Leveraging MRI radiomics signature for predicting the diagnosis of CXCL9 in breast cancer.

Objective: A non-invasive predictive model was developed using radiomic features to forecast CXCL9 expression level in breast cancer patients. Methods: CXCL9 expression data and MRI images of breast cancer patients were obtained from The Cancer Genome Atlas (TCGA) and The Cancer Imaging Archive (TCIA) databases, respectively. Local tissue samples from 20 breast cancer patients were collected to measure CXCL9 expression levels. Radiomic features were extracted from MRI images using 3DSlicer, and the minimum Redundancy Maximum Relevance and Recursive Feature Elimination (mRMR_RFE) method was employed to select the most pertinent radiomic features associated with CXCL9 expression levels. Support vector machine (SVM) and Logistic Regression (LR) models were utilized to construct the predictive model, and the area under the receiver operating characteristic curve (AUC) was calculated for performance evaluation. Results: CXCL9 was found to be upregulated in breast cancer patients and linked to breast cancer prognosis. Nine radiomic features were ultimately selected using the mRMR_RFE method, and SVM and LR models were trained and validated. The SVM model achieved AUC values of 0.748 and 0.711 in the training and validation sets, respectively. The LR model obtained AUC values of 0.771 and 0.724 in the training and validation sets, respectively. Conclusion: The utilization of MRI radiomic features for predicting CXCL9 expression level provides a novel non-invasive approach for breast cancer Prognostic research.

[Inhibition Effect of Resveratrol on Spontaneous Senescence of Human Bone Marrow Derived Mesenchymal Stem Cells].

OBJECTIVE: To explore whether Resveratrol (RSV) can inhibit the spontaneous senescence of human bone marrow-derived mesenchymal stem cells (MSC). METHODS: MSC were serially cultured to passage 13 and passage 15 to establish model groups exhibiting spontaneous senescence, respectively. MSC at passage 13 and passage 15 were treated with 5 nmol/L RSV for 48 h to establish the RSV-treated groups. SA-ß-Gal staining was used to detect cell senescence. MTT assay was used to detect cell proliferation. RT-PCR method was used to detect senescenceassociated telomerase activity. Western blot was used to detect the senescence-associated protein level of the phosphorylated-mTOR. RESULTS: SA-ß-Gal staining showed that the senescent cells of MSC in RSV-treated group was significantly less than those in the model group (RSV group compared with model group at passage 13, P < 0.05; RSV group compared with model group at passage 15, P < 0.01). The cell proliferation ability of MSC in RSV-treated group was significantly higher than those in model group, at 72 h in passage 13, there was significant difference between RSV-treated group and model group (P < 0.05). RT-PCR results showed that the hTERT mRNA expression of MSC in RSV-treated group was higher than that in model group, which was significantly different between RSV-treated group and model group at passage 13 (P < 0.05). Western blot results showed that the phosphorylated (Ser2448)-mTOR level of MSC in RSV-treated group was lower than that in model group, which was significantly different between RSV-treated group and model group at passage 13 (P < 0.05). CONCLUSION: RSV can inhibit the spontaneous senescence of human MSC by mediating mTOR activity.

Fabrication of Waterborne Silicone-Modified Polyurethane Nanofibers for Nonfluorine Elastic Waterproof and Breathable Membranes.

Waterproof and breathable membranes have a huge market demand in areas, such as textiles and medical protection. However, existing fluorinated nanofibrous membranes, while possessing good waterproof and breathable properties, pose health and environmental hazards. Consequently, fabricating fluorine-free, eco-friendly waterborne membranes by integrating outstanding waterproofing, breathability, and robust mechanical performance remains a significant challenge. Herein, we successfully prepared waterborne silicone-modified polyurethane nanofibrous membranes with excellent elasticity, waterproofing, and breathability properties through waterborne electrospinning, using a small quantity of poly(ethylene oxide) as a template polymer and in situ doping of the poly(carbodiimide) crosslinking agent, followed by a simple hot-pressing treatment. The silicone imparted the nanofibrous membrane with high hydrophobicity, and the crosslinking agent enabled its stable porous structure. The hot-pressing treatment (120 °C) further reduced the pore size and improved the water resistance. This environmentally friendly nanofibrous membrane showed a high elongation at break of 428%, an ultra-high elasticity of 67.5% (160 cycles under 400% tensile strain), an air transmission of 13.2 mm s-1, a water vapor transmission rate of 5476 g m-2 d-1, a hydrostatic pressure of 51.5 kPa, and a static water contact angle of 137.9°. The successful fabrication of these environmentally friendly, highly elastic membranes provides an important reference for applications in healthcare, protective textiles, and water purification.

Fast detection of tryptamine in meat products with azide-functionalized covalent organic frameworks confined in molecularly imprinted polymers.

Tryptamine is a biogenic amine that affects organoleptic quality through the generation of off-odours in foods. Herein, imine-based covalent organic frameworks (COFs) were synthesized via Schiff base reactions and postmodified with click chemistry to generate azide-functionalized COFs with tunable azide units on the walls. The combination of molecular imprinting with COFs enabled the specific recognition of the targets. The resulting optosensing system (azide-functionalized COFs@MIPs) was used as a sample-to-answer analyser for detecting tryptamine (detection time within 10 min). A linear relationship was observed for the fluorescence response to tryptamine concentrations in the range of 3-120 µg L-1, with a limit of detection of 1.74 µg L-1. The recoveries for spiked samples were satisfactory, with relative standard deviations <9.90%. The optosensing system is a potential tool for the quantitative detection of tryptamine in meat products because of its lower cost, shorter processing time, and simpler processing steps compared to conventional chromatographic techniques.

Highly Thermally Conductive Flexible Biomimetic APTES-BNNS/BC Nanocomposite Paper by Sol-Gel-Film Technology.

Owing to the evolution of 5G technology, new energy vehicles, flexible electronics, miniaturization and integration of microelectronic devices, high-frequency and high-power devices, and thermal management of materials must consider additional limitations such as electrical insulation, excellent transverse heat transfer, flexibility, and weight. Boron nitride nanosheets (BNNSs) are ideal insulating materials with high thermal conductivity. However, the problem of the 3D thermal conductivity pathway and toughness strength of nanocomposite paper loaded with inorganic thermal conductivity fillers remains a huge challenge. In this study, we propose a new method for preparing ultrathin, large, and uniformly thick BNNS for quantitative production. Bulk hexagonal boron nitride (hBN) layers were exfoliated using a simple and low-cost hydrothermal reaction, and large-scale fewer-layered BNNSs were efficiently prepared by ball milling with a high yield (up to 80%). Based on the aforementioned step, a flexible insulating composite film with high thermal conductivity and a natural "brick-mud" shell structure was constructed via the sol-gel-film conversion method. After prestretching and hot-pressing treatment, the hydrogels became denser, and the modified BNNS formed a three-dimensional (3D) network structure with an ordered orientation and interconnections in the bacterial cellulose (BC) matrix. After 100 folding cycles, the tensile strength of the nanofiber composite film reached 53 MPa, and the strength retention rate exceeded 42%. By optimizing the modified BNNS content, the thermal conductivity reached 24 W/(m·K). This simple approach has wide application potential in the next-generation electronic devices, providing options for designing thermal interface materials with excellent electrical insulation, high thermal stability, and flexibility.

Solution Blow Spinning Ultrafine Fiber Sponge-Loaded MOF-808 for Effective Adsorption and Degradation of Mustard Gas.

Functional materials that can quickly absorb and degrade mustard gas are essential for chemical warfare emergency response kits. In this study, a fiber membrane with excellent adsorption and catalytic degradation activity was developed by solution blow spinning polystyrene (PS)/polyurethane (PU) and hydrothermal in situ growth of a zirconium-based MOF (MOF-808). The mechanical properties of the PS/PU fibers were improved by adding a trimethylolpropane tris (2-methyl-1-aziridine propionate) (TTMA) cross-linking agent. Moreover, the C═O bonds in TTMA provided abundant growth sites for MOF-808 in the hydrothermal process, thereby greatly increasing the loading capacity. The fiber surface was completely covered with the MOF-808 particles within 24 h. The PS/PU/TTMA/MOF-808 fiber membrane was used for the catalytic degradation of 2-chloroethyl ethyl sulfide (CEES). The degradation efficiency reached 97.7% after 72 h, indicating its great application potential in emergency wiping cloths for mustard gas adsorption and degradation.

Predicting the Effect of the Loading Rate on the Fracture Toughness of Hydraulic Asphalt Concrete Based on the Weibull Distribution.

The cracking problem of asphalt concrete panels is a crucial consideration in the design of hydraulic asphalt concrete seepage control bodies. Panels experiencing uneven rises or falls of water levels during impoundment may exhibit loading rate effects. Investigating the fracture toughness value of asphalt concrete under varying loading rates is essential. This study employs a statistical method to calculate the fracture index KIC, using the semi-circular bending test (SCB) to examine the effect of loading rates on the Type I fracture mode of hydraulic asphalt concrete. The data are analyzed using the two-parameter Weibull distribution curve, offering insights into the minimum number of KIC test specimens. The results indicate an increase in KIC with loading rate, with greater data dispersion at faster rates. The Weibull distribution curve successfully fits the fracture behavior under different loading rates, providing valuable predictions. This study estimates the minimum number of SCB test specimens to be nine, based on a confidence level of 0.95 and a relative deviation not exceeding 5%.

Ion/Electron Co-Conductive Triple-Phase Interface Enabling Fast Redox Reaction Kinetics in Lithium-Sulfur Batteries.

Lithium-sulfur batteries (LSBs) are promising next-generation energy storage systems because of their high energy densities and high theoretical specific capacities. However, most catalysts in the LSBs are based on carbon materials, which can only improve the conductivity and are unable to accelerate lithium-ion transport. Therefore, it would be worthwhile to develop a catalytic electrode exhibiting both ion and electron conductivity. Herein, a triple-phase interface using lithium lanthanum titanate/carbon (LLTO/C) nanofibers to construct ion/electron co-conductive materials was used to afford enhanced adsorption of lithium polysulfides (LiPSs), high conductivity, and fast ion transport in working LSBs. The triple-phase interface accelerates the kinetics of the soluble LiPSs and promotes uniform Li2S precipitation/dissolution. Additionally, the LLTO/C nanofibers decrease the reaction barrier of the LiPSs, significantly improving the conversion of LiPSs to Li2S and promoting rapid conversion. Specifically, the LLTO promotes ion transport owing to its high ionic conductivity, and the carbon enhances the conductivity to improve the utilization rate of sulfur. Therefore, the LSBs with LLTO/C functional separators deliver stable life cycles, high rates, and good electrocatalytic activities. This strategy is greatly important for designing ion/electron conductivity and interface engineering, providing novel insight for the development of the LSBs.

Boosting Potassium Storage via Multifunctional Interface with High Lattice-Matching.

Atomic-scale interface engineering is a prominent strategy to address the large volume expansions and sluggish redox kinetics for reinforcing K-storage. Here, to accelerate charge transport and lower the activation energy, dual carbon-modified interfacial regions are synthesized with high lattice-matching degree, which is formed from a CoSe2 /FeSe2 heterostructure coated onto hollow carbon fibers. State-of-the-art characterization techniques and theoretical analysis, including ex-situ soft X-ray absorption spectroscopy, synchrotron X-ray tomography, ultrasonic transmission mapping, and density functional theory, are conducted to probe local atomic structure evolution, mechanical degradation mechanisms, and ion/electron migration pathways. The results suggest that the heterostructure composed of the same crystal system and space group can sharply regulate the redox kinetics of transition metal selenium and dual carbon-modified approach can tailor physicochemical degradation. Overall, this work presents the design of a stable heterojunction synergistic superior hollow carbon substrate, inspiring a pathway of interface engineering strategy toward high-performance electrode.

[Identification of a new C-23 metabolite in sterol degradation of Mycobacterium neoaurum HGMS2 and analysis of its metabolic pathways].

Mycobacterium neoaurum has the ability to produce steroidal intermediates known as 22-hydroxy-23, 24-bisnorchol-4-en-3-one (BA) upon the knockout of the genes for either the hydroxyacyl-CoA dehydrogenase (Hsd4A) or acyl-CoA thiolase (FadA5). In a previous study, we discovered a novel metabolite in the fermentation products when the fadA5 gene was deleted. This research aims to elucidate the metabolic pathway of this metabolite through structural identification, homologous sequence analysis of the fadA5 gene, phylogenetic tree analysis of M. neoaurum HGMS2, and gene knockout. Our findings revealed that the metabolite is a C23 metabolic intermediate, named 24-norchol-4-ene-3, 22-dione (designated as 3-OPD). It is formed when a thioesterase (TE) catalyzes the formation of a ß-ketonic acid by removing CoA from the side chain of 3, 22-dioxo-25, 26-bisnorchol-4-ene-24-oyl CoA (22-O-BNC-CoA), followed by spontaneously undergoing decarboxylation. These results have the potential to contribute to the development of novel steroid intermediates.

An Atypical Incontinentia Pigmenti Female with Persistent Mucocutaneous Hyperinflammation and Immunodeficiency Caused by a Novel Germline IKBKG Missense Mutation.

While most missense mutations of the IKBKG gene typically result in Ectodermal Dysplasia with Immunodeficiency, there have been rare reported instances of missense mutations of the IKBKG gene causing both Incontinentia Pigmenti (IP) and immunodeficiency in female patients. In this study, we described an atypical IP case in a 19-year-old girl, characterized by hyperpigmented and verrucous skin areas over the entire body. Remarkably, she experienced recurrent red papules whenever she had a feverish upper respiratory tract infection. Immunohistochemical staining unveiled a substantial accumulation of CD68+ macrophages alongside the TNF-α positive cells in the dermis tissue of new pustules, with increased apoptotic basal keratinocytes in the epidermis tissue of these lesions. Starting from the age of 8 years old, the patient suffered from severe and sustained chronic respiratory mucous membrane scar hyperplasia and occluded subglottic lumen. In addition to elevated erythrocyte sedimentation rate values, inflammatory cells were observed in the pathologic lesions of endobronchial biopsies and Bronchoalveolar Lavage Fluid (BALF) smear. Further histological analysis revealed a destructive bronchus epithelium integrity with extensive necrosis. Simultaneously, the patient experienced recurrent incomplete intestinal obstructions and lips contracture. The patient's BALF sample displayed an augmented profile of proinflammatory cytokines and chemokines, suggesting a potential link to systemic hyperinflammation, possibly underlying the pathogenic injuries affecting the subglottic, respiratory, and digestive systems. Furthermore, the patient presented with recurrent pneumonias and multiple warts accompanied by a T+BlowNKlow immunophenotype. Next generation sequencing showed that the patient carried a novel de novo germline heterozygous missense mutation in the IKBKG gene (c. 821T>C, p. L274P), located in the highly conserved CC2 domain. TA-cloning sequencing of patient's cDNA yielded 30 mutant transcripts out of 44 clones. In silico analysis indicated that the hydrogen bond present between Ala270 and Leu274 in the wild-type NEMO was disrupted by the Leu274Pro mutation. However, this mutation did not affect NEMO expression in peripheral blood mononuclear cells (PBMCs). Moreover, patient PBMCs exhibited significantly impaired TNF-α production following Lipopolysaccharide (LPS) stimulation. X-chromosome inactivation in T cells and neutrophils were not severely skewed. Reduced levels of IκBα phosphorylation and degradation in patient's PBMCs were observed. The NF-κB luciferase reporter assay conducted using IKBKG-deficient HEK293T cells revealed a significant reduction in NF-kB activity upon LPS stimulation. These findings adds to the ever-growing knowledge on female IP that might contribute to the better understanding of this challenging disorder.

Non-hypothetical projection pursuit regression for the prediction of hydration heat of Portland-cement-based cementitious system.

In this study, the non-hypothetical projection pursuit regression (NH-PPR) is proposed. The proposed NH-PPR model can predict the hydration heat based on the four cement phases, FA, SL, cement fineness and hydration time. The NH-PPR model is proposed by using the multiple layer iteration method and the non-hypothetical and non-parametric ridge functions to enhance accuracy and solve the problems caused by the parameter selection and the subjective hypothesis. The modeling data set is applied to train model, the testing data set is regressed and fitted into the model, and then the obtained results are compared with the BP model. To further validate the proposed model, another published data set is used to obtain a higher degree of confidence in the prediction. It is shown that the proposed model obtains the better accuracy, stability and versatility, and avoids the parameter selection and subjective hypothesis.

Construction of High Accuracy Machine Learning Interatomic Potential for Surface/Interface of Nanomaterials-A Review.

The inherent discontinuity and unique dimensional attributes of nanomaterial surfaces and interfaces bestow them with various exceptional properties. These properties, however, also introduce difficulties for both experimental and computational studies. The advent of machine learning interatomic potential (MLIP) addresses some of the limitations associated with empirical force fields, presenting a valuable avenue for accurate simulations of these surfaces/interfaces of nanomaterials. Central to this approach is the idea of capturing the relationship between system configuration and potential energy, leveraging the proficiency of machine learning (ML) to precisely approximate high-dimensional functions. This review offers an in-depth examination of MLIP principles and their execution and elaborates on their applications in the realm of nanomaterial surface and interface systems. The prevailing challenges faced by this potent methodology are also discussed.

[Biosynthesis of steroidal intermediates using Mycobacteria: a review].

Steroids are a class of medicines with important physiological and pharmacological effects. In pharmaceutical industry, steroidal intermediates are mainly prepared through Mycobacteria transformation, and then modified chemically or enzymatically into advanced steroidal compounds. Compared with the "diosgenin-dienolone" route, Mycobacteria transformation has the advantages of abundant raw materials, cost-effective, short reaction route, high yield and environmental friendliness. Based on genomics and metabolomics, the key enzymes in the phytosterol degradation pathway of Mycobacteria and their catalytic mechanisms are further revealed, which makes it possible for Mycobacteria to be used as chassis cells. This review summarizes the progress in the discovery of steroid-converting enzymes from different species, the modification of Mycobacteria genes and the overexpression of heterologous genes, and the optimization and modification of Mycobacteria as chassis cells.

Ag NW-Embedded Coaxial Nanofiber-Coated Yarns with High Stretchability and Sensitivity for Wearable Multi-Sensing Textiles.

The emerging intelligent piezoresistive yarn/textile-based sensors are of paramount importance for skin-interface electronics, owing to their unparalleled features including softness, breathability, and easy integration with functional devices. However, employing a facile way to fabricate 1D sensing yarns with mechanical robustness, multi-functional integration, and comfortability is still demanded for satisfying the practical applications. Herein, a facile one-step synchronous conjugated electrospinning and electrospraying technique is innovatively employed to continuously construct an Ag NW-embedded polyurethane (PU) nanofiber sensing yarn (AENSY) with hierarchical architecture. This 1D AENSY with weavability and stretchability can be woven into AENSY textile-based sensors integrated with functions of strain and pressure sensing. In this embedded multi-scale architecture, Ag NWs are evenly embedded and locked in the oriented and twisted PU nanofiber (PUNF) scaffold, forming the hierarchical mechanical sensing layer on the surface of the AENSY with favorable stability. Meanwhile, the presence of the elastic PUNFs enhances porosity, elasticity, and considerable deformation space, which in turn endow the AENSY textile-based sensor with a gauge factor (GF) up to 1010, a pressure sensitivity up to 16.7 N-1, high stretchability up to 160%, and high stability under long-term cycles. In addition, the AENSY textile-based sensor exhibits light weight and the unique advantage of skin-friendliness with the human body, which can be directly and conformally attached to the curved human skin to monitor the various human movements. Furthermore, the weavable AENSYs can be integrated into smart textiles with sensing arrays, which are capable for spatial pressure and strain mapping. Thus, the continuous one-step developing process and the stable embedded-twisted fiber structure provide a promising strategy to develop innovative smart yarns and textiles for personalized healthcare and human-machine interfaces.

Loop pathways are responsible for tuning the accumulation of C19- and C22-sterol intermediates in the mycobacterial phytosterol degradation pathway.

4-Androstene-3,17-dione (4-AD) and 22-hydroxy-23,24-bisnorchol-4-ene-3-one (BA) are the most important and representative C19- and C22-steroidal materials. The optimalization of sterol production with mycobacterial phytosterol conversion has been investigated for decades. One of the major challenges is that current industrial mycobacterial strains accumulate unignorable impurities analogous to desired sterol intermediates, significantly hampering product extractions and refinements. Previously, we identified Mycobacterium neoaurum HGMS2 as an efficient 4-AD-producing strain (Wang et al. in Microb Cell Fact. 19:187, 2020). Recently, we have genetically modified the HGMS2 strain to remove its major impurities including ADD and 9OH-AD (Li et al. in Microb Cell Fact. 20:158, 2021). Unexpectedly, the modified mutants started to significantly accumulate BA compared with the HGMS2 strain. In this work, while we attempted to block BA occurrence during 4-AD accumulation in HGMS2 mutants, we identified a few loop pathways that regulated metabolic flux switching between 4-AD and BA accumulations and found that both the 4-AD and BA pathways shared a 9,10-secosteroidial route. One of the key enzymes in the loop pathways was Hsd4A1, which played an important role in determining 4-AD accumulation. The inactivation of the hsd4A1 gene significantly blocked the 4-AD metabolic pathway so that the phytosterol degradation pathway flowed to the BA metabolic pathway, suggesting that the BA metabolic pathway is a complementary pathway to the 4-AD pathway. Thus, knocking out the hsd4A1 gene essentially made the HGMS2 mutant (HGMS2Δhsd4A1) start to efficiently accumulate BA. After further knocking out the endogenous kstd and ksh genes, an HGMS2Δhsd4A1 mutant, HGMS2Δhsd4A1/Δkstd1, enhanced the phytosterol conversion rate to BA in 1.2-fold compared with the HGMS2Δhsd4A1 mutant in pilot-scale fermentation. The final BA yield increased to 38.3 g/L starting with 80 g/L of phytosterols. Furthermore, we knocked in exogenous active kstd or ksh genes to HGMS2Δhsd4A1/Δ kstd1 to construct DBA- and 9OH-BA-producing strains. The resultant DBA- and 9OH-BA-producing strains, HGMS2Δhsd4A1/kstd2 and HGMS2Δkstd1/Δhsd4A1/kshA1B1, efficiently converted phytosterols to DBA- and 9OH-BA with the rates of 42.5% and 40.3%, respectively, and their final yields reached 34.2 and 37.3 g/L, respectively, starting with 80 g/L phytosterols. Overall, our study not only provides efficient strains for the industrial production of BA, DBA and 9OH-BA but also provides insights into the metabolic engineering of the HGMS2 strain to produce other important steroidal compounds.

Facile Synthesis of Aminated Graphene Quantum Dots for Promising and Selective Detection of Cobalt and Copper Ions in Aqueous Media.

Due to the rapid development of industrialization, various environmental problems such as water resource pollution are gradually emerging, among which heavy metal pollution is harmful to both human beings and the environment. As a result, there are many metal ion detection methods, among which fluorescence detection stands out because of its rapid, sensitive, low cost and non-toxic characteristics. In recent years, graphene quantum dots have been widely used and studied due to their excellent properties such as high stability, low toxicity and water solubility, and have a broad prospect in the field of metal ion detection. A novel high fluorescence Cu2+, Co2+ sensing probe produced by graphene quantum hydrothermal treatment is reported. After heat treatment with hydrazine hydrate, the small-molecule precursor nitronaphthalene synthesized by self-nitrification was transformed from blue fluorescent GQDs to green fluorescent amino-functionalized N-GQDs. Compared with other metal ions, N-GQDs are more sensitive to Cu2+ and Co2+ on the surface, and N-GQDs have much higher selectivity to Cu2+ and Co2+ than GQDs. The strategy proposed here is simple and economical in design.

Machine learning in a real-world PFO study: analysis of data from multi-centers in China.

PURPOSE: The association of patent foreman ovale (PFO) and cryptogenic stroke has been studied for years. Although device closure overall decreases the risk for recurrent stroke, treatment effects varied across different studies. In this study, we aimed to detect sub-clusters in post-closure PFO patients and identify potential predictors for adverse outcomes. METHODS: We analyzed patients with embolic stroke of undetermined sources and PFO from 7 centers in China. Machine learning and Cox regression analysis were used. RESULTS: Using unsupervised hierarchical clustering on principal components, two main clusters were identified and a total of 196 patients were included. The average age was 42.7 (12.37) years and 64.80% (127/196) were female. During a median follow-up of 739 days, 12 (6.9%) adverse events happened, including 6 (3.45%) recurrent stroke, 5 (2.87%) transient ischemic attack (TIA) and one death (0.6%). Compared to cluster 1 (n = 77, 39.20%), patients in cluster 2 (n = 119, 60.71%) were more likely to be male, had higher systolic and diastolic blood pressure, higher body mass index, lower high-density lipoprotein cholesterol and increased proportion of presence of atrial septal aneurysm. Using random forest survival (RFS) analysis, eight top ranking features were selected and used for prediction model construction. As a result, the RFS model outperformed the traditional Cox regression model (C-index: 0.87 vs. 0.54). CONCLUSIONS: There were 2 main clusters in post-closure PFO patients. Traditional cardiovascular profiles remain top ranking predictors for future recurrence of stroke or TIA. However, whether maximizing the management of these factors would provide extra benefits warrants further investigations.