1.Explainable Machine Learning Model for Predicting Prognosis in Patients with Malignant Tumors Complicated by Acute Respiratory Failure: Based on the eICU Collaborative Research Database in the United States
Zihan NAN ; Linan HAN ; Suwei LI ; Ziyi ZHU ; Qinqin ZHU ; Yan DUAN ; Xiaoting WANG ; Lixia LIU
Medical Journal of Peking Union Medical College Hospital 2026;17(1):98-108
To develop and validate a model for predicting intensive care unit (ICU) mortality risk in patients with malignant tumors complicated by acute respiratory failure (ARF) based on an explainable machine learning framework. Clinical data of patients with malignant tumors and ARF were extracted from the eICU Collaborative Research Database in the United States, including demographic characteristics, comorbidities, vital signs, laboratory test indicators, and major interventions within the first 24 hours after ICU admission.The study outcome was ICU death.Enrolled patients were randomly divided into a training set and a validation set at a ratio of 7:3.Predictor variables were selected using least absolute shrinkage and selection operator (LASSO) regression.Five machine learning algorithms-extreme gradient boosting (XGBoost), support vector machine (SVM), Logistic regression, multilayer perceptron (MLP), and C5.0 Decision Tree-were employed to construct predictive models.Model performance was evaluated based on the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and other metrics.The optimal model was further interpreted using the Shapley additive explanations (SHAP) algorithm. A total of 3196 patients with malignant tumors complicated by ARF were included.The training set comprised 2, 261 patients and the validation set 935 patients; 683 patients died during ICU stay, while 2513 survived.LASSO regression ultimately selected 12 variables closely associated with patient ICU outcomes, including sepsis comorbidity, use of vasoactive drugs, and within the first 24 hours after ICU admission: minimum mean arterial pressure, maximum heart rate, maximum respiratory rate, minimum oxygen saturation, minimum serum bicarbonate, minimum blood urea nitrogen, maximum white blood cell count, maximum mean corpuscular volume, maximum serum potassium, and maximum blood glucose.After model evaluation, the XGBoost model demonstrated the best performance.The AUCs for predicting ICU mortality risk in the training and validation sets were 0.940 and 0.763, respectively; accuracy was 88.3% and 81.2%;sensitivity was 98.5% and 95.9%.Its predictive performance also remained optimal in sensitivity analyses.SHAP analysis indicated that the top five variables contributing to the model's predictions were minimum oxygen saturation, minimum serum bicarbonate, minimum mean arterial pressure, use of vasoactive drugs, and maximum white blood cell count. This study successfully developed a mortality risk prediction model for ICU patients with malignant tumors complicated by ARF based on a large-scale dataset and performed explainability analysis.The model aids clinicians in early identification of high-risk patients and implementing individualized interventions.
2.Analysis of the causes of medical disputes and research on countermeasures: based on the practice of handling medical disputes in a certain hospital from 2018 to 2023
Xiaoting YAN ; Yi FAN ; Miao SHE
Chinese Medical Ethics 2025;38(1):131-138
ObjectiveTo explore the current situation, causes, and countermeasures of medical disputes in a tertiary A comprehensive hospital in Xi’an from 2018 to 2023. MethodsA total of 804 medical dispute cases that had been processed in a tertiary A hospital in Xi’an from January 2018 to December 2023 were collected. The occurrence, current characteristics, and cause distribution of medical disputes were retrospectively analyzed. ResultsThe cumulative incidence rate of medical disputes was 4.408/100,000, and the compensation rate was 31.59% (254/804). The incidence and compensation rate of medical disputes showed an overall increasing trend year by year from 2018 to 2023 (P<0.05). The proportion of deceased patients in dispute cases in the past 6 years showed an increasing trend annually (χ2=30.396, P=0.011). Medical dispute cases mainly came from hospitalized patients (65.17%) and patients undergoing surgical treatment (55.35%), and the top five departments with medical dispute cases were the hepatobiliary surgery department, cardiovascular medicine department, obstetrics and gynecology department, emergency department, and gastroenterology department. The number of in-hospital mediations showed a decreasing trend annually, while the number of people’s mediations showed an increasing trend annually from 2018 to 2023 (χ2=34.523, P<0.001). The top five causes of medical disputes in this hospital were surgical/operational complications (29.10%), disease condition assessment/disease diagnosis (10.45%), diagnosis and treatment plan (9.45%), diagnosis and treatment effectiveness (8.83%), and medical insurance reimbursement (7.59%). ConclusionThe number of medical disputes and compensation cases in comprehensive hospitals is increasing annually, and high-risk departments such as the surgical department and emergency department need to be given high attention. The handling of medical disputes has shown diversification, with people’s mediation gradually taking the dominant position. The causes of medical disputes mainly concentrated on diagnosis and treatment technology, doctor-patient communication, medical insurance, and other aspects. It is necessary to continuously improve the level of clinical diagnosis and treatment technology, strengthen doctor-patient communication, improve the medical insurance system, strengthen hospital connotation management, alleviate doctor-patient conflicts, and builds a harmonious doctor-patient relationship.
3.Lentivirus-modified hematopoietic stem cell gene therapy for advanced symptomatic juvenile metachromatic leukodystrophy: a long-term follow-up pilot study.
Zhao ZHANG ; Hua JIANG ; Li HUANG ; Sixi LIU ; Xiaoya ZHOU ; Yun CAI ; Ming LI ; Fei GAO ; Xiaoting LIANG ; Kam-Sze TSANG ; Guangfu CHEN ; Chui-Yan MA ; Yuet-Hung CHAI ; Hongsheng LIU ; Chen YANG ; Mo YANG ; Xiaoling ZHANG ; Shuo HAN ; Xin DU ; Ling CHEN ; Wuh-Liang HWU ; Jiacai ZHUO ; Qizhou LIAN
Protein & Cell 2025;16(1):16-27
Metachromatic leukodystrophy (MLD) is an inherited disease caused by a deficiency of the enzyme arylsulfatase A (ARSA). Lentivirus-modified autologous hematopoietic stem cell gene therapy (HSCGT) has recently been approved for clinical use in pre and early symptomatic children with MLD to increase ARSA activity. Unfortunately, this advanced therapy is not available for most patients with MLD who have progressed to more advanced symptomatic stages at diagnosis. Patients with late-onset juvenile MLD typically present with a slower neurological progression of symptoms and represent a significant burden to the economy and healthcare system, whereas those with early onset infantile MLD die within a few years of symptom onset. We conducted a pilot study to determine the safety and benefit of HSCGT in patients with postsymptomatic juvenile MLD and report preliminary results. The safety profile of HSCGT was favorable in this long-term follow-up over 9 years. The most common adverse events (AEs) within 2 months of HSCGT were related to busulfan conditioning, and all AEs resolved. No HSCGT-related AEs and no evidence of distorted hematopoietic differentiation during long-term follow-up for up to 9.6 years. Importantly, to date, patients have maintained remarkably improved ARSA activity with a stable disease state, including increased Functional Independence Measure (FIM) score and decreased magnetic resonance imaging (MRI) lesion score. This long-term follow-up pilot study suggests that HSCGT is safe and provides clinical benefit to patients with postsymptomatic juvenile MLD.
Humans
;
Leukodystrophy, Metachromatic/genetics*
;
Pilot Projects
;
Genetic Therapy/methods*
;
Hematopoietic Stem Cell Transplantation
;
Male
;
Follow-Up Studies
;
Female
;
Lentivirus/genetics*
;
Child
;
Child, Preschool
;
Hematopoietic Stem Cells/metabolism*
;
Cerebroside-Sulfatase/metabolism*
;
Adolescent
4.Research progress on male fertility damaged by marine work environment
Xiaoting LIN ; Jiaming GUO ; Xiwen YANG ; Guangya LU ; Hongli YAN
Journal of Navy Medicine 2025;46(5):525-532
Marine work environment is associated with unique risk factors,such as high salinity,high humidity,noise and vibration,and chemical pollution.Males make up the majority of marine workers.Prolonged exposure to these environmental factors may have adverse effects on male fertility,resulting in sperm quality reduction,endocrine disorder,and reproductive organ damage.Therefore,the potential harm of marine work environment to male reproductive health deserves attention.This review focused on the key exposure factors in marine work environment,and systematically explored the factors affecting male fertility,mechanisms,and physiological pathways.The aim is to provide scientific evidences to improve marine work environment,develop protective measures,and safeguard the reproductive health of workers,while also offering guidance for future research in this field.
5.Study on the epidemiological characteristics and cost structure of AECOPD patients in a class a tertiary hospital in Xi'an n from 2014 to 2023
Xiaoting YAN ; Jie REN ; Wen ZHANG ; Yugang LIU
Modern Hospital 2024;24(6):904-909,914
Objective To explore the epidemiological characteristics and cost structure of hospitalized patients with acute exacerbation of chronic obstructive pulmonary disease(AECOPD)in a tertiary comprehensive hospital in Xi'an from 2014 to 2023.Methods The first page information of AECOPD inpatients in a hospital from January 2014 to December 2023 was col-lected(n=6 748),and the distribution of their demographic characteristics and clinical characteristics as well as the composition of hospitalization expenses were retrospectively analyzed,and the structural change degree was used to analyze the structural change of hospitalization expenses in different years.Results AECOPD patients diagnosed and hospitalized accounted for 4.81 ‰ of all hospitalized patients in this institution from 2014-2023,the proportion of AECOPD inpatients showed a continuous downward trend during 10 years(x2=1 083.351,P<0.001);The distribution of gender,occupation,medical payment meth-ods,and sources of AECOPD inpatients in different years showed statistical differences(P<0.05);There is a statistical differ-ence in the seasonal distribution of AECOPD inpatients in different years(x2=157.328,P<0.001);There is a statistical difference in the average length of hospital stay among AECOPD inpatients in different years(x2=83.463,P<0.001);There were statistically significant differences in the proportion of hospitalizations,admission conditions,comorbidities or complications,history of drug allergies,and readmission plans among AECOPD patients in different years(P<0.05);There is a statistical difference in the outcomes of hospitalized AECOPD patients in different years(x2=153.367,P<0.001);The average hospital-ization cost of AECOPD patients gradually increased from 2014-2016,and gradually decreased from 2016-2023(P<0.01);The hospitalization expenses of AECOPD patients in different years are mainly based on drug expenses and diagnostic expenses;The overall degree of structural change(DSV)of average hospitalization expenses for AECOPD patients from 2014-2022 was 41.62%,with the highest DSV of hospitalization expenses in 2012-2013(22.69%)and the lowest DSV in 2017-2018(4.23%);The proportion of drug expenses in the average hospitalization cost of AECOPD patients continued to decline,while diagnostic expenses,hygiene material expenses,medical service expenses,and nursing expenses showed an overall increasing trend from 2014-2023.Conclusion The proportion and mortality rate of hospitalized AECOPD patients have shown an overall downward trend in the past 10 years;The proportion of hospitalized AECOPD patients in the age group of 70-79 is the highest,mainly distributed in winter;The overall average hospitalization cost over the past 10 years has shown a downward trend,with drug and diagnostic expenses as the main components.Drug expenses have been decreasing year by year,while diagnostic expen-ses have been increasing year by year.This indicates that medical policy reform plays a significant role in optimizing the hospitali-zation cost structure and controlling unreasonable cost growth.
6.Evaluation of complications after liver transplantation by CT and MRI
Xiaohua LI ; Min LI ; Wupeng WEI ; Yan ZHAO ; Jiecai LÜ ; Xiaoting WEN ; Weixiong LI
Journal of Practical Radiology 2024;40(11):1814-1817
Objective To evaluate the types and imaging features of early and late complications after liver transplantation.Methods A retro-spective analysis was conducted on the complications imaging data from 124 patients after liver transplantation.The CT and MRI characteristics of these complications was analyzed,categorized based on their occurrence time and type.Results The postoperative complications in the total 124 patients included vascular complication in 85 patients,biliary complication in 78 patients,and parenchy-mal complication in 21 patients.Additionally,the simple and complex complications were exhibited in 43 and 81 patients.Early vas-cular complication accounted for 69.8%,especially 21 cases(91.3%)were with early hepatic artery thrombosis.Bile leakage in 9 cases occurred in the early postoperative period,whereas biliary stenosis in 43 cases and biliary stones in 23 cases were observed in the late postoperative period.All ischemic infarction in 13 cases occurred in the early postoperative period.Among the 10 cases of liver abscess,8 cases were secondary to ischemic necrosis of liver parenchyma,and other 2 cases were secondary to suppurative cholangitis.Addi-tionally,6 cases of liver tumor recurrence were all detected in the late postoperative period.Conclusion The types of complications occurring after liver transplantation vary over time,and CT and MRI can be used to detect and evaluate postoperative complications.
7.Research progress on anti-glioma mechanism of natural sesquiter-pene lactones
Xiaoting YAN ; Xinye WANG ; Ming BAI ; Guodong YAO
Chinese Journal of Clinical Pharmacology and Therapeutics 2024;29(10):1174-1184
Glioma is a common primary intracra-nial tumor.Malignant glioma has a high mortality rate and an inferior prognosis.Despite various ther-apeutic interventions,the overall survival rate is still meager.Sesquiterpene lactone is a kind of natu-ral product containing α-methylene-γ-lactone,which has strong anti-tumor activity.In recent years,there have been many reports on the anti-gli-oma effect of sesquiterpene lactone compounds,such as ACT001,which is a structural modification of sesquiterpene lactone(Parthenolide)and has en-tered the clinical trial stage as a potential antican-cer drug.This paper reviews the activity and mecha-nism of sesquiterpene lactones with anti-glioma ef-fects,which have been studied in recent.years.
8.Application of different ways of ICG injection in the identification of pelvic lymph nodes under fluorescence laparoscopy
Xiaoting YAN ; Xiaoming CAO ; Bo WU ; Chao LIU ; Xiaofeng YANG
Chinese Journal of Urology 2024;45(8):592-597
Objective:To explore the effect of identifying pelvic lymph nodes by different injection of indocyanine green (ICG) during fluorescent laparoscopic radical cystectomy.Methods:The data of 54 male bladder cancer patients admitted to the First Hospital of Shanxi Medical University from September 2021 to September 2022 were analyzed. Preoperative cystoscopic biopsy all revealed a pathological diagnosis of urothelial carcinoma. All patients underwent fluorescent laparoscopic radical cystectomy plus pelvic lymph node dissection.They were divided into 3 groups according to the annotated route: 6 in the medial malleolus injection group, 68(61, 73) years; 4 in the perineal injection group, 67(59, 74) years; 44 in the medial malleolus and perineum, 64(45, 78) in the combined injection group. Under general anesthesia, patients were placed in the supine position with 0.3 to 0.5 ml of subcutaneous ICG solution (2.5 mg/ml) injected from 1.0 to 1.5 h before surgery. In the medial malleolus injection group, 0.3 to 0.5 ml (0.75 mg/place) was injected at 2 cm above the medial malleolus and 0.3 to 0.5 ml (0.75 mg/place) at two symmetrical sites above the anus. Intraoperatively, 4K fluorescent laparoscopy was used to observe the color development of the injection site and lymphatic vessels in vitro, and then to observe the development of pelvic lymph nodes in vivo.Results:In 3 groups, inguinal lymph nodes were developed about 30 min after ICG injection, and 1 h later, the longest (4.5±0.3) h. The external iliac and common iliac lymph nodes, the obturatorius lymph nodes in the perineal injection group, and the inguinal, obturatorius, external iliac, internal, anterior sacral, and common iliac lymph nodes in the medial malleolar and perineal injection group. In this study, in 54 cases, the postoperative pathological examination confirmed that the removed fluorescently labeled tissue was lymph node tissue, and the lymph node detection rate was 100%. The postoperative stage was pT 2, 14, 32 pT 3 and 8 pT 4; 4 pNx, 44 pN 0, 5 pN 1and 1 pN 2. In the medial malleolar injection group, 2 patients staged pT 2, 3 stage pT 3, 1 stage pT 4; 1stage pNx stage, 4 stage pN 0 and 1 stage pN 1 stage. In the perineal injection group, 4 patients had postoperative pathological stage pT 3N 0; in the medial malleolar and perineal injection group, 12 pT 2, 25 and 7 pT 3; 3 pNx, 36 pN 0, 4 pN 1 and 1pN 2. Conclusions:ICG injection through the perineum and subcutaneous labeling of pelvic lymph nodes, and the simple malleolus or perineal injection can not completely develop the pelvic lymph nodes. The combined injection could fully develop the pelvic lymph nodes, which may accurately guide the operator to locate, identify and remove pelvic lymph nodes.
9.Relationship between abnormal imaging findings of adipose tissue around primary colon cancer and prognosis
Xuefeng SUN ; Zhen GUAN ; Shuo YAN ; Xiaoting LI ; Yingshi SUN
Chinese Journal of Radiology 2023;57(8):904-911
Objective:To explore the relationship between the abnormal adipose tissue around the primary tumor of colon cancer and the prognosis.Methods:From January 2015 to December 2017, 448 patients with colon cancer in Peking University Cancer Hospital were retrospectively and consecutively collected. The scores were assigned to the severity, horizontal and vertical ranges of peritumoral adipose tissue (PAT) abnormalities, and the cumulative scores were calculated to establish the PAT grades from 1 to 3. We defined a score of 0 or 1 as PAT grade 1, a score of 2 as PAT grade 2, a score of 3 as PAT grade 3. The patients were followed up, and the overall survival (OS) and metastasis-free survival (MFS) were recorded. The Kaplan-Meier curve, log-rank test and Cox regression analysis were used to evaluate its impact on prognosis, and the hazard ratio (HR) and 95% confidence interval (CI) were calculated.Results:Among the 448 patients, patients with PAT grade 1, 2, and 3 accounted for 70.1% (314/448), 18.1% (81/448), and 11.8% (53/448), respectively. The Kaplan-Meier survival curve showed that patients with PAT grade 1 had the best OS, patients with grade 3 had the worst OS, and patients with grade 2 were in between, and the difference was statistically significant (χ 2=27.38, P<0.001). There were statistically significant differences between the grades in pairs ( P<0.05). There was no significant difference in MFS between different PAT grades (χ 2=2.85, P=0.240). The results of Cox regression analysis showed that PAT grade was an independent factor affecting the OS. Compared with PAT grade 1, the risk of death in patients with PAT grade 2 and 3 was significantly increased (HR 2.563, 95%CI 1.181-5.561; HR 2.269, 95%CI 1.005-5.121; P=0.034). PAT grade was not an independent factor of MFS ( P=0.253). Conclusion:The PAT grade established based on the degree and scope of abnormal PAT in colon cancer is an independent factor for poor prognosis of colon cancer.
10.Diagnostic value of a combined serology-based model for minimal hepatic encephalopathy in patients with compensated cirrhosis
Shanghao LIU ; Hongmei ZU ; Yan HUANG ; Xiaoqing GUO ; Huiling XIANG ; Tong DANG ; Xiaoyan LI ; Zhaolan YAN ; Yajing LI ; Fei LIU ; Jia SUN ; Ruixin SONG ; Junqing YAN ; Qing YE ; Jing WANG ; Xianmei MENG ; Haiying WANG ; Zhenyu JIANG ; Lei HUANG ; Fanping MENG ; Guo ZHANG ; Wenjuan WANG ; Shaoqi YANG ; Shengjuan HU ; Jigang RUAN ; Chuang LEI ; Qinghai WANG ; Hongling TIAN ; Qi ZHENG ; Yiling LI ; Ningning WANG ; Huipeng CUI ; Yanmeng WANG ; Zhangshu QU ; Min YUAN ; Yijun LIU ; Ying CHEN ; Yuxiang XIA ; Yayuan LIU ; Ying LIU ; Suxuan QU ; Hong TAO ; Ruichun SHI ; Xiaoting YANG ; Dan JIN ; Dan SU ; Yongfeng YANG ; Wei YE ; Na LIU ; Rongyu TANG ; Quan ZHANG ; Qin LIU ; Gaoliang ZOU ; Ziyue LI ; Caiyan ZHAO ; Qian ZHAO ; Qingge ZHANG ; Huafang GAO ; Tao MENG ; Jie LI ; Weihua WU ; Jian WANG ; Chuanlong YANG ; Hui LYU ; Chuan LIU ; Fusheng WANG ; Junliang FU ; Xiaolong QI
Chinese Journal of Laboratory Medicine 2023;46(1):52-61
Objective:To investigate the diagnostic accuracy of serological indicators and evaluate the diagnostic value of a new established combined serological model on identifying the minimal hepatic encephalopathy (MHE) in patients with compensated cirrhosis.Methods:This prospective multicenter study enrolled 263 compensated cirrhotic patients from 23 hospitals in 15 provinces, autonomous regions and municipalities of China between October 2021 and August 2022. Clinical data and laboratory test results were collected, and the model for end-stage liver disease (MELD) score was calculated. Ammonia level was corrected to the upper limit of normal (AMM-ULN) by the baseline blood ammonia measurements/upper limit of the normal reference value. MHE was diagnosed by combined abnormal number connection test-A and abnormal digit symbol test as suggested by Guidelines on the management of hepatic encephalopathy in cirrhosis. The patients were randomly divided (7∶3) into training set ( n=185) and validation set ( n=78) based on caret package of R language. Logistic regression was used to establish a combined model of MHE diagnosis. The diagnostic performance was evaluated by the area under the curve (AUC) of receiver operating characteristic curve, Hosmer-Lemeshow test and calibration curve. The internal verification was carried out by the Bootstrap method ( n=200). AUC comparisons were achieved using the Delong test. Results:In the training set, prevalence of MHE was 37.8% (70/185). There were statistically significant differences in AMM-ULN, albumin, platelet, alkaline phosphatase, international normalized ratio, MELD score and education between non-MHE group and MHE group (all P<0.05). Multivariate Logistic regression analysis showed that AMM-ULN [odds ratio ( OR)=1.78, 95% confidence interval ( CI) 1.05-3.14, P=0.038] and MELD score ( OR=1.11, 95% CI 1.04-1.20, P=0.002) were independent risk factors for MHE, and the AUC for predicting MHE were 0.663, 0.625, respectively. Compared with the use of blood AMM-ULN and MELD score alone, the AUC of the combined model of AMM-ULN, MELD score and education exhibited better predictive performance in determining the presence of MHE was 0.755, the specificity and sensitivity was 85.2% and 55.7%, respectively. Hosmer-Lemeshow test and calibration curve showed that the model had good calibration ( P=0.733). The AUC for internal validation of the combined model for diagnosing MHE was 0.752. In the validation set, the AUC of the combined model for diagnosing MHE was 0.794, and Hosmer-Lemeshow test showed good calibration ( P=0.841). Conclusion:Use of the combined model including AMM-ULN, MELD score and education could improve the predictive efficiency of MHE among patients with compensated cirrhosis.

Result Analysis
Print
Save
E-mail