07 Jun 2022
Computational approach combines multiple OCT volumes to reveal 3D features.
Developed at Duke University and described in Optica, the technique can image biomedical samples at higher contrast and resolution over a wider 3D field of view, potentially enabling more accurate diagnostic imaging.
Optical coherence refraction tomography (OCRT) is "a computational extension of OCT that synthesizes an incoherent contrast mechanism by combining multiple OCT volumes, acquired across two rotation axes, to form a resolution-enhanced, speckle-reduced, refraction-corrected 3D reconstruction," commented the Duke project team.
This approach is designed to combine the coherent detection sensitivity advantages of OCT with a speckle-free incoherent contrast mechanism analogous to that of incoherent microscopy, in addition to enhanced lateral resolution over an extended field of view.
"OCT is a volumetric imaging technique widely used in ophthalmology and other branches of medicine," said Duke University's Kevin Zhou. "We developed a new and exciting extension, featuring novel hardware combined with a new computational 3D image reconstruction algorithm to address some well-known limitations of the imaging technique."
Previous enhancements to OCT developed in Duke University labs have included the use of a dual-axis architecture to improve imaging in cases where a high degree of scattering limits the depth penetration of existing OCT devices. OCRT involves its own novel optomechanical design featuring a parabolic mirror as the imaging objective. This allows OCT volumes to be acquired from multiple views covering up to ±75 degrees without moving the sample itself.
"Our approach is the first experimental demonstration of a more general class of conic mirror-based methods that can in principle acquire images from multiple views over up to ±90 degrees across two rotation axes using low-inertia scanners as the only moving parts," commented the project in its paper.
New forms of image contrast
This approach has consequences for the computation operations required, since the imaging data sets are large and effectively 5D - three spatial dimensions plus two angular dimensions. As a result the project also needed to develop a novel 3D reconstruction algorithm that "leverages differentiable programming frameworks and optimization techniques developed in the deep learning community."
The team applied OCRT to biological samples including a zebrafish and fruit fly, along with mouse tissue samples of the trachea and esophagus to demonstrate the potential for medical diagnostic imaging. Improved image quality revealed new details in each case.
In mouse trachea imaging, for example, 3D OCRT "offers substantial improvements over conventional OCT revealing several layers not readily apparent," commented the project. "The large speckle grains in OCT obscure these layers."
The Duke team predicts that 3D OCRT could see wide use for in vivo biomedical imaging, requiring only conceptually straightforward improvements to existing platforms to open up new forms of image contrast.
"We envision this approach being applied in a wide variety of biomedical imaging applications, such as in vivo diagnostic imaging of the human eye or skin," commented project co-leader Joseph Izatt. "The hardware we designed to perform the technique can also be readily miniaturized into small probes or endoscopes, to access the gastrointestinal tract and other parts of the body."