Scholarlink

<p>Despite advances in therapeutic strategies, including biologic and small-molecule agents, many Crohns disease (CD) patients in remission or with mild activity based on Crohns Disease Activity Index still require abdominal surgery, yet current predictive tools remain insufficient. This study developed and validated a practical prediction model using routine clinical data to assess long-term surgical risk in this population. This retrospective cohort study analyzed 615 incident CD patients diagnosed between 2016 and 2022. Using backward stepwise Cox regression, we developed a prediction model incorporating only readily available clinical parameters. Model performance was evaluated through Harrells C-index, time-dependent ROC analysis, and calibration curves, with bootstrap validation for internal verification. Five surgical predictors emerged: C-reactive protein&#xa0;(HR = 1.06, 95% CI: 1.03–1.10), fibrinogen (HR = 0.71, 95% CI: 0.59–0.85), albumin (HR: 0.67, 95% CI: 0.50–0.90), Montreal B classification (HR of B2 = 2.28, 95% CI: 1.30–4.00, HR of B3 = 4.13, 95% CI: 2.56–6.65), the use of advanced therapy (HR 0.63, 95% CI 0.42–0.92). The model demonstrated excellent predictive performance, with a Harrells C-index of 0.728 (95% CI: 0.677–0.779), and maintained time-dependent AUC values of 0.729–0.766 across 0.5–5 years. Bootstrap validation showed consistent performance (mean C-index: 0.733; time-dependent AUC: 0.731–0.773) with excellent calibration, confirming the models stability and clinical utility for long-term prediction. Risk stratification based on the cox model effectively categorized patients into low-, intermediate-, and high-risk groups (<i>P</i> &lt; 0.001). Additionally, an online platform to enable clinical guidance: <a href="https://crohndisease.shinyapps.io/CDAI0-1/.The">https://crohndisease.shinyapps.io/CDAI0-1/.The</a> prediction model incorporating routinely available clinical variables to accurately predict abdominal surgery risk in CD patients in remission or with mild activity, supporting clinical decision-making.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A web-based prediction model using routinely available clinical variables to estimate surgical risk in Crohn’s disease during remission or mild activity

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月13日

查看全文
  • Kailing Xie,
  • Hengchang Yao,
  • Zhixian Jiang,
  • Hao liu,
  • Qi Sun,
  • Lichao Yang,
  • Lianwen Yuan

摘要

Despite advances in therapeutic strategies, including biologic and small-molecule agents, many Crohns disease (CD) patients in remission or with mild activity based on Crohns Disease Activity Index still require abdominal surgery, yet current predictive tools remain insufficient. This study developed and validated a practical prediction model using routine clinical data to assess long-term surgical risk in this population. This retrospective cohort study analyzed 615 incident CD patients diagnosed between 2016 and 2022. Using backward stepwise Cox regression, we developed a prediction model incorporating only readily available clinical parameters. Model performance was evaluated through Harrells C-index, time-dependent ROC analysis, and calibration curves, with bootstrap validation for internal verification. Five surgical predictors emerged: C-reactive protein (HR = 1.06, 95% CI: 1.03–1.10), fibrinogen (HR = 0.71, 95% CI: 0.59–0.85), albumin (HR: 0.67, 95% CI: 0.50–0.90), Montreal B classification (HR of B2 = 2.28, 95% CI: 1.30–4.00, HR of B3 = 4.13, 95% CI: 2.56–6.65), the use of advanced therapy (HR 0.63, 95% CI 0.42–0.92). The model demonstrated excellent predictive performance, with a Harrells C-index of 0.728 (95% CI: 0.677–0.779), and maintained time-dependent AUC values of 0.729–0.766 across 0.5–5 years. Bootstrap validation showed consistent performance (mean C-index: 0.733; time-dependent AUC: 0.731–0.773) with excellent calibration, confirming the models stability and clinical utility for long-term prediction. Risk stratification based on the cox model effectively categorized patients into low-, intermediate-, and high-risk groups (P < 0.001). Additionally, an online platform to enable clinical guidance: https://crohndisease.shinyapps.io/CDAI0-1/.The prediction model incorporating routinely available clinical variables to accurately predict abdominal surgery risk in CD patients in remission or with mild activity, supporting clinical decision-making.