Interpretable laboratory-data model for risk stratification of elevated NT-proBNP and its deployment in diagnostic support middleware.

IF 1.1 4区医学 Q3 MEDICAL LABORATORY TECHNOLOGY

Annals of Clinical Biochemistry Pub Date : 2026-09-01 Epub Date: 2026-03-17 DOI:10.1177/00045632261438009

Hidekazu Ishida, Noriko Ozawa, Masaya Tachikawa, Hiroki Nagasawa, Yohei Shirakami, Takatomo Watanabe, Hiroyuki Okura, Ryosuke Kikuchi

{"title":"Interpretable laboratory-data model for risk stratification of elevated NT-proBNP and its deployment in diagnostic support middleware.","authors":"Hidekazu Ishida, Noriko Ozawa, Masaya Tachikawa, Hiroki Nagasawa, Yohei Shirakami, Takatomo Watanabe, Hiroyuki Okura, Ryosuke Kikuchi","doi":"10.1177/00045632261438009","DOIUrl":null,"url":null,"abstract":"<p><p>BackgroundHeart failure (HF) is a growing global burden. Although N-terminal pro-B-type natriuretic peptide (NT-proBNP) guides diagnosis, assay cost and analyzer availability limit routine use. Routine laboratory data may offer a low-cost triage alternative.MethodsWe developed and validated an interpretable decision tree to stratify the risk of elevated NT-proBNP >300 pg/mL and assessed deployment in a diagnostic support system (DSS). We analyzed 19,889 encounters at Gifu University Hospital (Aug 2022-May 2024). All 20 candidate predictors were included without prior feature selection to capture non-linear associations. Hyperparameters were tuned by 10-fold cross-validation. Final classification used a fixed decision rule optimized for high sensitivity (≥0.90 in training) to support effective triage. Performance comprised AUROC (DeLong 95% CIs) and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy (Wilson 95% CIs) on an internal hold-out set (n = 3978) and a temporal external cohort (n = 14,903; Jun 2024-Jun 2025). Analyses were complete-case with no imputation.ResultsThe decision tree inherently utilized clinically relevant predictors including serum albumin, eGFR, and age. Internal test performance: AUROC 0.804 (0.791-0.818); sensitivity 0.879 (0.863-0.893); specificity 0.505 (0.484-0.526); and accuracy 0.674 (0.660-0.689). External performance within the DSS: AUROC 0.806 (0.799-0.813); sensitivity 0.882 (0.874-0.890); specificity 0.518 (0.508-0.529); and NPV 0.852 (0.842-0.861). Calibration and decision-curve analysis supported clinical utility.ConclusionsAn interpretable tree built from routine laboratories detects clinically relevant NT-proBNP elevation with high sensitivity and performs robustly after deployment. This scalable, low-cost approach could enable risk-directed triage and more efficient resource allocation.</p>","PeriodicalId":8005,"journal":{"name":"Annals of Clinical Biochemistry","volume":" ","pages":"435-444"},"PeriodicalIF":1.1000,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Annals of Clinical Biochemistry","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1177/00045632261438009","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2026/3/17 0:00:00","PubModel":"Epub","JCR":"Q3","JCRName":"MEDICAL LABORATORY TECHNOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

BackgroundHeart failure (HF) is a growing global burden. Although N-terminal pro-B-type natriuretic peptide (NT-proBNP) guides diagnosis, assay cost and analyzer availability limit routine use. Routine laboratory data may offer a low-cost triage alternative.MethodsWe developed and validated an interpretable decision tree to stratify the risk of elevated NT-proBNP >300 pg/mL and assessed deployment in a diagnostic support system (DSS). We analyzed 19,889 encounters at Gifu University Hospital (Aug 2022-May 2024). All 20 candidate predictors were included without prior feature selection to capture non-linear associations. Hyperparameters were tuned by 10-fold cross-validation. Final classification used a fixed decision rule optimized for high sensitivity (≥0.90 in training) to support effective triage. Performance comprised AUROC (DeLong 95% CIs) and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy (Wilson 95% CIs) on an internal hold-out set (n = 3978) and a temporal external cohort (n = 14,903; Jun 2024-Jun 2025). Analyses were complete-case with no imputation.ResultsThe decision tree inherently utilized clinically relevant predictors including serum albumin, eGFR, and age. Internal test performance: AUROC 0.804 (0.791-0.818); sensitivity 0.879 (0.863-0.893); specificity 0.505 (0.484-0.526); and accuracy 0.674 (0.660-0.689). External performance within the DSS: AUROC 0.806 (0.799-0.813); sensitivity 0.882 (0.874-0.890); specificity 0.518 (0.508-0.529); and NPV 0.852 (0.842-0.861). Calibration and decision-curve analysis supported clinical utility.ConclusionsAn interpretable tree built from routine laboratories detects clinically relevant NT-proBNP elevation with high sensitivity and performs robustly after deployment. This scalable, low-cost approach could enable risk-directed triage and more efficient resource allocation.

查看原文本刊更多论文

NT-proBNP升高风险分层的可解释实验室数据模型及其在诊断支持中间件中的部署。

背景：心力衰竭（HF）是一个日益严重的全球负担。虽然n端前b型利钠肽（NT-proBNP）指导诊断，但检测成本和分析仪的可用性限制了常规使用。常规的实验室数据可以提供低成本的分诊选择。方法：我们开发并验证了一个可解释的决策树，对NT-proBNP升高的风险进行分层，并评估在诊断支持系统（DSS）中的部署。我们分析了岐阜大学医院（2022年8月至2024年5月）的19,889例病例。所有20个候选预测因子都包括在内，没有事先选择特征以捕获非线性关联。超参数通过10倍交叉验证进行调整。最终分类使用了一个固定的决策规则，该规则优化为高灵敏度（训练≥0.90），以支持有效的分类。在内部保留组（n= 3978）和临时外部队列（n= 14903； 2024年6月至2025年6月）中，性能包括AUROC （DeLong 95% ci）和敏感性、特异性、阳性预测值（PPV）、阴性预测值（NPV）和准确性（Wilson 95% ci）。分析是完整的，没有归咎。结果：决策树固有地利用了临床相关的预测因子，包括血清白蛋白、eGFR和年龄。内测性能：AUROC 0.804 (0.791-0.818)；敏感性0.879 (0.863-0.893)；特异性0.505 (0.484-0.526)；精度0.674（0.660-0.689）。DSS内部外部绩效：AUROC为0.806 (0.799-0.813)；敏感性0.882 (0.874-0.890)；特异性0.518 (0.508-0.529)；净现值0.852（0.842-0.861）。校正和决策曲线分析支持临床应用。结论：从常规实验室建立的可解释树以高灵敏度检测临床相关NT-proBNP升高，并且在部署后表现稳健。这种可扩展的低成本方法可以实现风险导向的分类和更有效的资源分配。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Annals of Clinical Biochemistry Biochemistry, Genetics and Molecular Biology-Clinical Biochemistry

CiteScore

5.20

自引率

4.50%

发文量

期刊介绍： Annals of Clinical Biochemistry is the fully peer reviewed international journal of the Association for Clinical Biochemistry and Laboratory Medicine. Annals of Clinical Biochemistry accepts papers that contribute to knowledge in all fields of laboratory medicine, especially those pertaining to the understanding, diagnosis and treatment of human disease. It publishes papers on clinical biochemistry, clinical audit, metabolic medicine, immunology, genetics, biotechnology, haematology, microbiology, computing and management where they have both biochemical and clinical relevance. Papers describing evaluation or implementation of commercial reagent kits or the performance of new analysers require substantial original information. Unless of exceptional interest and novelty, studies dealing with the redox status in various diseases are not generally considered within the journal''s scope. Studies documenting the association of single nucleotide polymorphisms (SNPs) with particular phenotypes will not normally be considered, given the greater strength of genome wide association studies (GWAS). Research undertaken in non-human animals will not be considered for publication in the Annals. Annals of Clinical Biochemistry is also the official journal of NVKC (de Nederlandse Vereniging voor Klinische Chemie) and JSCC (Japan Society of Clinical Chemistry).