OptoCENTAL: a standardised, bench-testing platform based on phantoms for validating optical systems aimed at clinical monitoring of the placenta

The paper introduces OptoCENTAL, a standardized bench-testing platform utilizing diverse optical phantoms to validate and compare various optical imaging systems for the clinical monitoring of the human placenta, thereby facilitating the translation of these technologies into hospital settings and commercial markets.

The Gist

Imagine you are trying to listen to a baby's heartbeat, but the baby is hidden deep inside a complex, multi-layered house. The house has thick walls (skin and fat), a sturdy floor (muscle), and the baby is in a special room (the placenta) at the bottom. Now, imagine you are trying to use a flashlight to see inside this house, but the walls are foggy and the light gets scattered everywhere.

This is the challenge doctors face when trying to monitor a placenta during pregnancy using light-based technology (like Near-Infrared Spectroscopy or NIRS). They need to know if the baby is getting enough oxygen, but the light has to travel through layers of the mother's body to get there and come back.

The Problem:
Until now, there was no standard "test drive" for these new light-monitoring devices. It was like buying a new car without ever being able to test it on a track to see if the brakes work or if the engine can handle a hill. Manufacturers couldn't be sure their devices could actually see deep enough into the belly to give accurate readings, and doctors couldn't trust the data.

The Solution: OptoCENTAL
The researchers at University College London (UCL) built OptoCENTAL. Think of this as a "Simulator Track" or a "Training Ground" specifically designed to test these light-monitoring devices before they ever touch a real patient.

Instead of testing on real pregnant women (which is risky and hard to control), they built a series of "fake" bodies, called phantoms, that act exactly like real human tissue but are safe and repeatable.

Here is how their "Training Track" works, broken down into four simple levels:

Level 0: The Virtual Reality Simulator (Digital)

Before building anything physical, they use a super-computer to create a digital twin of a pregnant belly.

• The Analogy: Imagine a video game where you can build a house with perfectly measured walls. They used real ultrasound scans from 268 women to build this digital house.
• The Test: They drop a virtual flashlight into this digital house to see exactly how the light bounces around. This tells them, "Okay, theoretically, a device needs to be this strong to see the placenta."

Level 1: The "Stress Test" (Solid Phantoms)

Next, they build solid blocks that look like human tissue (made of resin, dye, and pigment).

• The Analogy: Think of these as calibration weights for a scale. You put a 1kg weight on a scale to see if it reads 1kg.
• The Test: They shine the light through these solid blocks to check the device's basic health:
  • Is the signal loud enough? (Signal-to-Noise Ratio)
  • Is the device steady, or does it jitter? (Stability)
  • If they make the block darker, does the device notice the change accurately? (Linearity)

Level 2: The "Pulse Simulator" (Liquid Phantom)

This is the most exciting part. They created a liquid mixture in a tank that acts like a living, breathing organ.

• The Analogy: Imagine a smoothie made of water, milk (to scatter light), and blood. They can pump oxygen into it to make the blood "happy" (oxygenated) or suck the oxygen out to make it "sad" (deoxygenated). They even add yeast, which eats the oxygen, to simulate the body's metabolism.
• The Test: They watch the device as the liquid changes from "happy" to "sad" and back again. This proves the device can actually track the changes in oxygen and metabolism in real-time, just like it would in a real placenta.

Level 3: The "Obstacle Course" (Hybrid Phantom)

Finally, they combine the solid blocks and the liquid tank. They put the "fake skin and muscle" (solid blocks) on top of the "fake placenta" (liquid tank).

• The Analogy: This is like putting thick winter coats over a person and asking the device to still hear their heartbeat.
• The Test: This is the ultimate challenge. The light has to penetrate the solid layers, hit the liquid layer, and bounce back. If the device can do this, it proves it has the "depth vision" needed to see a real placenta inside a real mother's belly.

Why Does This Matter?

The researchers tested four different devices on this track, from simple wearable sensors to complex laser systems.

• The Result: They found that some devices were great at seeing the surface but failed to see deep down. Others were accurate but too shaky.
• The Benefit: Now, instead of guessing, engineers can say, "Our device passed Level 3!" and doctors can say, "I trust this device because it survived the OptoCENTAL track."

In Summary:
OptoCENTAL is the gold standard for testing placental monitors. It turns a risky, unpredictable medical challenge into a controlled, repeatable science. It ensures that when these devices finally reach the hospital, they are ready to save lives by giving doctors a clear, real-time window into the health of the baby and the placenta.

Technical Summary

1. Problem Statement

The clinical monitoring of the human placenta during gestation is critical for assessing fetal health and predicting adverse outcomes (e.g., hypoxia, metabolic disruption). While Near-Infrared Spectroscopy (NIRS) and Diffuse Optical Tomography (DOT) offer promising, non-invasive, and wearable solutions for this purpose, their translation to clinical practice faces significant hurdles:

• Depth and Geometry Challenges: The placenta is located 1–4 cm deep beneath the skin, separated by variable layers of muscle, adipose tissue, and potentially bone or fluid. This geometry attenuates the optical signal and complicates the separation of placental signals from overlying tissue.
• Lack of Standardization: Unlike brain monitoring (which has protocols like MEDPHOT and NeurOPT), there is no standardized testing procedure to validate the performance, depth sensitivity, and accuracy of optical devices specifically designed for placental monitoring.
• Validation Gap: Without a rigorous, comparable bench-testing framework, it is difficult to determine if a specific photonics device can reliably recover physiological biomarkers (HbO₂, HHb, and Cytochrome-c-oxidase/CCO) from the placenta in vivo.

2. Methodology: The OptoCENTAL Platform

The authors developed OptoCENTAL, a standardized platform comprising four distinct protocols using digital, solid, liquid, and hybrid optical phantoms. The platform is designed to be completed within one week and allows for cross-comparison of diverse optical systems (CW-NIRS, TD-NIRS, broadband, and wearable).

Protocol 0: Computational Depth Sensitivity (Digital Phantom)

• Method: Uses the Monte Carlo eXtreme (MCX) simulator to model photon propagation.
• Phantom: A realistic, multi-layered digital twin of the human abdomen (skin, muscle, adipose, placenta) derived from ultrasound scans of 268 pregnant women.
• Function: Calculates the sensitivity profile of specific optical configurations (Source-Detector Separations) to the placenta layer, accounting for individual anatomical variations and skin pigmentation (Fitzpatrick scale).

Protocol 1: Basic Instrument Assessment (Solid Phantoms)

• Method: Tests fundamental hardware performance using custom-made, homogeneous solid phantoms (polyester resin with NIR dye and pigment).
• Metrics:
  • Signal-to-Noise Ratio (SNR): Signal quality vs. baseline noise.
  • Noise/Variance: Sensitivity to statistical fluctuations (Coefficient of Variation).
  • Linearity: Response to varying absorption coefficients.
  • Stability: Signal drift over 1 hour.
  • Repeatability: Consistency across three consecutive days.

Protocol 2: Physiological Monitoring Assessment (Homogeneous Liquid Phantom)

• Method: Uses a dynamic liquid phantom (PBS, Intralipid, blood, and yeast) housed in a waterproof container.
• Function: Simulates physiological changes by controlling oxygenation (via O₂/N₂ bubbling) and metabolism (via yeast consumption).
• Key Goal: Validates the system's ability to reconstruct dynamic changes in HbO₂, HHb, and oxidized Cytochrome-c-oxidase (oxCCO). Crucially, it tests for crosstalk between metabolic and oxygenation signals by comparing deoxygenation induced by N₂ (no metabolic change) vs. yeast (metabolic change).

Protocol 3: Depth Sensitivity Assessment (Hybrid Multilayer Phantom)

• Method: Combines the liquid phantom (representing the placenta) with solid phantom layers (representing skin, adipose, and muscle) placed on top.
• Function: Mimics the actual in vivo geometry (e.g., 1.5 cm total overlayer thickness). It tests the device's ability to penetrate the overlying tissue, detect physiological changes in the deep liquid layer, and accurately reconstruct them without significant attenuation or distortion.

3. Key Contributions

• First-of-its-Kind Standardization: OptoCENTAL is the first comprehensive framework specifically designed to validate optical systems for placental monitoring, bridging the gap between technical development and clinical certification.
• Multi-Modal Compatibility: The platform is flexible enough to test Continuous-Wave (CW), Time-Domain (TD), and Broadband NIRS systems, including wearable and wireless devices.
• Realistic Biomarker Simulation: Unlike simple dye-based phantoms, the liquid protocol uses real blood and yeast to generate authentic optical signatures for hemoglobin and CCO, allowing for rigorous crosstalk analysis.
• Cost-Effectiveness: The solid phantoms cost less than £20 per batch, and the liquid components use readily available commercial materials, making the platform accessible for widespread adoption.

4. Results and Validation

The authors applied OptoCENTAL to four NIRS instruments developed at UCL: microCYRIL (broadband CW), FetalSenseM V1.5 & V2 (multi-wavelength CW), and MAESTROS II (TD-NIRS).

• Protocol 0 (Simulation): Successfully mapped sensitivity profiles, showing how different Source-Detector Separations (SDS) and time-gating (for TD systems) affect the probability of photons reaching the placenta.
• Protocol 1 (Hardware):
  • Solid phantoms showed high reproducibility (measured properties within 10–15% of theoretical values).
  • Devices demonstrated varying SNR and linearity; for instance, FetalSenseM V2 showed higher linearity (R ≈ -0.98) than V1.5.
  • Stability tests confirmed devices could maintain readings within ±3% over an hour.
• Protocol 2 (Physiology & Crosstalk):
  • microCYRIL (Broadband) showed negligible crosstalk (<0.05 µM) between oxCCO and hemoglobin signals.
  • FetalSenseM V2 (Multi-wavelength CW) exhibited significant crosstalk (mean error -1.66 µM) when metabolic changes were not present, highlighting the limitations of CW systems without time-resolved data.
  • MAESTROS II (TD) demonstrated superior accuracy in absolute oxygen saturation (StO₂) reconstruction (98.34% max) compared to CW systems, which underestimated values by up to 35%.
• Protocol 3 (Depth):
  • Both FetalSenseM devices successfully detected dynamic changes through 1.5 cm of overlying tissue.
  • The protocol revealed that longer SDS (e.g., 5 cm) led to signal underestimation compared to shorter SDS (3 cm), likely due to pathlength differences not fully accounted for by standard algorithms. This highlights the need for protocol-specific corrections.

5. Significance

• Clinical Translation: OptoCENTAL provides the necessary "gatekeeping" validation to ensure that optical devices are safe, accurate, and reliable before being deployed in neonatal units or maternity wards.
• Regulatory Pathway: By establishing a standardized benchmark, the platform facilitates future certification and commercialization of placental monitoring technologies.
• Algorithm Development: The data generated (especially from Protocol 3) allows researchers to refine reconstruction algorithms (e.g., Differential Pathlength Factors) to better account for the complex geometry of the pregnant abdomen.
• Future Impact: The platform supports the development of wearable, continuous monitoring tools that could predict delivery complications (like preeclampsia or fetal distress) in real-time, potentially reducing perinatal mortality and morbidity.