Deep spectral learning for label-free optical imaging oximetry with uncertainty quantification

Quantitative Optical Imaging of Oxygen in Brain Vasculature

The intimate relationship between neuronal activity and cerebral oxygenation underpins fundamental brain functions like cognition, sensation, and motor control. Optical imaging offers a noninvasive approach to assess brain oxygenation and often serves as an indirect proxy for neuronal activity. However, deciphering neurovascular coupling─the intricate interplay between neuronal activity, blood flow, and oxygen delivery─necessitates independent, high spatial resolution, and high temporal resolution measurements of both microvasculature oxygenation and neuronal activation. This Perspective examines the established optical techniques employed for brain oxygen imaging, specifically functional near-infrared spectroscopy, photoacoustic imaging, optical coherence tomography, and two-photon phosphorescent lifetime microscopy, highlighting their fundamental principles, strengths, and limitations. Several other emerging optical techniques are also introduced. Finally, we discuss key technological challenges and future directions for quantitative optical oxygen imaging, paving the way for a deeper understanding of oxygen metabolism in the brain.

Enhanced Multiscale Human Brain Imaging by Semi-supervised Digital Staining and Serial Sectioning Optical Coherence Tomography

A major challenge in neuroscience is to visualize the structure of the human brain at different scales. Traditional histology reveals micro- and meso-scale brain features, but suffers from staining variability, tissue damage and distortion that impedes accurate 3D reconstructions. Here, we present a new 3D imaging framework that combines serial sectioning optical coherence tomography (S-OCT) with a deep-learning digital staining (DS) model. We develop a novel semi-supervised learning technique to facilitate DS model training on weakly paired images. The DS model performs translation from S-OCT to Gallyas silver staining. We demonstrate DS on various human cerebral cortex samples with consistent staining quality. Additionally, we show that DS enhances contrast across cortical layer boundaries. Furthermore, we showcase geometry-preserving 3D DS on cubic-centimeter tissue blocks and visualization of meso-scale vessel networks in the white matter. We believe that our technique offers the potential for high-throughput, multiscale imaging of brain tissues and may facilitate studies of brain structures.

A Low-Cost Algorithm for Uncertainty Quantification Simulations of Steady-State Flows: Application to Ocular Hemodynamics

An algorithm for the calculation of steady-state flowing under uncertain conditions is introduced in this work in order to obtain a probabilistic distribution of uncertain problem parameters. This is particularly important for problems with increased uncertainty, as typical deterministic methods are not able to fully describe all possible flow states of the problem. Standard methods, such as polynomial expansions and Monte Carlo simulations, are used for the formation of the generalized problem described by the incompressible Navier-Stokes equations. Since every realization of the uncertainty parameter space is coupled with non-linear terms, an incremental iterative procedure was adopted for the calculation. This algorithm adopts a Jacobi-like iteration methodology to decouple the equations and solve them one by one until there is overall convergence. The algorithm was tested in a typical artery geometry, including a bifurcation with an aneurysm, which consists of a well-documented biological flow test case. Additionally, its dependence on the uncertainty parameter space, i.e., the inlet velocity distribution, the Reynolds number variation, and parameters of the procedure, i.e., the number of polynomial expansions, was studied. Symmetry exists in probabilistic theories, similar to the one adopted by the present work. The results of the simulations conducted with the present algorithm are compared against the same but unsteady flow with a time-dependent inlet velocity profile, which represents a typical cardiac cycle. It was found that the present algorithm is able to correctly describe the flow field, as well as capture the upper and lower limits of the velocity field, which was made periodic. The comparison between the present algorithm and the typical unsteady one presented a maximum error of ≈2% in the common carotid area, while the error increased significantly inside the bifurcation area. Moreover, “sensitive” areas of the geometry with increased parameter uncertainty were identified, a result that is not possible to be obtained while using deterministic algorithms. Finally, the ability of the algorithm to tune the parameter limits was successfully tested.

Exploring uncertainty measures in convolutional neural network for semantic segmentation of oral cancer images

SIGNIFICANCE

Oral cancer is one of the most prevalent cancers, especially in middle- and low-income countries such as India. Automatic segmentation of oral cancer images can improve the diagnostic workflow, which is a significant task in oral cancer image analysis. Despite the remarkable success of deep-learning networks in medical segmentation, they rarely provide uncertainty quantification for their output.

AIM

We aim to estimate uncertainty in a deep-learning approach to semantic segmentation of oral cancer images and to improve the accuracy and reliability of predictions.

APPROACH

This work introduced a UNet-based Bayesian deep-learning (BDL) model to segment potentially malignant and malignant lesion areas in the oral cavity. The model can quantify uncertainty in predictions. We also developed an efficient model that increased the inference speed, which is almost six times smaller and two times faster (inference speed) than the original UNet. The dataset in this study was collected using our customized screening platform and was annotated by oral oncology specialists.

RESULTS

The proposed approach achieved good segmentation performance as well as good uncertainty estimation performance. In the experiments, we observed an improvement in pixel accuracy and mean intersection over union by removing uncertain pixels. This result reflects that the model provided less accurate predictions in uncertain areas that may need more attention and further inspection. The experiments also showed that with some performance compromises, the efficient model reduced computation time and model size, which expands the potential for implementation on portable devices used in resource-limited settings.

CONCLUSIONS

Our study demonstrates the UNet-based BDL model not only can perform potentially malignant and malignant oral lesion segmentation, but also can provide informative pixel-level uncertainty estimation. With this extra uncertainty information, the accuracy and reliability of the model’s prediction can be improved.

Cryptanalysis of an optical cryptosystem with uncertainty quantification in a probabilistic model

In this paper, a modified probabilistic deep learning method is proposed to attack the double random phase encryption by modeling the conditional distribution of plaintext. The well-trained probabilistic model gives both predictions of plaintext and uncertainty quantification, the latter of which is first introduced to optical cryptanalysis. Predictions of the model are close to real plaintexts, showing the success of the proposed model. Uncertainty quantification reveals the level of reliability of each pixel in the prediction of plaintext without ground truth. Subsequent simulation experiments demonstrate that uncertainty quantification can effectively identify poor-quality predictions to avoid the risk of unreliability from deep learning models.

Live-dead assay on unlabeled cells using phase imaging with computational specificity

Existing approaches to evaluate cell viability involve cell staining with chemical reagents. However, the step of exogenous staining makes these methods undesirable for rapid, nondestructive, and long-term investigation. Here, we present an instantaneous viability assessment of unlabeled cells using phase imaging with computation specificity. This concept utilizes deep learning techniques to compute viability markers associated with the specimen measured by label-free quantitative phase imaging. Demonstrated on different live cell cultures, the proposed method reports approximately 95% accuracy in identifying live and dead cells. The evolution of the cell dry mass and nucleus area for the labeled and unlabeled populations reveal that the chemical reagents decrease viability. The nondestructive approach presented here may find a broad range of applications, from monitoring the production of biopharmaceuticals to assessing the effectiveness of cancer treatments.

Common methods for characterising cell viability involve cell staining with chemical reagents. Here the authors report a method for cell viability assessment that does not require labelling; this uses quantitative phase imaging combined with deep learning.

Optical oxygen saturation imaging in cellular ex vivo lung perfusion to assess lobular pulmonary function

Ex vivo lung perfusion (EVLP) is an emerging tool to evaluate marginal lungs in lung transplantation. However, there is no objective metric to monitor lobular regional oxygenation during EVLP. In this study, we developed oxygen saturation (SaO₂) imaging to quantitatively assess the regional gas exchange potential of the lower lobes. Ten porcine lungs were randomly divided into control and donation after circulatory death (DCD) groups (n = 5, each). Lungs were perfused in cellular EVLP for 2 h, and multispectral images were continuously collected from the dorsal sides of the lower lobes. We examined whether lower lobe SaO₂ correlated with PaO₂/FiO₂ (P/F) ratios in lower pulmonary veins (PV). The wet/dry ratio in lower lobes was measured and Monte Carlo simulations were performed to investigate the method's feasibility. There was a significant correlation between lower lobe SaO₂ and the P/F ratio in lower PV (r = 0.855, P < 0.001). The DCD group was associated with lower SaO₂ and higher wet/dry ratio than the control group (P < 0.001). The error of estimated SaO₂ was limited according to Monte Carlo simulations. The developed technology provides a noninvasive and regional evaluative tool of quantitative lobular function in EVLP.

Bayesian deep learning for reliable oral cancer image classification

In medical imaging, deep learning-based solutions have achieved state-of-the-art performance. However, reliability restricts the integration of deep learning into practical medical workflows since conventional deep learning frameworks cannot quantitatively assess model uncertainty. In this work, we propose to address this shortcoming by utilizing a Bayesian deep network capable of estimating uncertainty to assess oral cancer image classification reliability. We evaluate the model using a large intraoral cheek mucosa image dataset captured using our customized device from high-risk population to show that meaningful uncertainty information can be produced. In addition, our experiments show improved accuracy by uncertainty-informed referral. The accuracy of retained data reaches roughly 90% when referring either 10% of all cases or referring cases whose uncertainty value is greater than 0.3. The performance can be further improved by referring more patients. The experiments show the model is capable of identifying difficult cases needing further inspection.

Deep Learning in Biomedical Optics

This article reviews deep learning applications in biomedical optics with a particular emphasis on image formation. The review is organized by imaging domains within biomedical optics and includes microscopy, fluorescence lifetime imaging, in vivo microscopy, widefield endoscopy, optical coherence tomography, photoacoustic imaging, diffuse tomography, and functional optical brain imaging. For each of these domains, we summarize how deep learning has been applied and highlight methods by which deep learning can enable new capabilities for optics in medicine. Challenges and opportunities to improve translation and adoption of deep learning in biomedical optics are also summarized. Lasers Surg. Med. © 2021 Wiley Periodicals LLC.

Uncertainty quantification implementations in human hemodynamic flows

BACKGROUND AND OBJECTIVE

Human hemodynamic modeling is usually influenced by uncertainties occurring from a considerable unavailability of information linked to the boundary conditions and the physical properties used in the numerical models. Calculating the effect of these uncertainties on the numerical findings along the cardiovascular system is a demanding process due to the complexity of the morphology of the body and the area dynamics. To cope with all these difficulties, Uncertainty Quantification (UQ) methods seem to be an ideal tool.

RESULTS

This study focuses on analyzing and summarizing some of the recent research efforts and directions of implementing UQ in human hemodynamic flows by analyzing 139 research papers. Initially, the suitability of applying this approach is analyzed and demonstrated. Then, an overview of the most significant research work in various fields of biomedical hemodynamic engineering is presented. Finally, it is attempted to identify any possible forthcoming directions for research and methodological progress of UQ in biomedical sciences.

CONCLUSION

This review concludes that by finding the best statistical methods and parameters to represent the propagated uncertainties, while achieving a good interpretation of the interaction between input-output, is crucial for implementing UQ in biomedical sciences.

Single-cell cytometry via multiplexed fluorescence prediction by label-free reflectance microscopy

Traditional imaging cytometry uses fluorescence markers to identify specific structures but is limited in throughput by the labeling process. We develop a label-free technique that alleviates the physical staining and provides multiplexed readouts via a deep learning-augmented digital labeling method. We leverage the rich structural information and superior sensitivity in reflectance microscopy and show that digital labeling predicts accurate subcellular features after training on immunofluorescence images. We demonstrate up to three times improvement in the prediction accuracy over the state of the art. Beyond fluorescence prediction, we demonstrate that single cell-level structural phenotypes of cell cycles are correctly reproduced by the digital multiplexed images, including Golgi twins, Golgi haze during mitosis, and DNA synthesis. We further show that the multiplexed readouts enable accurate multiparametric single-cell profiling across a large cell population. Our method can markedly improve the throughput for imaging cytometry toward applications for phenotyping, pathology, and high-content screening.