Imagine a highway with a single lane closed for construction. If only three cars are driving on the road, the closure causes no delay. However, during rush hour, that exact same bottleneck creates a massive traffic jam.
For decades, clinical diagnostics in vascular disease operated under a simple assumption: if a blood vessel looks narrow on a scan, it must be blocking blood flow. This is the anatomical approach. But as fluid dynamics and clinical outcomes have revealed, anatomy is an imperfect proxy for physiology. A physical narrowing (stenosis) may look severe on an X-ray or CT scan while maintaining perfectly adequate blood flow. Conversely, a vessel that appears only mildly obstructed might fail catastrophically under stress.
This discrepancy is known as the anatomical-functional diagnostic gap. Bridging this gap means moving away from purely visual assessments of structure and moving toward the quantification of actual physiological behavior.
The Illusion of Visual Severity
When clinicians look at a traditional two-dimensional angiogram, they estimate the severity of a blockage by comparing the narrowest part of the vessel to a nearby healthy segment. However, this visual estimation is notoriously unreliable.
In a post-hoc analysis of the FLAVOUR trial, [Jung‐Hee Lee] et al. (/work/discordance-between-angiographic-assessment-and-fractional) demonstrated that visual angiographic assessments disagree with objective physiological or structural measurements in approximately one-third of intermediate coronary lesions. This high rate of discordance means that relying solely on visual estimation frequently misclassifies how dangerous a blockage actually is. Furthermore, the authors noted that this relationship is highly dependent on the specific vessel architecture; for instance, blockages in the left anterior descending artery are far more likely to be functionally severe than they appear visually, due to the massive myocardial mass demanding blood downstream.
This mismatch is not unique to the heart. In a study evaluating computational models of intracranial arteries, [Xinyi Leng] et al. (/work/computational-fluid-dynamics-of-computed-tomography) pointed out that patients with identical percentages of physical arterial narrowing in the brain exhibit vastly different levels of physiological stability. Similarly, in pulmonology, [Aleksandr Medvedev] (/work/the-therapy-of-broncho-obstructive-syndrome-in-patients) demonstrated that the radiological staging of inflammatory lung diseases often fails to correlate with the actual degree of functional bronchial constriction, meaning that imaging alone can easily underestimate a patient's respiratory impairment.
The Fluid Dynamics of a Bottleneck
Why does anatomy lie? The answer lies in fluid mechanics. The pressure drop ($\Delta P$) across a vascular narrowing is not determined solely by the minimum diameter of the lumen. Instead, it is governed by a complex interplay of fluid velocity, flow separation, friction along the lesion length, and the geometry of the vessel entry and exit.
To capture this physics, cardiologists developed Fractional Flow Reserve (FFR). FFR is a physiological index calculated as the ratio of maximum blood pressure distal to a blockage versus the pressure proximal to it during maximum blood flow (hyperemia):
$$\text{FFR} = \frac{P_{\text{distal}}}{P_{\text{proximal}}}$$
If the ratio is $1.0$, blood flow is completely unimpeded. If it drops below $0.80$, the blockage is considered functionally significant, meaning it is actively depriving the downstream tissue of oxygen. As [Changling Li] et al. (/work/diagnostic-performance-of-angiography-based-fractional-flow) showed, evaluating these flow dynamics provides a significantly more reliable detection of true ischemia than traditional 2D anatomical measurements, confirming that visual narrowing is an insufficient predictor of physiological impact.
Historically, measuring FFR required threading a physical pressure wire into a patient's artery—an invasive, expensive, and technically demanding procedure.
Simulating Physiology Non-Invasively
To make functional assessments accessible, researchers have turned to Computational Fluid Dynamics (CFD) and machine learning. By applying the Navier-Stokes equations—the mathematical bedrock of fluid flow—to three-dimensional anatomical models derived from standard CT or angiographic scans, we can simulate blood flow and calculate "virtual" FFR without ever touching the patient with a pressure wire.
[Standard 3D Scan (CT/Angio)] ──> [Vessel Segmentation & Mesh] ──> [CFD / AI Flow Simulation] ──> [Virtual FFR / Hemodynamic Risk]
For example, [Yaodong Ding] et al. (/work/diagnostic-accuracy-of-noninvasive-fractional-flow-reserve) validated a non-invasive method that derives FFR directly from computed tomography angiography (CT-FFR). By incorporating physiological simulations into anatomical imaging, the authors achieved highly accurate ischemia diagnoses even in complex cases with severe calcification, which typically blinds traditional visual inspection.
Taking this a step further, [Catarina Oliveira] et al. (/work/non-invasive-derivation-of-instantaneous-free-wave-ratio-uynh) evaluated a deep learning model designed to instantly derive functional flow ratios from invasive coronary angiography images. Because the AI can model the underlying fluid dynamics faster than traditional human operators can visually guess them, it bypasses the unreliability of visual estimation entirely. Similarly, [Youchang Yang] et al. (/work/prediction-of-myocardial-ischemia-in-coronary-heart-disease) built a radiomic nomogram combining coronary CT angiography features with clinical data, significantly outperforming traditional clinical models in predicting whether a physical blockage was causing true myocardial ischemia.
A Systemic Diagnostic Gap
The necessity of looking past structure to see function extends across the entire human body:
- Stroke Recurrence: In patients with intracranial atherosclerosis, [Xinyi Leng] et al. (/work/computational-fluid-dynamics-modeling-of-symptomatic) demonstrated that CFD-derived hemodynamic metrics—such as high blood flow velocity and shear strain rates at the blockage site—are far better predictors of a repeat stroke than the physical percentage of arterial narrowing. Meanwhile, [Deidre Anne De Silva] et al. (/work/limb-shaking-transient-ischemic-attack-with-distal-micro) utilized Transcranial Doppler to show that measuring cerebrovascular reactivity (the brain's functional ability to dilate vessels in response to stress) identifies the exhaustion of compensatory mechanisms, providing a precise window into stroke risk that structural imaging misses.
- Pulmonary Hypertension: In conditions like chronic thromboembolic pulmonary hypertension, [Alexander Chernyavskiy] et al. (/work/chronic-thromboembolic-pulmonary-hypertension-epidemiology) highlighted how ventilation-perfusion mismatch screening acts as a vital functional tool, mapping the distribution of inhaled air against blood flow to catch organized clots that standard anatomical imaging overlooks.
- Peripheral and Tissue Perfusion: Traditional RGB imaging systems can show the structure of the skin, but as [Leonardo Ayala] et al. (/work/the-spectral-perfusion-arm-clamping-dataset-spectralpaca) noted, they cannot monitor real-time physiological states like tissue oxygenation. Emerging techniques like Photoacoustic Imaging, reviewed by [Yide Zhang] et al. (/work/emerging-photoacoustic-imaging-techniques-for-peripheral), bridge this gap by providing functional metrics like regional blood volume and oxygen saturation alongside structural details.
Even outside the vascular tree, the gap persists. [Lorenz Biggemann] et al. (/work/assessment-of-esophagogastric-junction-morphology-by-dynamic) compared dynamic real-time MRI to pressure-based manometry at the esophagogastric junction, finding a stark discrepancy between pressure signatures and physical morphology—proving that relying on a single modality can cause clinicians to completely miss anatomical hernias.
Shifting the Treatment Paradigm
Why does resolving this gap matter so deeply? Because treating anatomy instead of function leads to suboptimal patient care. If a surgeon places a stent into an artery that is physically narrowed but functionally stable, the patient gains no clinical benefit but assumes all the risks of surgery.
This realization is reshaping how we define successful treatment. Historically, "complete revascularization" meant fixing every physical narrowing seen on a scan. However, [Rui Zhang] et al. (/paper/art-ded6e0c4e13d49cab1909771fbbb970d/art_ded6e0c4e13d49cab1909771fbbb970d) demonstrated the value of a functional definition of completeness. By utilizing an optimized scoring system that accounts for angiographically mild but hemodynamically significant lesions, the authors showed that purely anatomical definitions underestimate patient risk and miss residual blockages that continue to cause ischemia.
Ultimately, the field is moving toward a consensus where physiology acts as the ultimate gatekeeper. Large-scale clinical trials, such as the FLAVOUR II trial outlined by [Jinlong Zhang] et al. (/work/rationale-and-design-of-a-comparison-of-angiography), are directly comparing whether guiding interventions using high-resolution internal anatomy or computer-simulated blood flow physiology yields better long-term patient outcomes. By anchoring clinical decisions in the functional reality of blood flow rather than the visual illusion of structure, modern medicine is finally closing the gap.
Key Foundational Work
Clinical Myocardial Perfusion PET/CT (2007) — contributed 2 concept elements to this direction.
Basic fibroblast growth factor improves myocardial function in chronically ischemic porcine hearts. (1994) — contributed 1 concept elements to this direction.