|
审核状态: Project audit state: |
通过审核 Successful |
|
注册号: Registration number: |
ChiCTR2600122238 |
|
最近更新日期: Date of Last Refreshed on: |
2026-04-10 11:51:00 |
|
注册时间: Date of Registration: |
2026-04-10 00:00:00 |
|
注册号状态: |
预注册 |
|
Registration Status: |
Prospective registration |
|
注册题目: |
基于可解释性机器学习与 Stacking 堆叠集成技术构建和验证肝硬化相关AKI 的预测模型 |
|
Public title: |
Development and validation of a predictive model for cirrhosis-related AKI based on interpretable machine learning and stacking ensemble technique |
|
注册题目简写: |
|
|
English Acronym: |
|
|
研究课题的正式科学名称: |
基于可解释性机器学习与 Stacking 堆叠集成技术构建和验证肝硬化相关AKI 的预测模型 |
|
Scientific title: |
Development and validation of a predictive model for cirrhosis-related AKI based on interpretable machine learning and stacking ensemble technique |
|
研究课题代号(代码): Study subject ID: |
|
|
在二级注册机构或其它机构的注册号: The registration number of the Partner Registry or other register: |
|
申请注册联系人: |
徐鹤翔 |
研究负责人: |
徐鹤翔 |
|
Applicant: |
Hexiang Xu |
Study leader: |
Xu Hexiang |
|
申请注册联系人电话: Applicant telephone: |
+86 15357912472 |
研究负责人电话: Study leader's telephone: |
+86 551 6218 3518 |
|
申请注册联系人传真 : Applicant Fax: |
研究负责人传真: Study leader's fax: |
||
|
申请注册联系人电子邮件: Applicant E-mail: |
xuhe860814@163.com |
研究负责人电子邮件: Study leader's E-mail: |
xuhe860814@163.com |
|
申请单位网址(自愿提供): Applicant website(voluntary supply): |
研究负责人网址(自愿提供): Study leader's website(voluntary supply): |
||
|
申请注册联系人通讯地址: |
安徽省合肥市淮河路390号感染性疾病科 |
研究负责人通讯地址: |
合肥市淮河路390号 |
|
Applicant address: |
Department of Infectious Diseases, No. 390 Huaihe Road, Hefei City, Anhui Province, China |
Study leader's address: |
390 Huaihe Road, Hefei |
|
申请注册联系人邮政编码: Applicant postcode: |
研究负责人邮政编码: Study leader's postcode: |
||
|
申请人所在单位: |
合肥市第一人民医院 |
||
|
Applicant's institution: |
Hefei First People's Hospital |
||
|
研究负责人所在单位: |
合肥市第一人民医院 |
||
|
Affiliation of the Leader: |
Hefei First People's Hospital |
||
|
是否获伦理委员会批准: |
是/Yes |
||
|
Approved by ethic committee: |
Yes |
||
|
伦理委员会批件文号: Approved No. of ethic committee: |
伦研批第2026-076-01号 |
伦理委员会批件附件: Approved file of Ethical Committee: |
查看附件View |
|
批准本研究的伦理委员会名称: |
合肥市第一人民医院伦理委员会 |
||
|
Name of the ethic committee: |
Ethics Committee of Hefei First People's hospital |
||
|
伦理委员会批准日期: Date of approved by ethic committee: |
2026-04-09 00:00:00 |
||
|
伦理委员会联系人: |
叶芝 |
||
|
Contact Name of the ethic committee: |
Ye Zhi |
||
|
伦理委员会联系地址: |
合肥市淮河路390号 |
||
|
Contact Address of the ethic committee: |
390 Huaihe Road, Hefei |
||
|
伦理委员会联系人电话: Contact phone of the ethic committee: |
+86 551 62183685 |
伦理委员会联系人邮箱: Contact email of the ethic committee: |
hfyykyc@163.com |
|
研究实施负责(组长)单位: |
合肥市第一人民医院 |
||||||||||||||||||||||
|
Primary sponsor: |
Hefei First People's Hospital |
||||||||||||||||||||||
|
研究实施负责(组长)单位地址: |
合肥市淮河路390号 |
||||||||||||||||||||||
|
Primary sponsor's address: |
390 Huaihe Road, Hefei |
||||||||||||||||||||||
|
试验主办单位(项目批准或申办者): Secondary sponsor: |
|
||||||||||||||||||||||
|
经费或物资来源: |
合肥市卫生健康科技项目 |
||||||||||||||||||||||
|
Source(s) of funding: |
Hefei Municipal Health Science and Technology Project |
||||||||||||||||||||||
|
Target disease: |
Cirrhosis-related AKI |
||||||||||||||||||||||
|
Target disease code: |
|
||||||||||||||||||||||
|
研究类型: |
观察性研究 |
||||||||||||||||||||||
|
Study type: |
Observational study |
||||||||||||||||||||||
|
研究所处阶段: |
其它 | ||||||||||||||||||||||
|
Study phase: |
N/A |
||||||||||||||||||||||
|
研究设计: |
队列研究 |
||||||||||||||||||||||
|
Study design: |
Cohort study |
||||||||||||||||||||||
|
研究目的: |
主要研究目的 1. 构建基于 Stacking 堆叠集成技术的肝硬化相关 AKI 高精度预测模型:融合Logistic 回归、随机森林、XGBoost 等多种机器学习算法,构建超越单一模型性能极限的集成预测模型,实现对肝硬化住院患者发生急性肾损伤(AKI)的早期、精准预警。 2. 系统验证模型的泛化能力与临床适用性:通过内部交叉验证、外部回顾性验证及外部前瞻性验证三层验证体系,全面评估模型的区分度(AUC-ROC、AUC-PR)、校准度(校准曲线、Brier 分数)及临床效用(决策曲线分析),确保模型在不同人群和真实临床环境中的稳健性与可靠性。 次要研究目的 1. 筛选并确定最优预测变量子集:通过 Spearman 相关性分析、LASSO 回归、Boruta 算法、递归特征消除(RFE)等多策略融合的特征选择方法,从多维度临床变量中筛选最具预测力的特征组合。 2. 实现模型的可解释性分析:运用 SHAP 框架对集成模型进行全局和局部解释,揭示关键预测因子及其非线性交互作用,打破“黑箱”困境,增强临床医生的信任度。 3. 开发在线风险预测工具并推动临床转化:基于 R Shiny 框架开发用户友好的网页版风险计算器,支持实时输入患者指标、输出个体化 AKI 风险概率及关键决策因子解读,并在本院及医联体单位试点应用。 |
||||||||||||||||||||||
|
Objectives of Study: |
Primary Research Objectives: 1. To construct a high-precision prediction model for cirrhosis-related AKI based on Stacking ensemble technology: By integrating multiple machine learning algorithms such as Logistic Regression, Random Forest, and XGBoost, develop an ensemble prediction model that surpasses the performance limits of individual models, achieving early and accurate warnings for acute kidney injury (AKI) in hospitalized cirrhosis patients. 2. To systematically validate the model's generalizability and clinical applicability: Through a three-tier validation system including internal cross-validation, external retrospective validation, and external prospective validation, comprehensively evaluate the model's discrimination (AUC-ROC, AUC-PR), calibration (calibration curves, Brier score), and clinical utility (decision curve analysis), ensuring the model's robustness and reliability across different populations and real clinical environments. Secondary Research Objectives: 1. To screen and determine the optimal subset of predictive variables: Using a multi-strategy feature selection approach that integrates Spearman correlation analysis, LASSO regression, Boruta algorithm, and recursive feature elimination (RFE), select the most predictive feature combinations from multidimensional clinical variables. 2. To achieve interpretability analysis of the model: Apply the SHAP framework to explain the ensemble model both globally and locally, revealing key predictive factors and their nonlinear interactions, overcoming the "black-box" dilemma, and enhancing clinician trust. 3. To develop an online risk prediction tool and promote clinical translation: Develop a user-friendly web-based risk calculator based on the R Shiny framework, supporting real-time input of patient indicators, output of individualized AKI risk probabilities, and interpretation of key decision factors, with pilot applications in the hospital and affiliated medical units. |
||||||||||||||||||||||
|
药物成份或治疗方案详述: |
|
||||||||||||||||||||||
|
Description for medicine or protocol of treatment in detail: |
|
||||||||||||||||||||||
|
纳入标准: |
1.年龄≥18周岁,性别不限; 2.临床诊断符合肝硬化标准,可基于组织学(肝活检证实假小叶形成)、影像学(超声、CT或MRI显示肝硬化征象); 3.住院时间≥48小时,具有完整的入院临床资料; 4.肝硬化病因不限(包括病毒性、酒精性、自身免疫性、隐源性等),且入选前无急性肾损伤诊断。 |
||||||||||||||||||||||
|
Inclusion criteria |
1.Age ≥ 18 years, no gender restriction; 2.The clinical diagnosis meets the criteria for liver cirrhosis, which can be based on histology (pseudolobule formation confirmed by liver biopsy), imaging (ultrasound, CT, or MRI showing signs of cirrhosis); 3.Hospital stay ≥ 48 hours, with complete clinical data at admission. 4. The cause of cirrhosis is not limited (including viral, alcoholic, autoimmune, cryptogenic, etc.), and there is no diagnosis of acute kidney injury before enrollment. |
||||||||||||||||||||||
|
排除标准: |
1.临床数据严重缺失(缺失率>20%); 2.合并严重心肺疾病或原发性血液系统疾病; 3.既往接受过肝移植或肾移植; 4.合并肝癌或其他部位的肿瘤; 5.既往已诊断为慢性肾脏病或已需要肾脏替代治疗; 6.近期明确使用过肾毒性药物; 7.入院前已被诊断为急性肾损伤或其他原因引起的急性肾损伤; 8.年龄<18周岁; 9.妊娠期或哺乳期妇女; 10.入院时间<48小时; 11.仅有 1 次血肌酐检测值。 |
||||||||||||||||||||||
|
Exclusion criteria: |
1.Severe missing clinical data (missing rate > 20%); 2.Concomitant severe cardiopulmonary disease or primary hematologic disease; 3.Previous liver or kidney transplantation; 4.Concomitant hepatocellular carcinoma or tumors in other sites; 5.Previously diagnosed with chronic kidney disease or already requiring renal replacement therapy; 6.Recently used nephrotoxic drugs definitively; 7.Diagnosed with acute kidney injury (AKI) prior to admission, or AKI caused by other etiologies; 8.Age < 18 years; 9.Pregnant or lactating women; 10. Admission time < 48 hours; 11. Only one blood creatinine test value. |
||||||||||||||||||||||
|
研究实施时间: Study execute time: |
从 From 2026-01-01 00:00:00至 To 2028-12-31 00:00:00 |
征募观察对象时间: Recruiting time: |
从From 2026-04-10 00:00:00 至 To 2028-12-31 00:00:00 |
|
干预措施: Interventions: |
|
|
研究实施地点: Countries of recruitment and research settings: |
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
测量指标: Outcomes: |
|
|
采集人体标本:
Collecting sample(s)
|
|
|
征募研究对象情况: Recruiting status: |
尚未开始 Not yet recruiting |
年龄范围: Participant age: |
|
||||||
|
性别: |
男女均可 |
Gender: |
Both |
||||||
|
随机方法(请说明由何人用什么方法产生随机序列): |
无 |
||||||||
|
Randomization Procedure (please state who generates the random number sequence and by what method): |
None |
||||||||
|
是否公开试验完成后的统计结果: Calculated Results after the Study Completed public access: |
不公开/Private |
|
盲法: |
无 |
|
Blinding: |
None |
|
是否共享原始数据: IPD sharing |
No |
|
共享原始数据的方式(说明:请填入公开原始数据日期和方式,如采用网络平台,需填该网络平台名称和网址): |
无 |
|
The way of sharing IPD”(include metadata and protocol, If use web-based public database, please provide the url): |
None |
|
数据采集和管理(说明:数据采集和管理由两部分组成,一为病例记录表(Case Record Form, CRF),二为电子采集和管理系统(Electronic Data Capture, EDC),如ResMan即为一种基于互联网的EDC: |
电子采集和管理 |
|
Data collection and Management (A standard data collection and management system include a CRF and an electronic data capture: |
Electronic Data Capture |
|
数据与安全监察委员会: Data and Safety Monitoring Committee: |
有/Yes |