A systematic review of machine learning models for management, prediction and classification of ARDS.

AIM: Acute respiratory distress syndrome or ARDS is an acute, severe form of respiratory failure characterised by poor oxygenation and bilateral pulmonary infiltrates. Advancements in signal processing and machine learning have led to promising solutions for classification, event detection and predictive models in the management of ARDS. METHOD: In this review, we provide systematic description of different studies in the application of Machine Learning (ML) and artificial intelligence for management, prediction, and classification of ARDS. We searched the following databases: Google Scholar, PubMed, and EBSCO from 2009 to 2023. A total of 243 studies was screened, in which, 52 studies were included for review and analysis. We integrated knowledge of previous work providing the state of art and overview of explainable decision models in machine learning and have identified areas for future research. RESULTS: Gradient boosting is the most common and successful method utilised in 12 (23.1%) of the studies. Due to limitation of data size available, neural network and its variation is used by only 8 (15.4%) studies. Whilst all studies used cross validating technique or separated database for validation, only 1 study validated the model with clinician input. Explainability methods were presented in 15 (28.8%) of studies with the most common method is feature importance which used 14 times. CONCLUSION: For databases of 5000 or fewer samples, extreme gradient boosting has the highest probability of success. A large, multi-region, multi centre database is required to reduce bias and take advantage of neural network method. A framework for validating with and explaining ML model to clinicians involved in the management of ARDS would be very helpful for development and deployment of the ML model.

Quantifying heterogeneity in an animal model of acute respiratory distress syndrome, a comparison of inspired sinewave technique to computed tomography.

The inspired sinewave technique (IST) is a non-invasive method to measure lung heterogeneity indices (including both uneven ventilation and perfusion or heterogeneity), which reveal multiple conditions of the lung and lung injury. To evaluate the reproducibility and predicted clinical outcomes of IST heterogeneity values, a comparison with a quantitative lung computed tomography (CT) scan is performed. Six anaesthetised pigs were studied after surfactant depletion by saline-lavage. Paired measurements of lung heterogeneity were then taken with both the IST and CT. Lung heterogeneity measured by the IST was calculated by (a) the ratio of tracer gas outputs measured at oscillation periods of 180 s and 60 s, and (b) by the standard deviation of the modelled log-normal distribution of ventilations and perfusions in the simulation lung. In the CT images, lungs were manually segmented and divided into different regions according to voxel density. A quantitative CT method to calculate the heterogeneity (the Cressoni method) was applied. The IST and CT show good Pearson correlation coefficients in lung heterogeneity measurements (ventilation: 0.71, and perfusion, 0.60, p < 0.001). Within individual animals, the coefficients of determination average ventilation (R2 = 0.53) and perfusion (R2 = 0.68) heterogeneity. Strong concordance rates of 98% in ventilation and 89% when the heterogeneity changes were reported in pairs measured by CT scanning and IST methods. This quantitative method to identify heterogeneity has the potential to replicate CT lung heterogeneity, and to aid individualised care in ARDS.

A comparison between different patient groups for diabetes management during phases of the COVID-19 pandemic: a retrospective cohort study in Ontario, Canada.

BACKGROUND: With the onset of the COVID-19 pandemic and the large uptake in virtual care in primary care in Canada, the care of patients with type 2 diabetes has been greatly affected. This includes decreased in-person visits, laboratory testing and in-person assessments such as blood pressure (BP). No studies have investigated if these changes persisted with pandemic progression, and it is unclear if shifts impacted patient groups uniformly. The purpose of this paper was to examine changes in diabetes care pre, early, and later pandemic across different patient groups. METHODS: A repeated cross-sectional design with an open cohort was used to investigate diabetes care in adults with type 2 diabetes for a 6-month interval from March 14 to September 13 over three consecutive years: 2019 (pre-pandemic period), 2020 (early pandemic period), and 2021 (later pandemic period). Data for this study were abstracted from the University of Toronto Practice-Based Research Network (UTOPIAN) Data Safe Haven, a primary care electronic medical records database in Ontario, Canada. Changes in diabetes care, which included primary care total visits, in-person visits, hemoglobin A1c (HbA1c) testing, and BP measurements were evaluated across the phases of the pandemic. Difference in diabetes care across patient groups, including age, sex, income quintile, prior HbA1c levels, and prior BP levels, were assessed. RESULTS: A total of 39,401 adults with type 2 diabetes were included in the study. Compared to the 6-month pre-pandemic period, having any in-person visits decreased significantly early pandemic (OR = 0.079 (0.076-0.082)), with a partial recovery later pandemic (OR = 0.162 (95% CI: 0.157-0.169). Compared to the pre-pandemic period, there was a significant decrease early pandemic for total visits (OR = 0.486 (95% CI: 0.470-0.503)), HbA1c testing (OR = 0.401 (95% CI: 0.389-0.413)), and BP measurement (OR = 0.121 (95% CI: 0.116-0.125)), with partial recovery later pandemic. CONCLUSIONS: All measures of diabetes care were substantially decreased early pandemic, with a partial recovery later pandemic across all patient groups. With the increase in virtual care due to the COVID-19 pandemic, diabetes care has been negatively impacted over 1-year after pandemic onset.

Complex chronic conditions among children born to women with schizophrenia.

PURPOSE: Maternal schizophrenia is linked to complications in offspring near the time of birth. Whether there is also a higher future risk of the child having a complex chronic condition (CCC) - a pediatric condition affecting any bodily system expected to last at least 12 months that is severe enough to require specialty care and/or a period of hospitalization - is not known. METHODS: In this population-based health administrative data cohort study (Ontario, Canada, 1995-2018), the risk for CCC was compared in 5066 children of women with schizophrenia (the exposed) vs. 2,939,320 unexposed children. Adjusted hazard ratios (aHR) were generated for occurrence of any CCC, by CCC category, and stratified by child sex, and child prematurity. RESULTS: CCC was more frequent in the exposed (7.7 per 1000 person-years [268 children]) than unexposed (4.2 per 100 person-years [124,452 children]) - an aHR of 1.25 (95% CI 1.10-1.41). aHRs were notably higher in 5 of 9 CCC categories: neuromuscular (1.73, 1.28-2.33), cardiovascular (1.94, 1.64-2.29), respiratory (1.83, 1.32-2.54), hematology/immunodeficiency (2.24, 1.24-4.05) and other congenital or genetic defect (1.59, 1.16-2.17). The aHR for CCC was more pronounced among boys (1.32, 1.13-1.55) than girls (1.16, 0.96-1.40), and of similar magnitude in term (1.22, 1.05-1.42) and preterm infants (1.18, 0.95-1.46). CONCLUSIONS: The risk for a CCC appears to be higher in children born to women with schizophrenia. This finding introduces opportunities for targeted preconception counselling, optimization of maternal risk factors, and intervention to support a vulnerable parent population who will experience unique challenges caring for a child with CCCs.

Nanotwin orientation on history-dependent stress decay in Cu nanopillar under constant strain.

Cu with nanotwin (NT) possesses great electrical, mechanical, and thermal properties and has potential for electronic applications. Various studies have reported the effect of NT orientation on Cu mechanical properties. However, its effect on Cu stress-relaxation behavior has not been clarified, particularly in nano-scale. In this study, Cu nanopillars with various orientations were examined by a picoindenter under constant strain and observed byin situTEM. The angles between the twin plane and the loading direction in the examined nanopillars were 0°, 60°, to 90°, and a benchmark pillar of single-crystal Cu without NT was examined. The stress drops were respectively 10%, 80%, 4%, and 50%. Owing to the interaction by NT, the dislocation behavior in nanopillars was different from that in bulk or in thin film samples. Especially, the rapid slip path of dislocations to go to the free surface of the nanopillar induced a dislocation-free zone in the 0° nanopillar, which led to work-softening. On the contrary, a high dislocation density was observed in the 90° nanopillar, which was generated by dislocation interaction and obstruction of dislocation slip by twin planes, and it led to work-hardening. The findings reveal the NT orientation in Cu nanopillars affected stress relaxation significantly.

Effect of Bonding Strength on Electromigration Failure in Cu-Cu Bumps.

In microelectronic packaging technology for three-dimensional integrated circuits (3D ICs), Cu-to-Cu direct bonding appears to be the solution to solve the problems of Joule heating and electromigration (EM) in solder microbumps under 10 µm in diameter. However, EM will occur in Cu-Cu bumps when the current density is over 106 A/cm2. The surface, grain boundary, and the interface between the Cu and TiW adhesion layer are the three major diffusion paths in EM tests, and which one may lead to early failure is of interest. This study showed that bonding strength affects the outcome. First, if the bonding strength is not strong enough to sustain the thermal mismatch of materials during EM tests, the bonding interface will fracture and lead to an open circuit of early failure. Second, if the bonding strength can sustain the bonding structure, voids will form at the passivation contact area between the Cu-Cu bump and redistribution layer (RDL) due to current crowding. When the void grows along the passivation interface and separates the Cu-Cu bump and RDL, an open circuit can occur, especially when the current density and temperature are severe. Third, under excellent bonding, when the voids at the contact area between the Cu-Cu bump and RDL do not merge together, the EM lifetime can be more than 5000 h.

Modeling of abnormal grain growth in (111) oriented and nanotwinned copper.

Uni-modal, not bi-modal, of abnormal grain growth has been observed in (111) oriented and nano-twinned Cu films. Because of the highly anisotropic microstructure, our kinetic analysis and calculation showed that it is the mobility which dominates the uni-modal growth, in which the lateral growth rate can be two orders of magnitude higher than the vertical growth rate. As a consequence, the abnormal grain growth has been converted from bi-modal to uni-modal.

Fabrication and characteristics of highly [Formula: see text]-oriented nanotwinned Au films.

Fine grained and nanotwinned Au has many excellent properties and is widely used in electronic devices. We have fabricated [Formula: see text] preferred-oriented Au thin films by DC plating at 5 mA/cm2. Microstructure analysis of the films show a unique fine grain structure with a twin formation. Hardness tests performed on electroplated [Formula: see text] Au films show a hardness 47% greater than random and untwinned Au. We then achieved direct bonding between two Au [Formula: see text] surfaces operating at 200 °C for an hour in a vacuum oven. The highly-oriented [Formula: see text] nanotwinned Au films could be an ideal material in many gold products.

Serious infections in patients with myasthenia gravis: population-based cohort study.

BACKGROUND AND PURPOSE: To characterize the frequency and risk of serious infections in patients with myasthenia gravis (MG) relative to age/sex/area-matched comparators. METHODS: This was a population-based cohort study in Ontario, Canada of patients with newly-diagnosed MG and 1:4 age/sex/area-matched general population comparators accrued from 1 April 2002 to 31 December 2015. The main outcome was a serious infection, defined by a primary diagnosis code on a hospitalization or emergency department record. We computed crude overall and sex-specific rates of infection among patients with MG and comparators, and the frequency of specific types of infection. Adjusted hazard ratios and 95% confidence intervals were estimated using Cox regression. RESULTS: Among 3823 patients with MG, 1275 (33.4%) experienced a serious infection compared with 2973/15 292 (19.4%) of comparators over a mean follow-up of over 5 years. Crude infection rates among patients with MG were twice those in comparators (72.5 vs. 35.0 per 1000 person-years, respectively). The most common infection types were respiratory infections, particularly bacterial pneumonia. After adjustment for potential confounders, MG was associated with a 39% increased infection risk (adjusted hazard ratio, 1.39; 95% confidence intervals, 1.28-1.51). CONCLUSIONS: Patients with MG are at a significantly higher absolute and relative risk of serious infections compared with age/sex/area-matched comparators. This needs to be considered when selecting MG treatments and when planning vaccination/prophylaxis. Determining whether this risk is due to the use of immunosuppressive medications (vs. MG itself) is an important area for future research.

Element Effects on High-Entropy Alloy Vacancy and Heterogeneous Lattice Distortion Subjected to Quasi-equilibrium Heating.

We applied Simmons-Balluffi methods, positron measurements, and neutron diffraction to estimate the vacancy of CoCrFeNi and CoCrFeMnNi high-entropy alloys (HEAs) using Cu as a benchmark. The corresponding formation enthalpies and associated entropies of the HEAs and Cu were calculated. The vacancy-dependent effective free volumes in both CoCrFeNi and CoCrFeMnNi alloys are greater than those in Cu, implying the easier formation of vacancies by lattice structure relaxation of HEAs at elevated temperatures. Spatially resolved synchrotron X-ray measurements revealed different characteristics of CoCrFeNi and CoCrFeMnNi HEAs subjected to quasi-equilibrium conditions at high temperatures. Element-dependent behavior revealed by X-ray fluorescence (XRF) mapping indicates the effect of Mn on the Cantor Alloy.

Effects of cold plasma, UV-C or aqueous ozone treatment on Botrytis cinerea and their potential application in preserving blueberry.

AIMS: This study aimed to compare the effects of cold plasma (CP), ultraviolet (UV-C) and aqueous ozone (AO) on Botrytis cinerea and explore their application in preserving blueberry. METHODS AND RESULTS: The effects of CP, UV-C or AO on B. cinerea were investigated. Results showed that three treatments effectively inhibited the growth of B. cinerea, increasing cell membrane penetrability and causing the leakage of cytoplasm and nucleic acid. Scanning electron microscopy showed that CP caused the mycelium fold and collapse depression; UV-C caused the mycelium shrivelled; mycelium treated with AO appeared to fold and break. In the in vivo test in blueberry, all treatments effectively inhibited the growth of micro-organisms, maintained qualities and enhanced antioxidant activities in blueberry during postharvest storage. CONCLUSION: Cold plasma, UV-C and AO effectively inhibited the growth of B. cinerea and significantly extended the shelf life of blueberry. Based on in vitro and in vivo tests, CP showed better effects to preserve blueberry compared with other treatments. SIGNIFICANCE AND IMPACT OF THE STUDY: Our findings revealed the effective mechanisms of CP, UV-C and AO to inhibit B. cinerea in vitro. CP exhibited better application potential to preserve fresh fruit than traditional sterilization methods such as UV-C and AO.

Preventing arteriovenous shunt failure in hemodialysis patients: a population-based cohort study.

Essentials Uncertainty remains about antiplatelets for vascular access patency in hemodialysis patients. 95 971 people under hemodialysis were followed in a claims database in Taiwan. Aspirin reduced vascular access failure rate and did not increase major bleeding rate. Clopidogrel, Aggrenox, and warfarin might increase major bleeding rate. SUMMARY: Background Dialysis adequacy is a major determinant of survival for patients with end-stage renal disease. Good vascular access is essential to achieve adequate dialysis. Objectives This study evaluated the impacts of different drugs on the vascular access failure rate of an arteriovenous fistula or an arteriovenous graft and the rate of major bleeding in hemodialysis patients. Patients and methods We studied patients with end-stage renal disease registered in the Taiwan National Health Insurance program from 1 January 1997 to 31 December 2012. A total of 95 971 patients were enrolled in our study. Vascular access dysfunction was defined as the need for thrombectomy or percutaneous angioplasty. Major bleeding was defined as emergency department visits or hospitalization with a primary diagnosis of gastrointestinal bleeding or intracerebral hemorrhage. The adjusted odds ratios between person-quarters with or without antiplatelet or oral anticoagulant use were calculated using a generalized estimating equation. Results The odds ratio of vascular access failure was 0.21 (0.11-0.39) for aspirin, 0.76 (0.74-0.79) for clopidogrel, 0.67 (0.59-0.77) for dipyridamole, 0.67 (0.53-0.86) for Aggrenox and 0.96 (0.90-1.03) for warfarin. The highest odds ratio for intracerebral hemorrhage was 5.33 (1.25-22.72) in younger patients using Aggrenox. The highest odds ratio for gastrointestinal bleeding was 1.34 (1.10-1.64) for clopidogrel. Conclusion Antiplatelet agents, but not warfarin, might reduce the vascular access thrombosis rate. The gastrointestinal bleeding rate was increased in the group using clopidogrel. Aggrenox should be used with caution in young individuals because it might increase the rate of intracerebral hemorrhage.

Correlation between the Microstructures of Bonding Interfaces and the Shear Strength of Cu-to-Cu Joints Using (111)-Oriented and Nanotwinned Cu.

Highly (111)-oriented Cu pillar-bumps were bonded to highly (111)-oriented Cu films at temperatures ranging from 200 °C/100 °C to 350 °C/100 °C in N2 ambient conditions. The microstructures of the bonded interfaces affected the shear strength performance of the bonded Cu joints. The bonded interfaces at 300 °C/100 °C and 350 °C/100 °C had far fewer voids than interfaces bonded at 200 °C/100 °C and 250 °C/100 °C. In addition, grain growth took place across the bonding interfaces at temperatures above 300 °C/100 °C. The corresponding orientation map (OIM) showed the preferred orientation of large grown grains to be <100>. Shear tests revealed that the fracture mode was brittle for joints bonded at 200 °C/100 °C, but became ductile after bonded above 300 °C/100 °C. Based on the results, we found that voids and grain growth behavior play import roles in the shear strength performance of bonded Cu joints.

Copper-to-copper direct bonding on highly (111)-oriented nanotwinned copper in no-vacuum ambient.

A vacuum-free Cu-to-Cu direct bonding by using (111)-oriented and nanotwinned Cu has been achieved. A fast bonding process occurs in 5 min under a temperature gradient between 450 and 100 °C. It is verified by grain growth across the bonded interface. To investigate the grain growth behavior, further annealing in the temperature gradient, as well as in a reversed temperature gradient, was performed. They showed similar recrystallization behavior with de-twinning. To analyze the de-twinning, we recall the classic model of annealing twin formation by Fullman and Fisher as comparison. Our case is opposite to the model of Fullman and Fisher. A mechanism of direct bonding by surface diffusion creep is proposed.

Comparison of oxidation in uni-directionally and randomly oriented Cu films for low temperature Cu-to-Cu direct bonding.

Cu-to-Cu direct bonding has attracted attention because it has been implemented in CMOS image sensors. Prior to the bonding, the oxides on the Cu surface needs to be removed, yet the surface may oxidize right after cleaning. Thus, oxidation is an inherent issue in the application of Cu direct bonding. Our previous study reported that Cu direct bonding can be achieved below 250 °C by using (111)-oriented nanotwinned Cu because it has the fastest surface diffusivity. However, the oxidation behavior of the nanotwinned Cu is unclear. Here, we examined the oxidation behavior of highly (111) and (200) oriented, and randomly-oriented Cu films at temperatures ranging from 120 to 250 °C. Transmission electron microscopy was used to measure the oxide thickness. The results show that the oxidation rate of (111)-oriented nanotwinned Cu has the lowest oxidation rate among them. Together, it is unique to possess the combination of the fastest surface diffusivity and the lowest oxidation rate.

Suppression of interdiffusion-induced voiding in oxidation of copper nanowires with twin-modified surface.

Cavitation and hollow structures can be introduced in nanomaterials via the Kirkendall effect in an alloying or reaction system. By introducing dense nanoscale twins into copper nanowires (CuNWs), we change the surface structure and prohibit void formation in oxidation of the nanowires. The nanotwinned CuNW exhibits faceted surfaces of very few atomic steps as well as a very low vacancy generation rate at copper/oxide interfaces. Together they lower the oxidation rate and eliminate void formation at the copper/oxide interface. We propose that the slow reaction rate together with the highly effective vacancy absorption at interfaces leads to a lattice shift in the oxidation reaction. Our findings suggest that the nanoscale Kirkendall effect can be manipulated by controlling the internal and surface crystal defects of nanomaterials.

Carvacrol and eugenol effectively inhibit Rhizopus stolonifer and control postharvest soft rot decay in peaches.

AIMS: This study aimed to investigate the antifungal mechanism of carvacrol and eugenol to inhibit Rhizopus stolonifer and the control of postharvest soft rot decay in peaches. METHODS AND RESULTS: To investigate the antifungal mechanism, the effects of carvacrol and eugenol on the mycelium growth, leakages of cytoplasmic contents, mycelium morphology, cell membrane and membrane composition of R. stolonifer were studied. Carvacrol and eugenol both exhibited dose-dependent antifungal activity against R. stolonifer, carvacrol at a concentration of 2 µl per plant and eugenol at a concentration of 4 µl per plant inhibited fungal growth completely. The two essential oils (EOs) increased cell membrane penetrability and caused the leakage of cytoplasm, nucleic acid and protein content. The observation using scanning electron microscopy and fluorescent microscopy showed modification of the hyphal morphology and breakage of the cell plasma membrane. Decreased ergosterol contents confirmed that the two EOs could destroy the membrane of R. stolonifer. For the in vivo test, the inhibition of soft rot disease and the induction of defence-related enzymes were investigated. Carvacrol and eugenol significantly reduced the incidence and severity of soft rot decay in inoculated peaches. The best treatments for controlling soft rot decay were obtained at 0·5 µl l-1 for carvacrol and 1 µl l-1 for eugenol. The activities of defence-related enzymes in peaches were also enhanced by fumigation with two EOs. CONCLUSION: This study showed that carvacrol and eugenol could effectively inhibit the growth of R. stolonifer in vitro and successfully control the incidence of soft rot decay in honey peaches. SIGNIFICANCE AND IMPACT OF THE STUDY: The above findings may be the main antifungal mechanism of carvacrol and eugenol on R. stolonifer. Furthermore, carvacrol and eugenol are helpful for their commercial application on the preservation of fresh fruit.

Identification and characterization of polymorphic microsatellite loci in the red-crowned crane.

We isolated and characterized microsatellite loci for the red-crowned crane (Grus japonensis) from a microsatellite-enriched database, which was obtained using high-throughput sequencing technology. We designed primer sets for 445 microsatellite loci and after initial screening, 34 loci were genotyped in 31 red-crowned cranes. The number of observed alleles ranged from 3 to 10. Observed and expected heterozygosities ranged from 0.197 to 0.935 and 0.453 to 0.887, respectively; the mean polymorphic information content was 0.663. Loci Lia10943, Lia60455, Lia48514, Lia62171, Lia1059, and Lia5286 deviated from expectation of the Hardy-Weinberg equilibrium; however, significant linkage disequilibrium was not observed among the 34 loci. Using these 34 markers, we successfully completed parental identification for 19 cranes. The probability of exclusion for 7 selected loci (Lia271333, Lia3745, Lia11091, Lia45761, Lia16468, Lia21909, and Lia22355) was >0.9977 and analyses with more loci increased the combination efficiency. These 34 markers were also proven to be efficient for individual identification. We recommend that this marker system be used in the systematic control of pedigree management and future genetic variation studies of red-crowned cranes.

Effect of Elastic Strain Fluctuation on Atomic Layer Growth of Epitaxial Silicide in Si Nanowires by Point Contact Reactions.

Effects of strain impact a range of applications involving mobility change in field-effect-transistors. We report the effect of strain fluctuation on epitaxial growth of NiSi2 in a Si nanowire via point contact and atomic layer reactions, and we discuss the thermodynamic, kinetic, and mechanical implications. The generation and relaxation of strain shown by in situ TEM is periodic and in synchronization with the atomic layer reaction. The Si lattice at the epitaxial interface is under tensile strain, which enables a high solubility of supersaturated interstitial Ni atoms for homogeneous nucleation of an epitaxial atomic layer of the disilicide phase. The tensile strain is reduced locally during the incubation period of nucleation by the dissolution of supersaturated Ni atoms in the Si lattice but the strained-Si state returns once the atomic layer epitaxial growth of NiSi2 occurs by consuming the supersaturated Ni.

Histone acetyltransferase PCAF accelerates apoptosis by repressing a GLI1/BCL2/BAX axis in hepatocellular carcinoma.

P300/CBP-associated factor (PCAF), a histone acetyltransferase (HAT), has been found to regulate numerous cell signaling pathways controlling cell fate by acetylating both histone and non-histone proteins. We previously reported that PCAF upregulates cell apoptosis by inactivating Serine/Threonine Protein Kinase 1 (AKT1) signaling and consequently inhibits hepatocellular carcinoma (HCC) cell growth. Here, we show that PCAF can directly acetylate cytoplasmic GLI1 protein at lysine 518, preventing its nuclear translocation and promoter occupancy, and consequently suppressing Hedgehog (Hh) signaling in HCC. Further, our results show that GLI1 can increase Bcl-2 expression and downregulate BAX. Interestingly, forced expression of PCAF reduced Bcl-2 expression, upregulated BAX and repressed cell apoptosis. Further, we provide evidence that knockdown of GLI1 abrogates the inhibitory effect of PCAF on the growth of HCC in vitro. PCAF was also found to sensitize HCC cells to 5-fluorouracil (5-FU) treatment by regulating GLI1/Bcl-2/BAX axis-dependent apoptosis. In vivo experiments also confirmed the regulatory effect of PCAF on the GLI1/Bcl-2/BAX axis and its synergistic antitumor effects with 5-FU. Gene expression microarray studies showed that PCAF was downregulated in HCC tissues compared with adjacent liver tissues and that PCAF expression was significantly associated with longer overall survival and recurrence-free survival after surgery. Together, these results show that PCAF can induce cell apoptosis by modulating a GLI1/Bcl-2/BAX axis that in turn suppresses HCC progression, and suggest that 5-FU may exert a stronger anti-tumor effect in patients with PCAF expression in HCC tumors.