Development of multi-layered skin phantoms with embedded perfusable microvasculature for Optical Coherence Tomography Angiography

Tissue-simulating phantoms are crucial in the development of optical imaging techniques, such as Optical Coherence Tomography (OCT), in order to test and compare systems performance, validate simulations and evaluate repeatability and reproducibility. One of the most successful OCT functional extensions is OCT Angiography (OCTA), capable of visualizing the blood flow. Thus, the ideal OCTA phantom should mimic the morphology and optical properties of tissue as well as the vasculature shapes and dimensions with the possibility to perfuse it. In this work, skin mimicking optical phantoms were developed involving both 3D-printing and casting techniques. Firstly, several vessel models were designed tracing the morphology of healthy and pathological microvasculature. Two photons polymerization (2PP) 3D-printing allowed to print hollow vessels with a resolution down to 0.5-8 µm, depending on the exploited magnification objective. Then, polydimethylsiloxane (PDMS) with different concentrations of black and white silicone pigments was used to simulate the multi-layered skin architecture mimicking the optical properties of both dermis and epidermis. Finally, through an infusion pump, the phantom was accurately perfused with diluted whole milk during OCT volume acquisitions so that OCTA volume could be extracted. Despite the challenge of faithfully reproducing the real microvasculature, capillary-mimicking vessels were printed with inner diameters from 600 µm down to 10 µm and different levels of morphological complexity were obtained, from a simple three-branches model to an entire capillary network. Also, designing branches at different depths, the penetration depth value of 1.5 mm could be assessed for the employed system. The resin vessel wall thickness was set to be 20 µm and its effect on OCT and OCTA was evaluated without showing significant distortion. Considering the absorption and reduced scattering coefficients values and the OCT imaging appearance for different concentrations of black and white pigments in PDMS, their percentages in weight were respectively set to be 0.18% and 0.36% for dermis and 0.20% and 0.72% for epidermis. In order to properly process the acquired signals and validate the OCTA results, the relationship between OCTA signal intensity and both flow velocity and interscan time (i.t.) was thoroughly investigated obtaining that, for each i.t., the OCTA signal averagely reaches a common saturation level, around the value of 0.3 for the intensity-based logarithmic OCTA. Moreover, the i.t. influences the flow velocity from which the saturation level is observed as well as the ability to discriminate velocities. Considering the results from a velocities range between 0.1 and 80 mm/s, an i.t. below 1 ms should be implemented to distinguish slow and fast velocities. Thus, since the A-scan rate of 222.22 kHz combined with the BM-scan protocol of the employed system leads to an i.t. of about 4.6 ms, no differences in velocities were observed for the full volume acquisitions. Also, intensity-based OCTA turned out to be more suitable for OCTA extraction on the developed phantoms. The OCTA volumes were then segmented and compared with the printed models in terms of overlapping percentage, showing the potential of this phantom fabrication approach to test OCT systems as well as OCTA algorithms. Further developments include the evaluation of these phantoms to be potentially used as multimodal imaging phantoms considering more imaging techniques, such as photoacoustic imaging(PAI).