Significance: Skin cancer is one of the most prevalent cancers worldwide. In the advent of medical digitization and telepathology, hyper/multispectral imaging (HMSI) allows for noninvasive, nonionizing tissue evaluation at a macroscopic level.

Aim: We aim to summarize proposed frameworks and recent trends in HMSI-based classification and segmentation of gross-level skin tissue.

Approach: A systematic review was performed, targeting HMSI-based systems for the classification and segmentation of skin lesions during gross pathology, including melanoma, pigmented lesions, and bruises. The review adhered to the 2020 Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines. For eligible reports published from 2010 to 2020, trends in HMSI acquisition, preprocessing, and analysis were identified.

Results: HMSI-based frameworks for skin tissue classification and segmentation vary greatly. Most reports implemented simple image processing or machine learning, due to small training datasets. Methodologies were evaluated on heavily curated datasets, with the majority targeting melanoma detection. The choice of preprocessing scheme influenced the performance of the system. Some form of dimension reduction is commonly applied to avoid redundancies that are inherent in HMSI systems.

Conclusions: To use HMSI for tumor margin detection in practice, the focus of system evaluation should shift toward the explainability and robustness of the decision-making process.

Significance: Photothrombotic stroke is an important and widely used model for ischemic stroke research. However, the significant scattering of the skull during the procedure limits the light’s ability to penetrate and focus on its target. Targeted photothrombosis uses surgery-based skull windows to obtain optical access to the brain, but it renders the brain’s environment unnatural even before a stroke is established.

Aim: To establish a targeted, controllable ischemic stroke model in mice through an intact skull.

Approach: The in vivo skull optical clearing technique provides a craniotomy-free “optical window” that allows light to penetrate. Alongside the local photodynamic effect, we have established targeted photothrombosis without skull removal, effectively controlling the degree of thrombotic occlusion by changing the light dose.

Results: Ex vivo and in vivo results demonstrated that skull optical clearing treatment significantly enhanced light’s ability to penetrate the skull and focus on its target, contributing to thrombotic occlusion. The skull optical clearing window was also used for continuous blood flow mapping, and the relationship between light dose and injury degree was evaluated over 14 days of monitoring. Per our findings, increasing the light dose was accompanied by more severe infarction, indicating that the model was easily controllable.

Conclusions: Herein, a targeted, controllable ischemic stroke model was established by combinedly running an in vivo skull optical clearing technique and a photothrombotic procedure, avoiding unnecessary damage or environmental changes to the brain caused by surgery on the skull. Our established model should offer significant value to research on ischemic stroke.

Significance: Growing levels of obesity and metabolic syndrome have driven demand for more advanced forms of body composition assessment. While various techniques exist to measure body composition, devices are typically expensive and not portable, involve radiation [in the case of dual-energy x-ray absorptiometry (DXA)], and are limited to analysis of adiposity while metabolic information from blood supply and oxygenation are not considered.

Aim: We evaluate whether diffuse optical spectroscopic imaging (DOSI) can be used to predict site-specific adiposity and percent fat (whole body) while simultaneously providing information about local tissue hemoglobin levels and oxygenation.

Approach: DOSI measures of tissue composition in gastrocnemius, quadriceps, abdomen, and biceps, DXA whole-body composition, and ultrasound-derived skin and adipose tissue thickness (SATT) in the quadriceps were obtained from 99 individuals aged 7 to 34 years old.

Results: Various DOSI-derived parameters were correlated with SATT and an optical method is proposed for estimating SATT using a newly defined parameter, the optical fat fraction (OFF), which considers all parameters that correlate with SATT. Broadband absorption and scattering spectra from study participants with the thinnest (SATT  ≈  0.25  ±  0.02  cm) and thickest SATT (SATT  ≈  1.55  ±  0.14  cm), representing best estimates for pure in vivo lean and fatty tissue, respectively, are reported. Finally, a trained prediction model is developed which allows DOSI assessment of OFF to predict DXA body-fat percentage, demonstrating that DOSI can be used to quantify body composition.

Conclusions: This study shows that DOSI can be used to assess the adiposity of specific tissues or the entire human body, and the OFF parameter is defined for corroboration and further evaluation in future research.

Significance: We present an approach to estimate with simple instrumentation the amount of red blood cells in the skin microvasculature, designated as parameter L_RBC. Variations of parameter L_RBC are shown to reflect local changes in the quantity of skin red blood cells during a venous occlusion challenge.

Aim: To validate a simple algebraic model of light transport in skin using the Monte Carlo method and to develop a measure of the red blood cell content in skin microvessels using the Monte Carlo predictions; to guide the development of an instrument to measure experimentally variations of the amount of red blood cells in the skin.

Approach: Monte Carlo simulations were carried out in a multilayer model of the skin to compute remitted light intensities as a function of distance from the illumination locus for different values of the skin blood content. The simulation results were used to compute parameter L_RBC and its variations with local skin blood content. An experimental setup was developed to measure parameter L_RBC in human volunteers in whom skin blood content of the forearm increased during temporary interruption of the venous outflow.

Results: In the simulations, parameter L_RBC was ∼16  μm in baseline conditions, and it increased in near proportion with the blood content of the skin layers. Measuring the diffusely reflected light intensity 0.5 to 1.2 mm away from the illumination locus was optimal to detect appreciable changes of the reflected light intensity as skin blood content was altered. Parameter L_RBC measured experimentally on the human forearm was 17  ±  2  μm in baseline conditions it increased at a rate of 4  ±  2  μm  /  min when venous outflow was temporarily interrupted.

Conclusion: Parameter L_RBC derived experimentally with a two-wavelength diffuse reflectometer can be used to measure local variations of the amount of red blood cells in skin microvessels.

Significance: Raman spectroscopy (RS) provides an automated approach for assisting Mohs micrographic surgery for skin cancer diagnosis; however, the specificity of RS is limited by the high spectral similarity between tumors and normal tissues structures. Reflectance confocal microscopy (RCM) provides morphological and cytological details by which many features of epidermis and hair follicles can be readily identified. Combining RS with deep-learning-aided RCM has the potential to improve the diagnostic accuracy of RS in an automated fashion, without requiring additional input from the clinician.

Aim: The aim of this study is to improve the specificity of RS for detecting basal cell carcinoma (BCC) using an artificial neural network trained on RCM images to identify false positive normal skin structures (hair follicles and epidermis).

Approach: Our approach was to build a two-step classification model. In the first step, a Raman biophysical model that was used in prior work classified BCC tumors from normal tissue structures with high sensitivity. In the second step, 191 RCM images were collected from the same site as the Raman data and served as inputs for two ResNet50 networks. The networks selected the hair structure and epidermis images, respectively, within all images corresponding to the positive predictions of the Raman biophysical model with high specificity. The specificity of the BCC biophysical model was improved by moving the Raman spectra corresponding to these selected images from false positive to true negative.

Results: Deep-learning trained on RCM images removed 52% of false positive predictions from the Raman biophysical model result while maintaining a sensitivity of 100%. The specificity was improved from 84.2% using Raman spectra alone to 92.4% by integrating Raman spectra with RCM images.

Conclusions: Combining RS with deep-learning-aided RCM imaging is a promising tool for guiding tumor resection surgery.

Significance: X-ray imaging serves as the mainstream imaging in dentistry, but it involves risk of ionizing radiation.

Aim: This study presents the feasibility of indocyanine green-assisted near-infrared fluorescence (ICG-NIRF) dental imaging with 785-nm NIR laser in the first (ICG-NIRF-I: 700 to 1000 nm) and second (ICG-NIRF-II: 1000 to 1700 nm) NIR wavelengths.

Approach: Sprague Dawley rats with different postnatal days were used as animal models. ICG, as a fluorescence agent, was delivered to dental structures by subcutaneous injection (SC) and oral administration (OA).

Results: For SC method, erupted and unerupted molars could be observed from ICG-NIRF images at a short imaging time (<1  min). ICG-NIRF-II could achieve a better image contrast in unerupted molars at 24 h after ICG injection. The OA could serve as a non-invasive method for ICG delivery; it could also cause the glow-in-dark effect in unerupted molars. For erupted molars, OA can be considered as mouthwash and exhibits outstanding performance for delivery of ICG dye; erupted molar structures could be observed at a short imaging time (<1  min) and low ICG dose (0.05  mg  /  kg).

Conclusions: Overall, ICG-NIRF with mouthwash could perform in-vivo dental imaging in two NIR wavelengths at a short time and low ICG dose.

Significance: Accurate early diagnosis of malignant skin lesions is critical in providing adequate and timely treatment; unfortunately, initial clinical evaluation of similar-looking benign and malignant skin lesions can result in missed diagnosis of malignant lesions and unnecessary biopsy of benign ones.

Aim: To develop and validate a label-free and objective image-guided strategy for the clinical evaluation of suspicious pigmented skin lesions based on multispectral autofluorescence lifetime imaging (maFLIM) dermoscopy.

Approach: We tested the hypothesis that maFLIM-derived autofluorescence global features can be used in machine-learning (ML) models to discriminate malignant from benign pigmented skin lesions. Clinical widefield maFLIM dermoscopy imaging of 41 benign and 19 malignant pigmented skin lesions from 30 patients were acquired prior to tissue biopsy sampling. Three different pools of global image-level maFLIM features were extracted: multispectral intensity, time-domain biexponential, and frequency-domain phasor features. The classification potential of each feature pool to discriminate benign versus malignant pigmented skin lesions was evaluated by training quadratic discriminant analysis (QDA) classification models and applying a leave-one-patient-out cross-validation strategy.

Results: Classification performance estimates obtained after unbiased feature selection were as follows: 68% sensitivity and 80% specificity with the phasor feature pool, 84% sensitivity, and 71% specificity with the biexponential feature pool, and 84% sensitivity and 32% specificity with the intensity feature pool. Ensemble combinations of QDA models trained with phasor and biexponential features yielded sensitivity of 84% and specificity of 90%, outperforming all other models considered.

Conclusions: Simple classification ML models based on time-resolved (biexponential and phasor) autofluorescence global features extracted from maFLIM dermoscopy images have the potential to provide objective discrimination of malignant from benign pigmented lesions. ML-assisted maFLIM dermoscopy could potentially assist with the clinical evaluation of suspicious lesions and the identification of those patients benefiting the most from biopsy examination.

Significance: The shortwave infrared (SWIR) optical window (∼900 to 2000 nm) has attracted interest for deep tissue imaging due to the lower scattering of light. SWIR spatial frequency domain imaging (SWIR SFDI) provides wide-field tissue optical property measurements in this wavelength band. Key design and performance characteristics, such as portability, wavelength selection, measurement resolution, and the effect of skin have not yet been addressed for SWIR SFDI.

Aim: To fabricate and characterize a SWIR SFDI system for clinical use.

Approach: The optimal choice of wavelengths was identified based on optical property uncertainty estimates and imaging depth. A compact light-emitting diode-based dual wavelength SWIR SFDI system was fabricated. A two-layer inverse model was developed to account for the layered structure of skin. Performance was validated using tissue-simulating phantoms and in-vivo measurements from three healthy subjects.

Results: The SWIR SFDI system had a μs′ resolution of at least 0.03  mm^−  1 at 880 nm and 0.02  mm^−  1 at 1100 nm. The two-layer inverse model reduced the error in deeper layer μs′ extractions by at least 24% in the phantom study. The two-layer model also increased the contrast between superficial vessels and the surrounding tissue for in-vivo measurements.

Conclusion: The clinic-ready SWIR SFDI device is sensitive to small optical property alterations in diffuse media, provides enhanced accuracy in quantifying optical properties in the deeper layers in phantoms, and provided enhanced contrast of subcutaneous blood vessels.

Significance: Bioluminescence imaging and tomography (BLT) are used to study biologically relevant activity, typically within a mouse model. A major limitation is that the underlying optical properties of the volume are unknown, leading to the use of a “best” estimate approach often compromising quantitative accuracy.

Aim: An optimization algorithm is presented that localizes the spatial distribution of bioluminescence by simultaneously recovering the optical properties and location of bioluminescence source from the same set of surface measurements.

Approach: Measured data, using implanted self-illuminating sources as well as an orthotopic glioblastoma mouse model, are employed to recover three-dimensional spatial distribution of the bioluminescence source using a multi-parameter optimization algorithm.

Results: The proposed algorithm is able to recover the size and location of the bioluminescence source while accounting for tissue attenuation. Localization accuracies of <1  mm are obtained in all cases, which is similar if not better than current “gold standard” methods that predict optical properties using a different imaging modality.

Conclusions: Application of this approach, using in-vivo experimental data has shown that quantitative BLT is possible without the need for any prior knowledge about optical parameters, paving the way toward quantitative molecular imaging of exogenous and indigenous biological tumor functionality.

Significance: In circular scanning photoacoustic tomography (PAT), it takes several minutes to generate an image of acceptable quality, especially with a single-element ultrasound transducer (UST). The imaging speed can be enhanced by faster scanning (with high repetition rate light sources) and using multiple-USTs. However, artifacts arising from the sparse signal acquisition and low signal-to-noise ratio at higher scanning speeds limit the imaging speed. Thus, there is a need to improve the imaging speed of the PAT systems without hampering the quality of the PAT image.

Aim: To improve the frame rate (or imaging speed) of the PAT system by using deep learning (DL).

Approach: For improving the frame rate (or imaging speed) of the PAT system, we propose a novel U-Net-based DL framework to reconstruct PAT images from fast scanning data.

Results: The efficiency of the network was evaluated on both single- and multiple-UST-based PAT systems. Both phantom and in vivo imaging demonstrate that the network can improve the imaging frame rate by approximately sixfold in single-UST-based PAT systems and by approximately twofold in multi-UST-based PAT systems.

Conclusions: We proposed an innovative method to improve the frame rate (or imaging speed) by using DL and with this method, the fastest frame rate of ∼3  Hz imaging is achieved without hampering the quality of the reconstructed image.

Significance: Optical coherence tomography (OCT) allows high-resolution volumetric three-dimensional (3D) imaging of biological tissues in vivo. However, 3D-image acquisition can be time-consuming and often suffers from motion artifacts due to involuntary and physiological movements of the tissue, limiting the reproducibility of quantitative measurements.

Aim: To achieve real-time 3D motion compensation for corneal tissue with high accuracy.

Approach: We propose an OCT system for volumetric imaging of the cornea, capable of compensating both axial and lateral motion with micron-scale accuracy and millisecond-scale time consumption based on higher-order regression. Specifically, the system first scans three reference B-mode images along the C-axis before acquiring a standard C-mode image. The difference between the reference and volumetric images is compared using a surface-detection algorithm and higher-order polynomials to deduce 3D motion and remove motion-related artifacts.

Results: System parameters are optimized, and performance is evaluated using both phantom and corneal (ex vivo) samples. An overall motion-artifact error of <4.61 microns and processing time of about 3.40 ms for each B-scan was achieved.

Conclusions: Higher-order regression achieved effective and real-time compensation of 3D motion artifacts during corneal imaging. The approach can be expanded to 3D imaging of other ocular tissues. Implementing such motion-compensation strategies has the potential to improve the reliability of objective and quantitative information that can be extracted from volumetric OCT measurements.

Significance: Three-dimensional (3D) visualization of multicellular tumor spheroids (MCTS) in fluorescence microscopy can rapidly provide qualitative morphological information about the architecture of these cellular aggregates, which can recapitulate key aspects of their in vivo counterpart.

Aim: The present work is aimed at overcoming the shallow depth-of-field (DoF) limitation in fluorescence microscopy while achieving 3D visualization of thick biological samples under study.

Approach: A custom-built fluorescence microscope with an electrically focus-tunable lens was developed to optically sweep in-depth the structure of MCTS. Acquired multifocus stacks were combined by means of postprocessing algorithms performed in the Fourier domain.

Results: Images with relevant characteristics as extended DoF, stereoscopic pairs as well as reconstructed viewpoints of MCTS were obtained without segmentation of the focused regions or estimation of the depth map. The reconstructed images allowed us to observe the 3D morphology of cell aggregates.

Conclusions: Computational multifocus fluorescence microscopy can provide 3D visualization in MCTS. This tool is a promising development in assessing the morphological structure of different cellular aggregates while preserving a robust yet simple optical setup.

Significance: Quantitative oblique back-illumination microscopy (qOBM) is a recently developed label-free imaging technique that enables 3D quantitative phase imaging of thick scattering samples with epi-illumination. Here, we propose dynamic qOBM to achieve functional imaging based on subcellular dynamics, potentially indicative of metabolic activity. We show the potential utility of this novel technique by imaging adherent mesenchymal stromal cells (MSCs) grown in bioreactors, which can help address important unmet needs in cell manufacturing for therapeutics.

Aim: We aim to develop dynamic qOBM and demonstrate its potential for functional imaging based on cellular and subcellular dynamics.

Approach: To obtain functional images with dynamic qOBM, a sample is imaged over a period of time and its temporal signals are analyzed. The dynamic signals display an exponential frequency response that can be analyzed with phasor analysis. Functional images of the dynamic signatures are obtained by mapping the frequency dynamic response to phasor space and color-coding clustered signals.

Results: Functional imaging with dynamic qOBM provides unique information related to subcellular activity. The functional qOBM images of MSCs not only improve conspicuity of cells in complex environments (e.g., porous micro-carriers) but also reveal two distinct cell populations with different dynamic behavior.

Conclusions: In this work we present a label-free, fast, and scalable functional imaging approach to study and intuitively display cellular and subcellular dynamics. We further show the potential utility of this novel technique to help monitor adherent MSCs grown in bioreactors, which can help achieve quality-by-design of cell products, a significant unmet need in the field of cell therapeutics. This approach also has great potential for dynamic studies of other thick samples, such as organoids.

Significance: “Diffuse in vivo flow cytometry” (DiFC) is an emerging technology for fluorescence detection of rare circulating cells directly in large deep-seated blood vessels in mice. Because DiFC uses highly scattered light, in principle, it could be translated to human use. However, an open question is whether fluorescent signals from single cells would be detectable in human-scale anatomies.

Aim: Suitable blood vessels in a human wrist or forearm are at a depth of ∼2 to 4 mm. The aim of this work was to study the impact of DiFC instrument geometry and wavelength on the detected DiFC signal and on the maximum depth of detection of a moving cell.

Approach: We used Monte Carlo simulations to compute fluorescence Jacobian (sensitivity) matrices for a range of source and detector separations (SDS) and tissue optical properties over the visible and near infrared spectrum. We performed experimental measurements with three available versions of DiFC (488, 640, and 780 nm), fluorescent microspheres, and tissue mimicking optical flow phantoms. We used both computational and experimental data to estimate the maximum depth of detection at each combination of settings.

Results: For the DiFC detection problem, our analysis showed that for deep-seated blood vessels, the maximum sensitivity was obtained with NIR light (780 nm) and 3-mm SDS.

Conclusions: These results suggest that—in combination with a suitable molecularly targeted fluorescent probes—circulating cells and nanosensors could, in principle, be detectable in circulation in humans.

Significance: Monitoring the movement and vital signs of patients in hospitals and other healthcare environments is a significant burden on healthcare staff. Early warning systems using smart bed sensors hold promise to relieve this burden and improve patient outcomes. We propose a scalable and cost-effective optical fiber sensor array that can be embedded into a mattress to detect movement, both sensitively and spatially.

Aim: Proof-of-concept demonstration that a multimode optical fiber (MMF) specklegram sensor array can be used to detect and image movement on a bed.

Approach: Seven MMFs are attached to the upper surface of a mattress such that they cross in a 3  ×  4 array. The specklegram output is monitored using a single laser and single camera and movement on the fibers is monitored by calculating a rolling zero-normalized cross-correlation. A 3  ×  4 image is formed by comparing the signal at each crossing point between two fibers.

Results: The MMF sensor array can detect and image movement on a bed, including getting on and off the bed, rolling on the bed, and breathing.

Conclusions: The sensor array shows a high sensitivity to movement, which can be used for monitoring physiological parameters and patient movement for potential applications in healthcare settings.