🔬 optics

Multi-contrast laser endoscopy for in vivo gastrointestinal imaging

This paper introduces Multi-contrast Laser Endoscopy (MLE), a tunable widefield imaging platform that integrates multispectral reflectance, laser speckle contrast, and photometric stereo to significantly enhance tissue contrast and reveal complementary features like blood flow and topography during in vivo gastrointestinal examinations, demonstrating a three-fold improvement in contrast over standard white light and narrow band imaging.

Original authors: Taylor L. Bobrow, Mayank Golhar, Suchapa Arayakarnkul, Anthony A. Song, Saowanee Ngamruengphong, Nicholas J. Durr

Published 2026-03-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Taylor L. Bobrow, Mayank Golhar, Suchapa Arayakarnkul, Anthony A. Song, Saowanee Ngamruengphong, Nicholas J. Durr

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 you are a detective trying to find a tiny, camouflaged thief hiding in a crowded, dimly lit room. The thief (a precancerous polyp) looks almost exactly like the innocent bystanders (healthy tissue). If you only have a standard flashlight (current medical endoscopy), you might miss the thief entirely because the shadows and colors blend together.

This paper introduces a new, super-powered detective tool called Multi-contrast Laser Endoscopy (MLE). Instead of just one flashlight, MLE is a "Swiss Army Knife" of light that can change its color, texture, and direction in the blink of an eye to make the thief stand out.

Here is how it works, broken down into simple concepts:

1. The Problem: The "Invisible" Thief

Currently, doctors use a standard camera with a white light to look inside your stomach or colon. They look for changes in color or shape.

The Issue: Many dangerous growths are very subtle. They are the same color as the surrounding tissue, just slightly different in texture. It's like trying to find a white cat in a pile of white snow.
The Consequence: Doctors miss about 26% of these dangerous growths, which means they can't be removed early enough to prevent cancer.

2. The Solution: The "Magic Flashlight" (MLE)

The researchers took a standard medical camera (a colonoscope) and gave it a custom upgrade. They didn't replace the camera; they just swapped out the light bulb for a high-tech laser system. This system can switch between different "modes" instantly, like changing channels on a TV.

The MLE system has three special "superpowers":

Superpower A: The "X-Ray Vision" (Spectral Imaging)

How it works: Instead of just showing Red, Green, and Blue, this light shines eight specific colors (wavelengths) that human eyes can't see.
The Analogy: Imagine looking at a painting. A normal light shows you the paint. This special light shows you the chemistry of the paint. It can tell the difference between healthy blood and sick blood based on how they absorb light.
The Result: It creates a "false color" map where the sick tissue turns a bright, obvious red, while healthy tissue stays blue. It's like putting on sunglasses that make the bad guys glow neon.

Superpower B: The "Flow Meter" (Laser Speckle)

How it works: It uses a laser that creates a grainy, shimmering pattern (called "speckle") on the tissue. If the tissue is moving (because blood is flowing), the shimmer changes.
The Analogy: Think of looking at a calm pond. If you throw a stone in, the ripples tell you something is moving. This laser looks at the "ripples" of blood flow.
The Result: It can map exactly where blood is flowing. Surprisingly, the study found that some polyps have less blood flow than healthy tissue. The MLE highlights these "quiet zones" so the doctor knows, "Hey, that spot is different."

Superpower C: The "Shadow Play" (Photometric Stereo)

How it works: The camera has three tiny lights around it. The system turns them on and off one by one very quickly.
The Analogy: Imagine looking at a bump on a wall. If the light comes from the left, the bump casts a shadow to the right. If the light comes from the right, the shadow flips. By watching how the shadows move, you can tell exactly how bumpy the surface is, even if the bumps are tiny.
The Result: It creates a 3D "topographic map" of the inside of your gut. It reveals tiny bumps and ridges that look flat under normal white light.

3. The Real-World Test

The team tested this on 31 real polyps in patients during standard colonoscopies.

The Comparison: They compared the new MLE images against the standard white light and the current "improved" light (Narrow Band Imaging).
The Score: The MLE images were three times clearer and showed five times more color difference between healthy and sick tissue.
The Best Part: It didn't slow down the doctor. The system works in the background. The doctor sees the normal video on their screen, but the computer is secretly analyzing the data and highlighting the trouble spots instantly.

4. Why This Matters

Think of this technology as upgrading from a black-and-white TV to a 4K HDR screen with night vision.

Current Tech: You might miss a thief because they are wearing a gray suit in a gray room.
MLE Tech: The system instantly paints the thief in bright orange and highlights their footprints.

The researchers proved that this system is safe, fits into existing hospital equipment, and can be used right now. By making the invisible visible, it promises to help doctors catch cancer much earlier, saving lives by ensuring fewer "thieves" slip through the net.

1. Problem Statement

Standard high-definition white light endoscopy (WLE) is the clinical gold standard for gastrointestinal (GI) screening but suffers from limited contrast. Many precancerous lesions (e.g., adenomas, dysplasia in Barrett's esophagus) exhibit subtle differences in color, texture, and shape compared to healthy tissue, leading to high miss rates (e.g., ~26% for adenomas in colonoscopy).

Limitations of Current Solutions:
- Narrow Band Imaging (NBI): Enhances vascular contrast but offers minimal improvement in adenoma detection rates (<3%) and fails to improve detection in the upper GI tract significantly.
- Microscopic/Probe-based Imaging (e.g., Confocal, OCT): Provides cellular-level detail but has a very small field of view (FoV), making them impractical for widefield surveillance.
- Exogenous Dyes: Require administration of fluorescent agents, introducing cost, safety concerns, and regulatory hurdles.
- Technical Barriers: Previous attempts to enhance widefield imaging often required replacing the endoscope's optics (reducing resolution) or using the instrument channel for probes (lengthening procedure time), hindering clinical adoption.

2. Methodology: Multi-contrast Laser Endoscopy (MLE)

The authors developed MLE, a research platform that retrofits a standard clinical colonoscope to enable multimodal, widefield imaging without compromising standard-of-care capabilities.

System Architecture & Hardware:

Retrofitted Colonoscope: The internal fiber optic light guide of an Olympus CF-HQ190L colonoscope was replaced with a custom fiber bundle. This bundle integrates three MLE fiber bundles alongside the clinical illumination fibers.
- Key Innovation: The MLE fibers are randomly mixed with clinical fibers at the distal tip, allowing the system to toggle between clinical WLE/NBI and MLE modes in <1 second without removing the scope.
Multi-contrast Illumination Source: A custom cart-mounted unit containing 15 laser diodes:
- 8 Narrowband Wavelengths (406–657 nm): For multispectral diffuse reflectance imaging.
- High-Coherence Laser (639 nm): For Laser Speckle Contrast Imaging (LSCI) to measure blood flow.
- Directional Illumination: Three point sources (RGB units) that can be toggled independently for Photometric Stereo Endoscopy (PSE).
Control & Processing: A custom light modulation controller synchronizes laser pulse widths with the colonoscope's CCD sensor frame rate (29.97 fps). A workstation handles real-time data acquisition, auto-exposure, and image processing (GPU-accelerated).

Imaging Modalities:

Multispectral Diffuse Reflectance: Captures reflectance at 8 wavelengths to map tissue chromophores (hemoglobin) and calculate total oxygen saturation ( $StO_2$ ).
Laser Speckle Contrast Imaging (LSCI): Uses high-coherence light to generate speckle patterns. Motion (blood flow) reduces speckle contrast, allowing for the quantification of perfusion.
Photometric Stereo Endoscopy (PSE): Sequentially toggles directional light sources to computationally reconstruct surface topography (height maps) and surface normals, enhancing the visibility of subtle shape changes.

Validation Strategy:

Benchtop: Validated using color calibration charts, microfluidic flow phantoms (simulating blood flow), and 3D-printed silicone colon models with known topography.
In Vivo Pilot Study: Conducted on 20 patients undergoing standard colonoscopy at Johns Hopkins Hospital. 31 precancerous polyps were imaged using both clinical WLE/NBI and MLE modalities.

3. Key Contributions

Clinical Integration: Demonstrated a method to add advanced optical imaging capabilities to a standard clinical endoscope without sacrificing resolution, field of view, or workflow efficiency.
Multimodal Fusion: Successfully combined three distinct contrast mechanisms (spectral, flow, and topographic) into a single platform.
Real-time Capability: Developed a system capable of video-rate acquisition and processing, allowing for seamless integration into clinical workflows.
Quantitative Biomarkers: Enabled the real-time mapping of functional biomarkers (oxygen saturation, blood flow) and structural features (surface height) in vivo.

4. Results

Benchtop Validation:

Spectral Accuracy: MLE measurements of a color chart showed a mean absolute reflectance error of $0.04 \pm 0.03$ compared to a reference spectrometer.
Oxygen Saturation: In a finger occlusion trial, MLE detected a drop in $StO_2$ from a baseline of 61.5% to 33.0% during ischemia, returning to baseline upon release, demonstrating high sensitivity to functional changes.
Flow Sensitivity: In microfluidic phantoms, LSCI successfully visualized flow channels indistinguishable in WLE. Flow contrast increased significantly with velocity and exposure time.
Topography: PSE reconstructed surface height maps of a silicone phantom with a mean absolute error of 0.11, accurately capturing 3D features invisible in standard WLE.

In Vivo Clinical Study (31 Polyps):

Contrast Enhancement: MLE provided an approximate three-fold improvement in RMS contrast and a five-fold improvement in color difference (CIEDE2000) compared to WLE and NBI.
- Spectral Enhancement (SE): By optimizing spectral transmission curves, the color difference between polyp and normal tissue increased from 3.1 (WLE) to 16.7 (SE).
Functional Imaging:
- Blood Flow: Polyps showed significantly lower blood flow compared to normal tissue ( $p=0.046$ ).
- Oxygen Saturation: While $StO_2$ differences were not statistically significant in this small cohort, the system successfully mapped the metric.
Topography: Surface height reconstruction revealed distinct topographical changes at lesion margins, offering a two-fold improvement in contrast over WLE hue/saturation channels ( $p < 0.01$ ).

5. Significance and Future Outlook

Improved Detection Potential: The significant increase in contrast across multiple modalities suggests MLE could reduce adenoma miss rates, potentially improving the efficacy of CRC screening.
Label-Free & Non-Invasive: The system relies on endogenous tissue properties (hemoglobin, structure) rather than exogenous dyes, making it safer and more cost-effective for widespread use.
Platform for AI: The rich, multi-dimensional data (spectral, flow, topographic) acquired by MLE provides an ideal dataset for training deep learning algorithms to improve automated lesion detection and classification.
Scalability: The modular design allows for future upgrades, such as adding UV wavelengths for autofluorescence or Near-Infrared (NIR) for molecular imaging.

Limitations:

The study was a pilot; it did not measure the actual clinical adenoma detection rate (ADR) improvement, which requires a larger randomized controlled trial.
The system relies on post-processing image registration to align frames acquired with different illumination modes, which can be challenged by rapid scope movement.
Pixel values are subject to the clinical video processor's internal non-linear transformations (gamma correction), though the authors note this can be mitigated in future iterations.

In conclusion, Multi-contrast Laser Endoscopy represents a significant step forward in GI imaging, bridging the gap between high-resolution clinical endoscopy and advanced optical sensing, offering a promising tool to enhance the detection of gastrointestinal diseases.