreme gradient boosting were implemented. The dataset was divided into 7:3 ratios for training and testing. Model performance was assessed using the R2 value. RESULTS: We analyzed total 2,427 measurements from 32 patients. The median (interquartile range) age was 16 (12-19.5) months, and 74.1% were female. The PaCO2 and EtCO2 were 63 (50-83) mmHg and 43 (35-54) mmHg, respectively. A significant discrepancy of 19 (12-31) mmHg existed between EtCO2 and the measured PaCO2. The R2 coefficient of determination for the developed models was 0.799 for the linear regression model, 0.851 for the multilayer perceptron model, and 0.877 for the extreme gradient boosting model. The correlations with PaCO2 were higher in all three models compared to EtCO2. CONCLUSIONS: We developed machine learning models to non-invasively estimate PaCO2 in pediatric patients receiving mechanical ventilation, demonstrating acceptable performance. Further research is needed to improve reliability and external validation.

Hemophagocytic Lymphohistiocytosis as Initial Presentation of Malignancy in Pediatric Patients: Rare but Not to Be Ignored.

It is complicated to establish a consensus on the management and diagnosis of malignancy-triggered hemophagocytic lymphohistiocytosis (M-HLH) in children, as an initial presentation of malignancy is complicated. In this paper, we analyze the clinical characteristics and outcomes of eight pediatric patients in which M-HLH was the initial presentation of malignancy. All patients had hematologic malignancies: three subcutaneous panniculitis-like T-cell lymphomas, two acute lymphoblastic leukemias, two anaplastic large cell lymphomas, and a systemic EBV + T-cell lymphoma of childhood. The incidence rate of M-HLH among leukemia and malignant lymphoma patients in our institution was 1.9%. From the initial diagnosis of HLH, the median time taken to be diagnosed as a malignancy was about 1.3 months. The majority of patients received HLH-targeted immunosuppression and/or etoposide at first. The patients' clinical response to treatment for HLH and malignancies were varied. Five out of the eight patients died, one of whom died due to HLH-related cerebral edema after the initiation of chemotherapy. The median overall survival was 1.6 years. In order to improve the survival rate, the early detection of M-HLH, rapid screening for malignancy, and complete control of M-HLH with HLH-directed therapy followed by a thorough response monitoring are required.

Uniform Ag Nanocubes Prepared by AgCl Particle-Mediated Heterogeneous Nucleation and Disassembly and Their Mechanism Study by DFT Calculation.

Uniform Ag nanocubes are reproducibly synthesized by a AgCl particle-mediated heterogeneous nucleation and disassembly process in polyol chemistry. By introducing N,N-dimethylformamide (DMF) in a conventional polyol method with HCl etchant, Ag nanocrystals (NCs) begin to be nucleated on the surface of AgCl-precipitated particles due to the promoted reduction reaction by DMF. The nucleated Ag NCs on the AgCl particles are grown to Ag nanocubes in shape by consuming Ag sources from the AgCl mother particles. Eventually the grown Ag nanocubes are disassembled from the mother AgCl particles because the AgCl particles are fully digested by the growing Ag nanocubes. Density functional theory calculation confirms that the Ag atoms can be favorably deposited on the (100) facet of AgCl particles and the Ag nuclei on the AgCl particles tend to reveal (100) facet.

Pyrene-Tagged Ionic Liquids: Separable Organic Catalysts for SN2 Fluorination.

We prepared pyrene-substituted imidazolium-based ionic liquids (PILs) as organic catalysts for the SN2 fluorination using alkali metal fluoride (MF). In this system, the PIL significantly enhanced the reactivity of MF due to the phase-transfer catalytic effect of the imidazolium moiety as well as the metal cation-π (pyrene) interactions. Furthermore, this homogeneous catalyst PIL was easily separated from the reaction mixture using reduced graphene oxide by π-π stacking with the pyrene of PIL.

Uniform Microgels Containing Agglomerates of Silver Nanocubes for Molecular Size-Selectivity and High SERS Activity.

Surface-enhanced Raman scattering (SERS) is a promising technique for molecular analysis as the molecular fingerprints (Raman spectra) are amplified to detectable levels compared with common spectroscopy. Metal nanostructures localize electromagnetic field on their surfaces, which can lead to dramatic increase of Raman intensity of molecules adsorbed. However, the metal surfaces are prone to contamination, thereby requiring pretreatment of samples to remove adhesive molecules. To avoid the pretreatment and potentially achieve point-of-care (POC) analysis, we have developed SERS-active microgels using the droplet-microfluidic system. As the microgels are composed of water-swollen network with consistent mesh size, they selectively allow diffusion of molecules smaller than the mesh, thereby excluding large adhesives. To render the microgels highly SERS-active, we destabilize silver nanocubes to form agglomerates, which are embedded in the matrix of microgels. The nanogaps in the agglomerates provide high sensitivity in Raman measurement and size-selective permeability of the microgel matrix obviates the pretreatment of samples. To validate the functions, we demonstrate the direct detection of Aspirin dissolved in whole blood without any pretreatment.

Efficient hysteresis-less bilayer type CH3NH3PbI3 perovskite hybrid solar cells.

Bilayer type CH3NH3PbI3 (MAPbI3) perovskite hybrid solar cells were fabricated via a one-step spin-coating process by using solubility controlled MAPbI3 solutions of MAPbI3-DMSO (dimethyl sulfoxide) and MAPbI3-DMF (N, N-dimethylformamide)-HI. The best DMSO-bilayer device showed 1.07 ± 0.02 V V(oc) (open-circuit voltage), 20.2 ± 0.1 mA cm(-2) J(sc) (short-circuit current density), 68 ± 2% FF (fill factor), and 15.2 ± 0.3% η (overall power conversion efficiency) under the forward scan direction and 1.07 ± 0.02 V V(oc), 20.4 ± 0.1 mA cm(-2) J(sc), 70 ± 3% FF, and 15.9 ± 0.4% η under the reverse scan direction. The best HI-bilayer device had 1.08 ± 0.02 V V(oc), 20.6 ± 0.1 mA cm(-2) J(sc), 75 ± 1% FF, and 17.2 ± 0.2% η under the forward scan direction and 1.08 ± 0.02 V V(oc), 20.6 ± 0.1 mA cm(-2) J(sc), 76 ± 2% FF, and 17.4 ± 0.3% η under the reverse scan direction. The deviation of average device efficiency (η(avg)) of 20 DMSO samples and 20 HI samples was 14.2 ± 0.95% and 16.2 ± 0.85%, respectively. Therefore, the HI-bilayer devices exhibited better device efficiency and smaller J-V (current density-voltage) hysteresis with respect to the scan direction than the DMSO-bilayer devices due to the reduced recombination and traps by the formation of a purer and larger MAPbI3 perovskite crystalline film.