Most surgeries are currently performed subjectively, with outcomes that are largely dependent on the experience of the surgeon. Providing objective information about tissue that need to be resected or avoided could reduce the variability in surgical outcomes. Spatial Frequency Domain Imaging (SFDI) is a novel diffuse optical imaging method that has the potential to identify tissue viability over a large field of view. In this method, a spatial sinusoidal pattern is projected onto the tissue to get the optical properties of the tissue at each pixel. More recently, Single Snapshot of Optical Properties (SSOP) hase been developed to provide such image feedback in real-time within the specific constraints of surgery. However, while SSOP has been shown to provide information about tissues at a single wavelength in real-time, during surgical applications, it is critical to obtain spectrally-resolved functional information that can be easily interpreted by the surgeons. Optical properties at multiple wavelengths are therefore needed to correlate the absorption and scattering coefficients with the tissue functional and structural information.
In this work, we propose a novel method relying on spatio-temporal modulation of light to obtain multispectral optical properties in real time. A temporal-encoding method is used to distinguish different wavelengths by modulating each wavelength at a one particular chosen frequency. The temporally-modulated light is then used to project sinusoidal patterns onto the tissues for SSOP processing. The scene is then recorded with a fast camera to get multiple information in 3 dimensions: two spatial dimensions for SSOP and the third temporal dimension for wavelength. Discrete Fourier Transform (DFT) is used to separate the modulation frequencies pixels by pixels, and each 2D image obtained for every wavelength is processed using SSOP to obtain absorption and scattering coefficients. Finally, the optical properties at each wavelength are used to provide functional and structural information about tissues. We validated this proof of concept using 2 wavelengths (665 and 860nm) during phantom measurements and in vivo by obtaining real-time oxygenation videos. This work lays the foundation for the clinical translation of real-time quantitative multispectral imaging.
|