⚡ electrical engineering

Basis for a hands free blood flow measurement with automated vessel focus

This paper presents a novel hands-free system featuring a specialized ultrasonic probe and automated software that accurately identifies blood vessels and quantifies blood flow during cardiopulmonary resuscitation (CPR), achieving high correlation with ground truth measurements to guide first responders in optimizing chest compression quality.

Original authors: Reinhard Fuchs, Nathalie Sumrah, Johannes Schwerdt, Michael Unger, Georg Stachel, Michael Schultz, Karsten Lenk, Thomas Neumuth

Published 2026-06-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Reinhard Fuchs, Nathalie Sumrah, Johannes Schwerdt, Michael Unger, Georg Stachel, Michael Schultz, Karsten Lenk, Thomas Neumuth

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a person suffering a cardiac arrest. Their heart has stopped, so emergency responders are performing CPR—rhythmic chest compressions—to manually pump blood to the brain. The problem is that right now, no one knows if those compressions are actually working well enough to keep the brain alive. It's like trying to fix a car engine in the dark; you can hear the hammering, but you can't see if the pistons are firing.

This paper presents a new "flashlight" for that situation: a hands-free, automated device that uses sound waves to measure blood flow in the neck during CPR.

Here is how the system works, broken down into simple concepts:

1. The Problem with Current Tools

Currently, doctors rely on guidelines for how hard and fast to push on the chest, but they can't actually see the blood moving. Existing ultrasound machines are like high-end cameras that require a professional photographer (a skilled sonographer) to hold them steady, aim them perfectly, and interpret the pictures. In a chaotic emergency, you don't have time for a photographer, and you can't hold a camera steady while your hands are busy pumping.

2. The New "Smart Probe"

The researchers built a special probe that acts like a smart, self-adjusting flashlight.

The Hardware: Instead of one camera lens, this probe has three tiny "ears" (ultrasound sensors) made of special ceramic. These ears are angled differently (one straight, one tilted left, one tilted right).
The Goal: To find the common carotid artery (the main blood highway to the brain) and measure how fast the blood is moving without anyone holding the device.

3. The Three-Step "Search Party"

The device doesn't just guess; it runs a three-step automated search to find the blood flow, similar to how you might tune a radio to find a clear station.

Step 1: The Wide Net (Initial Mode)
The device first casts a "wide net" of sound waves to see if it's even over a blood vessel. It's like casting a large fishing net to see if there are any fish in the water at all. If the sensors don't hear the rhythmic "whoosh" of blood, it knows it's in the wrong spot.
Step 2: The Fine-Tune (Iterative Mode)
Once it finds a signal, it switches to a "fine-tune" mode. It starts taking tiny, step-by-step measurements at different depths, like a diver slowly descending to find the perfect depth to see a fish. It calculates a "score" based on how rhythmic and strong the signal is. It keeps moving until it finds the exact depth where the blood flow looks the clearest.
Step 3: The Calculation (Monitoring Mode)
Now that it has found the perfect spot and the perfect angle, it starts the real work. Because the device has three sensors at different angles, it can use math to figure out the exact speed of the blood without needing to know the exact angle of the vessel (which is usually impossible to see). It's like triangulating a location using three cell towers to get a precise GPS fix.

4. The "Phantom" Test

Since they couldn't test this on real people in an emergency yet, they built a "fake neck" (a phantom).

They made a gel neck with a plastic tube inside representing the artery.
They pumped a special fluid through the tube at a steady rhythm (simulating a heartbeat).
They compared their new device's readings against a high-precision flow meter (the "gold standard" truth).

5. The Results

The test was a success.

Finding the Spot: The device correctly identified when it was over the "vessel" and when it wasn't.
The Math: When the device calculated the blood flow, it matched the "gold standard" meter almost perfectly. The paper reports a correlation of 0.98 (where 1.0 is a perfect match).
The Error: The average difference between their device and the real meter was very small (about 3.84 ml/s), which the authors consider a strong result for a prototype.

What the Paper Claims (and What It Doesn't)

The paper claims they have built a foundation for a device that can automatically find the artery and measure blood flow without a human operator. They proved this works in a lab setting with a fake neck.

Crucially, the paper does not claim:

That this device is ready for hospitals or ambulances today.
That it has been tested on real humans.
That it will definitely save lives (though they hope it will).

The authors describe this as a "basis for future developments." They envision a future where this technology is miniaturized and attached to a neck brace, allowing first responders to get instant feedback on whether their CPR is actually moving blood, all without lifting a finger. For now, it is a promising prototype that has passed its first major laboratory exam.

Technical Summary: Basis for a Hands-Free Blood Flow Measurement with Automated Vessel Focus

Problem Statement
Cardiopulmonary resuscitation (CPR) is critical for maintaining oxygen supply during cardiac arrest, yet the quality of chest compressions and the resulting blood flow are currently unquantifiable in real-time. While guidelines exist for compression frequency and depth, they do not guarantee adequate cerebral blood flow. Invasive monitoring is infeasible during resuscitation, and existing non-invasive ultrasound methods require significant manual expertise, time, and hands-on operation, which impedes their use in emergency scenarios. Current commercial solutions lack robust, hands-free mounting options for the neck during CPR and often rely on continuous wave (CW) Doppler, which cannot isolate specific sample volumes (SV) or account for vessel angle changes, leading to potential quantification errors. There is a lack of automated, non-invasive tools to measure common carotid artery (CCA) blood flow to guide first responders and provide prognostic data.

Methodology
The authors propose a system comprising a custom sensor probe and automated analysis software designed for Pulsed Wave (PW) Doppler measurements.

Hardware: The core innovation is a sensor probe containing three piezo-ceramic elements with non-symmetrically angled orientations (25°, 0°, and -12° relative to the probe surface normal). The probe casing is polyactide, with polyetheretherketon (PEEK) filling the sound path. The system connects to a computing unit via three ultrasound modules, allowing synchronized, independent triggering of pulses for each ceramic.
Experimental Setup: Evaluation was conducted using a neck phantom containing an 8 mm diameter vessel filled with Doppler fluid, driven by a mechanical pump at 60 BPM. A reference flow sensor (VISION 1005 2F66) provided ground truth data.
Measurement Protocol: The system operates in three sequential modes:
1. Initial Mode: Uses a large sample volume (SV) with the 25° transducer to rapidly scan for vessel coverage. If no periodic flow is detected, the position is deemed unsuitable.
2. Iterative Mode: Once a viable position is found, the system performs a depth-dependent scan using smaller SVs (2 µs duration, stepped by 2 µs). All three ceramics act as receivers while one acts as a transmitter, cycling through all combinations. A scoring algorithm ( $P_{Score}$ ) evaluates the periodicity and power of the Doppler signal to identify the optimal depth and transmitter-receiver pair.
3. Monitoring Mode: Using the optimal configuration identified in the iterative mode, the system calculates blood flow velocity. To address the unknown Doppler angle, the system employs a "Doppler angle substitution" method. By comparing frequency shifts between pairs of transducers with known fixed angles, the system mathematically derives the flow velocity without needing to know the exact vessel angle.
Data Analysis: Signals were processed using MATLAB to generate spectrograms and calculate envelopes via the Modified Geometric Method (MGM). Flow velocity was converted to volume flow using the phantom's vessel radius. Performance was evaluated using the coefficient of determination ( $R^2$ ), Pearson correlation coefficient, and Root Mean Square Error (RMSE) against the reference flow sensor.

Key Contributions

Automated Vessel Identification: The paper presents a method to automatically identify vessel coverage and optimal depth without manual parameter adjustment, utilizing a multi-angle transducer array.
Angle-Independent Flow Calculation: A novel mathematical approach is introduced to substitute the unknown Doppler angle by leveraging the fixed geometric relationship between the three transducers, enabling flow quantification without precise alignment knowledge.
Hands-Free Prototype: The development of a specific probe design intended for hands-free attachment, addressing the limitations of current portable ultrasound devices that require manual holding.

Results

Vessel Detection: The initial mode successfully distinguished between positions with and without vessel coverage. Positions 1 and 3 showed no significant reflections, while Position 2 showed strong correlation between the Doppler envelope and the reference flow.
Depth Optimization: The iterative mode successfully identified the vessel depth. The maximum power score occurred at an SV depth of 15.75 mm (12 mm after subtracting probe internal distance), which aligned with the vessel center observed in B-Mode imaging.
Flow Quantification: In the monitoring mode, the calculated flow curves demonstrated high agreement with the reference sensor.
- Correlation: The coefficient of determination ( $R^2$ ) ranged from 0.96 to 0.98 (average 0.97).
- Correlation Coefficient: Values ranged from 0.98 to 0.99 (average 0.98).
- Error: The average Root Mean Square Error (RMSE) was 3.84 ml/s.
Limitations Observed: The authors noted discrepancies in flow curve shapes (specifically during flow acceleration/deceleration) attributed to the "Windkessel" effect in the elastic phantom tubing compared to the rigid reference sensor, rather than solely measurement error.

Significance and Claims
The paper claims to establish a "basis for future developments" toward fully automated vessel identification and blood flow calculation during CPR. The authors state that the proposed hardware and software solutions successfully demonstrated the feasibility of:

Identifying valuable measurement positions for carotid PW Doppler.
Performing automatic parametrization during periodic flow activity.
Quantifying measured flow within a small sample volume.

The authors emphasize that while the current system is a laboratory demonstrator, the method distinguishes itself from existing solutions by enabling detailed automatic analysis of vessel location and direction through fixed, non-symmetric ceramic angles. They suggest that a miniaturized version integrated into a neck stabilizer could provide vital, real-time feedback on CPR efficiency, potentially improving therapy for cardiac arrest patients. The work is presented as a foundational step, acknowledging that further clinical validation and hardware miniaturization are required before clinical deployment.