Multispectral snapshot imaging of skin microcirculatory hemoglobin oxygen saturation using artificial neural networks trained on in vivo data

Optimizing clinical O2 saturation mapping using hyperspectral imaging and diffuse reflectance spectroscopy in the context of epinephrine injection

Clinical determination of oxygen saturation (sO2) in patients is commonly performed via non-invasive optical techniques. However, reliance on a few wavelengths and some form of pre-determined calibration introduces limits to how these methods can be used. One example involves the assessment of sO2 after injection of local anesthetic using epinephrine, where some controversy exists around the time it takes for the epinephrine to have an effect. This is likely caused by a change in the tissue environment not accounted for by standard calibrated instruments and conventional analysis techniques. The present study aims to account for this changing environment by acquiring absorption spectra using hyperspectral imaging (HSI) and diffuse reflectance spectroscopy (DRS) before, during, and after the injection of local anesthesia containing epinephrine in human volunteers. We demonstrate the need to account for multiple absorbing species when applying linear spectral unmixing in order to obtain more clinically relevant sO2 values. In particular, we demonstrate how the inclusion of water absorption greatly affects the rate at which sO2 seemingly drops, which in turn sheds light on the current debate regarding the time required for local anesthesia with epinephrine to have an effect. In general, this work provides important insight into how spectral analysis methods need to be adapted to specific clinical scenarios to more accurately assess sO2.

Multispectral indices for real-time and non-invasive tissue ischemia monitoring using snapshot cameras

An adequate supply of oxygen-rich blood is vital to maintain cell homeostasis, cellular metabolism, and overall tissue health. While classical methods of measuring tissue ischemia are often invasive, localized and require skin contact or contrast agents, spectral imaging shows promise as a non-invasive, wide field, and contrast-free approach. We evaluate three novel reflectance-based spectral indices from the 460 - 840 nm spectral range. With the aim of enabling real time visualization of tissue ischemia, information is extracted from only 2-3 spectral bands. Video-rate spectral data was acquired from arm occlusion experiments in 27 healthy volunteers. The performance of the indices was evaluated against binary Support Vector Machine (SVM) classification of healthy versus ischemic skin tissue, two other indices from literature, and tissue oxygenation estimated using spectral unmixing. Robustness was tested by evaluating these under various lighting conditions and on both the dorsal and palmar sides of the hand. A novel index with real-time capabilities using reflectance information only from 547 nm and 556 nm achieves an average classification accuracy of 88.48, compared to 92.65 using an SVM trained on all available wavelengths. Furthermore, the index has a higher accuracy compared to reference methods and its time dynamics compare well against the expected clinical responses. This holds promise for robust real-time detection of tissue ischemia, possibly contributing to improved patient care and clinical outcomes.

Switchable LED-based laparoscopic multispectral system for rapid high-resolution perfusion imaging

Significance

Multispectral imaging (MSI) is an approach for real-time, quantitative, and non-invasive tissue perfusion measurements. Current laparoscopic systems based on mosaic sensors or filter wheels lack high spatial resolution or acceptable frame rates.

Aim

To develop a laparoscopic system for MSI-based color video and tissue perfusion imaging during gastrointestinal surgery without compromising spatial or temporal resolution.

Approach

The system was built with 14 switchable light-emitting diodes in the visible and near-infrared spectral range, a 4K image sensor, and a 10 mm laparoscope. Illumination patterns were created for tissue oxygenation and hemoglobin content monitoring. The system was calibrated to a clinically approved laparoscopic hyperspectral system using linear regression models and evaluated in an occlusion study with 36 volunteers.

Results

The root mean squared errors between the MSI and reference system were 0.073 for hemoglobin content, 0.039 for oxygenation in deeper tissue layers, and 0.093 for superficial oxygenation. The spatial resolution at a working distance of 45 mm was 156    μ m . The effective frame rate was 20 fps.

Conclusions

High-resolution perfusion monitoring was successfully achieved. Hardware optimizations will increase the frame rate. Parameter optimizations through alternative illumination patterns, regression, or assumed tissue models are planned. Intraoperative measurements must confirm the suitability during surgery.

Oxygen saturation mapping during reconstructive surgery of human forehead flaps with hyperspectral imaging and spectral unmixing

BACKGROUND

Optical spectroscopy is commonly used clinically to monitor oxygen saturation in tissue. The most commonly employed technique is pulse oximetry, which provides a point measurement of the arterial oxygen saturation and is commonly used for monitoring systemic hemodynamics, e.g. during anesthesia. Hyperspectral imaging (HSI) is an emerging technology that enables spatially resolved mapping of oxygen saturation in tissue (sO2), but needs to be further developed before implemented in clinical practice. The aim of this study is to demonstrate the applicability of HSI for mapping the sO2 in reconstructive surgery and demonstrate how spectral analysis can be used to obtain clinically relevant sO2 values.

METHODS

Spatial scanning HSI was performed on cutaneous forehead flaps, raised as part of a direct brow lift, in eight patients. Pixel-by-pixel spectral analysis, accounting for the absorption from multiple chromophores, was performed and compared to previous analysis techniques to assess sO2.

RESULTS

Spectral unmixing using a broad spectral range, and accounting for the absorption of melanin, fat, collagen, and water, provided a more clinically relevant estimate of sO2 than conventional techniques, where typically only spectral features associated with absorption of oxygenated (HbO2) and deoxygenated (HbR) hemoglobin are considered. We demonstrate its clinical applicability by generating sO2 maps of partially excised forehead flaps showed a gradual decrease in sO2 along the length of the flap from 95 % at the flap base to 85 % at the flap tip. After being fully excised, sO2 in the entire flap decreased to 50 % within a few minutes.

CONCLUSIONS

The results demonstrate the capability of sO2 mapping in reconstructive surgery in patients using HSI. Spectral unmixing, accounting for multiple chromophores, provides sO2 values that are in accordance with physiological expectations in patients with normal functioning microvascularization. Our results suggest that HSI methods that yield reliable spectra are to be preferred, so that the analysis can produce results that are of clinical relevance.

Rapid extraction of skin physiological parameters from hyperspectral images using machine learning

Noninvasive assessment of skin structure using hyperspectral images has been intensively studied in recent years. Due to the high computational cost of the classical methods, such as the inverse Monte Carlo (IMC), much research has been done with the aim of using machine learning (ML) methods to reduce the time required for estimating parameters. This study aims to evaluate the accuracy and the estimation speed of the ML methods for this purpose and compare them to the traditionally used inverse adding-doubling (IAD) algorithm. We trained three models – an artificial neural network (ANN), a 1D convolutional neural network (CNN), and a random forests (RF) model – to predict seven skin parameters. The models were trained on simulated data computed using the adding-doubling algorithm. To improve predictive performance, we introduced a stacked dynamic weighting (SDW) model combining the predictions of all three individually trained models. SDW model was trained by using only a handful of real-world spectra on top of the ANN, CNN and RF models that were trained using simulated data. Models were evaluated based on the estimated parameters’ mean absolute error (MAE), considering the surface inclination angle and comparing skin spectra with spectra fitted by the IAD algorithm. On simulated data, the lowest MAE was achieved by the RF model (0.0030), while the SDW model achieved the lowest MAE on in vivo measured spectra (0.0113). The shortest time to estimate parameters for a single spectrum was 93.05 μs. Results suggest that ML algorithms can produce accurate estimates of human skin optical parameters in near real-time.

A non-contact oxygen saturation detection method based on dynamic spectrum

Blood oxygen saturation (SpO₂) is an important monitoring indicator for many respiratory diseases. Non-contact oximetry offers outstanding advantages in both coronavirus pandemic monitoring and sleep monitoring, but at the same time poses both challenges regarding technology and environment. Therefore, we propose a method for non-contact SpO₂ measurement based on the principle of DS (dynamic spectrum) in this paper. A multispectral camera with 24 wavelengths (range in 660 nm-950 nm) is used to capture video of the people's cheek region, and then the two-dimensional images are converted into a one-dimensional temporal PPG signal. After pre-processing the PPG signal, the 24 wavelengths DS values are extracted. The optimal wavelength combination is obtained by wavelength screening using the one-by-one elimination method, and a PLS (partial least squares) model is established using the SpO₂ values measured simultaneously by pulse oximetry as the modeled true values. The facial videos of eight healthy subjects were collected, and a total of 140 valid samples were obtained. By analyzing the modeling results, the regression coefficient (R) and root mean square error (RMSE) of the modeled set were 0.6366 and 0.9906, respectively. This method can significantly respond to the variation of SpO₂, and the prediction results are approaching to the prediction accuracy (±2%) of most pulse oximeters in the market. Using DS theory in this method eliminates in principle the interference of static tissue, individual differences, and environment. It fully meets the strong demand for non-contact oximetry and provides a new measurement idea.