OBJECTIVE: To evaluate the advantages and limitations of transcutaneous electrical acupoint stimulation (TEAS) in the treatment of patients with severe gastrointestinal function injury in intensive care unit (ICU) by analyzing dynamic changes of intestinal fatty acid binding protein (I-FABP), D-lactic acid and citrulline. METHODS: A prospective single-center randomized controlled trial was conducted. Patients with severe gastrointestinal function injury admitted to the ICU from February 2021 to January 2024 were enrolled [age > 18 years old, acute gastrointestinal injury (AGI) grade 2 to 3, stable hemodynamics]. Patients with different AGI grades were randomly assigned in a 1:1 ratio to the TEAS group and the control group using simple randomization. Both groups received conventional treatment and enteral nutrition (EN). In addition, the TEAS group underwent TEAS at the Neiguan and Zusanli points for 30 minutes per session, twice daily for 7 days. Baseline data, including age, gender, underlying diseases, and primary diagnoses, were recorded. Three intestinal biomarkers, such as I-FABP, D-lactic acid, and citrulline were measured before and after 7 days of treatment. EN tolerance indicators and 28 days survival status were documented. The differences in various indicators were compared between the two groups, subgroup analyses were conducted based on AGI grading, and interaction between AGI grade and TEAS were analyzed. The 28-day Kaplan-Meier survival curves were generated for both groups. RESULTS: Finally, 133 patients were included, with 68 in the TEAS group and 65 in the control group. Baseline characteristics were comparable between the two groups. A comparison of the dynamic changes in intestinal biomarkers revealed that the I-FABP level in both groups decreased after treatment compared to pre-treatment, with a more pronounced reduction in the TEAS group. The least square mean difference (LS Mean difference) for the corrected I-FABP level between the two groups during the observation period was -0.23 μg/L [95% confidence interval (95%CI) was -0.45 to -0.01], which was statistically significant (P = 0.041). Additionally, a significant interaction with AGI was observed (P = 0.004). Post-treatment, D-lactic acid level decreased in both groups compared to pre-treatment, with a more significant reduction in the TEAS group. The LS Mean difference for the corrected D-lactic acid level was -0.08 mmol/L (95%CI was -0.11 to -0.05), which was statistically significant (P < 0.001), and the interaction with AGI was also significant (P = 0.005). There was no significant change in citrulline levels between the two groups before and after treatment. The LS Mean difference for the corrected citrulline level was -0.17 μmol/L (95%CI was -1.87 to 1.53), which was not statistically significant (P = 0.845), and no significant interaction with AGI was observed (P = 0.913). Comparison of EN tolerance parameters between the two groups revealed that the TEAS group had a longer total EN time (hours: 72±31 vs. 60±28) and higher total EN calories (kJ: 11 469.23±7 237.34 vs. 6 638.76±5 098.37), as well as a higher 70% target caloric attainment rate (52.9% vs. 32.3%) compared to the control group (all P < 0.05). The incidence of abdominal distension after EN was lower in the TEAS group than that in the control group (23.5% vs. 43.1%, P < 0.05), while the incidence of diarrhea after EN was higher in the TEAS group (22.1% vs. 7.7%, P < 0.05). There were no significantly differences in AGI grade reduction rate, post-EN vomiting/gastric retention rate, incidence of feeding interruption, and 28-day survival rate between the two groups. Furthermore, there were no significantly interaction between these observation measures and AGI. Kaplan-Meier survival analysis showed that there was no significantly difference in 28-day cumulative survival rate between the TEAS group and the control group [Log-Rank test: P = 0.501, hazard ratio (HR) = 0.81, 95%CI was 0.43-1.51), and there was no significantly interaction with AGI (P = 0.702). CONCLUSIONS The advantage of TEAS in the treatment of ICU patients with severe gastrointestinal function injury lies in its ability to reverse intestinal cell necrosis and promote the reconstruction of intestinal barrier function. Additionally, gastrointestinal tolerance is significantly improved, and both the duration and total calories of EN are increased. However, the limitation of TEAS therapy is that it does not promote the recovery of intestinal cell absorption and synthesis function in the target patients. Moreover, it may lead to nutrient solution overload due to improved gastrointestinal tolerance. Furthermore, TEAS does not appear to improve 28-day cumulative survival rate in the target patients.
Humans ; Prospective Studies ; Intensive Care Units ; Acupuncture Points ; Fatty Acid-Binding Proteins/metabolism* ; Transcutaneous Electric Nerve Stimulation ; Male ; Female ; Citrulline/metabolism* ; Lactic Acid/metabolism* ; Gastrointestinal Diseases/therapy* ; Middle Aged ; Enteral Nutrition ; Adult

ObjectiveThe theory of homotherapy for heteropathy is one of the classical rules in traditional Chinese medicine. Taking this theory as a breakthrough point， this study employed gas chromatography-mass spectrometry （GC-MS） to elucidate the mechanism underlying the therapeutic effects of Zhuyuwan on both intrahepatic cholestasis （IC） and ulcerative colitis （UC） from the viewpoint of serum metabolic homeostasis. MethodsThe rat models of α-naphthylisothiocyanate （ANIT）-induced cholestasis and 2，4，6-trinitro-benzenesulfonic acid （TNBS）-induced UC were treated with low （0.6 g·kg^-1） and high （1.2 g·kg^-1） doses of Zhuyuwan by gavage. In the experiment regarding IC， 24 Sprague-Dawley （SD） rats were randomly assigned into four groups： normal， ANIT model， low-dose Zhuyuwan， and high-dose Zhuyuwan. In the experiment regarding UC， 24 SD rats were randomly allocated into four groups： normal， TNBS model， low-dose Zhuyuwan， and high-dose Zhuyuwan. Firstly， the two disease models and the intervention effects of Zhuyuwan on the two diseases were evaluated based on serum levels of biochemical indicators ［alanine aminotransferase （ALT）， aspartate transaminase （AST）， γ-glutamyltranspeptidase （γ-GT）， and total bile acid （TBA）］， colon damage score， colon weight index， disease activity index， and histopathological changes in rats. Secondly， the rat serum samples were analyzed by gas chromatography-mass spectrometry （GC-MS） to screen the common core pathways of the two disease models， and the expression of core genes in the pathways was determined by Real-time PCR， on the basis of which the biological mechanism of the treatment of the two disease models by Zhuyuwan was ultimately elucidated. ResultsThe results of the experiment regarding IC showed that the ANIT model group had higher ALT， AST， γ-GT， and TBA levels than the normal group （P<0.01）. Compared with the ANIT model group， the low-dose Zhuyuwan group showed declined ALT and TBA levels （P<0.01） and the high-dose Zhuyuwan group showed lowered ALT， TBA， AST， and γ-GT levels （P<0.01）. The results of the experiment regarding UC showed that compared with the normal group， the TNBS model group presented increases in the colonic damage score， colon weight index， and disease activity index （P<0.01）. Compared with the TNBS model group， the low-dose Zhuyuwan group showcased declines in colon weight index （P<0.01） and disease activity index （P<0.05）， and the high-dose Zhuyuwan group showed reductions in the colon damage score， colon weight index， and disease activity index （P<0.01）. GC-MS metabolomics analysis combined with qRT-PCR demonstrated that Zhuyuwan had a similar inverse regulatory effect on arginine metabolism disruption in the above two disease models. ConclusionZhuyuwan exhibited definite therapeutic effects on both IC and UC， and the regulation of arginine biosynthesis pathway is the core mechanism for the treatment of both diseases by Zhuyuwan.

ObjectiveThe theory of homotherapy for heteropathy is one of the classical rules in traditional Chinese medicine. Taking this theory as a breakthrough point， this study employed gas chromatography-mass spectrometry （GC-MS） to elucidate the mechanism underlying the therapeutic effects of Zhuyuwan on both intrahepatic cholestasis （IC） and ulcerative colitis （UC） from the viewpoint of serum metabolic homeostasis. MethodsThe rat models of α-naphthylisothiocyanate （ANIT）-induced cholestasis and 2，4，6-trinitro-benzenesulfonic acid （TNBS）-induced UC were treated with low （0.6 g·kg^-1） and high （1.2 g·kg^-1） doses of Zhuyuwan by gavage. In the experiment regarding IC， 24 Sprague-Dawley （SD） rats were randomly assigned into four groups： normal， ANIT model， low-dose Zhuyuwan， and high-dose Zhuyuwan. In the experiment regarding UC， 24 SD rats were randomly allocated into four groups： normal， TNBS model， low-dose Zhuyuwan， and high-dose Zhuyuwan. Firstly， the two disease models and the intervention effects of Zhuyuwan on the two diseases were evaluated based on serum levels of biochemical indicators ［alanine aminotransferase （ALT）， aspartate transaminase （AST）， γ-glutamyltranspeptidase （γ-GT）， and total bile acid （TBA）］， colon damage score， colon weight index， disease activity index， and histopathological changes in rats. Secondly， the rat serum samples were analyzed by gas chromatography-mass spectrometry （GC-MS） to screen the common core pathways of the two disease models， and the expression of core genes in the pathways was determined by Real-time PCR， on the basis of which the biological mechanism of the treatment of the two disease models by Zhuyuwan was ultimately elucidated. ResultsThe results of the experiment regarding IC showed that the ANIT model group had higher ALT， AST， γ-GT， and TBA levels than the normal group （P<0.01）. Compared with the ANIT model group， the low-dose Zhuyuwan group showed declined ALT and TBA levels （P<0.01） and the high-dose Zhuyuwan group showed lowered ALT， TBA， AST， and γ-GT levels （P<0.01）. The results of the experiment regarding UC showed that compared with the normal group， the TNBS model group presented increases in the colonic damage score， colon weight index， and disease activity index （P<0.01）. Compared with the TNBS model group， the low-dose Zhuyuwan group showcased declines in colon weight index （P<0.01） and disease activity index （P<0.05）， and the high-dose Zhuyuwan group showed reductions in the colon damage score， colon weight index， and disease activity index （P<0.01）. GC-MS metabolomics analysis combined with qRT-PCR demonstrated that Zhuyuwan had a similar inverse regulatory effect on arginine metabolism disruption in the above two disease models. ConclusionZhuyuwan exhibited definite therapeutic effects on both IC and UC， and the regulation of arginine biosynthesis pathway is the core mechanism for the treatment of both diseases by Zhuyuwan.

Objectives:To investigate the physical growth status of pediatric patients with transfusion-dependent thalassemia (TDT) and analyze the effects of treatment-related and socioeconomic factors on physical growth.Methods:Based on the specialized thalassemia database from gene therapy clinical research at the Institute of Hematology & Hospital of Blood Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, we collected data on height and weight development, family economic status, and medical records of 338 pediatric patients with TDT from October 2023 to May 2024. The length/height-for-age and body mass index (BMI) -for-age were classified based on the Growth Standard for Children under 7 Years of Age, Standard for Height Level Classification among Children and Adolescents Aged 7-18 Years, and Dietary Guidelines for Chinese Residents. Logistic regression analysis was conducted to assess the effects of family economic status and disease-related treatment on length/height-for-age and BMI-for-age.Results:Among the 338 patients, 118 were children and 220 were adolescents (192 males and 146 females), with a median age of 12 years (range: 0.8-18) and a median diagnosis duration of 10.3 years (range: 0.5-17.9). Subtypes included α-thalassemia [21 cases (6.2%) ], β-thalassemia [288 cases (85.2%) ], and combined αβ-thalassemia[29 cases (8.6%) ]. The monthly household income of patients was concentrated in 3 000-5 000 yuan (39.9%) and 5 001-10 000 yuan (34.9%), whereas 67.2% of the families had monthly medical expenses of <3 000 yuan. Of the patients, 75.5% received their first transfusion before 1 year of age. The proportions of children and adolescents with pretransfusion hemoglobin (HGB) of ≤70 g/L were 4.2% and 6.4%, respectively. Adolescents demonstrated significantly higher rates of transfusion frequency of <4 weeks/session, monthly red blood cell infusion of >2 U, serum ferritin (SF) of ≥5 000 μg/L, iron chelation therapy, and splenectomy compared with children (all P<0.05). Of the 338 patients, 26.0%, 22.8%, and 8.9% demonstrated stunted growth, underweight, and concurrent stunted growth with underweight, respectively. No significant difference was observed in the stunted growth rates between children (22.9%) and adolescents (27.7%) ( P=0.402). However, the underweight rate in adolescents (26.8%) was significantly higher than that in children (15.3%) ( P=0.023). The multivariate analysis determined the following risk factors for stunted growth: monthly household income of <10 000 yuan (5 001-10 000 yuan: OR=5.49, 95% CI: 1.48-35.76; 3 000-5 000 yuan: OR=6.87, 95% CI: 1.88-44.60; <3 000 yuan: OR=9.29, 95% CI: 2.20-64.77), pretransfusion HGB of ≤70 g/L ( OR=3.25, 95% CI: 1.07-10.18), and SF of ≥5 000 μg/L ( OR = 3.04, 95% CI: 1.20-7.70). Longer diagnostic duration was associated with underweight ( OR=1.10, 95% CI: 1.01-1.20) . Conclusions:Children and adolescents with TDT with pretransfusion SF of ≥5 000 μg/L, HGB of ≤70 g/L, low monthly household income, or longer diagnosis duration were significantly more likely to experience delayed physical growth.

Objective:To screen the optimal machine learning model for predicting the recurrence condition of hepatocellular carcinoma (HCC) at different time points post-surgery, based on the cutoff value of the Ki67 positive proliferation index condition calculated from recurrence-free survival and combined with various clinical features.Methods:retrospective study included initially treated patients with solitary HCC who underwent radical surgery at the Fifth Medical Center of the PLA General Hospital from January 2013 to March 2023. Data included general clinical data, preoperative laboratory parameters, and surgical pathology information about the subjects. The postoperative recurrence status was assessed by querying the medical record system or by telephone follow-up. The Ki67 positive index cutoff value was determined by the X-tile software based on the patient's recurrence-free survival status and time analysis. Survival rates were calculated using the Kaplan-Meier method, and survival curves were plotted. The study population was randomly divided into training and testing groups in a 7:3 ratio using a computer-generated random number method. The minimum redundancy maximum relevance (mRMR) method was used for feature variable selection. Predictive models for postoperative HCC recurrence conditions in patients with HCC were constructed using random forest, support vector machine, logistic regression, and gradient boosting decision tree machine learning algorithms. Inter-group comparisons for continuous data were performed using the t-test or Mann-Whitney U test. Inter-group comparisons of enumeration data were performed using the Pearson χ2 test, continuity-corrected χ2 test, or Fisher's exact test. Results:The cutoff values for the Ki67 positivity index were 0.3 and 0.5 in 510 cases, with a follow-up time ranging from 1.2 to 11.4 years (median: 6.2 years). The recurrence-free survival time was between 1 and 135 months (median: 32 months), with recurrence-free survival rates post-surgery at 1, 2, 3, and 5 years were 87.5%, 77.1%, 61.2%, and 54.5%, respectively. The top five variables predicted HCC recurrence and non-recurrence conditions following surgical follow-up at 6 months, 1 year, 2 years, and beyond 2 years, in accordance with information obtained by the mRMR screen out. The Ki67 positivity index screened a successfully constructed machine learning model to predict HCC recurrence and non-recurrence conditions following surgical follow-up at 6 months, 1 year, 2 years, and beyond 2 years. The machine learning model based on the gradient boosting decision tree algorithm had the best prediction performance among them (areas under the receiver operating characteristic curves for predicting HCC recurrence within six months in the training and validation sets were 0.996 and 0.946, and accuracies were 0.972 and 0.935, respectively).Conclusion:A machine learning model was successfully constructed using the Ki67 positivity index combined with four readily available clinical features to predict HCC recurrence. The machine learning model based on the gradient boosting decision tree algorithm demonstrated the best performance in terms of predicting HCC recurrence within six months after surgery.

Objective To investigate the association between rest-activity rhythm(RAR)and the risks of rheumatoid arthritis(RA),and to evaluate whether genetic susceptibility modifies this relationship.Methods This prospective cohort study utilized data from the UK Biobank,including 88 060 participants who did not have RA at baseline.RAR parameters(e.g.,relative amplitude)were calculated using data obtained through wrist-worn accelerometers.The participants'genetic susceptibility to RA was assessed using a polygenic risk score.Cox proportional hazards models were employed to analyze the association between RAR and RA risk,with interaction terms incorporated to evaluate the effect modification by genetic susceptibility.Results Over a median follow-up period of 7.97 years,660 incident RA cases were identified.After adjusting for age,sex,ethnicity,educational attainment,Townsend deprivation index,drinking status,smoking status,dietary score,body mass index,and polygenic risk score for incident RA,the dose-response analysis revealed a linear relationship between the RAR-related parameters,including the average amplitude during the most active 10 h(M10),interdaily stability(IS),intradaily variability(IV),and the risk of developing RA(P>0.05).In contrast,relative amplitude and the average amplitude during the least active 5 h(L5)showed a nonlinear relationship with the risk of developing RA(P<0.05).Compared to those in the the highest quartile of relative amplitude,participants in the lowest quartile had a 49%increase in the risk of developing RA(hazard ratio[HR]=1.49;95%CI,1.17-1.90).Compared to those in the lowest quartile,participants in the highest quartile of L5 had a 40%increased risk of developing RA(HR=1.40;95%CI,1.12-1.75).Every time M10 increased by one standard deviation,the risk of developing RA decreased by 12%(HR=0.88;95%CI,0.80-0.96).No evidence of effect modification by genetic susceptibility was observed in the RAR-RA association(P>0.05).Conclusion Disrupted rest-activity rhythm is associated with an increased risk of RA,which is independent of genetic susceptibility to RA.Our findings suggest that improving rest-activity rhythm may help reduce RA risks.

ObjectiveTo analyze the diversity and structural characteristics of microbial communities in the rhizosphere soil of healthy and diseased Andrographis paniculata and to explore the interactions of soil， plants， and microorganisms during the occurrence of diseases. MethodsThe physicochemical properties of the rhizosphere soil of healthy and diseased A.paniculata were determined， and the composition and diversity of bacterial and fungal communities in the rhizosphere soil were analyzed by Illumina high-throughput sequencing. Furthermore， the correlations between physicochemical properties and microorganisms of the rhizosphere soil were explored. ResultsThe content of total nitrogen， total potassium， and available potassium in the rhizosphere soil of diseased A. paniculata was significantly higher than that of healthy A. paniculata. The alpha diversity and richness （operational taxonomic units） of bacterial and fungal communities in the rhizosphere soil of diseased plants decreased compared with those of healthy plants. The microbial communities in the rhizosphere soil of healthy and diseased A. paniculata showed similar composition but different relative abundance. At the phylum level， the relative abundance of Proteobacteria and Chytridiomycota significantly increased， while that of Bacteroidota significantly decreased in the rhizosphere soil of diseased plants. At the genus level， the relative abundance of Sphingomonas， Pseudomonas， and Bryobacter significantly increased， while that of RB41 showed a significant decrease in the rhizosphere soil of diseased plants. The correlation analysis showed different correlations of microbial phyla with physicochemical properties of the rhizosphere soil between healthy and diseased plants. Organic matter， alkaline nitrogen， available phosphorus， and total potassium were correlated with the relative abundance of some dominant bacterial and fungal phyla in the rhizosphere soil of healthy plants， while available nitrogen and total phosphorus were correlated with the relative abundance of some dominant bacterial and fungal phyla in the rhizosphere soil of diseased plants. ConclusionThere are differences in the diversity and richness of microbial communities in the rhizosphere soil of healthy and diseased A. paniculata. The physicochemical properties of soil may have an impact on the rhizosphere microorganisms of A. paniculata， leading to the development of diseases. The results provide a scientific basis for the prevention and ecological management of A. paniculata diseases.

ObjectiveTo analyze the diversity and structural characteristics of microbial communities in the rhizosphere soil of healthy and diseased Andrographis paniculata and to explore the interactions of soil， plants， and microorganisms during the occurrence of diseases. MethodsThe physicochemical properties of the rhizosphere soil of healthy and diseased A.paniculata were determined， and the composition and diversity of bacterial and fungal communities in the rhizosphere soil were analyzed by Illumina high-throughput sequencing. Furthermore， the correlations between physicochemical properties and microorganisms of the rhizosphere soil were explored. ResultsThe content of total nitrogen， total potassium， and available potassium in the rhizosphere soil of diseased A. paniculata was significantly higher than that of healthy A. paniculata. The alpha diversity and richness （operational taxonomic units） of bacterial and fungal communities in the rhizosphere soil of diseased plants decreased compared with those of healthy plants. The microbial communities in the rhizosphere soil of healthy and diseased A. paniculata showed similar composition but different relative abundance. At the phylum level， the relative abundance of Proteobacteria and Chytridiomycota significantly increased， while that of Bacteroidota significantly decreased in the rhizosphere soil of diseased plants. At the genus level， the relative abundance of Sphingomonas， Pseudomonas， and Bryobacter significantly increased， while that of RB41 showed a significant decrease in the rhizosphere soil of diseased plants. The correlation analysis showed different correlations of microbial phyla with physicochemical properties of the rhizosphere soil between healthy and diseased plants. Organic matter， alkaline nitrogen， available phosphorus， and total potassium were correlated with the relative abundance of some dominant bacterial and fungal phyla in the rhizosphere soil of healthy plants， while available nitrogen and total phosphorus were correlated with the relative abundance of some dominant bacterial and fungal phyla in the rhizosphere soil of diseased plants. ConclusionThere are differences in the diversity and richness of microbial communities in the rhizosphere soil of healthy and diseased A. paniculata. The physicochemical properties of soil may have an impact on the rhizosphere microorganisms of A. paniculata， leading to the development of diseases. The results provide a scientific basis for the prevention and ecological management of A. paniculata diseases.

Objectives:To investigate the physical growth status of pediatric patients with transfusion-dependent thalassemia (TDT) and analyze the effects of treatment-related and socioeconomic factors on physical growth.Methods:Based on the specialized thalassemia database from gene therapy clinical research at the Institute of Hematology & Hospital of Blood Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, we collected data on height and weight development, family economic status, and medical records of 338 pediatric patients with TDT from October 2023 to May 2024. The length/height-for-age and body mass index (BMI) -for-age were classified based on the Growth Standard for Children under 7 Years of Age, Standard for Height Level Classification among Children and Adolescents Aged 7-18 Years, and Dietary Guidelines for Chinese Residents. Logistic regression analysis was conducted to assess the effects of family economic status and disease-related treatment on length/height-for-age and BMI-for-age.Results:Among the 338 patients, 118 were children and 220 were adolescents (192 males and 146 females), with a median age of 12 years (range: 0.8-18) and a median diagnosis duration of 10.3 years (range: 0.5-17.9). Subtypes included α-thalassemia [21 cases (6.2%) ], β-thalassemia [288 cases (85.2%) ], and combined αβ-thalassemia[29 cases (8.6%) ]. The monthly household income of patients was concentrated in 3 000-5 000 yuan (39.9%) and 5 001-10 000 yuan (34.9%), whereas 67.2% of the families had monthly medical expenses of <3 000 yuan. Of the patients, 75.5% received their first transfusion before 1 year of age. The proportions of children and adolescents with pretransfusion hemoglobin (HGB) of ≤70 g/L were 4.2% and 6.4%, respectively. Adolescents demonstrated significantly higher rates of transfusion frequency of <4 weeks/session, monthly red blood cell infusion of >2 U, serum ferritin (SF) of ≥5 000 μg/L, iron chelation therapy, and splenectomy compared with children (all P<0.05). Of the 338 patients, 26.0%, 22.8%, and 8.9% demonstrated stunted growth, underweight, and concurrent stunted growth with underweight, respectively. No significant difference was observed in the stunted growth rates between children (22.9%) and adolescents (27.7%) ( P=0.402). However, the underweight rate in adolescents (26.8%) was significantly higher than that in children (15.3%) ( P=0.023). The multivariate analysis determined the following risk factors for stunted growth: monthly household income of <10 000 yuan (5 001-10 000 yuan: OR=5.49, 95% CI: 1.48-35.76; 3 000-5 000 yuan: OR=6.87, 95% CI: 1.88-44.60; <3 000 yuan: OR=9.29, 95% CI: 2.20-64.77), pretransfusion HGB of ≤70 g/L ( OR=3.25, 95% CI: 1.07-10.18), and SF of ≥5 000 μg/L ( OR = 3.04, 95% CI: 1.20-7.70). Longer diagnostic duration was associated with underweight ( OR=1.10, 95% CI: 1.01-1.20) . Conclusions:Children and adolescents with TDT with pretransfusion SF of ≥5 000 μg/L, HGB of ≤70 g/L, low monthly household income, or longer diagnosis duration were significantly more likely to experience delayed physical growth.

Objective:To screen the optimal machine learning model for predicting the recurrence condition of hepatocellular carcinoma (HCC) at different time points post-surgery, based on the cutoff value of the Ki67 positive proliferation index condition calculated from recurrence-free survival and combined with various clinical features.Methods:retrospective study included initially treated patients with solitary HCC who underwent radical surgery at the Fifth Medical Center of the PLA General Hospital from January 2013 to March 2023. Data included general clinical data, preoperative laboratory parameters, and surgical pathology information about the subjects. The postoperative recurrence status was assessed by querying the medical record system or by telephone follow-up. The Ki67 positive index cutoff value was determined by the X-tile software based on the patient's recurrence-free survival status and time analysis. Survival rates were calculated using the Kaplan-Meier method, and survival curves were plotted. The study population was randomly divided into training and testing groups in a 7:3 ratio using a computer-generated random number method. The minimum redundancy maximum relevance (mRMR) method was used for feature variable selection. Predictive models for postoperative HCC recurrence conditions in patients with HCC were constructed using random forest, support vector machine, logistic regression, and gradient boosting decision tree machine learning algorithms. Inter-group comparisons for continuous data were performed using the t-test or Mann-Whitney U test. Inter-group comparisons of enumeration data were performed using the Pearson χ2 test, continuity-corrected χ2 test, or Fisher's exact test. Results:The cutoff values for the Ki67 positivity index were 0.3 and 0.5 in 510 cases, with a follow-up time ranging from 1.2 to 11.4 years (median: 6.2 years). The recurrence-free survival time was between 1 and 135 months (median: 32 months), with recurrence-free survival rates post-surgery at 1, 2, 3, and 5 years were 87.5%, 77.1%, 61.2%, and 54.5%, respectively. The top five variables predicted HCC recurrence and non-recurrence conditions following surgical follow-up at 6 months, 1 year, 2 years, and beyond 2 years, in accordance with information obtained by the mRMR screen out. The Ki67 positivity index screened a successfully constructed machine learning model to predict HCC recurrence and non-recurrence conditions following surgical follow-up at 6 months, 1 year, 2 years, and beyond 2 years. The machine learning model based on the gradient boosting decision tree algorithm had the best prediction performance among them (areas under the receiver operating characteristic curves for predicting HCC recurrence within six months in the training and validation sets were 0.996 and 0.946, and accuracies were 0.972 and 0.935, respectively).Conclusion:A machine learning model was successfully constructed using the Ki67 positivity index combined with four readily available clinical features to predict HCC recurrence. The machine learning model based on the gradient boosting decision tree algorithm demonstrated the best performance in terms of predicting HCC recurrence within six months after surgery.