血液多标志物组合区分阿尔茨海默病与健康对照：200 名参与者测 12 项指标、《科学报告》模型曲线下面积达 0.9999 且注明须外部验证
中国 — 《科学报告》2026 年 11 月 11 日刊出一项阿尔茨海默病血液标志物研究，作者为 Li、Sun、Jin 等人。研究旨在筛选有区分力的血液生物标志物，并评估其区分已确诊阿尔茨海默病患者与健康对照的表现。
样本与指标方面，研究共招募 200 名参与者，其中阿尔茨海默病患者 100 人、健康对照 100 人，共测量 12 项血液生物标志物，涵盖神经营养因子、tau 蛋白、神经元损伤标志物与外周炎症指标。
方法方面，研究采用嵌套分层十折交叉验证以避免数据泄漏，以 L1 惩罚的 LASSO 逻辑回归作为主要特征选择方法，并构建堆叠集成模型作对照。在 lambda.1se 标准下，模型保留了 8 项预测标志物；其中脑源性神经营养因子、神经生长因子、神经丝轻链与 pTau217 是两条分析管线一致选出的稳健标志物，pTau231 与 S100β 为第二梯队稳定标志物。
表现方面，LASSO 模型的折外曲线下面积与精确率召回曲线下面积均达到 0.9999，堆叠模型两项指标均为 0.999。论文明确指出，全部表现指标来自单中心内部交叉验证，临床应用前必须做外部多队列验证。本稿数字均以论文摘要正文为准，研究结论以该验证限定为准。
Blood Multi-Biomarker Panel Discriminates Alzheimer Disease from Healthy Controls: 200 Participants and 12 Markers, Scientific Reports Model Reaches AUROC of 0.9999 with External Validation Still Required
China — Scientific Reports published a blood biomarker study of Alzheimer disease on November 11, 2026, by Li, Sun, Jin and colleagues. The study aimed to screen discriminatory blood biomarkers and evaluate their performance in differentiating patients with established Alzheimer disease from healthy controls.
The study recruited 200 participants, 100 with Alzheimer disease and 100 healthy controls, and measured 12 blood biomarkers, including neurotrophic factors, tau proteins, neuronal injury markers and peripheral inflammatory indices.
Nested stratified ten-fold cross-validation was used to avoid data leakage. L1-penalized LASSO logistic regression served as the primary feature selection method, with a stacking ensemble model constructed for benchmarking. Under the lambda.1se criterion, eight predictive biomarkers were retained. Brain-derived neurotrophic factor, nerve growth factor, neurofilament light and pTau217 were robust markers consistently selected by both pipelines, while pTau231 and S100 beta represented second-tier stable markers.
The LASSO-derived model achieved an out-of-fold AUROC of 0.9999 and AUPRC of 0.9999; the stacking model reached an AUROC of 0.999 and an AUPRC of 0.999. The paper states that all performance metrics were derived from single-centre internal cross-validation and that external multi-cohort validation is mandatory before clinical application. All figures in this report follow the abstract of the paper.
发布：GSM·美国（发布链接）
图片来源：图：图中是一排血液检测样本管。图中是血液样本管的实物照片，为血液标志物研究题材通用配图，非本研究样本。Oleg Yunakov，CC BY-SA 4.0，经Wikimedia Commons。 / Photo: file photo. Oleg Yunakov, CC BY-SA 4.0, via Wikimedia Commons.
GSM Global · 本篇英文版 English edition of this story
