We Captured Speckle Dynamics with a Disposable Bronchoscope
Here’s Why That’s a Big Deal
In medical imaging, there’s a constant tension between sophistication and accessibility. The more advanced the technique, the more specialized — and expensive — the equipment tends to be. But what if we could flip that logic on its head?
In our latest preprint, Turning a Disposable Bronchoscope into a Dynamic Speckle Imaging Tool: Yes, It Works, we show that even a single-use, low-cost bronchoscope — the kind commonly used in hospitals to inspect airways — can be transformed into a functional imaging device based on dynamic speckle imaging, capable of visualizing blood flow in tissues.
Wait… What is Speckle Imaging?
When you illuminate by a coherent light (like that from a laser) on a surface, you get a grainy, shimmering pattern called a speckle. If the tissue is alive and perfused with blood, that pattern subtly fluctuates over time. Analyzing those fluctuations allows us to infer microvascular activity, making speckle imaging a non-invasive technique for assessing blood flow dynamics.
Speckle imaging is already well-established in experimental optics labs. But it typically requires high-end cameras, stable illumination, and precise control over acquisition parameters. That’s why it’s never been done —to our knowledge — with something as “unsophisticated” as a disposable bronchoscope.
Why Use a Bronchoscope?
Bronchoscopes are designed to explore the lungs and airways. They’re small, flexible, and already equipped with a camera and light source. Most importantly, they’re used every day in clinical settings. So if we can repurpose them for functional imaging, we could bring advanced diagnostics right into standard medical workflows. That’s exactly what we set out to test.
Turning Trash into Tech
To turn a bronchoscope into a speckle imaging device, we had to get a bit creative.
First, we introduced a coherent laser source into the system. Since the bronchoscope comes with its own white LED illumination — designed for standard visual inspection — we disabled that and instead inserted an optical fiber through the instrument channel. This allowed us to deliver coherent laser light directly to the tissue, a critical requirement for speckle imaging.
Next came the real challenge: processing the video data. Commercial bronchoscopes are not designed for scientific image processing. We had no access to the raw sensor signal; what we got was compressed video, processed by the internal electronics. That’s far from ideal — speckle dynamics are highly sensitive to temporal artifacts and compression losses.
So we had to improvise. In this first experiment, we extracted frames using simple screen captures of the video stream. Yes, really. It’s not elegant, but it worked.
Despite the noise, compression, and optical limitations, we were able to detect temporal speckle fluctuations in the image sequence. And from those fluctuations, we extracted a functional signal: evidence of vascular motion beneath the tissue surface.
In one of our early recordings, we imaged a small skin structure — and for the curious: don’t worry, it’s just a nevus. 😉
This is a true proof of concept: a demonstration that even with limited, noisy, and repurposed clinical hardware, it is possible to extract rich physiological information, provided the right optical setup and post-processing strategy.
To validate what we were seeing, we compared our speckle image from the bronchoscope setup with data obtained using our reference system — the Vasculoscope (www.itae.fr). This is a dedicated speckle imaging device, designed for high-resolution vascular mapping under controlled illumination and with access to raw image sequences, and also an orthogonal polarimetric filter.
The comparison was encouraging: despite the lower resolution, unknown frame rate, differences in imaging geometry, absence of the polarimetric filter, and the lossy compression inherent to the bronchoscope workflow, we were able to identify shared vascular structures across both modalities. Remarkably, the speckle patterns exhibited similar spatial organization, reinforcing the physiological relevance of the signal extracted from the bronchoscope.
This result reinforced our confidence: even in a severely constrained acquisition pipeline, the physiological signal persists — noisy, but visible. That’s a strong indication that functional endoscopy using commodity devices isn’t just a gimmick — it’s a real, emerging possibility.
What This Means for the Future of Endoscopy
This work opens the way for a future in which functional endoscopy — the ability to visualise not only anatomical structures but also tissue dynamics in real time — could be implemented using instruments already present in clinical environments. Of course, certain adaptations will be necessary: better access to raw video streams, better lighting control, and the integration of basic processing protocols adapted to speckle analysis. But above all, all this can be done within the ecosystem of inexpensive and widely deployed medical devices.
We’re not advocating a complete redesign, but smart, minimal interventions that could bring advanced physiological imaging into routine clinical practice, affordably and scalably.
🧠 Curious to dive deeper? Check out the full preprint on arXiv:
📄 https://arxiv.org/abs/2504.21469
