Most of the curves related to the dermis optical properties show high standard deviation bars that are overlapped for the consecutive values (unlike the epidermis), thus hardening the interpretation of the observable kinetic changes (Figure 4 and 5 (d, e, f), Table 2). This is possibly due to the combined effect of measurement uncertainties propagated through the Jacques model. Additional reason might be the unstable LC-OCT probe positioning, which was constantly reapplied to tested skin sites without the possibility to use any kind of adhesive position-tracking labels due to mechanical manipulations with the skin surface in the beginning of the experimental protocol, making it impossible to analyze exactly the same tissue volume throughout the experiment. Another reason might be the morphological variation between the dermal layers of the volunteers and a more significant contribution of multiple scattering to the dermal layer linear fit (as compared to epidermis).^[16]

The uncertainties mentioned in the previous paragraph are described in detail in the work of Waszczuk et al .^[16]First, there are uncertainties on the optical properties, estimated for a calibration phantom using double integrating spheres, which are then used in LC-OCT images processing. Second, there is an effect of both linear fit accuracy and variation between different images of the same sample on the µ _eff and ρ parameters. Finally, all these uncertainties affect the resulting LC-OCT estimation of scattering coefficient and scattering anisotropy factor. However, despite it was mentioned by Waszczuk et al . that when propagating through the model of Jacques the errors were amplified for high values of µ _s(λ ₈₀₀) andg(λ ₈₀₀) , this method is still well suited to samples with relatively high scattering anisotropy (g(λ ₈₀₀) factor ranged between 0.7 and 0.9) and scattering coefficientsµ _s(λ ₈₀₀) up to 12 mm^-1, i.e. biological tissues such as skin.

Epidermis scattering coefficient demonstrated ~ 20% decrease for the control condition. But with a mean standard deviation over time of 25% of the observed values (Table 2) and the significant overlapping, it can be concluded that control conditions in general do not significantly affect the skin optical properties. However, it can also be seen that some OCA caused noticeable clearing effect in dermis. For example, a mixture of both sugars with DMSO caused a decrease in epidermis scattering \(µ_{s}^{\text{epi}}\) (Figure 4(b)). Sucrose/DMSO mixture effect resulted into 28±18% decrease after 21 minutes of experimental protocol. Such decrease overcomes a moderate overall standard deviation (mean ~12% for all timepoints). Mixture of Glucose and DMSO demonstrated more pronounced effect – 40±3% of epidermal scattering decrease (from 8.3 ± 1.3 mm^-1 to 5 ± 0.9 mm^-1 for three volunteers), which is greater than the mean variation of data over the time for this OCA – 10%. Both mixtures demonstrated as well an increase in scattering anisotropy of a dermal layer. However, as it was mentioned above and can be seen in Table 2, dermis values have a relatively high standard deviation, making it onerous to conclude that there is a clearing effect on dermis. So, only the general behavior can be mentioned. The result for Glucose/DMSO mixture is in a good agreement with observations made in our previous publication, where this mixture was considered as one of the most efficient among nine OCA in terms of in-depth increase of LC-OCT image intensity and contrast.^[26] Such increase was assumed to be caused by the epidermis scattering decrease, that is quantitatively assessed in the present study. Other notable results are related to the mixtures of Glucose and PEG with PG as a permeation enhancer. Glucose/PG mixture (Figure 4(c)) caused 23% decrease of epidermal scattering with the mean ~11% data variation over time and PEG/PG mixture caused 33% decrease with ~10% mean data variation.

PEG/OA/PG mixture demonstrated the most pronounced\(µ_{s}^{\text{epi}}\) decrease (Figure 4(a), blue triangle data points). Already after 21 minutes of experimental protocol (10 minutes after ultrasound-assisted clearing was over), this parameter decreased by 33 ± 17% for 3 volunteers, from 9.9 ± 1.1 mm^-1down to 6.6 ± 1.7 mm^-1. It can be seen that the decrease pattern in this case is the most confident among the other OCA. Moreover, for the first 6 measurement points (up to t = 21 min), the mean standard deviation is only ~5% of the data (overall SD is ~9% of the data). Considering all the mentioned uncertainties and high standard deviation bars in the case of other mixtures, PEG/OA/PG mixture demonstrated the most significant clearing effect. That is as well in good agreement with previous observations,^[26] where PEG/OA/PG mixture demonstrated the best increase (40%) in image in-depth intensity and contrast after 10 minutes of ultrasound-assisted clearing.

Concerning the scattering anisotropy parameter, none of the OCA demonstrated significant changes in g ^epidermiswith relatively low SD values. Together with decrease in\(µ_{s}^{\text{epi}}\) one can conclude that OCA clearing effect on epidermis expressed and limited to matched RI of scattering particles and interstitial fluid. This RI matching effect is a well-investigated cause of reduction in tissue scattering mentioned in the literature.^[17] On the contrary, dermal scattering does not change significantly due to high SD, but the tendency of dermal scattering anisotropy g ^dermis towards increase indicates a different influence of tested OCA on skin dermis than to the epidermis in vivo . Mie theory explains that such behavior is related to an increase in the size of the scattering particles rather than to RI matching.^[19] This applies to the dermis as there are bundles of collagen, which are probably swollen under the action of OCA mixtures, causing light to scatter in the forward direction.

Our results obtained on in vivo human skin demonstrate the possibility of LC-OCT quantitative estimation of changes in skin scattering coefficientµ _s(λ ₈₀₀) and scattering anisotropy factor g(λ ₈₀₀) , caused by biocompatible optical clearing (involving the reduced concentrations of clearing agents used together with chemical and physical permeation enhancers). Moreover, the mixture of PEG/OA/PG has shown the best results. Although our current results were not validated with integrating spheres measurements (as the study was carried out on skin in vivo ), this was done on phantoms in the study of Waszczuket al. , demonstrating a good correlation between the values obtained using integrating spheres and application of Jacques’ model to LC-OCT images.^[16]