Spatial Frequency Domain Imaging of a Facial Tissue Simulating Phantom of Known Optical Properties

Abstract

Purpose : Spatial frequency domain imaging (SFDI) is a wide-field optical imaging modality that applies patterned illumination at multiple center wavelengths (470-851 nm) to quantify the properties of subsurface tissue. Its application to the periocular region facilitates understanding of inflammatory conditions by quantifying parameters such as tissue oxy- and deoxy hemoglobin concentrations. The periocular region is unique due to complex ocular adnexal anatomy. We explored the potential geometric impact of neighboring bony and soft tissue regions on the accuracy of SFDI measurements by imaging a fabricated facial phantom to identify robust and accurate regions of interest (ROI) to image in human patients.

Methods : We cast a polydimethylsiloxane-based tissue-simulating facial phantom having homogeneous known optical properties that mimic those of human skin. The phantom was placed on a chin rest and imaged nine times from an ‘en face’ or ‘side profile’ position. To test reproducibility in a realistic way, the phantom was removed and replaced on the chin rest between each pair of measurements. The flat back of the phantom was measured after being placed on a cart 15 times. Mean optical properties were calculated for ROIs: inferior temporal quadrant (ITQ), inferior nasal quadrant (INQ), central eyelid margin (CEM), rostral lateral nasal bridge (RLNB), and forehead (FH)

Results : The measured absorption and reduced scattering coefficients are reproducible and accurate when comparing the ITQ, INQ, and FH to its flat posterior surface (reference point that lacks curvature and height variation). The ITQ and INQ were similar when compared to each other and the FH (two sample test p-value >0.05), and the imaging orientation did not impact the accuracy of the measurements for these ROIs. Regions of extreme topographical variation, i.e., CEM and RLNB, did exhibit differences in measured optical properties, identifying areas to avoid conducting SFDI in patients.

Conclusions : This study suggests that the ITQ and INQ of a generalized face with its relative elevation and adjacent structures has minimal impact on SFDI measurement accuracy. Areas with heightened topographic variation, such as the RLNB and CEM, do exhibit variability in measurement. These findings confirm the need to carefully select an optimal ROI to minimize the effects of topography when measuring optical properties along the periocular region.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.