Qi Rui, Wang He-Bin, Luo Ren-Ying, Li Jing, Su Li
Department of Hepatology, Panzhihua Hospital of Integrated Chinese and Western Medicine, Panzhihua, China.
Department of Intensive Care Medicine, Panzhihua Hospital of Integrated Chinese and Western Medicine, Panzhihua, China.
Front Med (Lausanne). 2025 Aug 7;12:1636733. doi: 10.3389/fmed.2025.1636733. eCollection 2025.
This study aimed to develop and validate a clinical prediction model for identifying intra-abdominal infection (IAI) in patients with severe acute pancreatitis (SAP).
We conducted a retrospective cohort study of patients diagnosed with SAP at our institution between January 2020 and December 2023. A total of 415 eligible patients were enrolled and randomly allocated into a training set ( = 291) and a validation set ( = 124) in a 7:3 ratio for model development and internal validation. In the training cohort, candidate predictors were selected using least absolute shrinkage and selection operator (LASSO) regression to mitigate overfitting and retain the most clinically relevant variables. A multivariable logistic regression model was subsequently constructed, and a nomogram was developed to facilitate individualized risk assessment. Model performance was evaluated based on discrimination, calibration, and clinical utility. Discrimination was assessed using the area under the receiver operating characteristic curve (AUC) in both cohorts. Calibration was examined via calibration plots with bootstrapping (1,000 resamples) to correct for optimism. Decision curve analysis (DCA) was performed to determine the net clinical benefit across different risk thresholds.
The final cohort comprised 415 patients, with 291 in the training set and 124 in the validation set. LASSO regression identified four independent predictors with non-zero coefficients: hematocrit (HCT), procalcitonin (PCT), Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and neutrophil-to-lymphocyte ratio (NLR). The prediction model demonstrated robust discrimination, with an AUC of 0.853 (95% CI: 0.804-0.901) in the training set and 0.858 (95% CI: 0.786-0.930) in the validation set. Calibration plots indicated excellent agreement between predicted and observed probabilities. DCA confirmed significant clinical utility across a wide range of risk thresholds.
The proposed prediction model, incorporating HCT, PCT, APACHE II, and NLR, accurately stratifies the risk of IAI in SAP patients. This tool may facilitate early risk identification, guide timely antibiotic therapy, and optimize clinical decision-making to improve patient outcomes.
本研究旨在开发并验证一种用于识别重症急性胰腺炎（SAP）患者腹腔内感染（IAI）的临床预测模型。
我们对2020年1月至2023年12月在我院诊断为SAP的患者进行了一项回顾性队列研究。总共纳入415例符合条件的患者，并按照7:3的比例随机分为训练集（n = 291）和验证集（n = 124），用于模型开发和内部验证。在训练队列中，使用最小绝对收缩和选择算子（LASSO）回归选择候选预测因子，以减轻过拟合并保留最具临床相关性的变量。随后构建多变量逻辑回归模型，并开发列线图以促进个体化风险评估。基于区分度、校准度和临床实用性对模型性能进行评估。在两个队列中均使用受试者操作特征曲线下面积（AUC）评估区分度。通过带有自抽样（1000次重复抽样）的校准图检查校准度，以校正乐观偏差。进行决策曲线分析（DCA）以确定不同风险阈值下的净临床获益。
最终队列包括415例患者，其中训练集291例，验证集124例。LASSO回归确定了四个非零系数的独立预测因子：血细胞比容（HCT）、降钙素原（PCT）、急性生理与慢性健康状况评分系统II（APACHE II）评分和中性粒细胞与淋巴细胞比值（NLR）。预测模型显示出强大的区分度，训练集中的AUC为0.853（95%CI：0.804 - 0.901），验证集中的AUC为0.858（95%CI：0.786 - 0.930）。校准图表明预测概率与观察概率之间具有良好的一致性。DCA证实了在广泛的风险阈值范围内具有显著的临床实用性。
所提出的预测模型纳入了HCT、PCT、APACHE II和NLR，能够准确地对SAP患者发生IAI的风险进行分层。该工具可能有助于早期风险识别，指导及时的抗生素治疗，并优化临床决策以改善患者预后。
