目的探讨年龄校正的查尔森共病指数(age-adjusted Charlson comorbidity index,aCCI)用于评估气管插管的慢性阻塞性肺疾病急性加重期(acute exacerbation of chronic obstructive pulmonary disease,AECOPD)患者预后的价值,并分析共病...目的探讨年龄校正的查尔森共病指数(age-adjusted Charlson comorbidity index,aCCI)用于评估气管插管的慢性阻塞性肺疾病急性加重期(acute exacerbation of chronic obstructive pulmonary disease,AECOPD)患者预后的价值,并分析共病负担对28天院内死亡率的影响。方法本研究为回顾性队列研究,主分析队列来源于重症监护医学信息数据库Ⅳ(Medical Information Mart for Intensive CareⅣ,MIMIC-Ⅳ),共纳入2270例气管插管的AECOPD患者。另回顾性收集2021年1月至2026年2月就诊于复旦大学附属中山医院(厦门)的72例气管插管的AECOPD患者作为外部验证队列。主要结局为28天院内死亡。采用Cox比例风险回归模型评估aCCI与死亡风险的关系,并分析共病项的独立效应;通过X-tile确定aCCI最佳截断值并进行分层生存比较。进一步构建aCCI模型与联合模型,采用ROC曲线和AUC评价判别能力,结合校准曲线与决策曲线分析评估模型一致性和临床净获益,并在外部验证队列中进行验证。结果患者28天院内死亡率为27.14%(616/2270)。多因素分析显示,aCCI是28天院内死亡率的独立危险因素,调整后风险比(adjusted hazard ratio,aHR)为1.08。在共病项分析中,严重肝病(aHR=1.56)、肾病(aHR=1.24)和转移性实体瘤(aHR=1.70)与更高的死亡风险相关(P均10)较低共病负担组(aCCI≤10)的28天院内死亡率更高,风险比(hazard ratio,HR)为1.91(P<0.001)。预测效能方面,aCCI模型AUC为0.65,由aCCI、血氧饱和度、乳酸、碱剩余及序贯器官衰竭评估评分(Sequential Organ Failure Assessment,SOFA)构成的联合模型的AUC提升至0.73(P<0.001);该联合模型在外部验证队列中的AUC为0.75,且在建模队列和验证队列中均显示出较好的校准度,并在较宽阈值概率范围内具有较高的净获益。结论共病负担与气管插管的AECOPD患者28天院内死亡率独立相关,aCCI可用于快速风险分层。联合乳酸、SOFA及血氧饱和度等指标可进一步提高识别高危患者的能力,为早期干预和个体化管理提供依据。展开更多
Among plant hormones,cytokinins have been extensively studied for over a century due to their central roles in regulating plant development,physiology,and environmental responses.In this study,we chemically synthesize...Among plant hormones,cytokinins have been extensively studied for over a century due to their central roles in regulating plant development,physiology,and environmental responses.In this study,we chemically synthesized several Cytokinin-Active ingreDients(CADs)and evaluated their effects on the in vitro regeneration of Sansevieria trifasciata and seedling growth of Nicotiana tabacum,comparing their activities with established cytokinins,kinetin and zeatin.Our results show that CADs promote cell division and shoot proliferation in S.trifasciata,with the highest shoot numbers per explant observed at CAD-A(0.9 mg L-1),cycle-iP(0.1 mg L-1),and zeatin(0.3 mg L-1),all significantly different from kinetin(0.3 mg L-1)—the previously suggested cytokinin for in vitro shoot proliferation in this plant.CAD-A and CAD-B at 0.9 mg L-1increased leaf number by approximately 4.5,significantly outperforming kinetin at 0.3 mg L-1.In tobacco seedlings,CAD-B at 0.3 mg L-1significantly enhanced root number(7.3)comparable to kinetin(6.7)at the same concentration.Overall seedling height increased more than two-fold under both CAD-B and kinetin treatments compared to control plants.The bioactivity of CADs under in vitro conditions closely parallels that of kinetin,supporting their potential as effective cytokinin mimetics.Collectively,these results indicate that synthesized CADs can serve as potent alternatives to natural cytokinins for plant tissue culture and developmental studies,with promising applications in propagation and regeneration protocols.展开更多
目的基于机器学习(machine learning,ML)算法,建立并验证缺血性脑卒中(ischemic stroke,IS)危重患者发生院内再次转入重症监护病房(intensive care unit,ICU)的预测模型。方法从美国重症监护医学信息数据库Ⅳ(Medical Information Mart ...目的基于机器学习(machine learning,ML)算法,建立并验证缺血性脑卒中(ischemic stroke,IS)危重患者发生院内再次转入重症监护病房(intensive care unit,ICU)的预测模型。方法从美国重症监护医学信息数据库Ⅳ(Medical Information Mart for Intensive CareⅣ,MIMIC-Ⅳ)中纳入2929例IS患者的临床资料,采用最小绝对收缩与选择算子(least absolute shrinkageand selection operator,LASSO)回归确定预测因子,并使用合成少数类过采样技术(synthetic minority oversampling technique,SMOTE)形成包含2583例患者的推导队列,按照8∶2的比例随机分为训练集(n=2066)与测试集(n=517)。应用决策树、随机森林、自适应提升(adaptive boosting,AdaBoost)、梯度提升决策树(gradient boosting decision tree,GBDT)及支持向量机(support vector machine,SVM)5种ML算法建立预测模型。在训练集中,使用五折交叉验证评估模型性能。在测试集中,采用ROC曲线下面积(area under the curve,AUC)和决策曲线分析(decision curve analysis,DCA)对各模型进行评估与比较。采用Shapley加法解释方法对性能最优的模型进行解释。结果共纳入2929例患者,其中704例(24.0%)发生院内再次转入ICU。在5种ML模型中,随机森林模型的预测性能较好,其AUC值为0.839(95%CI 0.801~0.877)。特征重要性分析确定了影响模型预测最重要的5个特征,包括急性生理学评分Ⅲ(acute physiology scoreⅢ,APSⅢ)评分、白蛋白、年龄、心率及序贯器官衰竭评分。结论基于ML算法的模型能有效预测IS危重患者院内再次转入ICU的风险,且随机森林模型展现出更优的预测性能,在早期临床风险分层与干预中具有良好的潜在应用价值。展开更多
目的基于MIMIC-Ⅳ数据库的大型队列,系统探讨非ICU住院及入住重症监护病房(intensive care unit,ICU)的结节性多动脉炎(polyarteritis nodosa,PAN)患者1年死亡结局的相关风险因素。方法从MIMIC-Ⅳ数据库(版本3.0)中提取PAN患者的临床资...目的基于MIMIC-Ⅳ数据库的大型队列,系统探讨非ICU住院及入住重症监护病房(intensive care unit,ICU)的结节性多动脉炎(polyarteritis nodosa,PAN)患者1年死亡结局的相关风险因素。方法从MIMIC-Ⅳ数据库(版本3.0)中提取PAN患者的临床资料。采用Logistic回归分析、Cox比例风险回归模型及Kaplan-Meier(KM)生存曲线,对相关临床指标与非ICU住院及ICU收治PAN患者1年死亡率之间的关联性进行评估。结果本研究共纳入110例非ICU住院PAN患者及38例ICU收治PAN患者。多因素Logistic回归与Cox回归分析结果显示,血红蛋白水平降低是非ICU住院PAN患者1年死亡率的显著预测因素,且为独立危险因素。进一步分析表明,简化急性生理学评分Ⅱ(simplified acute physiology scoreⅡ,SAPSⅡ)是ICU收治PAN患者1年死亡率的显著预测指标及独立危险因素。KM生存曲线分析显示,高SAPSⅡ评分组ICU-PAN患者的生存率显著低于低评分组(P=0.012)。结论低血红蛋白水平是非ICU住院PAN患者1年死亡风险升高的独立危险因素。此外,本研究首次明确SAPSⅡ评分是ICU-PAN患者1年死亡的独立危险因素,且当SAPSⅡ评分超过45分时,患者死亡风险显著增加。展开更多
BACKGROUND Foreign body-associated reactive masses may appear months to decades after abdominal surgery and often present as well-circumscribed lesions on imaging studies.Their radiological features frequently mimic g...BACKGROUND Foreign body-associated reactive masses may appear months to decades after abdominal surgery and often present as well-circumscribed lesions on imaging studies.Their radiological features frequently mimic gastrointestinal stromal tumors(GISTs)or other malignant intra-abdominal neoplasms.While most reported postoperative foreign body reactions arise in the peritoneal cavity or mesentery,reactive masses originating from the colonic wall are exceptionally rare.Large colonic lesions presenting many years after surgery further complicate diagnosis.CASE SUMMARY A 38-year-old man with a history of open appendectomy 20 years prior was incidentally found to have a 9-10 cm intra-abdominal mass on a subsequent computed tomography(CT)screening.Contrast-enhanced CT presented a well-circumscribed lesion with marked central low attenuation and peripheral enhancement,strongly suggestive of a GIST in the small bowel.Exploratory laparotomy revealed a firm,heterogeneous mass densely adherent to adjacent structures and arising from the transverse colon.Thus,segmental transverse colectomy with stapled anastomosis was performed.Histopathological examination showed extensive fibrosis,necrosis,and old hemorrhage with a chronic foreign body-type reactive process and no neoplastic features.No grossly identifiable foreign material was detected,although microscopic findings supported a chronic foreign body reaction.Postoperative recovery was uneventful aside from a brief readmission for transient ileus.CONCLUSION Colonic foreign body-associated reactive masses can mimic GISTs decades after surgery and should be considered in patients with prior abdominal operations.展开更多
目的构建并验证急性缺血性脑卒中(acute ischemic stroke,AIS)静脉溶栓后出血转化(hemorrhagic transformation,HT)风险预测模型。方法回顾性分析2008年1月—2021年11月美国重症监护医学信息(medical information mart for intensive c...目的构建并验证急性缺血性脑卒中(acute ischemic stroke,AIS)静脉溶栓后出血转化(hemorrhagic transformation,HT)风险预测模型。方法回顾性分析2008年1月—2021年11月美国重症监护医学信息(medical information mart for intensive careⅣ,MIMIC-Ⅳ)数据库中收录的AIS静脉溶栓患者全病程中的生理和临床信息,挖掘样本的向量化特征,单因素分析找出溶栓后HT的高危因素,采用多种机器学习算法构建溶栓后HT风险预测模型,并采用受试者工作特征ROC曲线对预测模型进行检验。结果本研究共纳入910例AIS静脉溶栓患者。单因素分析结果显示:既往病史(房颤、糖尿病、脑梗、无氯吡格雷服药史)、年龄、卒中溶栓时间(door-to-needle,DNT)、无创收缩压、无创舒张压、平均血压、氧流量、肌酸酐、尿素氮、尿比重、尿液pH、天冬氨酸氨基转移酶、乳酸脱氢酶、碱性磷酸盐、白蛋白含量、红细胞、红细胞压积、红细胞平均分布宽度、红细胞分布宽度标准差、血钠浓度、钙离子、血液氯化物、凝血酶原时间(国际标准比值)、凝血酶原时间、部分凝血活酶时间差异有统计学意义(P<0.05)。通过多种机器学习算法,构建并验证预测模型的性能,得出年龄、DNT、糖尿病史、尿液pH、肌酐、白蛋白、血红蛋白、红细胞和无创收缩压10个最优因素组合下最近邻算法对溶栓后HT预测的灵敏度、特异度和准确度分别达到了0.789、0.724和0.731,预测效能较好;且该模型的ROC曲线及曲线下面积AUC值(0.74),均优于其他机器学习算法构建的预测模型。结论溶栓后HT预测模型能够优化临床资源配置,为临床医生提供较为准确的诊断和治疗决策支持,为临床护士提供针对性的护理干预指导。展开更多
摘要目的探讨年龄校正的查尔森共病指数(age-adjusted Charlson comorbidity index,aCCI)用于评估气管插管的慢性阻塞性肺疾病急性加重期(acute exacerbation of chronic obstructive pulmonary disease,AECOPD)患者预后的价值,并分析共病负担对28天院内死亡率的影响。方法本研究为回顾性队列研究,主分析队列来源于重症监护医学信息数据库Ⅳ(Medical Information Mart for Intensive CareⅣ,MIMIC-Ⅳ),共纳入2270例气管插管的AECOPD患者。另回顾性收集2021年1月至2026年2月就诊于复旦大学附属中山医院(厦门)的72例气管插管的AECOPD患者作为外部验证队列。主要结局为28天院内死亡。采用Cox比例风险回归模型评估aCCI与死亡风险的关系,并分析共病项的独立效应;通过X-tile确定aCCI最佳截断值并进行分层生存比较。进一步构建aCCI模型与联合模型,采用ROC曲线和AUC评价判别能力,结合校准曲线与决策曲线分析评估模型一致性和临床净获益,并在外部验证队列中进行验证。结果患者28天院内死亡率为27.14%(616/2270)。多因素分析显示,aCCI是28天院内死亡率的独立危险因素,调整后风险比(adjusted hazard ratio,aHR)为1.08。在共病项分析中,严重肝病(aHR=1.56)、肾病(aHR=1.24)和转移性实体瘤(aHR=1.70)与更高的死亡风险相关(P均10)较低共病负担组(aCCI≤10)的28天院内死亡率更高,风险比(hazard ratio,HR)为1.91(P<0.001)。预测效能方面,aCCI模型AUC为0.65,由aCCI、血氧饱和度、乳酸、碱剩余及序贯器官衰竭评估评分(Sequential Organ Failure Assessment,SOFA)构成的联合模型的AUC提升至0.73(P<0.001);该联合模型在外部验证队列中的AUC为0.75,且在建模队列和验证队列中均显示出较好的校准度,并在较宽阈值概率范围内具有较高的净获益。结论共病负担与气管插管的AECOPD患者28天院内死亡率独立相关,aCCI可用于快速风险分层。联合乳酸、SOFA及血氧饱和度等指标可进一步提高识别高危患者的能力,为早期干预和个体化管理提供依据。
摘要Among plant hormones,cytokinins have been extensively studied for over a century due to their central roles in regulating plant development,physiology,and environmental responses.In this study,we chemically synthesized several Cytokinin-Active ingreDients(CADs)and evaluated their effects on the in vitro regeneration of Sansevieria trifasciata and seedling growth of Nicotiana tabacum,comparing their activities with established cytokinins,kinetin and zeatin.Our results show that CADs promote cell division and shoot proliferation in S.trifasciata,with the highest shoot numbers per explant observed at CAD-A(0.9 mg L-1),cycle-iP(0.1 mg L-1),and zeatin(0.3 mg L-1),all significantly different from kinetin(0.3 mg L-1)—the previously suggested cytokinin for in vitro shoot proliferation in this plant.CAD-A and CAD-B at 0.9 mg L-1increased leaf number by approximately 4.5,significantly outperforming kinetin at 0.3 mg L-1.In tobacco seedlings,CAD-B at 0.3 mg L-1significantly enhanced root number(7.3)comparable to kinetin(6.7)at the same concentration.Overall seedling height increased more than two-fold under both CAD-B and kinetin treatments compared to control plants.The bioactivity of CADs under in vitro conditions closely parallels that of kinetin,supporting their potential as effective cytokinin mimetics.Collectively,these results indicate that synthesized CADs can serve as potent alternatives to natural cytokinins for plant tissue culture and developmental studies,with promising applications in propagation and regeneration protocols.
基金Supported by Biomedical Research Institute Grant,Kyungpook National University HospitalNo.2025Korea Health Technology R&D Project through the Korea Health Industry Development Institute(KHIDI),Funded by the Ministry of Health&Welfare,Republic of Korea,No.RS-2022-KH130590.
摘要BACKGROUND Foreign body-associated reactive masses may appear months to decades after abdominal surgery and often present as well-circumscribed lesions on imaging studies.Their radiological features frequently mimic gastrointestinal stromal tumors(GISTs)or other malignant intra-abdominal neoplasms.While most reported postoperative foreign body reactions arise in the peritoneal cavity or mesentery,reactive masses originating from the colonic wall are exceptionally rare.Large colonic lesions presenting many years after surgery further complicate diagnosis.CASE SUMMARY A 38-year-old man with a history of open appendectomy 20 years prior was incidentally found to have a 9-10 cm intra-abdominal mass on a subsequent computed tomography(CT)screening.Contrast-enhanced CT presented a well-circumscribed lesion with marked central low attenuation and peripheral enhancement,strongly suggestive of a GIST in the small bowel.Exploratory laparotomy revealed a firm,heterogeneous mass densely adherent to adjacent structures and arising from the transverse colon.Thus,segmental transverse colectomy with stapled anastomosis was performed.Histopathological examination showed extensive fibrosis,necrosis,and old hemorrhage with a chronic foreign body-type reactive process and no neoplastic features.No grossly identifiable foreign material was detected,although microscopic findings supported a chronic foreign body reaction.Postoperative recovery was uneventful aside from a brief readmission for transient ileus.CONCLUSION Colonic foreign body-associated reactive masses can mimic GISTs decades after surgery and should be considered in patients with prior abdominal operations.
摘要目的构建并验证急性缺血性脑卒中(acute ischemic stroke,AIS)静脉溶栓后出血转化(hemorrhagic transformation,HT)风险预测模型。方法回顾性分析2008年1月—2021年11月美国重症监护医学信息(medical information mart for intensive careⅣ,MIMIC-Ⅳ)数据库中收录的AIS静脉溶栓患者全病程中的生理和临床信息,挖掘样本的向量化特征,单因素分析找出溶栓后HT的高危因素,采用多种机器学习算法构建溶栓后HT风险预测模型,并采用受试者工作特征ROC曲线对预测模型进行检验。结果本研究共纳入910例AIS静脉溶栓患者。单因素分析结果显示:既往病史(房颤、糖尿病、脑梗、无氯吡格雷服药史)、年龄、卒中溶栓时间(door-to-needle,DNT)、无创收缩压、无创舒张压、平均血压、氧流量、肌酸酐、尿素氮、尿比重、尿液pH、天冬氨酸氨基转移酶、乳酸脱氢酶、碱性磷酸盐、白蛋白含量、红细胞、红细胞压积、红细胞平均分布宽度、红细胞分布宽度标准差、血钠浓度、钙离子、血液氯化物、凝血酶原时间(国际标准比值)、凝血酶原时间、部分凝血活酶时间差异有统计学意义(P<0.05)。通过多种机器学习算法,构建并验证预测模型的性能,得出年龄、DNT、糖尿病史、尿液pH、肌酐、白蛋白、血红蛋白、红细胞和无创收缩压10个最优因素组合下最近邻算法对溶栓后HT预测的灵敏度、特异度和准确度分别达到了0.789、0.724和0.731,预测效能较好;且该模型的ROC曲线及曲线下面积AUC值(0.74),均优于其他机器学习算法构建的预测模型。结论溶栓后HT预测模型能够优化临床资源配置,为临床医生提供较为准确的诊断和治疗决策支持,为临床护士提供针对性的护理干预指导。