Non-Invasive Functional Assessment of Vascular Stenosis

This direction is split into other directions: anatomical-functional diagnostic gap. Its last version stays readable and citable here.

Version 1 of 134 papers52 conceptsfrozen 2026-10-04

In clinical medicine, a "map" of the body is often mistaken for its "performance." For decades, the primary way to diagnose vascular disease has been anatomical imaging—CT scans, MRIs, and X-rays that show the physical narrowing (stenosis) of an artery. However, a growing body of research reveals a persistent "mismatch": the physical diameter of a vessel is a surprisingly poor predictor of whether blood is actually reaching the tissue it serves.

The emerging research direction focuses on functional hemodynamics: using computational fluid dynamics (CFD) and deep learning to derive physiological data—like pressure drops and flow velocity—directly from static anatomical images.

The Anatomy-Physiology Mismatch

The fundamental problem is that a 70% blockage in a coronary artery does not always cause a 70% drop in blood flow. As Sung‐A Chang et al. observed in high-risk surgical patients, many individuals exhibit significant physical stenosis without any corresponding perfusion defects. Conversely, Jung‐Hee Lee et al. found that in roughly one-third of intermediate cases, visual angiographic assessments disagree with physiological measurements.

This discrepancy occurs because blood flow is governed by complex fluid dynamics, not just geometry. Factors like blood viscosity, vessel wall friction, and the "internal weather" of a vascular tear (David Marlevi et al.) determine the actual risk. Relying solely on anatomy leads to "over-treatment" of benign narrowings and "under-treatment" of physiologically dangerous ones.

Physics as a Nonlinear Filter

To bridge this gap, researchers are treating the laws of physics as a "nonlinear filter" that translates static shapes into dynamic performance. Andrea Schillaci et al. demonstrated this by using CFD to extract "smarter" features for AI models, showing that the physics of airflow makes pathologies much easier to identify than geometry alone.

The "gold standard" for this functional assessment has traditionally been invasive: a surgeon must thread a pressure-wire into the heart to measure the Fractional Flow Reserve (FFR)—the ratio of maximum achievable blood flow in a diseased artery vs. a healthy one. The current research frontier aims to "virtualize" this procedure.

By applying CFD to standard CT scans (CT-FFR), Yaodong Ding et al. and Changling Li et al. have shown that we can achieve the accuracy of invasive tests non-invasively. This approach models the physiological impact of physical obstructions, allowing clinicians to see not just the "clog," but the resulting oxygen deprivation.

From Supercomputers to Digital Twins

While traditional CFD is computationally expensive and slow, new work is moving toward real-time "digital twins" of the vasculature. Xiaowu Sun et al. developed PINS-CAD, a framework that uses physics-informed self-supervised learning. Instead of solving complex fluid equations from scratch for every patient, the model is "taught" the laws of fluid dynamics using hundreds of thousands of synthetic arteries. This allows it to predict future cardiovascular events from a simple 2D X-ray more accurately than a human radiologist.

Similarly, Catarina Oliveira et al. have developed deep learning models that derive the "instantaneous free-wave ratio" (a pressure-based metric) directly from invasive coronary angiography images, effectively removing the need for additional invasive instrumentation.

Beyond the Heart

This shift from static mapping to functional assessment is expanding across medical disciplines:

  • Neurology: Xinyi Leng et al. found that high blood flow velocity and shear strain rates at the site of a brain artery narrowing are better predictors of repeat strokes than the physical degree of the blockage.
  • Pulmonology: Yuetan Chu et al. and Kaiwen Geng et al. are using deep learning to segment pulmonary vessels and predict lung function directly from CT scans, bypassing the need for traditional breathing tests.
  • Hepatology: Jiewen Chen et al. are using dual-energy CT to measure tissue perfusion in the liver, identifying physiological shifts that precede the physical enlargement of vessels in portal hypertension.

Why It Works

The reason these methods outperform traditional imaging is that they account for functional reserve. An organ can often compensate for a physical narrowing until a certain threshold is reached. Anatomical imaging only sees the narrowing; functional modeling sees the exhaustion of that compensation. As Junjie Yang et al. note, integrating these simulations into clinical workflows is projected to reduce unnecessary invasive diagnostic procedures by 30%, ensuring that interventions are reserved for lesions that truly restrict life-sustaining flow.

Papers in this version

History & citationVersion 1 · 2026-10-04

Split into other directions · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_non_invasive_functional_assessment_of_va_747aa2a8,
  title = {Non-Invasive Functional Assessment of Vascular Stenosis},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/non-invasive-functional-assessment-of-vascular-stenosis/rdr_747aa2a8bfde4e38a2e7e48a0a9a9313},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_non_invasive_functional_assessment_of_va_747aa2a8_v1,
  title = {Non-Invasive Functional Assessment of Vascular Stenosis},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/non-invasive-functional-assessment-of-vascular-stenosis/rdr_747aa2a8bfde4e38a2e7e48a0a9a9313/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}