Nano-sensitizer with self-amplified drug release and hypoxia normalization properties potentiates efficient chemoradiotherapy of pancreatic cancer.

The hypoxic nature of pancreatic cancer, one of the most lethal malignancies worldwide, significantly impedes the effectiveness of chemoradiotherapy. Although the development of oxygen carriers and hypoxic sensitizers has shown promise in overcoming tumor hypoxia. The heterogeneity of hypoxia-primarily caused by limited oxygen penetration-has posed challenges. In this study, we designed a hypoxia-responsive nano-sensitizer by co-loading tirapazamine (TPZ), KP372-1, and MK-2206 in a metronidazole-modified polymeric vesicle. This nano-sensitizer relies on efficient endogenous NAD(P)H quinone oxidoreductase 1-mediated redox cycling induced by KP372-1, continuously consuming periphery oxygen and achieving evenly distributed hypoxia. Consequently, the normalized tumor microenvironment facilitates the self-amplified release and activation of TPZ without requiring deep penetration. The activated TPZ and metronidazole further sensitize radiotherapy, significantly reducing the radiation dose needed for extensive cell damage. Additionally, the coloaded MK-2206 complements inhibition of therapeutic resistance caused by Akt activation, synergistically enhancing the hypoxic chemoradiotherapy. This successful hypoxia normalization strategy not only overcomes hypoxia resistance in pancreatic cancer but also provides a potential universal approach to sensitize hypoxic tumor chemoradiotherapy by reshaping the hypoxic distribution.

HOT-GAN: Hilbert Optimal Transport for Generative Adversarial Network.

Generative adversarial network (GAN) has achieved remarkable success in generating high-quality synthetic data by learning the underlying distributions of target data. Recent efforts have been devoted to utilizing optimal transport (OT) to tackle the gradient vanishing and instability issues in GAN. They use the Wasserstein distance as a metric to measure the discrepancy between the generator distribution and the real data distribution. However, most optimal transport GANs define loss functions in Euclidean space, which limits their capability in handling high-order statistics that are of much interest in a variety of practical applications. In this article, we propose a computational framework to alleviate this issue from both theoretical and practical perspectives. Particularly, we generalize the optimal transport-based GAN from Euclidean space to the reproducing kernel Hilbert space (RKHS) and propose Hilbert Optimal Transport GAN (HOT-GAN). First, we design HOT-GAN with a Hilbert embedding that allows the discriminator to tackle more informative and high-order statistics in RKHS. Second, we prove that HOT-GAN has a closed-form kernel reformulation in RKHS that can achieve a tractable objective under the GAN framework. Third, HOT-GAN's objective enjoys the theoretical guarantee of differentiability with respect to generator parameters, which is beneficial to learn powerful generators via adversarial kernel learning. Extensive experiments are conducted, showing that our proposed HOT-GAN consistently outperforms the representative GAN works.

Cross-Domain Social Rumor-Propagation Model Based on Transfer Learning.

Rumors in different topic domains have different text characteristics but similar emotional tendencies. To resolve the scarce-data problem in some rumor-topic domains, this study proposes a cross-domain rumor-propagation model, which is based on transfer learning. First, given the diversity and complexity of the rumor-propagation landscape, this study introduces a novel method, User-Retweet-Rumor2vec (URR2vec), which leverages the power of representation learning to uncover latent features within rumor topics. It also displays the forwarding relationship between users and rumors, user node information, and rumor-topic information in low-dimensional space. To capture the impact of human emotional cognition during rumor spreading, we also introduce a deep-learning model based on the natural language texts of rumor topics, which analyzes the sentiment in the text and uncovers the emotional correlations among users. Furthermore, a rumor-propagation prediction model based on the text-sentiment analysis-graph convolutional network (TSA-GCN) is proposed and pre-trained on existing rumor-topic data to ensure its prediction accuracy. Finally, considering the data sparsity at a rumor-topic outbreak, the trained propagation model is transferred to the rumor topic for prediction. Meanwhile, the rumor topic in different domains has different edges and conditional distribution, similar emotional characteristics, and network structure among the rumor topics. After fine-tuning the parameter and adding a domain adaptation layer in TSA-GCN, a domain adaptation model based on parameter and graph-structure migration is obtained.

Visceral fat area is more strongly associated with arterial stiffness than abdominal subcutaneous fat area in Chinese patients with type 2 diabetes.

BACKGROUND: Few studies have compared the correlation between visceral fat area (VFA) and abdominal subcutaneous fat area (SFA) with arterial stiffness (AS) in patients with type 2 diabetes (T2D). In addition, there is currently controversy regarding the correlation between VFA and SFA with AS. We aimed to investigate the relationship between VFA and SFA with AS in patients with T2D. METHODS: In this cross-sectional study, 1475 Chinese T2D patients with an average age of 52.32 ± 10.96 years were included. VFA and SFA were determined by a dual bioelectrical impedance analyzer, and AS was determined by measurement of brachial-ankle pulse wave conduction velocity (baPWV). Atherosclerosis was deemed present in study participants with baPWV values higher than 75th percentile (1781 cm/s). Independent correlations of logVFA and logSFA with AS were assessed using multiple linear regression and multivariate logistic regression. RESULTS: The baPWV was linked with VFA, waist circumference, and women's SFA in a general linear correlation study (P < 0.05), but not with body mass index (P = 0.3783) or men's SFA (P = 0.1899). In both men and women, VFA and SFA were positively correlated with AS, according to the generalized additive model (GAM). After fully adjusting for confounders, multiple linear regression analyses showed that for every 1-unit increase in logVFA, the beta coefficient of baPWV increased by 63.1 cm/s (95% CI: 18.4, 107.8) (P < 0.05). logSFA did not correlate significantly with baPWV (P = 0.125). In the multiple logistic regression analysis, the odds ratio (OR) of elevated baPWV was 1.8 (95% CI: 1.1, 3.1) (P = 0.019) per 1-unit increase in logVFA. logSFA did not correlate significantly with AS (P = 0.091). In the subgroup analysis, the correlation between logVFA and baPWV did not interact across subgroups (P-interaction > 0.05). CONCLUSIONS: Compared with SFA, VFA had a stronger independent positive correlation with AS in Chinese T2D patients. Patients with T2D should pay more attention to monitoring VFA and lowering it to minimize cardiovascular events.

Redox dynamics and surface structures of an active palladium catalyst during methane oxidation.

Catalysts based on palladium are among the most effective in the complete oxidation of methane. Despite extensive studies and notable advances, the nature of their catalytically active species and conceivable structural dynamics remains only partially understood. Here, we combine operando transmission electron microscopy (TEM) with near-ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) and density functional theory (DFT) calculations to investigate the active state and catalytic function of Pd nanoparticles (NPs) under methane oxidation conditions. We show that the particle size, phase composition and dynamics respond appreciably to changes in the gas-phase chemical potential. In combination with mass spectrometry (MS) conducted simultaneously with in situ observations, we uncover that the catalytically active state exhibits phase coexistence and oscillatory phase transitions between Pd and PdO. Aided by DFT calculations, we provide a rationale for the observed redox dynamics and demonstrate that the emergence of catalytic activity is related to the dynamic interplay between coexisting phases, with the resulting strained PdO having more favorable energetics for methane oxidation.

Single-Molecule Assessment of DNA Hybridization Kinetics on Dye-Loaded DNA Nanostructures.

DNA nanostructures offer a versatile platform for precise dye assembly, making them promising templates for creating photonic complexes with applications in photonics and bioimaging. However, despite these advancements, the effect of dye loading on the hybridization kinetics of single-stranded DNA protruding from DNA nanostructures remains unexplored. In this study, the DNA points accumulation for imaging in the nanoscale topography (DNA-PAINT) technique is employed to investigate the accessibility of functional binding sites on DNA-templated excitonic wires. The results indicate that positively charged dyes on DNA frameworks can accelerate the hybridization kinetics of protruded ssDNA through long-range electrostatic interactions. Furthermore, the impacts of various charged dyes and binding sites are explored on diverse DNA frameworks with varying cross-sizes. The research underscores the crucial role of electrostatic interactions in DNA hybridization kinetics within DNA-dye complexes, offering valuable insights for the functionalization and assembly of biomimetic photonic systems.

Multiomics analysis of a resistant European turnip ECD04 during clubroot infection reveals key hub genes underlying resistance mechanism.

The clubroot disease has become a worldwide threat for crucifer crop production, due to its soil-borne nature and difficulty to eradicate completely from contaminated field. In this study we used an elite resistant European fodder turnip ECD04 and investigated its resistance mechanism using transcriptome, sRNA-seq, degradome and gene editing. A total of 1751 DEGs were identified from three time points after infection, among which 7 hub genes including XTH23 for cell wall assembly and two CPK28 genes in PTI pathways. On microRNA, we identified 17 DEMs and predicted 15 miRNA-target pairs (DEM-DEG). We validated two pairs (miR395-APS4 and miR160-ARF) by degradome sequencing. We investigated the miR395-APS4 pair by CRISPR-Cas9 mediated gene editing, the result showed that knocking-out APS4 could lead to elevated clubroot resistance in B. napus. In summary, the data acquired on transcriptional response and microRNA as well as target genes provide future direction especially gene candidates for genetic improvement of clubroot resistance on Brassica species.

ARID1A loss promotes RNA editing of CDK13 in an ADAR1-dependent manner.

BACKGROUND: ARID1A, a subunit of the SWI/SNF chromatin remodeling complex, is thought to play a significant role both in tumor suppression and tumor initiation, which is highly dependent upon context. Previous studies have suggested that ARID1A deficiency may contribute to cancer development. The specific mechanisms of whether ARID1A loss affects tumorigenesis by RNA editing remain unclear. RESULTS: Our findings indicate that the deficiency of ARID1A leads to an increase in RNA editing levels and alterations in RNA editing categories mediated by adenosine deaminases acting on RNA 1 (ADAR1). ADAR1 edits the CDK13 gene at two previously unidentified sites, namely Q113R and K117R. Given the crucial role of CDK13 as a cyclin-dependent kinase, we further observed that ADAR1 deficiency results in changes in the cell cycle. Importantly, the sensitivity of ARID1A-deficient tumor cells to SR-4835, a CDK12/CDK13 inhibitor, suggests a promising therapeutic approach for individuals with ARID1A-mutant tumors. Knockdown of ADAR1 restored the sensitivity of ARID1A deficient cells to SR-4835 treatment. CONCLUSIONS: ARID1A deficiency promotes RNA editing of CDK13 by regulating ADAR1.

Analysis of the pharmacokinetics and efficacy of RBD1016 - A GalNAc-siRNA targeting Hepatitis B Virus X gene using semi-mechanistic PK/PD model.

Small interference RNA (siRNA) is a class of short double-stranded RNA molecules that cause mRNA degradation through an RNA interference mechanism and is a promising therapeutic modality. RBD1016 is a siRNA drug in clinical development for the treatment of chronic Hepatitis B Virus (HBV) infection, which contains a conjugated with N-acetylglucosamine moiety that can facilitate its hepatic delivery. We aimed to construct a semi-mechanistic model of RBD1016 in pre-clinical animals, to elucidate the pharmacokinetic/pharmacodynamic (PK/PD) profiles in mice and PK profiles in monkeys, which can lay the foundation for potential future translation of RBD1016 PK and PD from the pre-clinical stage to the clinic stage. The proposed semi-mechanistic PK/PD model fitted PK and PD data in HBV transgenic mice well and described plasma and liver concentrations in the monkeys well. The simulation results showed that our model has a reasonable predictive ability for Hepatitis B surface antigen (HBsAg) levels after multiple dosing in mice. Further PK and PD data for RBD1016, including clinical data, will assist in refining the model presented here. Our current effort focused on model building for RBD1016, we anticipate that the model could apply to other GalNAc-siRNA drugs.

Dexmedetomidine with different concentrations added to local anesthetics in erector spinae plane block: a meta-analysis of randomized controlled trials.

Background: Dexmedetomidine has been used as a perineural local anesthetic (LA) adjuvant to facilitate the potency of erector spinal plane block (ESPB). This quantitative review aimed to evaluate whether perineural dexmedetomidine for ESPB can improve the effects of analgesia compared to LA alone. Methods: Randomized controlled trials (RCTs) that investigated the addition of dexmedetomidine to LA compared to LA alone in ESPB were included. The pain scores, duration of sensory block, the time to first analgesia requirement, postoperative morphine consumption, rescue analgesia, and dexmedetomidine-related side effects were analyzed and combined using random-effects models. Results: A total of 823 patients from 13 RCTs were analyzed. Dexmedetomidine was used at the concentration of 0.5 µg/kg in three trials and 1 µg/kg in nine trials, and both in one trial. Both concentrations of dexmedetomidine perineurally administrated significantly reduced the rest VAS scores postoperatively at 12 h (0.5 µg/kg dexmedetomidine: MD = -0.86; 95% CI: -1.59 to -0.12; p = 0.02; 1 µg/kg dexmedetomidine: MD = -0.49; 95% CI: -0.83 to -0.16; p = 0.004), and 24 h (0.5 µg/kg dexmedetomidine: MD = -0.43; 95% CI: -0.74 to -0.13; p = 0.005; 1 µg/kg dexmedetomidine: MD = -0.62; 95% CI: -0.84 to -0.41; p < 0.00001). Both concentrations of dexmedetomidine added in LAs improved the dynamic VAS scores postoperatively at 12 h (0.5 µg/kg dexmedetomidine: MD = -0.55; 95% CI: -0.95 to -0.15; p = 0.007; 1 µg/kg dexmedetomidine: MD = -0.66; 95% CI: -1.05 to -0.28; p = 0.0006) and 24 h (0.5 µg/kg dexmedetomidine: MD = -0.52; 95% CI: -0.94 to -0.10; p = 0.01; 1 µg/kg dexmedetomidine: MD = -0.46; 95% CI: -0.75 to -0.16; p = 0.002). Furthermore, perineural dexmedetomidine prolonged the duration of the sensory block and the time to first analgesia requirement, reduced postoperative morphine consumption, and lowered the incidence of rescue analgesia and chronic pain. Conclusion: The meta-analysis showed that using perineural dexmedetomidine at either 0.5 µg/kg or 1 µg/kg doses in ESPB can effectively and safely enhance pain relief. Systematic review registration: PROSPERO (CRD42023424532: https://www.crd.york.ac.uk/PROSPERO/).

Capsular Repair Versus No Repair After Hip Arthroscopy in Patients Without Dysplasia or Generalized Ligamentous Laxity: A Systematic Review and Meta-analysis.

Background: Many recent studies have shown that patients who undergo capsular repair after hip arthroscopy achieve superior clinical outcomes compared with those who do not. However, patients with dysplasia or generalized ligamentous laxity (GLL) were not excluded from most of these studies, which may have affected the outcomes. Purpose: To determine whether capsular repair influences the outcomes of hip arthroscopy for patients without dysplasia or GLL. Study Design: Systematic review; Level of evidence, 1. Methods: Under the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, randomized controlled trials comparing the outcomes of capsulotomy with versus without repair were included, but studies that included patients with dysplasia or GLL were excluded. The study outcomes were patient-reported outcome measures (PROMs) at 6 months and 2 years postoperatively-including the modified Harris Hip Score (mHHS), Hip Outcome Score-Activities of Daily Living (HOS-ADL), and Hip Outcome Score-Sport-Specific Subscale (HOS-SSS)- and were compared between the repair and no-repair groups. A narrative analysis and meta-analysis were performed to integrate and compare the results of the 2 groups. In the meta-analysis of the outcome measures, studies with significant differences in the preoperative scores between the repair and no-repair groups were excluded because previous studies have shown that these can affect the outcomes. Results: A total of 761 studies were initially identified, of which 3 were included. Of the 322 included patients, 136 underwent capsular repair, and 186 underwent capsulotomy with no repair. The meta-analysis showed that capsular repair was associated with significantly higher postoperative PROMs: the mHHS at 2 years (P = .03), the HOS-ADL at 6 months (P = .02) and 2 years (P < .0001), and the HOS-SSS at 6 months (P = .02) and 2 years (P = .001). Conclusion: Capsular repair after hip arthroscopy was associated with superior clinical outcomes when compared with no capsular repair in patients without dysplasia or GLL.

Burkholderia pseudomallei BipD modulates host mitophagy to evade killing.

Mitophagy is critical for mitochondrial quality control and function to clear damaged mitochondria. Here, we found that Burkholderia pseudomallei maneuvered host mitophagy for its intracellular survival through the type III secretion system needle tip protein BipD. We identified BipD, interacting with BTB-containing proteins KLHL9 and KLHL13 by binding to the Back and Kelch domains, recruited NEDD8 family RING E3 ligase CUL3 in response to B. pseudomallei infection. Although evidently not involved in regulation of infectious diseases, KLHL9/KLHL13/CUL3 E3 ligase complex was essential for BipD-dependent ubiquitination of mitochondria in mouse macrophages. Mechanistically, we discovered the inner mitochondrial membrane IMMT via host ubiquitome profiling as a substrate of KLHL9/KLHL13/CUL3 complex. Notably, K63-linked ubiquitination of IMMT K211 was required for initiating host mitophagy, thereby reducing mitochondrial ROS production. Here, we show a unique mechanism used by bacterial pathogens that hijacks host mitophagy for their survival.

New Glires materials from the East Mesa, Erlian Basin (Nei Mongol, China).

East Mesa, an Eocene locality situated east of the Shara Murun River in the Erlian Basin, is characterized by basal lower red mudstones. However, a distinct depositional hiatus has been observed in these red mudstones, leading to their division into layers 1 and 2. Excitingly, recent discoveries respectively in layers 1 and 2 of the lower red mudstones at the East Mesa include new Glires (Gomphos sp.) and rodents such as Asiomys dawsoni, Gobiocylindrodon cf. G. ulausuensis, and Yuomys sp., marking their first occurrences within these horizons. A comparative analysis of small mammal faunas suggests that layer 1 may correspond to the Arshantan age, while layer 2 indicates a late Irdinmanhan age based on the combination of taxa found.

N-p-coumaroyloctopamine ameliorates hepatic glucose metabolism and oxidative stress involved in a PI3K/AKT/GSK3ß pathway.

Type 2 diabetes mellitus is regarded as a chronic metabolic disease characterized by hyperglycemia. Long-term hyperglycemia may result in oxidative stress, damage pancreatic ß-cell function and induce insulin resistance. Herein we explored the anti-hypoglycemic effects and mechanisms of action of N-p-coumaroyloctopamine (N-p-CO) in vitro and in vivo. N-p-CO exhibited high antioxidant activity, as indicated by the increased activity of SOD, GSH and GSH-Px in HL-7702 cells induced by both high glucose (HG) and palmitic acid (PA). N-p-CO treatment significantly augmented glucose uptake and glycogen synthesis in HG/PA-treated HL-7702 cells. Moreover, administration of N-p-CO in diabetic mice induced by both high-fat diet (HFD) and streptozotocin (STZ) not only significantly increased the antioxidant levels of GSH-PX, SOD and GSH, but also dramatically alleviated hyperglycemia and hepatic glucose metabolism in a dose-dependent manner. More importantly, N-p-CO upregulated the expressions of PI3K, AKT and GSK3ß proteins in both HG/PA-induced HL-7702 cells and HFD/STZ-induced mice. These findings clearly suggest that N-p-CO exerts anti-hypoglycemic and anti-oxidant effects, most probably via the regulation of a PI3K/AKT/GSK3ß signaling pathway. Thus, N-p-CO may have high potentials as a new candidate for the prevention and treatment of diabetes.

Patient perceptions and preferences of minimally invasive treatment modalities in varicose veins: a cross-sectional survey.

Purpose: Minimally invasive therapies (MIT) have gained popularity due to their capacity to reduce trauma, enhance aesthetic outcomes, and shorten recovery periods. This article explores patients' perceptions and preferences regarding MIT for varicose veins (VVs) while analyzing associated influencing factors to provide a better understanding. Patients and methods: A cross-sectional survey at Zhejiang Rongjun Hospital was performed from January 2022 to June 2023, involving 305 participants with VVs. The questionnaire assessed patient demographics, VVs severity, prior treatment experiences, and treatment preferences. Statistical analyses, including chi-square and Kruskal-Wallis tests, were conducted to explore the correlations between patient characteristics, treatment preferences, and factors influencing these choices. Results: Nearly half of the participants (44.3%) lacked information on any surgical options, whereas a slight majority (55.7%) possessed familiarity with at least one treatment modality, and only 9.8% knew of all six treatment methods presented. Patient surveys discerned that the majority (68.5%) declared an inadequate grasp of treatment methodologies to articulate a treatment preference. Among the 96 patients who made a treatment choice, 24.0% opted for traditional surgery, while 76.0% chose MIT and a higher preference for MIT among male patients compared to female patients (p = 0.006). The patients preferred treatment options for VVs significantly affected by vascular surgeon recommendations and the number of follow-up visits (r = 0.129, p = 0.024; r = 0.122, p = 0.033). Conclusion: The study highlights limited awareness of MIT among Chinese patients with VVs. The insights emphasize the influential role of vascular surgeons' recommendations and suggest a growing predilection for less invasive treatments due to their advantages in recovery and aesthetics. Provider-patient communication, including education about available treatments and shared decision-making, is essential to align treatment plans with patient expectations and improve outcomes.

Association between hydroxyethyl starch 130/0.4 administration during noncardiac surgery and postoperative acute kidney injury: A propensity score-matched analysis of a large cohort in China.

STUDY OBJECTIVE: The use of hydroxyethyl starch 130/0.4 has been linked to renal injury in critically ill patients, but its impact on surgical patients remains uncertain. DESIGN: A retrospective cohort study. SETTING: This study was conducted at one tertiary care hospital in China. PATIENTS: We evaluated the records of 51,926 Chinese adults who underwent noncardiac surgery from 2013 to 2022. Patients given a combination of hydroxyethyl starch 130/0.4 and crystalloids were propensity-matched at a 1: 1 ratio of baseline characteristics to patients given only crystalloids (11,725 pairs). INTERVENTIONS: Eligible patients were divided into those given a combination of hydroxyethyl starch 130/0.4 and crystalloid during surgery and a reference crystalloid group consisting of patients who were not given any colloid. MEASUREMENTS: The primary outcome was the incidence of acute kidney injury. Secondarily, acute kidney injury stage, need for renal replacement therapy, intensive care unit transfer rate, and duration of postoperative hospitalization were considered. MAIN RESULTS: After matching, hydroxyethyl starch use [8.5 (IQR: 7.5-10.0) mL/kg] did not increase the incidence of acute kidney injury compared with that in the crystalloid group [2.0 vs. 2.2%, OR: 0.90 (0.74-1.08), P = 0.25]. Nor did hydroxyethyl starch use worsen acute kidney injury stage [OR 0.90 (0.75-1.08), P = 0.26]. No significant differences between the fluid groups were observed in renal replacement therapy [OR 0.60 (0.41-0.90), P = 0.02)] or intensive care unit transfers [OR 1.02 (0.95-1.09), P = 0.53] after Bonferroni correction. Even in a subset of patients at high risk of renal injury, hydroxyethyl starch use was not associated with worse outcomes. CONCLUSIONS: Hydroxyethyl starch 130/0.4 use was not significantly associated with a greater incidence of postoperative acute kidney injury compared to receiving crystalloid solutions only.

Hybrid feature selection in a machine learning predictive model for perioperative myocardial injury in noncoronary cardiac surgery with cardiopulmonary bypass.

BACKGROUND: Perioperative myocardial injury (PMI) is associated with increased mobility and mortality after noncoronary cardiac surgery. However, limited studies have developed a predictive model for PMI. Therefore, we used hybrid feature selection (FS) methods to establish a predictive model for PMI in noncoronary cardiac surgery with cardiopulmonary bypass (CPB). METHODS: This was a single-center retrospective study conducted at the Fuwai Hospital in China. Patients aged 18-70 years who underwent elective noncoronary surgery with CPB at our institution from December 2018 to April 2021 were enrolled. The primary outcome was PMI, defined as the postoperative cardiac troponin I (cTnI) levels exceeding 220 times of upper reference limit (URL). Statistical analyses were conducted by Python (Python Software Foundation, version 3.9.7 and integrated development environment Jupyter Notebook 1.1.0) and SPSS software version 26.0 (IBM Corp., Armonk, New York, USA). RESULTS: A total of 1130 patients were eventually eligible for this study. The incidence of PMI was 20.3% (229/1130) in the overall patients, 20.6% (163/791) in the training dataset, and 19.5% (66/339) in the testing dataset. The logistic regression model performed the best AUC of 0.6893 (95 CI%: 0.6371-0.7382) by the traditional selection method, and the random forest model performed the best AUC of 0.6937 (95 CI%: 0.6416-0.7423) by the union of Wrapper and Embedded method, and the CatBoost model performed the best AUC of 0.6828 (95 CI%: 0.6304-0.7320) by the union of Embedded and forward logistic regression technique, and the Naïve Bayes model achieved the best AUC with 0.7254 (95 CI%: 0.6746-0.7723) by forwarding logistic regression method. Moreover, the decision tree, KNeighborsClassifier, and support vector machine models performed the worse AUC in all selection forms. Furthermore, the SHapley Additive exPlanations plot showed that prolonged CPB, aortic clamp time, and preoperative low platelets count were strongly related to the PMI risk. CONCLUSIONS: In total, four category feature selection methods were utilized, comprising five individual selection techniques and 15 combined methods. Notably, the combination of logistic regression and embedded methods demonstrated outstanding performance in predicting PMI risk. We also concluded that the machine learning model, including random forest, catboost, and Naive Bayes, were suitable candidates for establishing PMI predictive model. Nevertheless, additional investigation and validation are imperative for substantiating these finding.