The in vitro dynamics of pseudo-vascular network formation.

BACKGROUND/OBJECTIVES: Pseudo-vascular network formation in vitro is considered a key characteristic of vasculogenic mimicry. While many cancer cell lines form pseudo-vascular networks, little is known about the spatiotemporal dynamics of these formations. METHODS: Here, we present a framework for monitoring and characterising the dynamic formation and dissolution of pseudo-vascular networks in vitro. The framework combines time-resolved optical microscopy with open-source image analysis for network feature extraction and statistical modelling. The framework is demonstrated by comparing diverse cancer cell lines associated with vasculogenic mimicry, then in detecting response to drug compounds proposed to affect formation of vasculogenic mimics. Dynamic datasets collected were analysed morphometrically and a descriptive statistical analysis model was developed in order to measure stability and dissimilarity characteristics of the pseudo-vascular networks formed. RESULTS: Melanoma cells formed the most stable pseudo-vascular networks and were selected to evaluate the response of their pseudo-vascular networks to treatment with axitinib, brucine and tivantinib. Tivantinib has been found to inhibit the formation of the pseudo-vascular networks more effectively, even in dose an order of magnitude less than the two other agents. CONCLUSIONS: Our framework is shown to enable quantitative analysis of both the capacity for network formation, linked vasculogenic mimicry, as well as dynamic responses to treatment.

Moving Beyond Simulation: Data-Driven Quantitative Photoacoustic Imaging Using Tissue-Mimicking Phantoms.

Accurate measurement of optical absorption coefficients from photoacoustic imaging (PAI) data would enable direct mapping of molecular concentrations, providing vital clinical insight. The ill-posed nature of the problem of absorption coefficient recovery has prohibited PAI from achieving this goal in living systems due to the domain gap between simulation and experiment. To bridge this gap, we introduce a collection of experimentally well-characterised imaging phantoms and their digital twins. This first-of-a-kind phantom data set enables supervised training of a U-Net on experimental data for pixel-wise estimation of absorption coefficients. We show that training on simulated data results in artefacts and biases in the estimates, reinforcing the existence of a domain gap between simulation and experiment. Training on experimentally acquired data, however, yielded more accurate and robust estimates of optical absorption coefficients. We compare the results to fluence correction with a Monte Carlo model from reference optical properties of the materials, which yields a quantification error of approximately 20%. Application of the trained U-Nets to a blood flow phantom demonstrated spectral biases when training on simulated data, while application to a mouse model highlighted the ability of both learning-based approaches to recover the depth-dependent loss of signal intensity. We demonstrate that training on experimental phantoms can restore the correlation of signal amplitudes measured in depth. While the absolute quantification error remains high and further improvements are needed, our results highlight the promise of deep learning to advance quantitative PAI.

Effects of skin tone on photoacoustic imaging and oximetry.

Significance: Photoacoustic imaging (PAI) provides contrast based on the concentration of optical absorbers in tissue, enabling the assessment of functional physiological parameters such as blood oxygen saturation (sO2). Recent evidence suggests that variation in melanin levels in the epidermis leads to measurement biases in optical technologies, which could potentially limit the application of these biomarkers in diverse populations. Aim: To examine the effects of skin melanin pigmentation on PAI and oximetry. Approach: We evaluated the effects of skin tone in PAI using a computational skin model, two-layer melanin-containing tissue-mimicking phantoms, and mice of a consistent genetic background with varying pigmentations. The computational skin model was validated by simulating the diffuse reflectance spectrum using the adding-doubling method, allowing us to assign our simulation parameters to approximate Fitzpatrick skin types. Monte Carlo simulations and acoustic simulations were run to obtain idealized photoacoustic images of our skin model. Photoacoustic images of the phantoms and mice were acquired using a commercial instrument. Reconstructed images were processed with linear spectral unmixing to estimate blood oxygenation. Linear unmixing results were compared with a learned unmixing approach based on gradient-boosted regression. Results: Our computational skin model was consistent with representative literature for in vivo skin reflectance measurements. We observed consistent spectral coloring effects across all model systems, with an overestimation of sO2 and more image artifacts observed with increasing melanin concentration. The learned unmixing approach reduced the measurement bias, but predictions made at lower blood sO2 still suffered from a skin tone-dependent effect. Conclusion: PAI demonstrates measurement bias, including an overestimation of blood sO2, in higher Fitzpatrick skin types. Future research should aim to characterize this effect in humans to ensure equitable application of the technology.

Distribution-informed and wavelength-flexible data-driven photoacoustic oximetry.

Significance: Photoacoustic imaging (PAI) promises to measure spatially resolved blood oxygen saturation but suffers from a lack of accurate and robust spectral unmixing methods to deliver on this promise. Accurate blood oxygenation estimation could have important clinical applications from cancer detection to quantifying inflammation. Aim: We address the inflexibility of existing data-driven methods for estimating blood oxygenation in PAI by introducing a recurrent neural network architecture. Approach: We created 25 simulated training dataset variations to assess neural network performance. We used a long short-term memory network to implement a wavelength-flexible network architecture and proposed the Jensen-Shannon divergence to predict the most suitable training dataset. Results: The network architecture can flexibly handle the input wavelengths and outperforms linear unmixing and the previously proposed learned spectral decoloring method. Small changes in the training data significantly affect the accuracy of our method, but we find that the Jensen-Shannon divergence correlates with the estimation error and is thus suitable for predicting the most appropriate training datasets for any given application. Conclusions: A flexible data-driven network architecture combined with the Jensen-Shannon divergence to predict the best training data set provides a promising direction that might enable robust data-driven photoacoustic oximetry for clinical use cases.