Current medical imaging technologies allow visualization of tissue anatomy in the human body at resolutions ranging from 100 micrometers to 1 millimeter. These technologies are generally not sensitive enough to detect early-stage tissue abnormalities associated with diseases such as cancer and atherosclerosis, which require micrometer-scale resolution. Here, optical coherence tomography was adapted to allow high-speed visualization of tissue in a living animal with a catheter-endoscope 1 millimeter in diameter. This method, referred to as "optical biopsy," was used to obtain cross-sectional images of the rabbit gastrointestinal and respiratory tracts at 10-micrometer resolution.

In Vivo Endoscopic Optical Biopsy with Optical Coherence Tomography

Deep learning approaches to biomedical image segmentation

In this paper, we review the current state of technology development and clinical applications of endoscopic optical coherence tomography (OCT). Key design and engineering considerations are discussed for most OCT endoscopes, including side-viewing and forward-viewing probes, along with different scanning mechanisms (proximal-scanning versus distal-scanning). Multi-modal endoscopes that integrate OCT with other imaging modalities are also discussed. The review of clinical applications of endoscopic OCT focuses heavily on diagnosis of diseases and guidance of interventions. Representative applications in several organ systems are presented, such as in the cardiovascular, digestive, respiratory, and reproductive systems. A brief outlook of the field of endoscopic OCT is also discussed.

Endoscopic optical coherence tomography: Technologies and clinical applications [invited]

In recent years, Light Detection and Ranging (LiDAR) has been drawing extensive attention both in academia and industry because of the increasing demand for autonomous vehicles. LiDAR is believed to be the crucial sensor for autonomous driving and flying, as it can provide high-density point clouds with accurate three-dimensional information. This review presents an extensive overview of Microelectronechanical Systems (MEMS) scanning mirrors specifically for applications in LiDAR systems. MEMS mirror-based laser scanners have unrivalled advantages in terms of size, speed and cost over other types of laser scanners, making them ideal for LiDAR in a wide range of applications. A figure of merit (FoM) is defined for MEMS mirrors in LiDAR scanners in terms of aperture size, field of view (FoV) and resonant frequency. Various MEMS mirrors based on different actuation mechanisms are compared using the FoM. Finally, a preliminary assessment of off-the-shelf MEMS scanned LiDAR systems is given.

MEMS Mirrors for LiDAR: A review.

Intrinsic, three-dimensionally resolved, microscopic imaging of dynamical structures and biochemical processes in living preparations has been realized by nonlinear laser scanning fluorescence microscopy. The search for useful two-photon and three-photon excitation spectra, motivated by the emergence of nonlinear microscopy as a powerful biophysical instrument, has now discovered a virtual artist's palette of chemical indicators, fluorescent markers, and native biological fluorophores, including NADH, flavins, and green fluorescent proteins, that are applicable to living biological preparations. More than 25 two-photon excitation spectra of ultraviolet and visible absorbing molecules reveal useful cross sections, some conveniently blue-shifted, for near-infrared absorption. Measurements of three-photon fluorophore excitation spectra now define alternative windows at relatively benign wavelengths to excite deeper ultraviolet fluorophores. The inherent optical sectioning capability of nonlinear excitation provides three-dimensional resolution for imaging and avoids out-of-focus background and photodamage. Here, the measured nonlinear excitation spectra and their photophysical characteristics that empower nonlinear laser microscopy for biological imaging are described.

Multiphoton fluorescence excitation: New spectral windows for biological nonlinear microscopy

This Letter reports a fully packaged microelectromechanical system (MEMS) endoscopic catheter for forward imaging optical coherence tomography (OCT). Two-dimensional optical scanning of Lissajous patterns was realized by the orthogonal movement of two commercial aspherical glass lenses laterally mounted on two resonating electrostatic MEMS microstages at low operating voltages. The MEMS lens scanner was integrated on a printed circuit board and packaged with an aluminum housing, a gradient index fiber collimator, and an objective lens. A home-built spectral-domain OCT system with 60 kHz A-line acquisition rate was combined with the endoscopic MEMS catheter. Three-dimensional images of 256×256×995 voxels were directly reconstructed by mapping the A-line datasets along the Lissajous patterns. The endoscopic catheter can provide a new direction for forward endoscopic OCT imaging.

Forward imaging OCT endoscopic catheter based on MEMS lens scanning.

We report a fully packaged and compact forward viewing endomicroscope by using a resonant fiber scanner with two dimensional Lissajous trajectories. The fiber scanner comprises a single mode fiber with additional microstructures mounted inside a piezoelectric tube with quartered electrodes. The mechanical cross-coupling between the transverse axes of a resonant fiber with a circular cross-section was completely eliminated by asymmetrically modulating the stiffness of the fiber cantilever with silicon microstructures and an off-set fiber fragment. The Lissajous fiber scanner was fully packaged as endomicroscopic catheter passing through the accessory channel of a clinical endoscope and combined with spectral domain optical coherence tomography (SD-OCT). Ex-vivo 3D OCT images were successfully reconstructed along Lissajous trajectory. The preview imaging capability of the Lissajous scanning enables rapid 3D imaging with high temporal resolution. This endoscopic catheter provides many opportunities for on-demand and non-invasive optical biopsy inside a gastrointestinal endoscope.

/pdf/lissajous-fiber-scanning-for-forward-viewing-optical-9a1402tc00.pdf

Lissajous fiber scanning for forward viewing optical endomicroscopy using asymmetric stiffness modulation

We report a novel MEMS fiber scanner with an electrothermal silicon microactuator and a directly mounted optical fiber. The microactuator comprises double hot arm and cold arm structures with a linking bridge and an optical fiber is aligned along a silicon fiber groove. The unique feature induces separation of resonant scanning frequencies of a single optical fiber in lateral and vertical directions, which realizes Lissajous scanning during the resonant motion. The footprint dimension of microactuator is 1.28 x 7 x 0.44 mm3. The resonant scanning frequencies of a 20 mm long optical fiber are 239.4 Hz and 218.4 Hz in lateral and vertical directions, respectively. The full scanned area indicates 451 μm x 558 μm under a 16 Vpp pulse train. This novel laser scanner can provide many opportunities for laser scanning endomicroscopic applications.

/pdf/electrothermal-mems-fiber-scanner-for-optical-endomicroscopy-3vvm289i39.pdf

Electrothermal MEMS fiber scanner for optical endomicroscopy.

We report parallel-trained deep neural networks for automated endoscopic OCT image segmentation feasible even with a limited training data set. These U-Net-based deep neural networks were trained using a modified dice loss function and manual segmentations of ultrahigh-resolution cross-sectional images collected by an 800 nm OCT endoscopic system. The method was tested on in vivo guinea pig esophagus images. Results showed its robust layer segmentation capability with a boundary error of 1.4 µm insensitive to lay topology disorders. To further illustrate its clinical potential, the method was applied to differentiating in vivo OCT esophagus images from an eosinophilic esophagitis (EOE) model and its control group, and the results clearly demonstrated quantitative changes in the top esophageal layers’ thickness in the EOE model.

Parallel deep neural networks for endoscopic OCT image segmentation

This work reports micromachined tethered silicon oscillators (MTSOs) for endoscopic Lissajous fiber scanners. An MTSO comprises an offset silicon spring for stiffness modulation of a scanning fiber and additional mass for modulation of resonant scanning frequency in one body. MTSOs were assembled with a resonant fiber scanner and enhanced scanning reliability of the scanner by eliminating mechanical cross coupling. The fiber scanner with MTSOs was fully packaged as an endomicroscopic catheter and coupled with a conventional laparoscope and spectral domain OCT system. The endomicroscope was maneuvered with the integrated laparoscope and in vivo swine tissue OCT imaging was successfully demonstrated during open surgery. This new component serves as an important element inside an endoscopic Lissajous fiber scanner for early cancer detection or on-demand minimum lesional margin decision during noninvasive endoscopic biopsy.

Hyeon Cheol Park

Papers

Forward imaging OCT endoscopic catheter based on MEMS lens scanning.

Lissajous fiber scanning for forward viewing optical endomicroscopy using asymmetric stiffness modulation

Electrothermal MEMS fiber scanner for optical endomicroscopy.

Parallel deep neural networks for endoscopic OCT image segmentation

Micromachined tethered silicon oscillator for an endomicroscopic Lissajous fiber scanner