南方医科大学南方医院研究入选《Stroke》杂志2 月亮点文章，有望精准预警缺血性脑卒中风险

2025 年 2 月，南方医科大学南方医院在脑血管病领域权威期刊《Stroke》杂志第 2 期发表了题为《Large-Scale Plasma Proteomics Profiles for Predicting Ischemic Stroke Risk in the General Population》（大规模血浆蛋白质组学分析用于预测一般人群的缺血性脑卒中风险）的研究论文。此研究凭借其创新性与显著的临床价值，被《Stroke Alert Podcast》节目评选为当期亮点文章，并进行了深度专题报道。

在《Stroke Alert Podcast》第 49 期节目中，主持人 Negar Asdaghi 博士对该论文进行了重点推介。

On Episode 49 of the Stroke Alert Podcast, host Dr. Negar Asdaghi highlights an article from the February 2025 issue of Stroke:“Large-Scale Plasma Proteomics Profiles for Predicting Ischemic Stroke Risk in the General Population.”

Negar Asdaghi 博士：我们先从这些问题开始。
Dr. Negar Asdaghi: Let's start with some questions.

1) 我们能否通过检测某些血浆蛋白来预测后续患缺血性中风的风险呢？

1) Can certain plasma proteins predict the future risk of ischemic stroke?

2) 如果通过血液检测能在缺血性中风症状出现之前，就判断出是否有患病风险，这会带来怎样的影响呢？
2) What if your blood could reveal whether you're at risk for an ischemic stroke long before symptoms appear?

在本月的期刊里，来自中国广州南方医科大学南方医院国家肾脏病临床医学研究中心的秦献辉教授、甘小琴博士及他们的同事，在一篇题为《大规模血浆蛋白质组学分析用于预测一般人群的缺血性脑卒中风险》的文章中，探讨了缺血性脑卒中风险和血浆蛋白质组学之间的有趣联系。下面，让我们一起来深入了解一下，看看这些研究者是如何把血液变成预测未来的 “水晶球” 的。为了验证这个想法，研究者们采用了英国生物银行的数据。经常收听《Stroke Alert》播客的朋友，应该对英国生物银行不陌生。在我们的播客系列中，已经介绍过好几项利用该生物银行数据开展的研究。

In this month's issue of the journal, Dr. Xiaoqin Gan from National Clinical Research Center for Kidney Disease at Southern Medical University at Guangzhou, China, and colleagues explore a fascinating connection between the risk of ischemic stroke and the world of plasma proteomics in an article titled "Large-Scale Plasma Proteomics Profiles for Predicting Ischemic Stroke Risk in the General Population." Let's dive in and see how the authors are turning blood into a crystal ball. To explore this idea, the authors use data from the UK Biobank. Our Stroke Alert Podcast listeners are now well aware of the UK Biobank. We've covered several studies in our podcast series that have used data from this biobank.

我们知道，英国生物银行是一个大规模的医学数据库，包含了近50万名年龄在40至70岁之间的参与者的去识别化健康信息。其最大的优势在于，不仅是海量医学数据的集合，更是一个对所有研究人员开放的共享资源平台，为全球医学科研发展提供了有力支撑。

We know that the UK Biobank is a large-scale medical database containing de-identified health information from close to half a million participants that are between the ages of 40 and 70. The very important point about the UK Biobank is not only a great research resource, but it's also available to any researcher to use.

让我们聚焦本期期刊中的这篇蛋白质组学研究论文。在这项研究里，作者们用到了英国生物银行的一个子集，叫 “UK Biobank Pharma Proteomics Project，简称 UKB - PPP” 。这个项目涵盖了超过 53,000 名参与者，他们在加入英国生物银行时，都接受了蛋白质组学检测。在这篇论文中，作者仅保留了英国白人血统的个体，排除了有缺血性脑卒中史或心房颤动史的患者，最终纳入 43,000 多例患者用于分析。

So, let's go back to our proteomic paper in this issue of the journal. For this study, the authors use a subset of the UK Biobank known as the Pharma Proteomics Project. This project included over 53,000 participants who underwent proteomic measurement testing as part of their enrollment in the UK Biobank. For this paper, when they applied various exclusion criteria for inclusion in this study, importantly only keeping individuals of British White descent and excluding patients with a prior history of ischemic stroke or history of atrial fibrillation, they were left with over 43,000 patients for this analysis.

那么，研究团队具体开展了哪些研究工作呢？首先，他们依托强大的蛋白质组学数据库，从近 3000 种已测量的血浆蛋白中，运用 Lasso Cox 回归分析精准识别出 17 种与缺血性脑卒中密切相关的蛋白质，并以此为基础创建了一个蛋白质风险评分体系。在这 17 种蛋白质中，部分蛋白质与未来发生缺血性脑卒中的风险呈现正相关关系，也就是说，这些蛋白质可能会导致缺血性脑卒中风险增加；而另一部分蛋白质则呈现负相关关系，意味着它们在一定程度上能够起到保护作用，降低缺血性脑卒中发生的可能性。为了评估这个蛋白质风险评分体系的预测性能，研究人员引入了 C 统计量这一关键指标。C 统计量是衡量预测模型准确性的重要标准，进一步分析显示，这个新构建的蛋白质风险评分体系在预测缺血性脑卒中时，C 统计量达到了 0.76，这表明该评分体系具有较高的预测价值。值得注意的是，与传统的仅基于经典血管风险因素（如高血压、高血脂、高血糖等）、年龄和其他临床特征得出的缺血性脑卒中临床风险评分相比，新的蛋白质风险评分体系的 C 统计量更高；同时，也超过了英国生物银行此前发布的针对未来缺血性脑卒中风险的多基因风险评分。这充分表明，新构建的蛋白质风险评分体系在预测缺血性脑卒中风险方面，具有更为显著的优势，有望为临床预防和早期诊断提供更有力的支持。

So, what did they do? First off, using the proteomic data bank, they identified 17 proteins out of close to 3,000 measured plasma proteins to create a protein risk score. Some of these proteins were positively associated with a higher future risk of development of ischemic stroke, and some had a negative association. In other words, some were protective and some were, in a sense, a risk for future stroke. So, using Lasso Cox regression, they came up with the protein risk score, and it turns out that the protein risk score had a pretty decent C statistic. In other words, a pretty decent predictive value for predicting the odds of future ischemic stroke. The C score was 0.76. This was certainly higher than the ischemic stroke clinical risk score that uses classic vascular risk factors and age and other characteristics, or the polygenic risk score that had previously been released by the UK Biobank on the same outcome, which is the future risk of development of ischemic stroke.

更重要的是，在这 17 种蛋白质里，有 5 种在回归分析具有最高的绝对系数，表明这 5 种蛋白质在预测缺血性脑卒中方面发挥着最为关键的作用，贡献度最强。令人兴奋的是，研究人员仅凭借这 5 种蛋白质，再融合参与者的年龄和性别这些基础信息，就成功构建出一个更为简易的蛋白质组预测风险评分。经测试，这个新评分与最初复杂的包括17种蛋白质的风险评分相比，C 统计量近乎一致，预测能力也相差无几。这意味着复杂的蛋白质组学风险评分（需要更高的检测能力来测量多种血浆蛋白）可以被一个更简单的评分所取代。新的评分仅需测量五种血浆蛋白，再加上参与者的年龄和性别，便能在预测参与者未来缺血性脑卒中风险方面，展现出与复杂评分极为相似的价值。这一发现极具意义，为后续的临床应用开辟了新路径，带来了更多便利与可能，有望在缺血性脑卒中的预防和诊疗领域发挥重要作用。

Now, this is exciting. Better yet, they found that out of these 17 proteins, five had the most predictive value given that they had the highest absolute coefficient in the regression analysis. So, what made it really nice was that taking these five proteins alone and adding just simply age and sex of the participants, they were able to create a simpler proteomic predictive risk score that had essentially the same C statistics or predictive power as the original proteomic risk score. In essence, that would mean that the complex proteomic risk score, which would require higher capabilities to measure various plasma proteins, can be replaced by a much simpler score with only five plasma protein measurements and additions of participants' age and sex, to have very similar values in terms of predicting the future risk of ischemic stroke in any participant. Very interesting.

综上所述，这项研究成果表明，在不久的将来，我们或许只需进行一次快速的血液检测，再结合一些简单的计算，就能精准预测出哪些人群存在较高的缺血性脑卒中发病风险。这项研究充分地展现了蛋白质组学技术的巨大潜力，使我们朝着个性化医学和精准风险评估的目标又迈进了一大步。通过对特定蛋白质的分析，我们能够更深入地了解个体的健康状况，提前识别潜在的风险因素。这意味着，在缺血性脑卒中发生之前，我们就有足够的时间和能力制定并实施有效的预防措施，从而降低疾病的发生率，提高患者的生活质量。相信随着蛋白质组学技术的不断发展和完善，它将在临床实践中发挥更为重要的作用，为人类的健康事业做出更大的贡献。

So, in the future, we may be able to do a quick blood test and some simple calculation, and predict with a lot more certainty who will or who will not develop an ischemic stroke. This study is an example of how proteomic technology can get us closer to personalized medicine and risk ascertainment, and perhaps the ability to do corrective measures way before an incident ischemic stroke occurs.

参考文献：
Gan X, Yang S, Zhang Y, Ye Z, Zhang Y, Xiang H, Huang Y, Wu Y, Zhang Y, Qin X. Large-Scale Plasma Proteomics Profiles for Predicting Ischemic Stroke Risk in the General Population.Stroke. 2025 Feb;56(2):456-464. doi: 10.1161/STROKEAHA.124.048654.

Stroke Alert Podcast网页链接:
https://www.ahajournals.org/do/10.1161/podcast.20250212.411369

编辑| 甘小琴 蔡湘连
审核| 秦献辉 张园园