Journal of Biomedical Optics最新文献

Significance: Monte Carlo (MC) methods are the gold standard for modeling light-tissue interactions due to their accuracy. Mesh-based MC (MMC) offers enhanced precision for complex tissue structures using tetrahedral mesh models. Despite significant speedups achieved on graphics processing units (GPUs), MMC performance remains hindered by the computational cost of frequent ray-boundary intersection tests.

Aim: We propose a highly accelerated MMC algorithm, RT-MMC, which leverages the hardware-accelerated ray traversal and intersection capabilities of ray-tracing cores (RT-cores) on modern GPUs.

Approach: Implemented using NVIDIA's OptiX platform, RT-MMC extends graphics ray-tracing pipelines toward volumetric ray-tracing in turbid media, eliminating the need for challenging tetrahedral mesh generation while delivering significant speed improvements through hardware acceleration. It also intrinsically supports wide-field sources without complex mesh retessellation.

Results: RT-MMC demonstrates excellent agreement with traditional software-ray-tracing MMC algorithms while achieving 1.5× to 4.5× speedups across multiple GPU architectures. These performance gains significantly enhance the practicality of MMC for routine simulations.

Conclusion: Migration from software- to hardware-based ray tracing not only greatly simplifies MMC simulation workflows but also results in significant speedups that are expected to increase further as ray-tracing hardware rapidly gains adoption. Adoption of graphics ray-tracing pipelines in quantitative MMC simulations enables leveraging of emerging hardware resources and benefits a wide range of biophotonics applications.

Significance: Hyperspectral imaging (HSI) is an advanced spectral imaging technique that captures spatial and spectral information across numerous wavelength bands. This capability allows tissue characterization, disease detection and diagnosis, surgical guidance, and digital histopathology, making it an increasingly valuable tool with wide biological and medical applications.

Aim: We aim to provide readers with (1) an understanding of the principles and technological advancements in HSI, (2) a comprehensive overview of HSI data processing and analysis methods, and (3) an updated survey of biomedical applications, from disease detection, intraoperative imaging, to histopathology.

Approach: A systematic literature search was conducted using PubMed and Google Scholar with the keyword "hyperspectral imaging." We previously published a comprehensive review paper on medical HSI in 2014, which was widely cited in the field. Therefore, this updated review focused on new technology advancements and emerging applications. Based on their biological and medical relevance, 612 HSI papers were included and analyzed in this review.

Results: Recent advances in HSI span both hardware and computational techniques, including improvements in sensor technology, data processing and analysis, short-wave near-infrared imaging, and deep-learning and AI tools. HSI is actively explored for various applications in oncology, neurology, ophthalmology, dermatology, cardiology, gastroenterology, hepatology, wound care, endocrinology, dentistry, infectious disease, plastic and reconstructive surgery, general surgery, intraoperative guidance, histopathology, microbiology, nanopathology, and pharmacology.

Conclusions: HSI has become an emerging imaging modality in biomedical research and clinical settings. Continued advancements in hardware miniaturization, computational efficiency, and clinical validation will further solidify the role of next-generation HSI in biomedicine.

Significance: Craniofacial implants are prone to skin necrosis and exposure, likely due to implant-induced stress and ischemia. However, this relationship has not been quantitatively validated. Identifying an imaging biomarker of skin vascular health could confirm this mechanism and help predict skin failure to mitigate implant complications.

Aim: Photoacoustic tomography (PAT) was used to monitor changes in the skin vasculature surrounding subcutaneous implants with the goal of obtaining a quantitative metric predictive of implant exposure.

Approach: Three designs of 3D-printed porous polycaprolactone (PCL) constructs-unimodal block, bimodal block, and unimodal dome-were implanted in 16 hairless mice. PAT was performed biweekly for 16 weeks, and a skeletonization algorithm was applied to quantify vascular density in skin overlying the implants.

Results: Mice that developed implant exposure ( $N=6$ ) exhibited a progressive decline in vascular density beginning 6 weeks before visible exposure, whereas nonexposed mice ( $N=10$ ) remained stable. Group differences were significant 4 weeks ( $p=0.031$ ) and 2 weeks ( $p=0.001$ ) before exposure onset.

Conclusions: These findings establish a quantitative temporal relationship between vascular ischemia and implant exposure. PAT-derived vascular density serves as a predictive biomarker of skin failure, which can be used to enable interventional treatment and improve implant designs.

Significance: Black patients are at greater risk than White patients for occult hypoxemia due to a melanin-associated bias in ${\mathrm{SpO}}_2$ readings from commercially available pulse oximeters that employ polychromatic light sources.

Aim: We aim to demonstrate how melanin-associated inaccuracies in commercially available pulse oximeters increase the likelihood of occult hypoxemia for all patients and how monochromatic light sources could minimize melanin-associated occult hypoxemia.

Approach: Published values of mean error ( $M$ $[{\mathrm{SpO}}_2-{\mathrm{SaO}}_2]$ ) and uncertainty (SD $[{\mathrm{SpO}}_2-{\mathrm{SaO}}_2]$ ) for Black and White patients were used to analytically model the risk of occult hypoxemia. Mean errors and uncertainties were also estimated for hypothetical patients with no skin melanin, which are directly comparable to values for pulse oximeters employing monochromatic light sources.

Results: The analytically predicted risk of occult hypoxemia for Black patients relative to White patients was 2.1, consistent with published empirical findings. Oximeters unaffected by melanin in the skin would have smaller mean errors and uncertainties than current oximeters, significantly reducing the risk of occult hypoxemia in both Black patients (83% reduction) and White patients (65% reduction).

Conclusions: Because everyone has melanin in their skin, the use of monochromatic light sources in pulse oximeters could significantly reduce the risk of occult hypoxemia for everyone.

Significance: Hyperspectral imaging is a noninvasive, cost-effective modality with transformative clinical potential. Its adoption is limited by the lack of accurate and efficient methods that relate spectra to tissue parameters, essential for both AI training and validation of imaging methods, as gold standard Monte Carlo (MC) simulations remain prohibitively computationally expensive.

Aim: We aim to develop a scalable and accurate method for generating realistic tissue reflectance spectra in support of AI development and validation in biomedical imaging.

Approach: We trained a general-purpose neural surrogate model on $>50$ million MC simulations based on a flexible multilayer tissue model. We validated our model against $>5000$ open surgery in vivo hyperspectral images, annotated with 23 tissue classes for stratified performance analysis. In addition, we qualitatively evaluated clinical potential by testing whether surrogate-generated spectra enable recovery of organ-specific oxygenation dynamics in a controlled porcine aortic clamping experiment.

Results: The surrogate model achieved accuracy matching MC simulations with 5-10 million photons while delivering inference five orders of magnitude faster. Across 140 million human tissue spectra, it improved spectral recall by 13-48 percentage points over existing surrogate models. Scaling analyses revealed a power law relationship between training dataset size and test error, enabling the prediction of training data requirements for target accuracy. Our porcine study suggests that the synthetic data generated with the surrogate model is suitable for recovering organ-specific $s_t{O}_2$ trajectories.

Conclusion: Neural surrogate models can achieve MC-level accuracy and in vivo realism at negligible inference cost, enabling large-scale, compute-efficient data generation for biomedical optics and robust AI development for clinical applications.

Significance: Stomach (gastric) cancer survival depends significantly on the stage in which it is detected, and surveillance with white light endoscopy exhibits poor contrast between gastric cancer and healthy tissue, especially at early stages. Early gastric cancer can exhibit changes in epithelial microstructure, including loss of regular gastric pit structure and collagen alterations which increase tissue stiffness.

Aim: To improve contrast between early cancer and normal tissue, we investigate the use of optical coherence tomography (OCT) and elastography (OCE) to visualize changes in tissue structure and stiffness consistent with gastric cancer.

Approach: Images of eight samples of ex vivo human stomach tissue from three patients were collected with a benchtop OCT system. OCT was performed for qualitative visualization of tissue structure. OCE was then performed on 17 regions of interest using a simplified optical palpation method to extract relative stiffness measurements. A transparent silicone reference layer was placed on the tissue, and axial compression was applied. The resulting deformation (strain) of the reference layer was measured, and the corresponding stress applied to the sample surface was extracted from the characteristic stress-strain curve of the reference material. Spatially resolved stress measurements were mapped and overlaid on en face OCT images. Tissue classification was confirmed by pathology.

Results: OCT image volumes showed more distinct gastric pit and tissue layer structure, as well as less optical attenuation, in normal tissue compared to gastric metaplasia and focal signet ring cell carcinoma (SRCC). Exemplary OCE-derived stress maps showed a trend of increasing measured stress with progression of precancer (metaplasia and dysplasia) and SRCC, suggesting increased tissue stiffness.

Conclusions: This proof-of-concept study provides evidence that OCT and OCE may be capable of visualizing differences in tissue structure and stiffness between normal, metaplastic, dysplastic, and early cancerous gastric tissue, potentially providing the basis for improved screening tools with higher sensitivity.

Significance: Photoacoustic tomography (PAT) holds promise for non-invasive functional imaging in ovarian cancer diagnostics. However, accurate estimation of oxygen saturation ( ${\%\mathrm{sO}}_2$ ) and total hemoglobin concentration (THb) is hindered by wavelength- and depth-dependent fluence variations.

Aim: We aim to improve the accuracy and clinical utility of ${\%\mathrm{sO}}_2$ and THb quantification in transvaginal ultrasound-guided PAT (US-PAT) by developing an integrated spectral and depth compensation (ISDC) method that corrects for both spectral distortion and depth-dependent attenuation.

Approach: We introduce a spectral compensation strategy derived from Monte Carlo simulations and integrate it with depth-wise fluence correction to construct the proposed ISDC method. The approach has been validated using phantoms with known optical properties and applied to clinical PAT data from 82 ovarian lesions (67 benign and 15 malignant). Diagnostic performance was evaluated using logistic regression and receiver operating characteristic analysis.

Results: In phantom experiments, ISDC improved ${\%\mathrm{sO}}_2$ estimation accuracy compared with linear unmixing (LU) and enhanced uniformity of THb estimates across depth. In clinical data, ISDC has increased ${\%\mathrm{sO}}_2$ values by $\sim 5\%$ in both benign and malignant lesions, enhanced contrast of THb between malignant and benign lesion groups (mean THb ratio $R_{\mathrm{THb}}$ has increased from 1.4 to 1.9), and achieved higher classification performance (AUC = 0.93 versus 0.88 for LU) when combining ${\%\mathrm{sO}}_2$ and THb features.

Conclusions: The ISDC approach significantly enhances the quantitative accuracy and diagnostic performance of PAT by compensating for both spectral and depth fluence variations within biological tissue. These improvements support the integration of ISDC into US-PAT systems for ovarian lesion characterization and future clinical applications.

Significance: Abnormal placental development is a major cause of adverse pregnancy outcomes, but current methods for placenta monitoring are not suitable for bedside use. Continuous-wave near-infrared spectroscopy (CW-NIRS) is an optical technique that takes advantage of the near-infrared light to provide functional measurements such as tissue oxygenation at the bedside. However, the placenta is an organ located beneath several layers of tissue, making robust measurement of placental oxygenation with a CW-NIRS device a complex task.

Aim: We propose a framework based on light propagation simulations to evaluate the sensitivity of CW-NIRS devices for placenta detection, along with tools to support NIRS instrument development for engineers.

Approach: The maternal abdomen was modeled as a four-layer structure (i.e., skin, adipose tissue, muscle, and placenta). We used a numerical solution of the diffusion equation using a finite-element method to assess the sensitivity to measure placental function under various conditions (tissue layer thickness, skin tone, tissue oxygen saturation). We used a calibration procedure to evaluate the probability of acquiring a sufficient irradiation with a CW-NIRS device. We collected ultrasound abdomen images from 142 healthy pregnant participants that we segmented and digitized to demonstrate our approach.

Results: With a Mini-CYRIL CW-NIRS device, we showed that placenta monitoring is not possible when using short integration time with a subject having a deep placenta ( $\ge 20\mathrm{mm}$ ) and dark skin tones. With an integration time of 10 s and a temporal binning of 10 points, simulations indicated that subjects with very fair skin tone have a placenta-scanning probability of 12% at a placenta depth of 20 mm and 39% at a depth of 10 mm, using a 50 mm source-detector separation. Thick skin and dark skin tones act as a filter on the NIRS signal, blocking backscattered light and leading to greater absorption in deeper tissues. The spatially resolved spectroscopy method can be used to monitor placental oxygenation with a placenta close to the surface and an oxygen saturation in the muscle layer lower than that of the placenta. The simulation of a realistic cohort of 142 maternal abdomens aimed to identify the optimal acquisition conditions for CW-NIRS devices to be used in placental monitoring.

Conclusions: We proposed a framework to evaluate and optimize CW-NIRS sensitivity for placenta detection. Further work is needed to improve the reliability of placental tissue oxygenation.

Significance: Performing a ratiometric analysis of the fluorescence signals noninvasively measured at two different wavelengths can provide depth estimates of subsurface inner structures in a simple and fast manner, allowing for real-time applications in clinical settings. This can be done using the initially proposed single-excitation-multiple-emission wavelengths approach or by implementing a modified multiple-excitation-single-emission approach; the latter being sometimes preferred due to the larger variation of tissue optical properties at shorter wavelengths. However, previous works validating this method with Monte Carlo (MC) simulations, experiments on tissue-mimicking phantoms, and in vivo measurements on small animal models have reported different degrees of accuracy.

Aim: We tested the influence of factors not generally accounted for in the analytical model used for data interpretation (e.g., tissue geometry and boundaries, inclusion size and shape, and spectral characteristics of the excitation source). To address these limitations, we developed an improved theoretical framework that explicitly accounts for these factors during data interpretation.

Approach: Model validation was carried out with MC simulations and with phantom experiments using indocyanine green as the fluorescence contrast agent. The aimed tissue optical properties were those characteristic of the prostate in a wide range of wavelengths (from 550 to 900 nm).

Results: The aforementioned factors have a strong influence when changing the original single-excitation-multiple-emission approach to a multiple-excitation-single-emission approach. Though this might make the latter a less preferable method, the low variability of the optical properties in the multiple emission approach (as it happens with prostate tissue) negatively impacts the depth reconstruction process.

Conclusions: Regardless of the ratiometric strategy employed, accurate depth estimation requires that the theoretical model closely replicate the experimental conditions. Careful matching of model assumptions to the measurement environment is essential to achieve reliable data interpretation.

Significance: Port-wine stains (PWSs) are congenital capillary malformations with the incidence of in newborns of $\sim 0.8\%$ to 2.1%. Hematoporphyrin monomethyl ether-mediated photodynamic therapy (HMME-PDT) has been widely applied in China for PWS. However, there remains substantial room for improvement in both the phototherapeutic selectivity coefficient (PSC) and pain management.

Aim: We investigated the feasibility of modulating transcutaneous oxygen delivery during photodynamic therapy of PWS to enhance therapeutic efficacy and reduce pain.

Approach: A three-dimensional (3D) computational biophysical model was employed to elucidate the mechanisms through which transcutaneous oxygen modulation enhances the therapeutic efficacy of HMME-PDT and improves pain management. The model was constructed to simulate the light propagation, photosensitizer kinetics, oxygen diffusion, and reactive oxygen species (ROS) generation. A treatment optimization strategy based on epidermal oxygen regulation was proposed and evaluated in computational studies. The spatiotemporal distributions of singlet oxygen under normoxic, hypoxic, and anoxic conditions were evaluated, and their effects on treatment-induced pain and lesion-targeted cytotoxicity were analyzed.

Results: Computational analysis showed that compared with normoxic conditions, hypoxia and anoxia significantly enhanced PSC, with improvements of 48% and 61%, respectively. Furthermore, these oxygen-modulated regimens attenuated treatment-associated pain, reducing photochemical pain duration of 17% (hypoxia) and 30% (anoxia). Choosing the right combination of light source irradiance and surface oxygen supply rate amplified therapeutic performance and patient comfort, achieving a 213% increase in PSC and a 57% reduction in photochemical pain duration. These findings establish a mechanistic framework for advancing precision PDT protocols with minimized iatrogenic discomfort.

Conclusions: Established in this computational study, strategic epidermal oxygen restriction critically augments PDT PSC while improving patient tolerance. Computational modeling demonstrates that controlled epidermal hypoxia spatially redistributes oxygen gradients, thereby suppressing superficial ROS generation in nontargeted epidermal layers and selectively concentrating ROS within PWS vasculature. This dual mechanism-simultaneously enhancing therapeutic precision and attenuating treatment-induced pain-presents a pioneering strategy centered on an active oxygen control strategy for enhancing HMME-PDT clinical outcomes. Future research will progress from preclinical validation in animal models to clinical studies to evaluate the therapeutic efficacy and translational potential of this strategy.