Computational Imaging Tools Improve Lateral OCT Resolution

A new technique called optical coherence refraction tomography (OCRT) is able to increase the resolution of OCT down to a single micrometer in all directions, even in a living patient.

Because OCT provides better depth resolution than lateral direction, it has traditionally worked best when the tissue being imaged contains mostly flat layers. To extend the power of OCT for live imaging of tissues throughout the body, researchers at Duke University needed to overcome the trade-off between lateral resolution and depth of imaging. The researchers, led by professor Joseph Izatt, combined OCT images acquired from multiple angles to extend the depth resolution to the lateral dimension. In doing so, they found that individual OCT images became distorted by light refracted through irregularities in the cells and other tissue components. To compensate for these altered paths of light, they modeled how the light was bending as it passed through the sample. To develop an accurate model, they used machine learning.

The researchers used “gradient-based optimization” to infer the refractive index within the different areas of tissue based on multi-angle images. This approach allowed the researchers to determine the direction in which the refractive index needed to be adjusted for a better image to be created.

After many iterations, the algorithm created the map of the tissue’s refractive index that could best compensate for refractions in the light. The method was implemented using TensorFlow, a software library created by Google for deep learning applications. In correcting refraction-induced distortions to register OCT images, the new technique also achieves spatially resolved refractive index imaging.

The image quality of a normal OCT scan (left) and a new OCRT scan (right) are demonstrated with a mouse vas deferens sample. Note how the OCT scan quickly deteriorates with depth while the OCTR scan produces a complete image (top), and note the increase in fine detail and reduction in noise between the two (bottom). Courtesy of Kevin Zhou, Duke University.

For their proof-of-concept experiments, the researchers took tissue samples from the bladder and trachea of a mouse, placed the sample in a tube, and rotated the sample 360° beneath an OCT scanner. The algorithm successfully created a map of each sample’s refractive index, increasing the lateral resolution of the scan by more than 300% while reducing the background noise in the final image. While the study used samples already removed from the body, the researchers believe OCRT could be adapted to work in a living organism. “Rather than rotating the tissue, a scanning probe developed for this technique could rotate the angle of the beam on the tissue surface,” researcher Kevin Zhou said.

The team is investigating how much a corneal scan could be improved by the technology with less than a 180° sweep, and the results appear promising, Zhou said. If successful, the OCRT technique could be applicable to many medical imaging needs beyond ophthalmology.

“OCT has already revolutionized ophthalmic diagnostics by advancing noninvasive microscopic imaging of the living human retina,” Izatt said. “We believe that with further advances such as OCRT, the high impact of this technology may be extended not only to additional ophthalmic diagnostics, but to imaging of pathologies in tissues accessible by endoscopes, catheters, and bronchoscopes throughout the body.”

“Capturing high-resolution images of the conventional outflow tissues in the eye is a long sought-after goal in ophthalmology,” professor Sina Farsiu said. “Having an OCT scanner with this type of lateral resolution would be very important for early diagnosis and finding new therapeutic targets for glaucoma.”

Previous attempts at creating OCT images with high lateral resolution have relied on holography, the researchers said. The holographic approach requires the sample and imaging apparatus to remain completely still down to the nanometer scale throughout the measurement process — a challenge to achieve in living, moving tissues.

The research was published in Nature Photonics (https://doi.org/10.1038/s41566-019-0508-1).

As the tissue sample on the left rotates under a traditional OCT scan, computational imaging gradually builds the OCRT image on the right until the resolution has peaked in all directions. Courtesy of Kevin Zhou, Duke University.