Identification of two key biomarkers CD93 and FGL2 associated with survival of acute myeloid leukaemia by weighted gene co-expression network analysis.

Acute myeloid leukaemia (AML) is a biologically heterogeneous haematological malignancy. This study was performed to identify the potential biomarkers for the prognosis and treatment of AML. We applied weighted gene co-expression network analysis to identify key modules and hub genes related to the prognosis of AML using data from The Cancer Genome Atlas (TCGA). In total, 1581 differentially expressed genes (1096 upregulated and 485 downregulated) were identified between AML patients and healthy controls, with the blue module being the most significant among 14 modules associated with AML morphology. Through functional enrichment analysis, we identified 217 genes in the blue module significantly enriched in 'neutrophil degranulation' and 'neutrophil activation involved in immune response' pathways. The survival analysis revealed six genes (S100A9, S100A8, HK3, CD93, CXCR2 and FGL2) located in the significantly enriched pathway that were notably related to AML survival. We validated the expression of these six genes at gene and single-cell levels and identified methylation loci of each gene, except for S100A8. Finally, in vitro experiments were performed to demonstrate whether the identified hub genes were associated with AML survival. After knockdown of CD93 and FGL2, cell proliferation was significantly reduced in U937 cell line over 5 days. In summary, we identified CD93 and FGL2 as key hub genes related to AML survival, with FGL2 being a novel biomarker for the prognosis and treatment of AML.

YAP suppresses human T-cell leukemia virus type 1 transcription.

Human T-cell leukemia virus type 1 (HTLV-1) is an oncogenic retrovirus that causes adult T-cell leukemia/lymphoma (ATL). HTLV-1 encodes Tax protein that activates transcription from viral long terminal repeats (LTR). Multiple cofactors are involved in the regulation of HTLV-1 transcription via association with Tax. Yes-associated protein (YAP), which is the key effector of Hippo pathway, is elevated and activated in ATL cells. In this study, we reported that YAP protein suppressed Tax activation of HTLV-1 5' LTR but not 3' LTR. The activation of the 5' LTR by Tax was potentiated when YAP was depleted. Moreover, overexpression of YAP repressed HTLV-1 plus-strand viral gene expression and virion production, whereas compromising YAP by RNA inference augmented the expression of HTLV-1 protein. As mechanisms of YAP-mediated viral transcription inhibition, we found that YAP interacted with Tax, and prevented the association between Tax and p300. It finally led to the inhibition of recruitment of Tax to the Tax-responsive element in the 5' LTR of HTLV-1. Taken together, our results demonstrate the negative regulatory function of YAP in Tax activation of HTLV-1 transcription. It may achieve sufficient transcriptional repression to maintain persistent infection and long-term latency of HTLV-1 in the host cells.

Photodynamic Effect of Riboflavin on Chitosan Coatings and the Application in Pork Preservation.

Riboflavin (RF) was considered to be possessed of photoactivity to generate reactive oxygen species (ROS) under ultraviolet (UV) light, which is thought to be a favorable antibacterial candidate. Herein, RF was incorporated into chitosan (CS) coatings and treated under UV with different exposure times (2, 4, and 6 h) to improve the physicochemical and antibacterial properties. The results showed that the light transmittance and antibacterial performance of chitosan coatings gradually increased with the extension of the UV irradiation time. The antibacterial ability of chitosan coatings correlated with the generation of ROS: ∙OH and H2O2, which achieved 1549.08 and 95.48 µg/g, respectively, after 6 h irradiation. Furthermore, the chitosan coatings with UV irradiation also reduced the pH value, total volatile basic nitrogen (TVB-N), ΔE, and total viable counts (TVC) and improved sensory attributes of pork. In conclusion, the UV irradiated chitosan coatings could be used as an environmentally friendly antimicrobial packaging material to effectively delay the spoilage of pork, maintain its sensory quality and prolong its shelf life.

De Ritis Ratio as a Significant Prognostic Factor in Patients with Sepsis: A Retrospective Analysis.

PURPOSE: This study was performed to investigate the relationship between the aspartate transaminase and/or alanine transaminase ratio (DRR) and long-term mortality of patients diagnosed with sepsis or septic shock. MATERIALS AND METHODS: We conducted a retrospective study among adult septic patients who were admitted to the surgical intensive care unit (ICU) of the Chinese People's Liberation Army (PLA) General Hospital from January 2014 to December 2018. Baseline characteristics were compared between survivors and non survivors. We performed univariate and multivariate Cox regression analyses to evaluate the relation of DRR with 180-day mortality. The potential prognostic value of DRR in predicting mortality rate was assessed by receiver operating characteristic (ROC) curve analysis. In addition, we conducted subgroup analysis by the optimal DRR cutoff value. RESULTS: We included a total of 183 patients in the current study, and 44 (24%) patients died within 180 days of hospitalization. Univariate and multivariate Cox analyses revealed that DRR was an independent predictor of 180-day mortality (hazard ratio [HR] 1.421, 95% confidence interval [CI] 1.073-1.883, P = 0.014). The predictive accuracy of DRR for 180-day mortality was presented as an ROC curve, which had an area under the curve (AUC) of 0.708 (95% CI 0.629-0.786, P < 0.001). After we stratified all enrolled patients into two groups by using the optimal cutoff value of 1.29, we observed a significantly higher mortality in patients with a relatively high DRR. CONCLUSIONS: An elevated DRR was associated with higher 180-day mortality among septic patients, and DRR might be an optimal marker for predicting the long-term mortality of sepsis. More prospective and randomized trials are needed to confirm the prognostic value of DRR.

Anderson polyoxometalates with intrinsic oxidase-mimic activity for "turn on" fluorescence sensing of dopamine.

Anderson-type polyoxometalate containing Fe3+ and Mo6+, (NH4)3[H6Fe(III)Mo6O24] (FeMo6), was found to work as an oxidase-mimicking nanoenzyme for the first time, exhibiting the ability of catalytic oxidation of o-phenylenediamine (OPD), 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTs), and 3,3',5,5'-tetramethylbenzidine (TMB), which features easy synthesis, low cost, simple operation, and low consumption. Attributed to the nature of FeMo6 and Fenton-like effect, a novel sensor based on two consecutive "turn on" fluorescence was developed for detecting dopamine (DA) by employing the FeMo6-OPD system, and the linear range was from 1 to 100 µM with the detection limit 0.0227 µM (3σ/s). Moreover, to increase oxidase-mimic activity of FeMo6, reduced graphene oxide (rGO) loading FeMo6 composites (FeMo6@rGO (n), n = 5%, 10%, 15%) was fabricated, and results show that oxidase-like activities of FeMo6@rGO (n) are dependent on the mass ratio of FeMo6/rGO, and FeMo6@rGO (10%) exhibits the highest oxidase-mimic activity and the fastest respond time (4 min) among all reported oxidase mimic of DA to date. Graphical abstract Anderson-type Mo-POMs FeMo6 was found to work as an oxidase-mimicking nanoenzyme for the first time and was used to detect DA for two consecutive "turn on" fluorescence sensor modes.

Analysis of risk factors and prognosis of 253 lymph node metastasis in colorectal cancer patients.

BACKGROUND: This study aimed to explore potential risk factors for 253 lymph node metastasis, and to identify the prognostic impact of 253 lymph node metastasis in colorectal cancer patients. METHODS: A retrospective study was conducted of 391 colorectal cancer patients who underwent surgical treatments that included 253 lymph node dissection. Clinicopathological features, molecular indexes and 1-year overall survival rates were analyzed. RESULTS: Univariate analyses revealed the following risk factors for 253 lymph node metastasis: high preoperative levels of CEA, large tumour max diameters, and numbers of harvested lymph nodes, presence of vessel carcinoma emboli, low level of MSH6 and MLH1 immunohistochemical staining intensity. Multivariate analysis showed that elevated MLH1 immunohistochemical staining intensity was an independent protective factor for 253 lymph node metastasis (OR: 0.969, 95% CI 0.945, 0.994, P = 0.015). A significant difference was found in 1-year overall survival rate between 253 lymph node-positive and lymph node-negative colorectal cancer patients (88.9% vs.75.0%, P < 0.001). CONCLUSIONS: 253 lymph node-positive colorectal cancer patients had a worse prognosis than the 253 lymph node-negative patients. 253 lymph node dissection may improve the prognosis of colorectal cancer patients with high risk factors for 253 lymph node metastasis.

Al-Doped Ga-Based Liquid Metal: Modification Strategy and Controllable High-Temperature Lubricity through Frictional Interface Regulation.

Ga-based liquid metal (GLM) may be a potential heat-transfer material in many industries such as electronic devices and nuclear reactors, and thus, improving its lubricity is conducive to solving the friction and corrosion problems of the motion pairs served in the GLM medium. This paper proposes a modification strategy based on the conventional and large-scale mechanical milling to prepare Al-doped GLM successfully and regulates the GLM lubricity through controlling the frictional interface properties. It is found that at room temperature (∼20 °C), the Al element in the Al-doped GLM medium has a stronger adsorption capacity on the T91 surface, so the Al-rich film is formed on the frictional interfaces and reduces the wear of the T91 disk effectively compared to that in the GLM medium. However, the doping of Al is detrimental to the wear resistance of the T91 steel significantly at 400 °C by inhibiting the formation of the protective FeGa3 film on the frictional interfaces, which readily appears at T91 frictional interfaces in the GLM medium. At 600 °C, Al atoms participate in the formation of a multilayer intermetallic film, thus improving the wear resistance again. This paper provides a new idea for preparing the metal-doped GLM and improving the lubricity of GLM, and it drives forward our understanding of the lubrication mechanisms of liquid lubricants.

On Low-Rank Directed Acyclic Graphs and Causal Structure Learning.

Despite several advances in recent years, learning causal structures represented by directed acyclic graphs (DAGs) remains a challenging task in high-dimensional settings when the graphs to be learned are not sparse. In this article, we propose to exploit a low-rank assumption regarding the (weighted) adjacency matrix of a DAG causal model to help address this problem. We utilize existing low-rank techniques to adapt causal structure learning methods to take advantage of this assumption and establish several useful results relating interpretable graphical conditions to the low-rank assumption. Specifically, we show that the maximum rank is highly related to hubs, suggesting that scale-free (SF) networks, which are frequently encountered in practice, tend to be low rank. Our experiments demonstrate the utility of the low-rank adaptations for a variety of data models, especially with relatively large and dense graphs. Moreover, with a validation procedure, the adaptations maintain a superior or comparable performance even when graphs are not restricted to be low rank.

USP26 promotes colorectal cancer tumorigenesis by restraining PRKN-mediated mitophagy.

Deubiquitinating enzymes (DUBs) are promising targets for cancer therapy because of their pivotal roles in various physiological and pathological processes. Among these, ubiquitin-specific peptidase 26 (USP26) is a protease with crucial regulatory functions. Our study sheds light on the upregulation of USP26 in colorectal cancer (CRC), in which its increased expression correlates with an unfavorable prognosis. Herein, we evidenced the role of USP26 in promoting CRC tumorigenesis in a parkin RBR E3 ubiquitin-protein ligase (PRKN) protein-dependent manner. Our investigation revealed that USP26 directly interacted with PRKN protein, facilitating its deubiquitination, and subsequently reducing its activity. Additionally, we identified the K129 site on PRKN as a specific target for USP26-mediated deubiquitination. Our research highlights that a K-to-R mutation at the site on PRKN diminishes its potential for activation and ability to mediate mitophagy. In summary, our findings underscore the significance of USP26-mediated deubiquitination in restraining the activation of the PRKN-mediated mitophagy pathway, ultimately driving CRC tumorigenesis. This study not only elucidated the multifaceted role of USP26 in CRC but also introduced a promising avenue for therapeutic exploration through the development of small molecule inhibitors targeting USP26. This strategy holds promise as a novel therapeutic approach for CRC.

SOD1-high fibroblasts derived exosomal miR-3960 promotes cisplatin resistance in triple-negative breast cancer by suppressing BRSK2-mediated phosphorylation of PIMREG.

Platinum-based neoadjuvant therapy represented by cisplatin is widely employed in treating Triple-Negative Breast Cancer (TNBC), a particularly aggressive subtype of breast cancer. Nevertheless, the emergence of cisplatin resistance presents a formidable challenge to clinical chemotherapy efficacy. Herein, we revealed the critical role of tumor microenvironment (TME) derived exosomal miR-3960 and phosphorylation at the S16 site of PIMREG in activating NF-κB signaling pathway and promoting cisplatin resistance of TNBC. Detailed regulatory mechanisms revealed that SOD1-upregulated fibroblasts secrete miR-3960 and are then transported into TNBC cells via exosomes. Within TNBC cells, miR-3960 targets and inhibits the expression of BRSK2, an AMPK protein kinase family member. Furthermore, we emphasized that BRSK2 contributes to ubiquitination degradation of PIMREG and modulates subsequent activation of the NF-κB signaling pathway by mediating PIMREG phosphorylation at the S16 site, ultimately affects the cisplatin resistance of TNBC. In conclusion, our research demonstrated the crucial role of SOD1high fibroblast, exosomal miR-3960 and S16 site phosphorylated PIMREG in regulating the NF-κB signaling pathway and cisplatin resistance of TNBC. These findings provided significant potential as biomarkers for accurately diagnosing cisplatin-resistant TNBC patients and guiding chemotherapy strategy selection.

The physiochemical and photodynamic inactivation properties of corn starch/erythrosine B composite film and its application on pork preservation.

This study explored the effect of erythrosine B (EB) as a photosensitizer in corn starch (CS) film and its physicochemical properties and photodynamic bacteriostatic ability against Staphylococcus aureus, Escherichia coli, and Salmonella both in vitro and inoculated on pork under the irradiation of D65 light-emitting diode (LED) (400-800 nm). The study revealed that the physiochemical properties of CS films: moisture content, water solubility, and water vapor transmission were improved with the addition of EB. In addition, the elasticity and the thermal stability of the film were enhanced. The results showed that the CS-EB films stimulated a maximum of 26.36 µg/mL hydrogen peroxide and 74.5 µg/g hydroxyl radical under irradiation. The CS composite films with a 5 % concentration of EB inhibited the bacterial growth by 4.7 Log CFU/mL in vitro after 30 min of illumination, and 2.4 Log CFU/mL on the pork samples under the same experimental condition. Moreover, the antibacterial ability was enhanced with the increase in EB concentration. Overall, the CS-EB composite films can inhibit the growth of bacteria through photodynamic inactivation and has the potential to become a new type of environmentally friendly packaging material.

Effect of pores on microscopic wear properties and deformation behavior of Ni-Cr alloy coating.

CONTEXT: Molecular dynamics (MD) was carried out to simulate the friction behavior of Ni-Cr alloy coating containing pores. The mechanical properties, displacement, abrasion depth, and defect change patterns of the coating under nano-friction were studied. It was found that the stacking fault would extend to the pores, and both tangential and normal forces decreased when the grinding ball was above the pores. Meanwhile, the pores changed the extension direction of shear strain inside the coating, and stress concentrations were generated at the pores. In addition, the deformation behavior inside the coating was influenced by the processing depth, the smaller the relative height of the grinding ball and the pore, the greater the atomic deformation around the pore. The pores changed the path of atomic movement, resulting in less deformation of the coating below the pores. The presence of pores promoted the generation of surface steps and increased the amount of wear on the coating. It was also found that pores facilitated energy release and provided space for dislocation extension, and the large accumulation of dislocations led to frictional strengthening near the pores, which enhanced the properties of the material below the pores. It was found that the increase of the pore size caused the normal force decrease and the wear performance of the coating decrease, but the thermal insulation performance would be improved. METHODS: In this paper, nanoscale modeling was performed in the large-scale atomic/molecular parallel simulator (LAMMPS) simulation environment. The model was visualized and analyzed in three dimensions by Open Visualization Tool (OVITO), the common neighbor analysis (CNA) method was used to obtain the atomic structure information, and the dislocation analysis (DXA) method was applied to obtain the dislocations.

WToE: Learning When to Explore in Multiagent Reinforcement Learning.

Existing multiagent exploration works focus on how to explore in the fully cooperative task, which is insufficient in the environment with nonstationarity induced by agent interactions. To tackle this issue, we propose When to Explore (WToE), a simple yet effective variational exploration method to learn WToE under nonstationary environments. WToE employs an interaction-oriented adaptive exploration mechanism to adapt to environmental changes. We first propose a novel graphical model that uses a latent random variable to model the step-level environmental change resulting from interaction effects. Leveraging this graphical model, we employ the supervised variational auto-encoder (VAE) framework to derive a short-term inferred policy from historical trajectories to deal with the nonstationarity. Finally, agents engage in exploration when the short-term inferred policy diverges from the current actor policy. The proposed approach theoretically guarantees the convergence of the Q -value function. In our experiments, we validate our exploration mechanism in grid examples, multiagent particle environments and the battle game of MAgent environments. The results demonstrate the superiority of WToE over multiple baselines and existing exploration methods, such as MAEXQ, NoisyNets, EITI, and PR2.

Remarkable Wear Resistance in a Complex Concentrated Alloy with Nanohierarchical Architecture and Composition Undulation.

Sustained wear damages on the sliding surfaces of alloys are generally the culprit responsible for the failure of various mechanical systems. Inspired by high-entropy effects, here we deliberately deploy nanohierarchical architecture with composition undulation in a Ni50(AlNbTiV)50 complex concentrated alloy, which yields ultralow wear rate within the order of 10-7 to 10-6 mm3/Nm between room temperature and 800 °C. Such remarkable wear resistance heretofore represents one of the highest wear resistance reported for the bulk alloys or composites, and originates from the multi-type adaptive friction interface protection governed by intrinsically nano-coupled grains and nanoprecipitates. This cooperative heterostructure releases gradient frictional stress in stages upon wear at room temperature through the coexistence of multiple deformation pathways while activating a dense nanocrystalline glaze layer upon wear at 800 °C to minimize adhesive and oxidative wear. Our work uncovers a practical avenue for tailoring wear properties with multicomponent heterostructures over a wide temperature range.

Eukaryotic ribosome quality control system: a potential therapeutic target for human diseases.

Protein homeostasis is well accepted as the prerequisite for proper operation of various life activities. As the main apparatus of protein translation, ribosomes play an indispensable role in the maintenance of protein homeostasis. Nevertheless, upon stimulation of various internal and external factors, malfunction of ribosomes may be evident with the excessive production of aberrant proteins, accumulation of which can result in deleterious effects on cellular fate and even cell death. Ribosomopathies are characterized as a series of diseases caused by abnormalities of ribosomal compositions and functions. Correspondingly, cell evolves several ribosome quality control mechanisms in maintaining the quantity and quality of intracellular ribosomes, namely ribosome quality control system (RQCS). Of note, RQCS can tightly monitor the entire process from ribosome biogenesis to its degradation, with the capacity of coping with ribosomal dysfunction, including misassembled ribosomes and incorrectly synthesized ribosomal proteins. In the current literature review, we mainly introduce the RQCS and elaborate on the underlying pathogenesis of several ribosomopathies. With the in-depth understanding of ribosomal dysfunction and molecular basis of RQCS, therapeutic strategy by specifically targeting RQCS remains a promising option in treating patients with ribosomopathies and other ribosome-associated human diseases.

Sepsis-Associated Coagulopathy Predicts Hospital Mortality in Critically Ill Patients With Postoperative Sepsis.

BACKGROUND: The incidence of coagulopathy, which was responsible for poor outcomes, was commonly seen among patients with sepsis. In the current study, we aim to determine whether the presence of sepsis-associated coagulopathy (SAC) predicts the clinical outcomes among critically ill patients with postoperative sepsis. METHODS: We conducted a single-center retrospective cohort study by including patients with sepsis admitted to surgical ICU of Chinese PLA General Hospital from January 1, 2014 to December 31, 2018. Baseline characteristics and clinical outcomes were compared with respect to the presence of SAC. Kaplan-Meier analysis was applied to calculate survival rate, and Log-rank test was carried out to compare the differences between two groups. Furthermore, multivariable Cox and logistic and linear regression analysis were performed to assess the relationship between SAC and clinical outcomes, including hospital mortality, development of septic shock, and length of hospital stay (LOS), respectively. Additionally, both sensitivity and subgroup analyses were performed to further testify the robustness of our findings. RESULTS: A total of 175 patients were included in the current study. Among all included patients, 41.1% (72/175) ICU patients were identified as having SAC. In-hospital mortality rates were significantly higher in the SAC group when compared to that of the No SAC group (37.5% vs. 11.7%; p < 0.001). By performing univariable and multivariable regression analyses, presence of SAC was demonstrated to significantly correlate with an increased in-hospital mortality for patients with sepsis in surgical ICU [Hazard ratio (HR), 3.75; 95% Confidence interval (CI), 1.90-7.40; p < 0.001]. Meanwhile, a complication of SAC was found to be the independent predictor of the development of septic shock [Odds ratio (OR), 4.11; 95% CI, 1.81-9.32; p = 0.001], whereas it was not significantly associated with prolonged hospital LOS (OR, 0.97; 95% CI, 0.83-1.14; p = 0.743). CONCLUSION: The presence of SAC was significantly associated with increased risk of in-hospital death and septic shock among postoperative patients with sepsis admitted to ICU. Moreover, there was no statistical difference of hospital LOS between the SAC and no SAC groups.

The application of photodynamic inactivation to microorganisms in food.

Nowadays, food safety issues have drawn increased attention due to the continual occurrence of infectious diseases caused by foodborne pathogens, which is an important factor causing food safety hazard. Meanwhile, the emergence of an increasing number of antibiotic-resistant pathogens is a worrisome phenomenon. Therefore, it is imperative to find new technologies with low-cost to inactivate pathogenic microorganisms and prevent cross-contamination. Compared with traditional preservatives, photodynamic inactivation (PDI) has emerged as a novel and promising strategy to eliminate foodborne pathogens with advantages such as non-toxic and low microbial resistance, which also meets the demand of current consumers for green treatment. Over the past few years, reports of using this technology for food safety have increased rapidly. This review summarizes recent progresses in the development of photodynamic inactivation of foodborne microorganisms. The mechanisms, factors influencing PDI and the application of different photosensitizers (PSs) in different food substrates are reviewed.

The Role and Regulatory Mechanism of Transcription Factor EB in Health and Diseases.

Transcription factor EB (TFEB) is a member of the microphthalmia-associated transcription factor/transcription factor E (MiTF/TFE) family and critically involved in the maintenance of structural integrity and functional balance of multiple cells. In this review, we described the effects of post-transcriptional modifications, including phosphorylation, acetylation, SUMOylation, and ubiquitination, on the subcellular localization and activation of TFEB. The activated TFEB enters into the nucleus and induces the expressions of targeted genes. We then presented the role of TFEB in the biosynthesis of multiple organelles, completion of lysosome-autophagy pathway, metabolism regulation, immune, and inflammatory responses. This review compiles existing knowledge in the understanding of TFEB regulation and function, covering its essential role in response to cellular stress. We further elaborated the involvement of TFEB dysregulation in the pathophysiological process of various diseases, such as the catabolic hyperactivity in tumors, the accumulation of abnormal aggregates in neurodegenerative diseases, and the aberrant host responses in inflammatory diseases. In this review, multiple drugs have also been introduced, which enable regulating the translocation and activation of TFEB, showing beneficial effects in mitigating various disease models. Therefore, TFEB might serve as a potential therapeutic target for human diseases. The limitation of this review is that the mechanism of TFEB-related human diseases mainly focuses on its association with lysosome and autophagy, which needs deep description of other mechanism in diseases progression after getting more advanced information.

Lysosomal quality control of cell fate: a novel therapeutic target for human diseases.

In eukaryotic cells, lysosomes are digestive centers where biological macromolecules are degraded by phagocytosis and autophagy, thereby maintaining cellular self-renewal capacity and energy supply. Lysosomes also serve as signaling hubs to monitor the intracellular levels of nutrients and energy by acting as platforms for the assembly of multiple signaling pathways, such as mammalian target of rapamycin complex 1 (mTORC1) and adenosine 5'-monophosphate (AMP)-activated protein kinase (AMPK). The structural integrity and functional balance of lysosomes are essential for cell function and viability. In fact, lysosomal damage not only disrupts intracellular clearance but also results in the leakage of multiple contents, which pose great threats to the cell by triggering cell death pathways, including apoptosis, necroptosis, pyroptosis, and ferroptosis. The collapse of lysosomal homeostasis is reportedly critical for the pathogenesis and development of various diseases, such as tumors, neurodegenerative diseases, cardiovascular diseases, and inflammatory diseases. Lysosomal quality control (LQC), comprising lysosomal repair, lysophagy, and lysosomal regeneration, is rapidly initiated in response to lysosomal damage to maintain lysosomal structural integrity and functional homeostasis. LQC may be a novel but pivotal target for disease treatment because of its indispensable role in maintaining intracellular homeostasis and cell fate.

A Machine Learning-Based Prediction of Hospital Mortality in Patients With Postoperative Sepsis.

Introduction: The incidence of postoperative sepsis is continually increased, while few studies have specifically focused on the risk factors and clinical outcomes associated with the development of sepsis after surgical procedures. The present study aimed to develop a mathematical model for predicting the in-hospital mortality among patients with postoperative sepsis. Materials and Methods: Surgical patients in Medical Information Mart for Intensive Care (MIMIC-III) database who simultaneously fulfilled Sepsis 3.0 and Agency for Healthcare Research and Quality (AHRQ) criteria at ICU admission were incorporated. We employed both extreme gradient boosting (XGBoost) and stepwise logistic regression model to predict the in-hospital mortality among patients with postoperative sepsis. Consequently, the model performance was assessed from the angles of discrimination and calibration. Results: We included 3,713 patients who fulfilled our inclusion criteria, in which 397 (10.7%) patients died during hospitalization, and 3,316 (89.3%) patients survived through discharge. Fluid-electrolyte disturbance, coagulopathy, renal replacement therapy (RRT), urine output, and cardiovascular surgery were important features related to the in-hospital mortality. The XGBoost model had a better performance in both discriminatory ability (c-statistics, 0.835 vs. 0.737 and 0.621, respectively; AUPRC, 0.418 vs. 0.280 and 0.237, respectively) and goodness of fit (visualized by calibration curve) compared to the stepwise logistic regression model and baseline model. Conclusion: XGBoost model has a better performance in predicting hospital mortality among patients with postoperative sepsis in comparison to the stepwise logistic regression model. Machine learning-based algorithm might have significant application in the development of early warning system for septic patients following major operations.