Search

Abstract Stroke is a neurological complication that can occur in patients admitted to the intensive care unit (ICU) for non-neurological conditions, leading to increased mortality and prolonged hospital stays. The incidence of stroke in ICU settings is notably higher compared to the general population, and delays in diagnosis can lead to irreversible neurological damage. Early diagnosis of stroke is critical to protect brain tissue and treat neurological defects. Therefore, we developed a machine learning model to diagnose stroke in patients with acute neurological manifestations in the ICU. We retrospectively collected data on patients’ underlying diseases, blood coagulation tests, procedures, and medications before neurological symptom onset from 206 patients at the Chungbuk National University Hospital ICU (July 2020–July 2022) and 45 patients at Chungnam National University Hospital between (July 2020–March 2023). Using the Categorical Boosting (CatBoost) algorithm with Bayesian optimization for hyperparameter selection and k-fold cross-validation to mitigate overfitting, we analyzed model-feature relationships with SHapley Additive exPlanations (SHAP) values. Internal model validation yielded an average accuracy of 0.7560, sensitivity of 0.8959, specificity of 0.7000, and area under the receiver operating characteristic curve (AUROC) of 0.8201. External validation yielded an accuracy of 0.7778, sensitivity of 0.7500, specificity of 0.7931, and an AUROC of 0.7328. These results demonstrated the model’s effectiveness in diagnosing stroke in non-neurological ICU patients with acute neurological manifestations using their electronic health records, making it valuable for the early detection of stroke in ICU patients.

Background Acute ischemic stroke is a disease with multiple etiologies. Therefore, identifying the mechanism of acute ischemic stroke is fundamental to its treatment and secondary prevention. The Trial of Org 10172 in Acute Stroke Treatment classification is currently the most widely used system, but it often has a limitations of classifying unknown causes and inadequate inter-rater reliability. Therefore, we attempted to develop a three-dimensional (3D)-convolutional neural network (CNN)-based algorithm for stroke lesion segmentation and subtype classification using only the diffusion and apparent diffusion coefficient information of patients with acute ischemic stroke. Methods This study included 2,251 patients with acute ischemic stroke who visited our hospital between February 2013 and July 2019. Results The segmentation model for lesion segmentation in the training set achieved a Dice score of 0.843±0.009. The subtype classification model achieved an average accuracy of 81.9%, with accuracies of 81.6% for large artery atherosclerosis, 86.8% for cardioembolism, 72.9% for small vessel occlusion, and 86.3% for control. Conclusion We developed a model to predict the mechanism of cerebral infarction using diffusion magnetic resonance imaging, which has great potential for identifying diffusion lesion segmentation and stroke subtype classification. As deep learning systems are gradually developing, they are becoming useful in clinical practice and applications.

Abstract Respiratory distress is a common chief complaint in neonates admitted to the neonatal intensive care unit. Despite the increasing use of non-invasive ventilation in neonates with respiratory difficulty, some of them require advanced airway support. Delayed intubation is associated with increased morbidity, particularly in urgent unplanned cases. Early and accurate prediction of the need for intubation may provide more time for preparation and increase safety margins by avoiding the late intubation at high-risk infants. This study aimed to predict the need for intubation within 3 h in neonates initially managed with non-invasive ventilation for respiratory distress during the first 48 h of life using a multimodal deep neural network. We developed a multimodal deep neural network model to simultaneously analyze four time-series data collected at 1-h intervals and 19 variables including demographic, physiological and laboratory parameters. Evaluating the dataset of 128 neonates with respiratory distress who underwent non-invasive ventilation, our model achieved an area under the curve of 0.917, sensitivity of 85.2%, and specificity of 89.2%. These findings demonstrate promising results for the multimodal model in predicting neonatal intubation within 3 h.

We present a global optimization routine for the variational quantum algorithms, which utilizes the dynamic tunneling flow. Originally designed to leverage information gathered by a gradient-based optimizer around local minima, we adapt the conventional dynamic tunneling flow to exploit the distance measure of quantum states, resolving issues of extrinsic degeneracy arising from the parametrization of quantum states. Our global optimization algorithm is applied to the variational quantum eigensolver for the transverse-field Ising model to demonstrate the performance of our routine while comparing it with the conventional dynamic tunneling method, which is based on the Euclidean distance measure on the parameter space.

Papillary thyroid carcinoma (PTC) is the most common endocrine carcinoma and has frequent recurrence instances. Although PTC recurrence has been predicted using predictors established using various features and techniques, its early detection is still challenging. To address this issue, it is aimed to develop a deep‐learning model that utilizes not only the initial medical records but also the thyroid function tests (TFTs) performed periodically post‐surgery. Herein, a novel multimodal prediction model, called the hybrid architecture for multimodal analysis (HAMA), that can analyze numeric and time‐series data simultaneously, is proposed. For quantitative evaluation, fourfold cross validation is conducted on data of 1613 PTC patients including 63 locoregional recurrence patients, and the HAMA is achieved the following performance: sensitivity (0.9688); specificity (0.9781); F1‐score (0.7943); and area under the receiver‐operating characteristic curve, AUROC (0.9863). Furthermore, a real‐time prediction simulation is conducted at 6‐month intervals by reconstructing the data of each patient into real‐time data. It is demonstrated in the real‐time simulation results that the HAMA predicts PTC recurrence at least 1.5 years in advance by recalculating the recurrence probability using the additional follow‐up data. To the best of the knowledge, the HAMA is the first deep‐learning model to reflect continuous change in the physical condition of a patient post‐surgery.

X-ray imaging is one of the most widely used security measures for maintaining airport and transportation security. Conventional X-ray imaging systems typically apply tone-mapping (TM) algorithms to visualize high-dynamic-range (HDR) X-ray images on a standard 8-bit display device. However, X-ray images obtained through traditional TM algorithms often suffer from halo artifacts or detail loss in inter-object overlapping regions, which makes it difficult for an inspector to detect unsafe or hazardous objects. To alleviate these problems, this article proposes a deep learning-based TM method for X-ray inspection. The proposed method consists of two networks called detail-recovery network (DR-Net) and TM network (TM-Net). The goal of DR-Net is to restore the details in the input HDR image, whereas TM-Net aims to compress the dynamic range while preserving the restored details and preventing halo artifacts. Since there are no standard ground-truth images available for the TM of X-ray images, we propose a novel loss function for unsupervised learning of TM-Net. We also introduce a dataset synthesis technique using the Beer-Lambert law for supervised learning of DR-Net. Extensive experiments comparing the performance of our proposed method with state-of-the-art TM methods demonstrate that the proposed method not only achieves visually compelling results but also improves the quantitative performance measures such as FSITM and HDR-VDP-2.2.

For indirect contrast enhancement, researchers have proposed various transformation functions based on histogram equalization and gamma correction. However, these transformation functions tend to result in over-enhancement artifacts such as noise amplification, mean brightness change, and detail loss. To overcome the limitations of conventional transformation functions, this paper introduces a novel sigmoid function based on the contrast sensitivity of human brightness perception. In the proposed method, the contrast sensitivity of the human retina is modeled as an exponential function of the log-intensity, and a transformation function is derived using the sensitivity model as the exponent of Steven's power law. We also present a parameter optimization method that maintains the mean brightness of the input image and stretches the image histogram while minimizing information loss. Experimental results demonstrate that the proposed method has low computational complexity and outperforms the state-of-the-art methods in terms of contrast enhancement performance, mean brightness preservation, and detail preservation.

Objective: High-risk human papillomavirus (HR-HPV) infection is a leading cause of cervical cancer, of which human papillomavirus (HPV)-16 and HPV-18 account for about 70% of cases. Since HPV infection is common, it is important to focus on the HPV genotypes that pose the highest risk for effective cervical cancer screening. In this study, we evaluated the clinical usefulness of HPV-16/HPV-18 genotyping for cervical cancer screening. Methods: A total of 86,022 women aged 25 years or older was analyzed in this study. Sensitivity, specificity, positive predictive value, and negative predictive value of HPV genotyping and cytology were analyzed. In addition, we subdivided participants into two groups according to cytology results, negative for intraepithelial lesion of malignancy (NILM) and atypical squamous cells of undetermined significance (ASC-US), and analyzed absolute risk (AR) and relative risk (RR) of cervical intraepithelial neoplasia (CIN) 3 or worse according to HPV genotype. Results: The AR of CIN 3 or worse was 77.0 times higher in HR-HPV-positive compared to HR-HPV-negative. Compared to 12 other HR-HPV-positive, the AR of CIN 3 or worse was 4.2 times higher in HPV-16 and/or HPV-18 positive. This finding was more evident in women with NILM than in women with ASC-US. The RR of CIN 3 or worse was 7.0 in women with NILM and 4.5 in women with ASC-US. Conclusion: Regardless of the cytology results, the risk of CIN 3 or worse was higher in HPV-16/HPV-18 than in other HR-HPV. HPV-16/HPV-18 genotyping is recommended to screen women with a high risk of cervical cancer. [ABSTRACT FROM AUTHOR]

Acute myeloid leukemia (AML) is an aggressive blood cancer with a high rate of relapse associated with adverse survival outcomes, especially in elderly patients. An aberrant expression of cyclin dependent kinase 7 (CDK7) is associated with poor outcomes and CDK7 inhibition has showed antitumor activities in various cancers. We investigated the efficacy of YPN-005, a CDK7 inhibitor in AML cell lines, xenograft mouse model, and primary AML cells. YPN-005 effectively inhibited the proliferation of AML cells by inducing apoptosis and reducing phosphorylation of RNA polymerase II. The c-MYC expression decreased with treatment of YPN-005, and the effect of YPN-005 was negatively correlated with c-MYC expression. YPN-005 also showed antileukemic activities in primary AML cells, especially those harboring FMS-like tyrosine kinase 3–internal tandem duplication (FLT3-ITD) mutation and in in vivo mouse model. Phosphorylated FLT3/Signal transducer and activator of transcription 5 (STAT5) was decreased and FLT3/STAT5 was downregulated with YPN-005 treatment. Our data suggest that YPN-005 has a role in treating AML by suppressing c-MYC and FLT3.

Objective : Post-stroke shoulder pain (PSSP) is a common complication that limits the range of motion (ROM) of the shoulder, the patient’s rehabilitation and in turn, affects the patients’ quality of life (QoL). Several treatment modalities such as sling, positioning, strapping, functional electrical stimulation, and nerve block have been suggested in literatures, however none of the treatments had long-term effects for PSSP. In this study, the authors evaluated clinical efficacy of pulsed radiofrequency (PRF) neuromodulation on the suprascapular nerve for PSSP, and suggested it as a potential treatment with long-term effect. Methods : This retrospective case series was conducted at a single center, a private practice institution. From 2013 to 2021, 13 patients with PSSP underwent PRF neuromodulation of the suprascapular nerve. The primary outcome measure was the Visual analog scale (VAS) score. The secondary outcome measurements included the shoulder ROM, Disability assessment scale (DAS), modified Ashworth scale, modified Rankin scale (mRS), and EuroQol-5 dimension-3L questionnaire (EQ-5D-3L) scores. These parameters were evaluated before PRF modulation, immediately after PRF modulation, and every 3 months until the final follow-up visit. Results : Six men and seven women were enrolled, and all patients were followed-up for a minimum of 12 months. The mean VAS score was 7.07 points before PRF neuromodulation and 2.38 points immediately post-procedure. Shoulder ROM for abduction and flexion, DAS for pain, mRS, and EQ-5D-3L demonstrated marked improvement. No complications were reported. Conclusion : PRF neuromodulation of the suprascapular nerve is an effective modality in patients with PSSP, and has long-term effect of pain relief, improvement of QoL. [ABSTRACT FROM AUTHOR]

The diversity of secondary metabolites of individual plants results from multiple enzymatic processes in planta and various environmental factors, such as temperature, moisture, and soil conditions. Chemical composition analysis of plants can lead to a new method to understand relationship among comparable plants along with biological classification such as genetic and anatomical method. In this study, the chemical diversity of nine different Lauraceae species was investigated, and the plant samples were chemically analyzed and classified. Multivariate analysis methods, such as PLS-DA, were used to select important metabolites distinguishing the nine Lauraceae species. The selected metabolites were identified through preparative LC-MS or MS/MS fragment pattern analysis. In addition, the chemical dendrogram for the nine Lauraceae species was interpreted through molecular network analysis and compared with the genetic dendrogram. This approach enabled us to compare the complete chemical compositions of multiple plant samples to identify relationships among plants.