OBJECTIVE To investigate the actual construction and operation status of established and under-construction centralized （cloud） prescription review centers （shortened for “prescription review center”） of close-knit county-level medical consortium in a certain city， so as to provide reference for improving the construction quality of the prescription review center. METHODS An online questionnaire survey was conducted to collect the data from 51 established and under-construction prescription review center in the city， covering basic information， funding sources， talent management， system construction， review rule maintenance， prescription review practices， prescription evaluation， data utilization， and current challenges. The collected data were summarized and analyzed. RESULTS A total of 51 valid questionnaires were retrieved， covering 32 established and 19 under-construction prescription review center. Among the 32 established prescription review centers， the main funding sources for their construction came from government financial allocations， accounting for 56.25%. Only 25.00% of prescription review center had review pharmacists who fully met national qualification requirements， and just 55.00% updated more than 10 review rule entries per month on average. Outpatient prescription verification realized full coverage， but 37.50% of prescription review centers only supported rationality verification of single prescriptions， and 50.00% could not retrieve laboratory and examination results to assist in prescription review. Additionally， 40.62% of prescription review center had not regularly conducted prescription evaluations for primary care institutions. The data from prescription review center was mainly used to support medication monitoring. Among the 19 prescription review centers currently in the planning stage， 63.16% had no identified funding sources. CONCLUSIONS The operation and construction of prescription review center in the city face challenges， such as funding shortages， absence of collaborative incentive mechanisms， and insufficient manpower.It is suggested that the state should issue a unified standard for the construction of the prescription review center as soon as possible， and local health administrative departments should formulate supporting policies and clarify assessment indicators in combination with the actual situation of the region.

OBJECTIVE This study aims to establish a prompt engineering system for large language models in pharmaceutical popularization， and provide references for pharmacists to carry out efficient and standardized science popularization work. METHODS This study systematically expounded the principles and classifications of prompt engineering， as well as its effect on alleviating problems including model output hallucinations and poor interpretability. The design and optimization strategies of prompt engineering were defined for two core scenarios， namely text-to-text and text-to-image. Typical examples were adopted to compare the output effects before and after application. In addition， the limitations of prompt engineering applied in pharmaceutical popularization at the current stage were summarized. RESULTS In the two major scenarios of pharmaceutical popularization， well-designed prompt engineering improved the accuracy， readability and efficiency of outputs generated by large language models， and produced personalized popularization content adapted to clinical practice. CONCLUSIONS Prompt engineering can effectively improve the output quality of pharmaceutical popularization. The formulated standardized prompt engineering templates tailored for pharmaceutical popularization， can help pharmacists improve the efficiency and quality of popularization content creation.

OBJECTIVE To investigate the roles and mechanisms of γ-bungarotoxin(γ-BGT)in inducing respiratory distress in mice.METHODS Six male Kunming mice were selected and anesthe-tized before tracheal intubation and respiratory recording.After stabilizing respiration,the mice were intraperitoneally injected with γ-BGT at a dose of 6 mg·kg-1.Once a decrease in respiratory frequency was observed,the mice were intravenously injected with nikethamide at a dose of 12.5 mg·kg-1.Respi-ratory frequency was monitored using the BL420 signal acquisition and processing system.Male Kunming mice were randomly divided into the normal control group(saline,ip),γ-BGT group(6 mg·kg-1,ip),and γ-BGT+nikethamide group(γ-BGT 6 mg·kg-1,ip,followed by nikethamide 12.5 mg·kg-1,ip,when shal-low breathing and enhanced abdominal respiration were observed).The levels of Glu and GABA in the medulla oblongata were measured using ELISA.The protein expression levels of GAD65 and GAD67 in the medulla oblongata were determined by Western blotting.Primary mouse medullary neurons were cultured in vitro and divided into the following groups:cell control group,γ-BGT group,carbachol group,gallamine group,γ-BGT+H-89 group,and γ-BGT+Y-27632 group.The γ-BGT group,carbachol group,and gallamine group were incubated with γ-BGT(40 mg·L-1),carbachol(100 mmol·L-1),and gallamine(100 mmol·L-1),respectively,for 4 h.The γ-BGT+H-89 and γ-BGT+Y-27632 groups were pretreated with γ-BGT(40 mg·L-1)for 4 h,followed by incubation with the protein kinase A(PKA)inhibitor H-89(50 mmol·L-1)and the Ca2+channel inhibitor Y-27632(50 mmol·L-1)for another 2 h,respectively.ELISA was used to measure the levels of Glu,GABA,cAMP,and calpain in the primary mouse medul-lary neurons.Western blotting was employed to assess the protein expression levels of GAD65 and GAD67,and PKA phosphorylation levels.Fluo-4 fluorescent probe was used to detect the intracellular Ca2+level.RESULTS The respiratory rate of mice significantly decreased after iv administration of γ-BGT(γ-BGT group)(P<0.05).After treatment with nikethamide(nikethamide group),the respiratory rate significantly recovered(P<0.05).Compared with the normal control group,the γ-BGT group exhib-ited a significant decrease in Glu content(P<0.05),a significant increase in GABA content(P<0.05),and a significant decrease in the Glu/GABA ratio.Additionally,the protein expression levels of GAD65 and GAD67 were significantly elevated(P<0.05).Compared with the γ-BGT group,the γ-BGT+niketh-amide group showed a significant increase in Glu content(P<0.05),a significant decrease in GABA content(P<0.05),a significant increase in the Glu/GABA ratio,and a significant reduction in GAD65 and GAD67 protein expression levels(P<0.05).Compared to the cell control group,the γ-BGT group demonstrated a significant decrease in Glu content(P<0.05),a significant increase in GABA content(P<0.05),and a significant reduction in the Glu/GABA ratio.Furthermore,the protein expression levels of GAD65 and GAD67 were significantly elevated(P<0.05).Additionally,cAMP content,PKA phosphor-ylation levels,Ca2+levels,and calpain activity were all significantly increased(all P<0.05).Glu,GABA,Glu/GABA ratio,and GAD expression levels in the γ-BGT group changed in the same way as in the gallamine group;In the γ-BGT+Y-27632 group,calpain activity and expression levels of GAD65 and GAD67 were all significantly decreased(all P<0.05).In the γ-BGT+H-89 group,Ca2+levels and calpain activity were significantly reduced(all P<0.05).CONCLUSION γ-BGT-induced poisoning can lead to respiratory distress in mice,possibly through the antagonism of M2 muscarinic acetylcholine receptors in medullary neurons,activation of the cAMP/PKA signaling pathway,elevation of intracellular Ca2+levels,and increased expression and activity of GAD,resulting in an imbalance of Glu and GABA in the medulla.

ObjectiveTo clarify the graded quantitative diagnostic characteristics of lung qi deficiency syndrome in chronic obstructive pulmonary disease （COPD） based on latent class analysis combined with a hidden structure model. MethodsClinical data， including the four diagnostic methods of traditional Chinese medicine （TCM）， were collected from 745 COPD patients with lung qi deficiency syndrome. Latent class modeling was performed using R 4.1.2 software， and each patient was classified into one of three severity categories （mild， moderate， or severe） based on probabilistic parameterization， parameter estimation， and model fitting. A database was established for different severity levels of lung qi deficiency syndrome. Based on this， Lantern 5.0 software was used to construct hidden structure models for mild， moderate， and severe lung qi deficiency syndrome， and syndrome differentiation rules were developed through comprehensive clustering. ResultsA latent class model was constructed using 28 symptoms and signs with a frequency greater than 10%. Considering TCM theory and model simplicity， the optimal model was determined when the number of latent classes was three， categorizing lung qi deficiency syndrome into mild （298 cases）， moderate （164 cases）， and severe （283 cases）. Hidden structure models were separately developed for each severity level， and syndrome differentiation rules were established. A comparison of common symptoms in the syndrome differentiation rules for mild and moderate lung qi deficiency syndrome showed no statistically significant differences in diagnostic values and weights （P＞0.05）， leading to their combined analysis and the development of a unified syndrome differentiation rule. Value and weight of quantitative diagnosis of mild-to-moderate lung qi deficiency syndrome were as followed： shortness of breath （diagnostic value 9.3， diagnostic weight 86.92%）， dyspnea on exertion （8.2， 76.64%）， low voice and reluctance to speak （6.7， 62.62%）， poor appetite （4.0， 37.38%）， loose stools （4.0， 37.38%）， weak cough sound （2.9， 27.10%）， wheezing （2.3， 21.50%）， fatigue （1.8， 16.82%）， spontaneous sweating （1.7， 15.89%）， susceptibility to colds （1.6， 14.95%）， swollen tongue （1.4， 13.08%）， teeth marks on the tongue edge （1.2， 11.21%）， deep pulse （1.6， 14.95%）， with a diagnostic threshold of 10.3. Value and weight of quantitative diagnosis of severe lung qi deficiency syndrome were as followed： weak cough sound （15.1， 61.13%）， soreness and weakness of the waist and knees （12.6， 51.01%）， shortness of breath （11.1， 44.94%）， low voice and reluctance to speak （8.3， 33.60%）， frequent nocturia （6.1， 24.70%）， spontaneous sweating （3.7， 14.98%）， susceptibility to colds （3.5， 14.17%）， teeth marks on the tongue edge （7.8， 31.58%）， pale tongue body （1.9， 7.69%）， white tongue coating （5.5， 22.27%）， thin pulse （1.5， 6.07%）， with a diagnostic threshold of 23.7. ConclusionThe combination of latent class analysis and a hideen structure model effectively clarified the graded quantitative diagnostic characteristics of lung qi deficiency syndrome， providing a reference for the quantitative diagnosis of other fundamental syndromes in TCM.

Objective:To explore the effect of psoas muscle index (PMI) and construct a machine learning model to validate the 180-day prognosis in patients with decompensated liver cirrhosis.Methods:Retrospective data were collected from patients with decompensated liver cirrhosis at Henan Provincial People's Hospital from January 2022 to November 2022. The area of the psoas muscle index (PMI) at the level of the third lumbar vertebra was measured and calculated based on the abdominal X-ray computed tomography images stored in the Eastern China Hospital Information System (HIS). Patients were divided into low PMI and normal PMI groups according to the receiver operating characteristic curve. Patients clinical data and complication status were collected.The general conditions of both groups were compared using a t-test, chi-square test, and Mann-Whitney U test. The Kaplan-Meier method was applied for survival analysis. The outcome variable was 180-day mortality, and variables were selected using Cox and LASSO regression. The dataset was divided into training and testing sets in a 7∶3 ratio. Machine learning algorithms were used to build models in the training set, and model performance was validated by the test set. The model for MELD-Na score was compared with the model for End-Stage Liver Disease score. Results:A total of 298 patients with decompensated liver cirrhosis were included.The MELD scores, Child-Pugh classification, and NRS2002 scores, along with the incidence rate of complications such as ascites, hepatic encephalopathy, infections, and gastrointestinal bleeding, were significantly higher in the low PMI than the normal PMI group, with statistically significant differences ( P<0.05). The area under a receiver operating characteristic curve for the extreme gradient boosting model was higher than traditional clinical scores (MELD score 0.658, MELD_Na score 0.719) in the machine learning model. Furthermore, the application of SHAP results model indicated that PMI, hemoglobin, NRS2002 score, direct bilirubin, and blood ammonia were important factors in predicting the prognosis of patients with decompensated liver cirrhosis. Conclusion:A low PMI is closely related to poorer survival rates and the development of complication rates in patients with decompensated liver cirrhosis. The machine learning prediction model based on this construction, especially extreme gradient boosting, has favorable predictive performance, which is superior to the traditional clinical scoring system and can provide patients with the most accurate risk assessment and individualized treatment plan.

Objective:To explore the features of levels of lung cancer biomarkers in peripheral blood of adults in Beijing and surrounding areas, and establish personalized reference intervals for these biomarkers.Methods:A cross sectional study was conducted. The lung cancer biomarker data, including carcinoembryonic antigen (CEA), cytokeratin 19 fragment antigen 21-1 (CYFRA21-1), neuron specific enolase (NSE), progastrin-releasing peptide (ProGRP), and squamous cell carcinoma antigen (SCC-Ag), collected from adults who underwent cancer prevention examinations at the Cancer Hospital of the Chinese Academy of Medical Sciences from July 2021 to July 2022 were retrospectively analyzed. The interquartile range method was used to eliminate outliers, and the P95 value was calculated. Upper limit of 5 lung cancer biomarkers in different gender and age groups were obtained by referring to the reference intervals of quantitative analytes in the clinical laboratory (WS/T 402-2024). By analyzing the data of 208 adults who underwent cancer prevention physical examinations at the same center in June 2021 and 140 patients with benign lung masses confirmed by surgical resection pathology from January 2016 to June 2022, the established reference intervals for biomarkers were validated. Results:Two thousand six hundred and twenty-six cases of apparently healthy physical examiners were included for constructing reference intervals, including 1 456 males (55.4%) and 1 170 females (44.6%); the age range was 20-88 years old. The serum levels [ M ( Q1, Q3)] of CEA, NSE, ProGRP, SCC-Ag and CYFRA21-1 in 2 626 cases were 1.63 (1.07, 2.43) ng/ml, 13.08 (11.44, 14.77) ng/ml, 34.93 (29.02, 42.19) pg/ml, 0.80 (0.60, 1.00) ng/ml and 1.96 (1.48, 2.63) ng/ml, respectively. The serum levels of CEA [1.88 (1.22, 2.76) ng/ml vs. 1.41 (0.93, 2.02) ng/ml], NSE [13.31 (11.87, 15.00) ng/ml vs. 12.69 (10.96, 14.53) ng/ml], SCC-Ag [0.9 (0.7, 1.1) ng/ml vs. 0.7 (0.6, 0.9) ng/ml], and CYFRA21-1 [2.02 (1.53, 2.71) ng/ml vs. 1.87 (1.40, 2.51) ng/ml] in males were higher than those in females, and ProGRP [34.00 (28.25, 41.55) pg/ml vs. 36.12 (29.97, 42.98) pg/ml] was lower than that in females, and the differences were statistically significant (all P < 0.001). There were statistically significant differences in serum CEA levels between the groups of ≤ 40 years old (458 cases), >40-50 years old (827 cases), >50-60 years old (783 cases), >60-70 years old (412 cases), and >70 years old (146 cases) in pairwise comparison (all P < 0.05). Except for the age groups of ≤ 40 years old and >40-50 years old and the age groups of >60-70 years old and >70 years old, there were statistically significant differences in serum NSE levels among the other age groups in pairwise comparison (all P < 0.05). There were statistically significant differences in serum ProGRP levels between the 5 age groups (all P < 0.05). There were statistically significant differences when comparing the serum SCC-Ag level in the >40-50 age group, >50-60 age group and >60-70 age group with that in the ≤40 age group and >70 age group, respectively (all P < 0.05). However, there was no statistically significant difference between the other age groups in pairwise comparison (all P > 0.05). There were statistically significant differences in serum CYFRA21-1 levels between the 5 age groups (all P < 0.05). When gender and age were not distinguished, the P95 values of serum CEA, NSE, ProGRP, SCC-Ag and CYFRA21-1 levels were 4.44 ng/ml, 16.61 ng/ml, 57.65 pg/ml, 1.50 ng/ml, and 4.21 ng/ml, respectively. Considering gender and age, except for the >70 age group with no statistically significant difference in the P95 value of serum CEA level between males and females ( P > 0.05), the P95 value of serum CEA level in males was higher than that in females in all other age groups (all P < 0.001); the P95 values of serum CEA level in both males and females increased with age, but showed a decreasing trend in males over the age of 70. The P95 value of serum NSE level in males was higher than that in females in the age groups of ≤ 40 years and >40-50 years (both P < 0.05), while there was no statistically significant difference in the P95 value of serum NSE level between males and females in other age groups (all P > 0.05). The P95 values of serum NSE level in both males and females decreased firstly and increased later with age, reaching their highest levels at the age of >70. The P95 values of serum ProGRP level in females aged ≤ 40 and >50-60 were higher than those in males (both P < 0.05), while there was no statistically significant difference in the P95 value of serum ProGRP level between genders in other age groups (all P > 0.05); the P95 values of serum ProGRP level in both males and females increased with age. There was no statistically significant difference in the P95 value of serum SCC-Ag level between males and females in the ≤ 40 age group ( P > 0.05), while the P95 value of serum SCC-Ag level in males was higher than that in females in all other age groups (all P < 0.05). The P95 values of serum SCC-Ag level in males increased with age, while they were stable in females. There was no statistically significant difference in the P95 value of serum CYFRA21-1 level between males and females in the >60-70 age group ( P > 0.05), while the P95 value of serum CYFRA21-1 level in males was higher than those in females in all other age groups (all P < 0.05); the P95 values of serum CYFRA21-1 level in both males and females increased with age. Based on data from 2 626 apparently healthy physical examiners, reference intervals for the levels of 5 lung cancer biomarkers were constructed in different age groups of different genders. Validation was conducted on 208 physical examiners and 140 patients with benign lung lesions, and it was found that the compliance rate of using newly created reference intervals for different gender and age groups to interpret detection results was >90%, and the validation was passed. Conclusions:There are gender and age differences in the reference intervals of CEA, CYFRA21-1, NSE, ProGRP, and SCC-Ag in peripheral blood of adults in Beijing and surrounding areas. The constructed reference intervals of gender and age for biomarkers have been validated and shown good results, providing reference for optimizing the clinical application of lung cancer-related biomarkers.

Objective To provide methodological examples for related research,the influencing factors of the evolution of Qi deficiency syndrome in chronic obstructive pulmonary disease(COPD)based on a nonlinear mixed effects model was explored.Methods A research questionnaire on the influencing factors of the evolution of Qi deficiency syndrome in chronic obstructive pulmonary disease was developed,and clinical data of 650 COPD patients on the 1st and 14th day of acute exacerbation,the 1st and 28th day of risk window,the first day of stable period,and the 90th day were dynamically collected from 10 tertiary hospitals across the country.8 baseline data including gender and age were collected through the PROC NLMIXED process by SAS 9.4 software.Coronary heart disease,diabetes and hypertension accounted for the highest proportion.Nine concurrent syndromes including wind cold syndrome and phlegm heat syndrome were used as fixed effects,and individual level was used as random effects to gradually fit the model and screen the influencing factors of Qi deficiency syndrome in the entire process of disease occurrence and development.Results A total of 637 eligible cases were included,and clinical datas were dynamically collected on the 1st and 14th day of acute exacerbation,the 1st and 28th day of the risk window,the 1st and 90th day of the stable period.It was found that the number of acute exacerbations,alcohol consumption,concomitant hypertension,coronary heart disease,blood stasis syndrome,yin deficiency syndrome,yang deficiency syndrome,6-minute walking distance,and the modified Medical Research Council Dyspnea Questionnaire(mMRC)had an impact on the evolution of Qi deficiency syndrome in the previous year(P<0.05).Conclusion The use of a nonlinear mixed effects model revealed the relevant factors affecting the evolution of Qi deficiency syndrome from complex multi temporal dynamic data,providing methodological references for other related studies.

Objective To provide methodological examples for related research,the influencing factors of the evolution of Qi deficiency syndrome in chronic obstructive pulmonary disease(COPD)based on a nonlinear mixed effects model was explored.Methods A research questionnaire on the influencing factors of the evolution of Qi deficiency syndrome in chronic obstructive pulmonary disease was developed,and clinical data of 650 COPD patients on the 1st and 14th day of acute exacerbation,the 1st and 28th day of risk window,the first day of stable period,and the 90th day were dynamically collected from 10 tertiary hospitals across the country.8 baseline data including gender and age were collected through the PROC NLMIXED process by SAS 9.4 software.Coronary heart disease,diabetes and hypertension accounted for the highest proportion.Nine concurrent syndromes including wind cold syndrome and phlegm heat syndrome were used as fixed effects,and individual level was used as random effects to gradually fit the model and screen the influencing factors of Qi deficiency syndrome in the entire process of disease occurrence and development.Results A total of 637 eligible cases were included,and clinical datas were dynamically collected on the 1st and 14th day of acute exacerbation,the 1st and 28th day of the risk window,the 1st and 90th day of the stable period.It was found that the number of acute exacerbations,alcohol consumption,concomitant hypertension,coronary heart disease,blood stasis syndrome,yin deficiency syndrome,yang deficiency syndrome,6-minute walking distance,and the modified Medical Research Council Dyspnea Questionnaire(mMRC)had an impact on the evolution of Qi deficiency syndrome in the previous year(P<0.05).Conclusion The use of a nonlinear mixed effects model revealed the relevant factors affecting the evolution of Qi deficiency syndrome from complex multi temporal dynamic data,providing methodological references for other related studies.

OBJECTIVE To investigate the roles and mechanisms of γ-bungarotoxin(γ-BGT)in inducing respiratory distress in mice.METHODS Six male Kunming mice were selected and anesthe-tized before tracheal intubation and respiratory recording.After stabilizing respiration,the mice were intraperitoneally injected with γ-BGT at a dose of 6 mg·kg-1.Once a decrease in respiratory frequency was observed,the mice were intravenously injected with nikethamide at a dose of 12.5 mg·kg-1.Respi-ratory frequency was monitored using the BL420 signal acquisition and processing system.Male Kunming mice were randomly divided into the normal control group(saline,ip),γ-BGT group(6 mg·kg-1,ip),and γ-BGT+nikethamide group(γ-BGT 6 mg·kg-1,ip,followed by nikethamide 12.5 mg·kg-1,ip,when shal-low breathing and enhanced abdominal respiration were observed).The levels of Glu and GABA in the medulla oblongata were measured using ELISA.The protein expression levels of GAD65 and GAD67 in the medulla oblongata were determined by Western blotting.Primary mouse medullary neurons were cultured in vitro and divided into the following groups:cell control group,γ-BGT group,carbachol group,gallamine group,γ-BGT+H-89 group,and γ-BGT+Y-27632 group.The γ-BGT group,carbachol group,and gallamine group were incubated with γ-BGT(40 mg·L-1),carbachol(100 mmol·L-1),and gallamine(100 mmol·L-1),respectively,for 4 h.The γ-BGT+H-89 and γ-BGT+Y-27632 groups were pretreated with γ-BGT(40 mg·L-1)for 4 h,followed by incubation with the protein kinase A(PKA)inhibitor H-89(50 mmol·L-1)and the Ca2+channel inhibitor Y-27632(50 mmol·L-1)for another 2 h,respectively.ELISA was used to measure the levels of Glu,GABA,cAMP,and calpain in the primary mouse medul-lary neurons.Western blotting was employed to assess the protein expression levels of GAD65 and GAD67,and PKA phosphorylation levels.Fluo-4 fluorescent probe was used to detect the intracellular Ca2+level.RESULTS The respiratory rate of mice significantly decreased after iv administration of γ-BGT(γ-BGT group)(P<0.05).After treatment with nikethamide(nikethamide group),the respiratory rate significantly recovered(P<0.05).Compared with the normal control group,the γ-BGT group exhib-ited a significant decrease in Glu content(P<0.05),a significant increase in GABA content(P<0.05),and a significant decrease in the Glu/GABA ratio.Additionally,the protein expression levels of GAD65 and GAD67 were significantly elevated(P<0.05).Compared with the γ-BGT group,the γ-BGT+niketh-amide group showed a significant increase in Glu content(P<0.05),a significant decrease in GABA content(P<0.05),a significant increase in the Glu/GABA ratio,and a significant reduction in GAD65 and GAD67 protein expression levels(P<0.05).Compared to the cell control group,the γ-BGT group demonstrated a significant decrease in Glu content(P<0.05),a significant increase in GABA content(P<0.05),and a significant reduction in the Glu/GABA ratio.Furthermore,the protein expression levels of GAD65 and GAD67 were significantly elevated(P<0.05).Additionally,cAMP content,PKA phosphor-ylation levels,Ca2+levels,and calpain activity were all significantly increased(all P<0.05).Glu,GABA,Glu/GABA ratio,and GAD expression levels in the γ-BGT group changed in the same way as in the gallamine group;In the γ-BGT+Y-27632 group,calpain activity and expression levels of GAD65 and GAD67 were all significantly decreased(all P<0.05).In the γ-BGT+H-89 group,Ca2+levels and calpain activity were significantly reduced(all P<0.05).CONCLUSION γ-BGT-induced poisoning can lead to respiratory distress in mice,possibly through the antagonism of M2 muscarinic acetylcholine receptors in medullary neurons,activation of the cAMP/PKA signaling pathway,elevation of intracellular Ca2+levels,and increased expression and activity of GAD,resulting in an imbalance of Glu and GABA in the medulla.

Objective:To explore the effect of psoas muscle index (PMI) and construct a machine learning model to validate the 180-day prognosis in patients with decompensated liver cirrhosis.Methods:Retrospective data were collected from patients with decompensated liver cirrhosis at Henan Provincial People's Hospital from January 2022 to November 2022. The area of the psoas muscle index (PMI) at the level of the third lumbar vertebra was measured and calculated based on the abdominal X-ray computed tomography images stored in the Eastern China Hospital Information System (HIS). Patients were divided into low PMI and normal PMI groups according to the receiver operating characteristic curve. Patients clinical data and complication status were collected.The general conditions of both groups were compared using a t-test, chi-square test, and Mann-Whitney U test. The Kaplan-Meier method was applied for survival analysis. The outcome variable was 180-day mortality, and variables were selected using Cox and LASSO regression. The dataset was divided into training and testing sets in a 7∶3 ratio. Machine learning algorithms were used to build models in the training set, and model performance was validated by the test set. The model for MELD-Na score was compared with the model for End-Stage Liver Disease score. Results:A total of 298 patients with decompensated liver cirrhosis were included.The MELD scores, Child-Pugh classification, and NRS2002 scores, along with the incidence rate of complications such as ascites, hepatic encephalopathy, infections, and gastrointestinal bleeding, were significantly higher in the low PMI than the normal PMI group, with statistically significant differences ( P<0.05). The area under a receiver operating characteristic curve for the extreme gradient boosting model was higher than traditional clinical scores (MELD score 0.658, MELD_Na score 0.719) in the machine learning model. Furthermore, the application of SHAP results model indicated that PMI, hemoglobin, NRS2002 score, direct bilirubin, and blood ammonia were important factors in predicting the prognosis of patients with decompensated liver cirrhosis. Conclusion:A low PMI is closely related to poorer survival rates and the development of complication rates in patients with decompensated liver cirrhosis. The machine learning prediction model based on this construction, especially extreme gradient boosting, has favorable predictive performance, which is superior to the traditional clinical scoring system and can provide patients with the most accurate risk assessment and individualized treatment plan.