利用数据驱动方法改进脑血管病动态成像中的脑血流测量
回顾性分析16例单侧狭窄闭塞性疾病患者的乙酰唑胺负荷血氧水平依赖MRI与DSC灌注数据,训练三种随机森林分类器模型以识别CVR在正常半球中位数10%以内的候选体素。模型预测参考CVR与真实值的中位绝对百分比差异为全脑12.8%、灰质11.3%、白质9.8%,阈值体积分类除最低白质阈值外无统计学显著差异。在双侧病变小样本试点中展示了不依赖对侧参考的体素级CVR评估潜力。
要点
- 随机森林模型可脑全域预测参考CVR,白质模型中位绝对百分比差异9.8%优于全脑12.8%。
- 基于模型参考CVR的阈值体积分类与真实值高度一致,仅最低白质CVR阈值除外。
- 小样本双侧狭窄闭塞性疾病试点显示可不依赖对侧半球参考进行体素级CVR评估。
核心结论
- 提出一种数据驱动方法,利用随机森林分类器为双侧或不确定分布狭窄闭塞性疾病患者提供个体化全脑参考CVR,扩展CVR评估超越单侧病变限制,并可能应用于SPECT或PET血流动力学研究等类似受限场景。
研究设计与样本
- 回顾性分析16例单侧狭窄闭塞性疾病患者,行乙酰唑胺负荷血氧水平依赖MRI和DSC灌注;训练三种随机森林分类器模型(全脑体素、仅灰质、仅白质),采用留一交叉验证,输入特征为DSC的达峰时间、平均通过时间、CBF和CBV。
诊断标准与关键数值
- 以CVR在正常半球中位数10%以内的体素为候选参考体素;模型预测参考CVR与真实值中位绝对百分比差异:全脑12.8%(IQR 5.0%-18.9%)、灰质11.3%(IQR 9.3%-16.1%)、白质9.8%(IQR 4.4%-16.9%);阈值体积分类除最低白质CVR阈值外,模型与真实值无统计学显著差异(P > .01)。
临床应用与阅片要点
- 作者认为该方法可提供个体化全脑参考CVR,使体素级CVR评估不依赖对侧半球参考,并可能应用于SPECT或PET血流动力学研究等类似受限场景。
局限性
- 原文为回顾性分析。
- 样本量为16例单侧狭窄闭塞性疾病患者。
- 双侧狭窄闭塞性疾病病例仅为小样本试点部署。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
BACKGROUND AND PURPOSE: Cerebrovascular reactivity (CVR) is a widely studied biomarker of cerebral hemodynamics, commonly used in risk stratification and treatment planning in patients with steno-occlusive disease (SOD). Conventional use relies on normalization of estimates to contralateral hemisphere reference values, which is unsuitable for bilateral or indeterminate distributions of disease. We report upon a custom data-driven approach leveraging random forest classifiers (RFc) to identify candidate voxels for normalization to facilitate interrogation outside conditions of known unilateral SOD.
MATERIALS AND METHODS: We retrospectively analyzed 16 patients with unilateral SOD who underwent acetazolamide-augmented blood oxygen level- dependent-MRI and DSC perfusion. Three RFc models were trained by using leave-one-out cross-validation (LOOCV) to identify candidate voxels brain-wide whose CVR was within 10% of the normal hemispheric median: 1) all voxels; 2) gray matter only; and 3) white matter only. Model input features included time-to-maximum, mean transit time, CBF, and CBV from contemporaneous DSC. The median model-predicted reference cerebrovascular reactivity (CVRref) was compared with ground-truth medians in LOOCV, and its impact on threshold-based volumetric classification of CVR reduction assessed.
RESULTS: RFc models effectively predicted ground-truth CVR voxels, achieving median absolute percent differences of 12.8% (interquartile range [IQR]: 5.0%-18.9%) by using all voxels, 11.3% (IQR: 9.3%-16.1%) for gray matter, and 9.8% (IQR: 4.4%-16.9%) for white matter. Volumetric estimates of CVR reduction across thresholds for the models revealed excellent agreement between ground-truth and model estimates without statistically significant differences (P > .01), excepting lowest white matter CVR thresholds. Model use in a small pilot deployment of bilateral SOD cases demonstrated the potential utility, enabling voxelwise CVR assessment without reliance on contralateral reference.
CONCLUSIONS: We present a novel data-driven approach for normalizing CVR maps in patients with bilateral or indeterminate SOD. Using an RFc, our method provides an individualized, brain-wide reference CVR, expanding the utility of CVR estimates beyond the typical constraints of unilateral disease, and with potential application to other, similarly constrained scenarios such as for SPECT or PET hemodynamic studies.
引用
- Dogra S, Wang X, Gee JM, et al. Using Data-Driven Methods to Improve Brain Blood Flow Measurements in Cerebrovascular Disease with Dynamic Imaging. AJNR Am J Neuroradiol. 2025. PMID: 40262947. doi:10.3174/ajnr.A8813
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。