Scholarlink

Purpose <p>This study aimed to develop and validate a clinical-biomarker nomogram integrating heparin-binding protein (HBP), soluble suppression of tumorigenicity-2 (sST2), monocyte-to-lymphocyte ratio (MLR), and routine clinical variables for identifying baseline HF status during hospitalization.</p> Methods <p>This retrospective two-district study included 1279 hospitalized patients. A total of 737 patients from the Central District were randomly split into training and internal validation cohorts at a 7:3 ratio, and 542 patients from the South District served as the external validation cohort. Predictors were selected using LASSO logistic regression and multivariable logistic regression. Model performance was assessed by discrimination, calibration, clinical utility, and reclassification analyses. Training-derived thresholds were used for probability stratification, and a web-based calculator was developed.</p> Results <p>The final model included HBP, sST2, MLR, age, creatinine, aspartate aminotransferase, and albumin. The AUCs were 0.895 (95% CI 0.869–0.921), 0.901 (95% CI 0.860–0.942), and 0.816 (95% CI 0.782–0.851) in the training, internal validation, and external validation cohorts, respectively. The corresponding Brier scores were 0.132, 0.124, and 0.185. Compared with the prespecified clinical reference model, adding HBP, sST2, and MLR improved model performance, with a continuous net reclassification improvement of 1.140 (95% CI 1.001–1.277, <i>P</i> &lt; 0.001) and an integrated discrimination improvement of 0.311 (95% CI 0.276–0.351, <i>P</i> &lt; 0.001). Observed HF proportions increased progressively across low-, intermediate-, and high-probability groups in all cohorts.</p> Conclusion <p>This two-district retrospective study developed and externally validated a practical clinical-biomarker nomogram for estimating individualized prevalent HF probability during hospitalization. By integrating HBP, sST2, MLR, and routine clinical variables, this tool may support preliminary screening and probability-based triage using complementary non-imaging information. It should be used as an auxiliary tool rather than a replacement for guideline-based HF diagnosis, and further prospective multicenter validation is warranted.</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

Development and validation of a multi-biomarker nomogram for baseline heart failure

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月9日

查看全文
  • YuLin Liu,
  • KeKe Ding,
  • Bing Zhao,
  • XiZhen Fan,
  • ZiQiu Tang,
  • HongQi Li,
  • Li Chen

摘要

Purpose

This study aimed to develop and validate a clinical-biomarker nomogram integrating heparin-binding protein (HBP), soluble suppression of tumorigenicity-2 (sST2), monocyte-to-lymphocyte ratio (MLR), and routine clinical variables for identifying baseline HF status during hospitalization.

Methods

This retrospective two-district study included 1279 hospitalized patients. A total of 737 patients from the Central District were randomly split into training and internal validation cohorts at a 7:3 ratio, and 542 patients from the South District served as the external validation cohort. Predictors were selected using LASSO logistic regression and multivariable logistic regression. Model performance was assessed by discrimination, calibration, clinical utility, and reclassification analyses. Training-derived thresholds were used for probability stratification, and a web-based calculator was developed.

Results

The final model included HBP, sST2, MLR, age, creatinine, aspartate aminotransferase, and albumin. The AUCs were 0.895 (95% CI 0.869–0.921), 0.901 (95% CI 0.860–0.942), and 0.816 (95% CI 0.782–0.851) in the training, internal validation, and external validation cohorts, respectively. The corresponding Brier scores were 0.132, 0.124, and 0.185. Compared with the prespecified clinical reference model, adding HBP, sST2, and MLR improved model performance, with a continuous net reclassification improvement of 1.140 (95% CI 1.001–1.277, P < 0.001) and an integrated discrimination improvement of 0.311 (95% CI 0.276–0.351, P < 0.001). Observed HF proportions increased progressively across low-, intermediate-, and high-probability groups in all cohorts.

Conclusion

This two-district retrospective study developed and externally validated a practical clinical-biomarker nomogram for estimating individualized prevalent HF probability during hospitalization. By integrating HBP, sST2, MLR, and routine clinical variables, this tool may support preliminary screening and probability-based triage using complementary non-imaging information. It should be used as an auxiliary tool rather than a replacement for guideline-based HF diagnosis, and further prospective multicenter validation is warranted.