机器学习优化利钠肽在急性心力衰竭诊断中的应用。

Machine learning to optimise use of natriuretic peptides in the diagnosis of acute heart failure.

作者信息

Doudesis Dimitrios, Lee Kuan Ken, Anwar Mohamed, Singer Adam J, Hollander Judd E, Chenevier-Gobeaux Camille, Claessens Yann-Erick, Wussler Desiree, Weil Dominic, Kozhuharov Nikola, Strebel Ivo, Sabti Zaid, deFilippi Christopher, Seliger Stephen, Mesquita Evandro Tinoco, Wiemer Jan C, Möckel Martin, Coste Joel, Jourdain Patrick, Kimiaki Komukai, Yoshimura Michihiro, Ibrahim Irwani, Ooi Shirley Beng Suat, Sen Kuan Win, Gegenhuber Alfons, Mueller Thomas, Hanon Olivier, Vidal Jean-Sébastien, Cameron Peter, Lam Louisa, Freedman Ben, Chung Tommy, Collins Sean P, Lindsell Christopher J, Newby David E, Japp Alan G, Shah Anoop S V, Villacorta Humberto, Richards A Mark, McMurray John J V, Mueller Christian, Januzzi James L, Mills Nicholas L

机构信息

British Heart Foundation (BHF) Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, UK.

Usher Institute, University of Edinburgh, Edinburgh, UK.

出版信息

Eur Heart J Acute Cardiovasc Care. 2025 Apr 12. doi: 10.1093/ehjacc/zuaf051.

DOI:10.1093/ehjacc/zuaf051

PMID:40219913

Abstract

AIMS

B-type natriuretic peptide (BNP) and mid-regional pro-atrial natriuretic peptide (MR-proANP) testing are guideline-recommended to aid in the diagnosis of acute heart failure. Nevertheless, the diagnostic performance of these biomarkers is uncertain.

METHODS

We performed a systematic review and individual patient-level data meta-analysis to evaluate the diagnostic performance of BNP and MR-proANP. We subsequently developed and externally validated a decision-support tool called CoDE-HF that combines natriuretic peptide concentrations with clinical variables using machine learning to report the probability of acute heart failure.

RESULTS

Fourteen studies from 12 countries provided individual patient-level data in 8,493 patients for BNP and 3,899 patients for MR-proANP, in whom, 48.3% (4,105/8,493) and 41.3% (1,611/3,899) had an adjudicated diagnosis of acute heart failure, respectively. The negative predictive value (NPV) of guideline-recommended thresholds for BNP (100 pg/mL) and MR-proANP (120 pmol/L) was 93.6% (95% confidence interval 88.4-96.6%) and 95.6% (92.2-97.6%), respectively, whilst the positive predictive value (PPV) was 68.8% (62.9-74.2%) and 64.8% (56.3-72.5%). Significant heterogeneity in the performance of these thresholds was observed across important subgroups. CoDE-HF was well calibrated with excellent discrimination in those without prior acute heart failure for both BNP and MR-proANP (area under the curve of 0.914 [0.906-0.921] and 0.929 [0.919-0.939], and Brier scores of 0.110 and 0.094, respectively). CoDE-HF with BNP and MR-proANP identified 30% and 48% as low-probability (NPV of 98.5% [97.1-99.3%] and 98.5% [97.7-99.0%]), and 30% and 28% as high-probability (PPV of 78.6% [70.4-85.0%] and 75.1% [70.9-78.9%]), respectively, and performed consistently across subgroups.

CONCLUSION

The diagnostic performance of guideline-recommended BNP and MR-proANP thresholds for acute heart failure varied significantly across patient subgroups. A decision-support tool that combines natriuretic peptides and clinical variables was more accurate and supports more individualised diagnosis.

STUDY REGISTRATION

PROSPERO number, CRD42019159407.

摘要

目的

B型利钠肽（BNP）和中段心房利钠肽原（MR-proANP）检测是指南推荐用于辅助急性心力衰竭诊断的方法。然而，这些生物标志物的诊断性能尚不确定。

方法

我们进行了一项系统评价和个体患者水平的数据荟萃分析，以评估BNP和MR-proANP的诊断性能。随后，我们开发并外部验证了一种名为CoDE-HF的决策支持工具，该工具使用机器学习将利钠肽浓度与临床变量相结合，以报告急性心力衰竭的概率。

结果

来自12个国家的14项研究提供了8493例患者的BNP个体患者水平数据和3899例患者的MR-proANP个体患者水平数据，其中分别有48.3%（4105/8493）和41.3%（1611/3899）经判定诊断为急性心力衰竭。指南推荐的BNP（100 pg/mL）和MR-proANP（120 pmol/L）阈值的阴性预测值（NPV）分别为93.6%（95%置信区间88.4-96.6%）和95.6%（92.2-97.6%），而阳性预测值（PPV）分别为68.8%（62.9-74.2%）和64.8%（56.3-72.5%）。在重要亚组中观察到这些阈值性能存在显著异质性。对于BNP和MR-proANP，CoDE-HF在无既往急性心力衰竭的患者中校准良好且具有出色的鉴别能力（曲线下面积分别为0.914[0.906-0.921]和0.929[0.919-0.939]，Brier评分分别为0.110和0.094）。结合BNP和MR-proANP的CoDE-HF分别将30%和48%识别为低概率（NPV为98.5%[97.1-99.3%]和98.5%[97.7-99.0%]），将30%和28%识别为高概率（PPV为78.6%[70.4-85.0%]和75.1%[70.9-78.9%]），并且在各亚组中表现一致。