血管几何决定脉象频谱:1-D血流模型的闭环验证

Written by

in

Language

Evidence level: B

中医”脉诊”读取的波形,本质上是心脏搏动在特定血管几何网络中传播、反射后叠加出来的频谱——而这一点,现代生物力学已用”1-D血流模型”给出了可闭环验证的定量答案。

正文

中医把脉,三指搭在寸关尺,感知的是桡动脉上一段波形的形态、节律与力度。古人把这套读法总结成”脉证”,并与脏腑气血的盛衰对应起来。问题是:这段波形到底由什么决定?如果它主要由每个人血管的”几何形态”决定,那么”同病不同脉””同脉不同病”就获得了物理学的解释——脉象不是玄学,而是血管几何与心脏输出共同作用的结果。

现代生物力学给出了一条可量化的路径:一维(1-D)血流模型。它把人体动脉树抽象成一组串联的弹性管道,每条管道有其长度、半径、管壁弹性与分支位置,心脏的搏动作为入口边界条件输入,就能数值求解出压力波与血流波在整棵动脉树中传播的波形。关键在于:模型的输入参数——血管的几何结构——可以直接由无创成像测量获得,模型输出的波形又可以和体表测得的脉搏波直接对照。这就是一个”几何进、频谱出”的闭环。

这条闭环的可行性已被多项研究证实。Alastruey 等人在一个 37 条主传导动脉的 1:1 硅胶复制模型上,用完全由实测几何参数(无任何拟合)驱动的 1-D 模型,成功复现了 70 个测量位置的脉搏波形,压力波形的平均相对均方根误差低于 4%,血流误差低于 19%;加入管壁粘弹性后,压力误差进一步从 3.0% 降到 2.5%,血流误差从 15.7% 降到 10.8%(PMID 17640653, PMID 21724188)。这说明只要几何参数准确,1-D 模型就能还原波形的核心特征——波形不是”编”出来的,是被血管结构”决定”出来的。

更贴近临床的是患者特异的验证。Reymond 等用 MR 血管造影重建某个具体受试者的动脉树几何,再以同一受试者的压平眼压计与相位对比 MRI 实测波形做对照,结果表明患者特异模型在具体动脉位点对压力与血流波形的预测显著优于泛化的通用模型(PMID 21622820)。也就是说,个体血管几何的差异,确实携带了决定其脉搏频谱的独立信息量。这也直接呼应了”水法”与”脉诊”的科学内核:水的通道几何决定水波,血的通道几何决定脉波。

VascAgeNet 的权威综述进一步把散落的证据整合起来:1-D 模型对压力、血流、管腔横截面积的波形预测,相对误差可低至 1.2%、2.1% 与 2.6%,其精度足以与三维流固耦合模型相媲美,而计算成本低几个数量级(PMID 37000606)。这为”用无创血管几何测量 + 1-D 模型预测个体脉搏频谱”的临床应用铺平了道路。

放到易学-科学桥接的视角看,这条闭环的意义在于:中医脉诊的物理载体被明确地锚定在”血管几何”这个可测量、可建模、可预测的实体上。 “脉象=脏腑状态”的古老断言,其可检验的现代形式是”脉波频谱=F(血管几何, 心脏输出, 管壁弹性)”。前两项可用超声/MRI 无创测出,后一项可通过脉搏波速度(PWV)间接估计——三者合成一个可反演的预测模型。剩下的工作,就是像 VascAgeNet 综述所指出的那样,把”血管年龄””个体血管几何指纹”与经典脉象分类做系统对照。

当然,闭环尚有关键缺口:管壁弹性的个体化标定仍需借助文献先验而非直接测量;末梢小动脉-微循环边界条件存在简化;脉搏波形的形态学相似不等于频谱能量分布的逐谐波吻合。而这些缺口,恰恰是下一轮实验可以直接检验的课题。

证据与局限

已验证: 1-D 血流模型能在几何参数完全实测(无数据拟合)的条件下,以 <4% 压力误差复现 37 条动脉网络的脉搏波形(PMID 17640653, PMID 21724188);患者特异模型显著优于通用模型(PMID 21622820);VascAgeNet 综述确认 1-D 模型精度可媲美 3-D(PMID 37000606)。
部分验证: 管壁粘弹性显著降低高频过冲误差(压力 3.0%→2.5%,血流 15.7%→10.8%),说明弹性参数是第二阶重要输入;”血管几何指纹→个体脉搏频谱”的预测链条在方法学上成立,但尚未与经典脉象分类做系统盲法对照。
待验证: 无创血管几何测量与 1-D 模型预测频谱能否在大样本人群上稳定复现”同病不同脉/同脉不同病”;管壁弹性个体化标定是否成为闭环精度的主要瓶颈;预测频谱的逐谐波能量分布(而非仅波形形态)与实测的吻合度边界。

相关阅读

证据等级: B(1-D 模型的预测能力有 A 级直接支撑,但”血管几何→脉象频谱→脏腑辨证”的完整桥接链仍属科学类比/间接证据)
参考文献: PMID 37000606, PMID 21622820, PMID 21724188, PMID 17640653, PMID 22206422, PMID 29334240


免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。

Vascular Geometry Determines Pulse Wave Spectra: Closed-Loop Validation of a 1-D Hemodynamic Model

Evidence level: B

The waveform read by Traditional Chinese Medicine (TCM) “pulse diagnosis” is essentially a spectrum generated by the propagation and reflection of cardiac pulses within a specific vascular geometric network. Modern biomechanics has provided a quantifiable, closed-loop verifiable answer to this phenomenon using the “1-D hemodynamic model.”

Main Text

In TCM, pulse diagnosis involves placing three fingers on the Cun, Guan, and Shi positions to perceive the morphology, rhythm, and force of a specific waveform on the radial artery. Ancient practitioners summarized this reading into “pulse syndromes” and correlated them with the abundance or deficiency of organ zhangfu and qi-blood. The question remains: what determines this waveform? If it is primarily determined by the individual’s “geometric morphology” of the vasculature, then the phenomena of “same disease, different pulse” and “same pulse, different disease” gain a physical explanation—the pulse is not esoteric but the result of the interaction between vascular geometry and cardiac output.

Modern biomechanics provides a quantifiable path: the one-dimensional (1-D) hemodynamic model. This abstracts the arterial tree into a series of elastic pipes, each characterized by length, radius, wall elasticity, and branching location. With cardiac pulsation as the inlet boundary condition, numerical solutions can be derived for pressure and flow waves propagating throughout the arterial tree. The key lies in: the model’s input parameters—vascular geometry—can be directly measured via non-invasive imaging, while the model’s output waveforms can be directly compared with pulse waves measured at the body surface. This constitutes a “geometry in, spectrum out” closed loop.

The feasibility of this closed loop has been confirmed by multiple studies. Alastruey et al. successfully reproduced pulse waveforms at 70 measurement sites on a 1:1 silicone replica model comprising 37 major conducting arteries using a 1-D model driven entirely by measured geometric parameters (with no fitting). The mean relative root-mean-square error for pressure waveforms was below 4%, and flow error was below 19%; incorporating wall viscoelasticity further reduced pressure error from 3.0% to 2.5% and flow error from 15.7% to 10.8% (PMID 17640653, PMID 21724188). This demonstrates that provided geometric parameters are accurate, the 1-D model can reconstruct the core features of the waveform—the waveform is not “constructed” but is “determined” by vascular structure.

More clinically relevant are patient-specific validations. Reymond et al. reconstructed the arterial tree geometry of a specific subject using MR angiography and compared it against measured waveforms from an applanation tonometry and phase-contrast MRI of the same subject. The results indicated that patient-specific models significantly outperform generalized universal models in predicting pressure and flow waveforms at specific arterial sites (PMID 21622820). In other words, inter-individual differences in vascular geometry indeed carry independent information determining their pulse spectra. This directly resonates with the scientific core of “Water Method” and “Pulse Diagnosis”: the geometry of water channels determines water waves, and the geometry of blood channels determines pulse waves.

An authoritative review by VascAgeNet further integrates scattered evidence: the 1-D model predicts waveforms for pressure, flow, and lumen cross-sectional area with relative errors as low as 1.2%, 2.1%, and 2.6% respectively, a precision comparable to three-dimensional fluid-structure interaction models but with computational costs lower by several orders of magnitude (PMID 37000606). This paves the way for clinical applications involving “non-invasive vascular geometry measurement + 1-D model prediction of individual pulse spectra.”

From the perspective of bridging Yijing Studies and science, the significance of this closed loop lies in: the physical carrier of TCM pulse diagnosis is explicitly anchored to “vascular geometry,” a measurable, modelable, and predictable entity. The ancient assertion that “pulse equals organ state” finds its testable modern form as “pulse wave spectrum = F(vascular geometry, cardiac output, wall elasticity).” The first two terms can be non-invasively measured via ultrasound/MRI, while the latter can be indirectly estimated via pulse wave velocity (PWV)—three factors synthesized into an invertible predictive model. The remaining task is to systematically correlate “vascular age” and “individual vascular geometry fingerprints” with classical pulse classifications, as highlighted by the VascAgeNet review.

Of course, the closed loop still has key gaps: individualized calibration of wall elasticity still relies on literature priors rather than direct measurement; boundary conditions at the distal small arteries and microcirculation are simplified; morphological similarity of pulse waveforms does not equate to harmonic-by-harmonic agreement in spectral energy distribution. These gaps, however, are precisely the subjects that can be directly tested in the next round of experiments.

Evidence and Limitations

Verified: The 1-D hemodynamic model can reproduce pulse waveforms in a 37-artery network with <4% pressure error under conditions of fully measured geometric parameters (no data fitting) (PMID 17640653, PMID 21724188); patient-specific models significantly outperform universal models (PMID 21622820); the VascAgeNet review confirms 1-D model precision is comparable to 3-D (PMID 37000606).
Partially Verified: Wall viscoelasticity significantly reduces high-frequency overshoot errors (pressure 3.0%→2.5%, flow 15.7%→10.8%), indicating elasticity parameters are a second-order important input; the predictive chain from “vascular geometry fingerprint → individual pulse spectrum” is methodologically sound but has not yet undergone systematic blind comparison with classical pulse classifications.
To Be Verified: Whether non-invasive vascular geometry measurement combined with 1-D model prediction can stably reproduce “same disease, different pulse/same pulse, different disease” across large population cohorts; whether individualized wall elasticity calibration becomes the primary bottleneck for closed-loop precision; the boundary of agreement between predicted spectral energy distributions (rather than just waveform morphology) and measured data.

Related Reading

Evidence Level: B (The predictive capability of the 1-D model has direct A-level support, but the complete bridging chain from “vascular geometry → pulse spectrum → organ differentiation” remains a scientific analogy/indirect evidence).
References: PMID 37000606, PMID 21622820, PMID 21724188, PMID 17640653, PMID 22206422, PMID 29334240


Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *