Expanding density-correlation machine learning representations for anisotropic coarse-grained particles.

Physics-based, atom-centered machine learning (ML) representations have been instrumental to the effective integration of ML within the atomistic simulation community. Many of these representations build off the idea of atoms as having spherical, or isotropic, interactions. In many communities, there is often a need to represent groups of atoms, either to increase the computational efficiency of simulation via coarse-graining or to understand molecular influences on system behavior. In such cases, atom-centered representations will have limited utility, as groups of atoms may not be well-approximated as spheres. In this work, we extend the popular Smooth Overlap of Atomic Positions (SOAP) ML representation for systems consisting of non-spherical anisotropic particles or clusters of atoms. We show the power of this anisotropic extension of SOAP, which we deem AniSOAP, in accurately characterizing liquid crystal systems and predicting the energetics of Gay-Berne ellipsoids and coarse-grained benzene crystals. With our study of these prototypical anisotropic systems, we derive fundamental insights on how molecular shape influences mesoscale behavior and explain how to reincorporate important atom-atom interactions typically not captured by coarse-grained models. Moving forward, we propose AniSOAP as a flexible, unified framework for coarse-graining in complex, multiscale simulation.

Targeted isolation of dimeric sesquiterpene lactones and sesquiterpene derivatives from the roots of Inula helenium.

LC-HRMS/MS-based molecular networking was applied to investigate the bioactive sesquiterpene lactones and their analogs contained in Inula helenium, enabling the isolation of four undescribed eudesmane-eudesmane sesquiterpene dimers (1-4), a sesquiterpene-amino acid adduct (5), and 17 known sesquiterpenes (6-22). The structures were determined based on 1D, 2D NMR, and HRESIMS data analysis, as well as DP4+ probability analyses and ECD calculations. The biosynthetic pathway of the sesquiterpene dimers is postulated to proceed via the Diels-Alder reaction as the key step. The inhibitory activity of all isolates on LPS-induced nitric oxide (NO) production in RAW 264.7 macrophages was evaluated. Compounds 4, 17, and 20 exhibited significant inhibitory activity against NO production, with IC50 values of 8.4, 5.5, and 9.1 µM, respectively.

Risk factors and treatment strategies for local recurrence of locally advanced rectal cancer treated with neoadjuvant chemoradiotherapy followed by total mesorectal excision.

BACKGROUND: Despite advancements in total mesorectal excision (TME) and neoadjuvant radiotherapy, locally advanced rectal cancer remains challenging, impacting patient quality of life and mortality. This study aimed to identify the risk factors for local recurrence in locally advanced rectal cancer treated with neoadjuvant chemoradiotherapy (CRT) and assess treatment strategies for recurrence. METHODS: This retrospective analysis included 682 patients diagnosed with locally advanced rectal cancer who were treated with neoadjuvant CRT and TME at Samsung Medical Center from 2008 to 2017. The exclusion criteria ensured a homogenous cohort. Clinical staging involved colonoscopies, computed tomography, magnetic resonance imaging, and digital rectal exam. Risk factors, treatment modalities, and oncological outcomes for local recurrence were evaluated. RESULT: During a median 62-month follow-up, 47 patients (6.9 %) experienced local recurrence. The risk factors for local recurrence included a positive circumferential resection margin (CRM), venous invasion, and perineural invasion. Of the 47 patients with local recurrence, 25 (53.2 %) were considered resectable. Out of these, 23 patients underwent curative resections, and 15 (65.2 %) achieved R0 resection. Patients with R0 resections exhibited superior 5-year survival rates compared to R1-2 resection or non-surgical treatment, and there was no survival difference between R1-2 resection and non-surgical treatment. CONCLUSION: In locally advanced rectal cancer, positive CRM, venous invasion, and perineural invasion were associated with local recurrence. R0 resection showed favorable outcomes, emphasizing the importance of surveillance in high-risk patients. Treatment decisions should consider these factors for improved oncologic outcomes and quality of life.

Multi-Load Topology Optimization Design for the Structural Safety Maintenance of Low- and Intermediate-Level Radioactive Waste Packaging Containers in the Case of a Collision.

This paper presents an optimized design approach using nonlinear dynamic analysis and finite element methods to ensure the structural integrity of square-shaped containers made from ductile cast iron for intermediate- and low-level radioactive waste packaging. Ductile cast iron, with its spherical graphite structure, effectively distributes stress throughout the material, leading to a storage capacity increase of approximately 18%. Considering the critical need for containers that maintain integrity under extreme conditions like earthquakes, the design focuses on mitigating stress concentrations at the corners of square structures. Nonlinear dynamic analyses were conducted in five drop directions: three specified by ASTM-D5276 standards and two additional directions to account for different load patterns. Fractures were observed in four out of the five scenarios. For each direction where fractures occurred, equivalent loads causing similar displacement fields were applied to linear static models, which were then used for multi-load topology optimization. Three optimized models were derived, each increasing the volume by 1.4% to 1.6% compared to the original model, and the design that best met the structural integrity requirements during drop scenarios was selected. To further enhance the optimization process, weights were assigned to different load conditions based on numerical analysis results, balancing the impact of maximum stress, average stress, and plastic deformation energy. The final model, with its increased storage capacity and enhanced structural integrity, offers a practical solution for radioactive waste management, overcoming limitations in previous designs by effectively addressing complex load conditions.

Physicochemical and rheological properties of ultrasonic-assisted pregelatinized rice flour.

This study evaluated the physical and rheological properties of whole rice flour treated for different sonication times (0-15 min). Ultrasonication reduces the particle size of rice flour and improves its solubility. Viscosity tests using RVA and steady shear showed a notable decrease in the viscosity of the rehydrated pregelatinized rice flour. Although no unusual patterns were observed in the XRD analysis, the FT-IR and microstructure morphology findings suggest that ultrasonication led to structural changes in the rice flour. Overall, the study indicates that ultrasonication is a practical and clean method for producing plant-based drinks from rice flour, which could expand its limited applications in the beverage industry.

Protocol for screening and validating antibodies specific to protein phosphorylation sites using a set of yeast biopanning approaches.

Developing antibodies with high specificity against post-translationally modified epitopes remains a challenge. Yeast biopanning is well suited in screening for high-specificity binders. Here, we present a protocol for screening and validating antibodies specific to protein phosphorylation sites using a set of yeast biopanning approaches. We describe steps for screening a yeast surface display library for antibodies and other binders. We then detail procedures for validating the antibodies found by analyzing their specificity through whole-well image analysis in 96-well plates. For complete details on the use and execution of this protocol, please refer to Arbaciauskaite et al.1.

Cytotoxic Peptaibols from Trichoderma guizhouense, a Fungus Isolated from an Urban Soil Sample.

Soil sustains human life by nourishing crops, storing food sources, and housing microbes, which may affect the nutrition and biosynthesis of secondary metabolites, some of which are used as drugs. To identify lead compounds for a new class of drugs, we collected soil-derived fungal strains from various environments, including urban areas. As various human pathogens are assumed to influence the biosynthetic pathways of metabolites in soil fungi, leading to the production of novel scaffolds, we focused our work on densely populated urban areas and tourist attractions. A soil-derived fungal extract library was screened against MDA-MB-231 cells to derive their cytotoxic activity. Notably, 10 µg/mL of the extract of Trichoderma guizhouense (DS9-1) was found to exhibit an inhibitory effect of 71%. Fractionation, isolation, and structure elucidation efforts led to the identification of nine new peptaibols, trichoguizaibols A-I (1-9), comprising 14 amino acid residues (14-AA peptaibols), and three new peptaibols, trichoguizaibols J-L (10-12), comprising 18 amino acid residues (18-AA peptaibols). The chemical structures of 1-12 were determined based on their 1D and 2D NMR spectra, HRESIMS, electronic circular dichroism data, and results of the advanced Marfey's method. The 18-AA peptaibols were found to exhibit cytotoxicity against MDA-MB-231, SK-Hep1, SKOV3, DU145, and HCT116 cells greater than that of the 14-AA peptaibols. Among these compounds, 10-12 exhibited potent sub-micromolar IC50 values. These results are expected to shed light on a new direction for developing novel scaffolds as anticancer agents.

Associations Between Changes in Maximum Daily Atrial Fibrillation Duration, Ischemic Stroke, and Mortality.

BACKGROUND: Atrial fibrillation (AF) outcomes are strongly associated with continuous measures of AF burden. OBJECTIVES: This study sought to assess the association between changes in maximum daily AF duration (MDAFD) and stroke or mortality in patients with cardiac implantable electronic devices (CIEDs). METHODS: The Optum deidentified electronic health record data set (2007-2021) was linked with the Medtronic CareLink heart rhythm database. Patients with CIEDs and health care activity recorded in the electronic health record were included, excluding those with oral anticoagulation prescription. MDAFD was assessed 30 days post implant (baseline period) and 30 days before censoring or an event. HRs for the primary analysis were adjusted for components of CHA2DS2-VASc, baseline MDAFD category, and chronic kidney disease. RESULTS: Of 26,400 patients (age 68 ± 13 years; follow-up 2.6 ± 1.6 years) analyzed, 2,544 (9.6%) had AF during baseline. Increased (vs stable or decreased) MDAFD category in follow-up was associated with a higher adjusted rate of stroke and mortality (HR: 1.80; 95% CI: 1.61-2.01). There was no association between decreased MDAFD in follow-up and the combined endpoint (HR: 0.82; 95% CI: 0.68-1.00). Subgroup analysis by baseline MDAFD category demonstrated that increased MDAFD in follow-up was associated with a greater risk of stroke or mortality among patients with no AF at baseline, and decreased MDAFD in follow-up was associated with a lower risk of stroke or mortality among patients with baseline MDAFD of 1 to <5.5 hours and 5.5 to <23.5 hours. CONCLUSIONS: In CIED patients not on oral anticoagulation, increased MDAFD in follow-up was associated with a higher rate of stroke and mortality. These results suggest that AF burden, and associated risk, s not stable over time.

Engineered exosomes with a photoinducible protein delivery system enable CRISPR-Cas-based epigenome editing in Alzheimer's disease.

Effective intracellular delivery of therapeutic proteins can potentially treat a wide array of diseases. However, efficient delivery of functional proteins across the cell membrane remains challenging. Exosomes are nanosized vesicles naturally secreted by various types of cells and may serve as promising nanocarriers for therapeutic biomolecules. Here, we engineered exosomes equipped with a photoinducible cargo protein release system, termed mMaple3-mediated protein loading into and release from exosome (MAPLEX), in which cargo proteins can be loaded into the exosomes by fusing them with photocleavable protein (mMaple3)-conjugated exosomal membrane markers and subsequently released from the exosomal membrane by inducing photocleavage with blue light illumination. Using this system, we first induced transcriptional regulation by delivering octamer-binding transcription factor 4 and SRY-box transcription factor 2 to fibroblasts in vitro. Second, we induced in vivo gene recombination in Cre reporter mice by delivering Cre recombinase. Last, we achieved targeted epigenome editing in the brains of 5xFAD and 3xTg-AD mice, two models of Alzheimer's disease. Administration of MAPLEXs loaded with ß-site amyloid precursor protein cleaving enzyme 1 (Bace1)-targeting single guide RNA-incorporated dCas9 ribonucleoprotein complexes, coupled with the catalytic domain of DNA methyltransferase 3A, resulted in successful methylation of the targeted CpG sites within the Bace1 promoter. This approach led to a significant reduction in Bace1 expression, improved recognition memory impairment, and reduced amyloid pathology in 5xFAD and 3xTg-AD mice. These results suggest that MAPLEX is an efficient intracellular protein delivery system that can deliver diverse therapeutic proteins for multiple diseases.

Artificial intelligence algorithm for neoplastic cell percentage estimation and its application to copy number variation in urinary tract cancer.

Background: Bladder cancer is characterized by frequent mutations, which provide potential therapeutic targets for most patients. The effectiveness of emerging personalized therapies depends on an accurate molecular diagnosis, for which the accurate estimation of the neoplastic cell percentage (NCP) is a crucial initial step. However, the established method for determining the NCP, manual counting by a pathologist, is time-consuming and not easily executable. Methods: To address this, artificial intelligence (AI) models were developed to estimate the NCP using nine convolutional neural networks and the scanned images of 39 cases of urinary tract cancer. The performance of the AI models was compared to that of six pathologists for 119 cases in the validation cohort. The ground truth value was obtained through multiplexed immunofluorescence. The AI model was then applied to 41 cases in the application cohort that underwent next-generation sequencing testing, and its impact on the copy number variation (CNV) was analyzed. Results: Each AI model demonstrated high reliability, with intraclass correlation coefficients (ICCs) ranging from 0.82 to 0.88. These values were comparable or better to those of pathologists, whose ICCs ranged from 0.78 to 0.91 in urothelial carcinoma cases, both with and without divergent differentiation/ subtypes. After applying AI-driven NCP, 190 CNV (24.2%) were reclassified with 66 (8.4%) and 78 (9.9%) moved to amplification and loss, respectively, from neutral/minor CNV. The neutral/minor CNV proportion decreased by 6%. Conclusions: These results suggest that AI models could assist human pathologists in repetitive and cumbersome NCP calculations.

Validation of neuron-specific enolase in cardiac arrest patients with limited withdrawal of life-sustaining therapy.

Aim: We validated the prognostic performance of neuron-specific enolase (NSE) according to the recommended values in cardiac arrest (CA) survivors. Methods: We analyzed the data of adult CA survivors who underwent targeted temperature management between January 2014 and December 2020. We measured the NSE level 48 h and 72 h after CA. We performed receiver operating characteristics (ROC) and used the reference value (17 µg/L) and the guidelines-suggested value (60 µg/L) as thresholds. The primary outcome was 6-month neurological outcomes with Cerebral Performance Category (CPC), dichotomized into good (CPC 1 or 2) or poor (CPC 3-5). Results: Of the 513 included patients, 346 (67.4 %) patients had poor neurological outcomes. The area under ROC (AUC) of NSE at 48 h was 0.887 (95 % confidence intervals [CIs], 0.851-0.909) with the Youden index of 35.6 µg/L. A false positive rate (FPR) of <2 % was observed (54.1 µg/L). The thresholds values (17, 60) had a sensitivity of 86.1% and 56.7 % and a specificity of 66.7%and 98.8 %, respectively. The AUC of NSE at 72 h was 0.892 (95 % CIs, 0.849-0.920) with the Youden index of 30.4 µg/L. The threshold values (17, 60) had a sensitivity of 86.0%and 59.4 % with a specificity of 72.2%and 98.3 %, respectively. An FPR of <2 % was observed (53.6 µg/L). Among the 156 patients and 113 patients with NSE at 48 h and at 72 h ≤ 17 µg/L, respectively, 109 and 83 patients had good neurological outcomes. Conclusions: The cut-off value of NSE (60 µg/L) was acceptable to predict poor neurological outcomes with an FPR <2 % in cardiac arrest survivors, irrespective of at 48 or 72 h. NSE (17 µg/L) can function as mitigating factor to deter early WLST.

Acute kidney injury as a prognostic predictor of in-hospital mortality and neurological outcomes in patients after extracorporeal cardiopulmonary resuscitation.

INTRODUCTION: Extracorporeal cardiopulmonary resuscitation (ECPR) is increasingly being applied to patients with refractory cardiac arrest, but the survival rate to hospital discharge is only approximately 29%. Because ECPR requires intensive resources, it is important to predict outcomes. We therefore investigated the prognostic association between acute kidney injury (AKI) and ECPR to confirm the performance of AKI as a prognostic predictor of in-hospital mortality and neurological outcomes in ECPR. METHODS: We conducted a retrospective observational study on patients undergoing ECPR for cardiac etiology at Chonnam National University Hospital from 2015 to 2021. The group diagnosed with AKI in any KDIGO category within the first 48 h after ECPR was compared to that without AKI, and the primary outcome of the study was in-hospital mortality. RESULTS: Of 138 enrolled patients, 83 were studied. Hospital mortality occurred in 49 patients (59%), and 55 (66.3%) showed poor neurological outcomes. The AKI group displayed significantly elevated in-hospital mortality (77.8% vs 24.1%) and poor neurological outcomes (81.5% vs 37.9%) compared to the non-AKI group (p < 0.001). Regression analysis showed that AKI was associated with significantly higher rates of both in-hospital mortality (odds ratio (OR) range 10.75-12.88) and neurologic outcomes (OR range 5.9-6.22). CONCLUSIONS: There was a significant association of AKI with both in-hospital mortality and poor neurologic outcome in patients after ECPR, and AKI can be used as an early prognostic predictor in these patients.

Alterations in functional brain connectivity following treatment for restless legs syndrome: The role of symptom improvement in restoring functional connectivity.

The pathophysiology of restless legs syndrome (RLS) remains incompletely understood. Although several studies have investigated the alterations of brain connectivity as one of the pathophysiological mechanisms of RLS, there are only few reports on functional connectivity changes after RLS treatment. Forty-nine patients with newly diagnosed RLS and 50 healthy controls were prospectively enrolled. The patients underwent resting-state functional magnetic resonance imaging (rs-fMRI) at baseline, and 39 patients underwent follow-up rs-fMRI, 3 months after treatment with pramipexole or pregabalin. Patients were divided into good or poor medication response groups. Functional brain connectivity was analysed using rs-fMRI and graph theoretical analysis. Significant differences in functional connectivity were observed between the RLS patients and healthy controls. The average path length, clustering coefficient, transitivity, and local efficiency were lower (2.02 vs. 2.30, p < 0.001; 0.45 vs. 0.56, p < 0.001; 3.08 vs. 4.21, p < 0.001; and 0.71 vs. 0.76, p < 0.001, respectively) and the global efficiency was higher (0.53 vs. 0.50, p < 0.001) in patients with RLS than in healthy controls. Differences in functional connectivity at the global level were also observed between post- and pre-treatment RLS patients who showed a good medication response. Transitivity in the post-treatment group was higher than that in the pre-treatment group (3.22 vs. 3.04, p = 0.007). Global efficiency was positively correlated with RLS severity (r = 0.377, p = 0.007). This study demonstrates that RLS is associated with distinct alterations in brain connectivity, which can be partially normalised following symptom management. These findings suggest that therapeutic interventions for RLS modulate brain function, emphasising the importance of symptom-focussed treatment in managing RLS.

Efficient Ionovoltaic Energy Harvesting via Water-Induced p-n Junction in Reduced Graphene Oxide.

Water motion-induced energy harvesting has emerged as a prominent means of facilitating renewable electricity from the interaction between nanostructured materials and water over the past decade. Despite the growing interest, comprehension of the intricate solid-liquid interfacial phenomena related to solid state physics remains elusive and serves as a hindrance to enhancing energy harvesting efficiency up to the practical level. Herein, the study introduces the energy harvester by utilizing inversion on the majority charge carrier in graphene materials upon interaction with water molecules. Specifically, various metal electrode configurations are employed on reduced graphene oxide (rGO) to unravel its distinctive charge carriers that experience the inversion in semiconductor type upon water contact, and exploit this characteristic to leverage the efficacy of generated electricity. Through the strategic arrangement of the metal electrodes on rGO membrane, the open-circuit voltage (Voc) and short-circuit current (Isc) have exhibited a remarkable augmentation, reaching 1.05 V and 31.6 µA, respectively. The demonstration of effectively tailoring carrier dynamics via electrode configuration expands the practicality by achieving high power density and elucidating how the water-induced carrier density modulation occurs in 2D nanomaterials.

Selective delivery of imaging probes and therapeutics to the endoplasmic reticulum or Golgi apparatus: Current strategies and beyond.

To maximize therapeutic effects and minimize unwanted effects, the interest in drug targeting to the endoplasmic reticulum (ER) or Golgi apparatus (GA) has been recently growing because two organelles are distributing hubs of cellular building/signaling components (e.g., proteins, lipids, Ca2+) to other organelles and the plasma membrane. Their structural or functional damages induce organelle stress (i.e., ER or GA stress), and their aggravation is strongly related to diseases (e.g., cancers, liver diseases, brain diseases). Many efforts have been developed to image (patho)physiological functions (e.g., oxidative stress, protein/lipid-related processing) and characteristics (e.g., pH, temperature, biothiols, reactive oxygen species) in the target organelles and to deliver drugs for organelle disruption using organelle-targeting moieties. Therefore, this review will overview the structure, (patho)physiological functions/characteristics, and related diseases of the organelles of interest. Future direction on ER or GA targeting will be discussed by understanding current strategies and investigations on targeting, imaging/sensing, and therapeutic systems.

An update on the sample preparation and analytical methods for synthetic food colorants in food products.

Colorants, especially synthetic colorants, play a crucial role in enhancing the aesthetic qualities of food owing to their cost-effectiveness and stability against environmental factors. Ensuring the safe and regulated use of colorants is essential for maintaining consumer trust in food safety. Various preparation and analytical technologies, which are continuously undergoing improvement, are currently used to quantify of synthetic colorants in food products. This paper reviews recent developments in analytical techniques for synthetic food colorants, detection and compares the operational principles, advantages, and disadvantages of each technology. Additionally, it also explores advancements in these technologies, discussing several invaluable tools of analysis, such as high-performance liquid chromatography, liquid chromatography-tandem mass spectrometry, electrochemical sensors, digital image analysis, near-infrared spectroscopy, and surface-enhanced Raman spectroscopy. This comprehensive overview aims to provide valuable insights into current progress and research in the field of food colorant analysis.

Fasudil and viscosity of gelatin promote hepatic differentiation by regulating organelles in human umbilical cord matrix-mesenchymal stem cells.

BACKGROUND: Human mesenchymal stem cells originating from umbilical cord matrix are a promising therapeutic resource, and their differentiated cells are spotlighted as a tissue regeneration treatment. However, there are limitations to the medical use of differentiated cells from human umbilical cord matrix-mesenchymal stem cells (hUCM-MSCs), such as efficient differentiation methods. METHODS: To effectively differentiate hUCM-MSCs into hepatocyte-like cells (HLCs), we used the ROCK inhibitor, fasudil, which is known to induce endoderm formation, and gelatin, which provides extracellular matrix to the differentiated cells. To estimate a differentiation efficiency of early stage according to combination of gelatin and fasudil, transcription analysis was conducted. Moreover, to demonstrate that organelle states affect differentiation, we performed transcription, tomographic, and mitochondrial function analysis at each stage of hepatic differentiation. Finally, we evaluated hepatocyte function based on the expression of mRNA and protein, secretion of albumin, and activity of CYP3A4 in mature HLCs. RESULTS: Fasudil induced endoderm-related genes (GATA4, SOX17, and FOXA2) in hUCM-MSCs, and it also induced lipid droplets (LDs) inside the differentiated cells. However, the excessive induction of LDs caused by fasudil inhibited mitochondrial function and prevented differentiation into hepatoblasts. To prevent the excessive LDs formation, we used gelatin as a coating material. When hUCM-MSCs were induced into hepatoblasts with fasudil on high-viscosity (1%) gelatin-coated dishes, hepatoblast-related genes (AFP and HNF4A) showed significant upregulation on high-viscosity gelatin-coated dishes compared to those treated with low-viscosity (0.1%) gelatin. Moreover, other germline cell fates, such as ectoderm and mesoderm, were repressed under these conditions. In addition, LDs abundance was also reduced, whereas mitochondrial function was increased. On the other hand, unlike early stage of the differentiation, low viscosity gelatin was more effective in generating mature HLCs. In this condition, the accumulation of LDs was inhibited in the cells, and mitochondria were activated. Consequently, HLCs originated from hUCM-MSCs were genetically and functionally more matured in low-viscosity gelatin. CONCLUSIONS: This study demonstrated an effective method for differentiating hUCM-MSCs into hepatic cells using fasudil and gelatin of varying viscosities. Moreover, we suggest that efficient hepatic differentiation and the function of hepatic cells differentiated from hUCM-MSCs depend not only on genetic changes but also on the regulation of organelle states.

Stem Cell-Derived Extracellular Vesicle-Bearing Injectable Hydrogel for Collagen Generation in Dermis.

Despite the remarkable advances of dermal fillers that reduce wrinkles caused by dermis thickness reduction, they still lack effective hydrogel systems that stimulate collagen generation along with injection convenience. Here, we develop a stem cell-derived extracellular vesicle (EV)-bearing thermosensitive hydrogel (EVTS-Gel) for effective in vivo collagen generation. The TS-Gel undergoes sol-gel transition at 32.6 °C, as demonstrated by the storage and loss moduli crossover. Moreover, the TS-Gel and the EVTS-Gel have comparable rheological properties. Both hydrogels are injected in a sol state; hence, they require lower injection forces than conventional hydrogel-based dermal fillers. When locally administered to mouse skin, the TS-Gel extends the retention time of EVs by 2.23 times. Based on the nature of the controlled EV release, the EVTS-Gel significantly inhibits the dermis thickness reduction caused by aging compared to the bare EV treatment for 24 weeks. After a single treatment, the collagen layer thickness of the EVTS-Gel-treated dermis becomes 2.64-fold thicker than that of the bare EV-treated dermis. Notably, the collagen generation efficacy of the bare EV is poorer than that of the EVTS-Gel of a 10× lesser dose. Overall, the EVTS-Gel shows potential as an antiaging dermal filler for in vivo collagen generation.

Automated segmentation of MRI white matter hyperintensities in 8,421 patients with acute ischemic stroke.

BACKGROUND AND PURPOSE: To date, only a few small studies have attempted deep learning-based automatic segmentation of white matter hyperintensity (WMH) lesions in patients with cerebral infarction, which is complicated because stroke-related lesions can obscure WMH borders. We developed and validated deep learning algorithms to segment WMH lesions accurately in patients with cerebral infarction, using multisite datasets involving 8,421 patients with acute ischemic stroke. MATERIALS AND METHODS: We included 8,421 stroke patients from 9 centers in Korea. 2D UNet and SE-Unet models were trained using 2,408 FLAIR MRI from 3 hospitals and validated using 6,013 FLAIR MRIs from 6 hospitals. WMH segmentation performance was assessed by calculating DSC, correlation coefficient, and concordance correlation coefficient compared to a human-segmented gold standard. In addition, we obtained an uncertainty index that represents overall ambiguity in the voxel classification for WMH segmentation in each patient based on the Kullback-Leibler divergence. RESULTS: In the training dataset, the mean age was 67.4±13.0 years and 60.4% were men. The mean (95% CI) DSCs for Unet in internal testing and external validation were respectively 0.659 (0.649-0.669) and 0.710 (0.707-0.714), which were slightly lower than the reliability between humans (DSC=0.744; 95% CI=0.738-0.751; P=.031). Compared with the Unet, the SE-Unet demonstrated better performance, achieving a mean DSC of 0.675 (0.666-0.685; P<.001) in the internal testing and 0.722 (0.719-0.726; P<.001) in the external validation; moreover, it achieved high DSC values (ranging from 0.672 to 0.744) across multiple validation datasets. We observed a significant correlation between WMH volumes that were segmented automatically and manually for the Unet (r=0.917, P<.0001) and even stronger for the SE-Unet (r=0.933, P<.0001). The SE-Unet also attained a high concordance correlation coefficient (ranging from 0.841 to 0.956) in external test datasets. In addition, the uncertainty indices in the majority of patients (86%) in the external datasets were below 0.35, with an average DSC of 0.744 in these patients. CONCLUSIONS: We developed and validated deep learning algorithms to segment WMH in patients with acute cerebral infarction using the largest-ever MRI datasets. In addition, we showed that the uncertainty index can be used to identify cases where automatic WMH segmentation is less accurate and requires human review. ABBREVIATIONS: WMH = white matter hyperintensity; CNN = convolutional neural networks; SE = squeeze-and-excitation; KL = Kullback-Leibler; ReLU = rectified linear unit; LKW = last known well; mRS = modified Rankin Scale; NIHSS = National Institute of Health Stroke Scale; LAA = large artery atherosclerosis; SVO = small vessel occlusion; CE = cardioembolism.