Researchers from the Applied Optics Group at the University of Kent and the Hamlyn Centre for Robotic Surgery at Imperial College London have recently published in the Journal of Biomedical Optics the results of collaborative development and validation of a robotic multi-modal imaging system.

The system was designed for surgical imaging, has a narrow tip with a diameter of 3mm, and integrates full 3-D optical coherence tomography (OCT) imaging alongside fluorescence endomicroscopy. Fibre bundle endomicroscopes typically have excellent transversal resolution but limited ability to image below the surface, whereas OCT systems can resolve scattering centres (layers) with micrometric resolution over a range of 1 to 2 mm in depth, but have more limited transversal resolution, making the combination of the techniques particularly advantageous for real time tissue analysis.

The compact system features a 2-D robotic scanner developed at the Hamlyn Centre. The scanner is high speed, has accurate, repeatable positioning, and has a large lateral range of up to 3 mm. It includes two working channels through which the en-face OCT imaging probe and the fluorescence endomicroscope were inserted, allowing simultaneous imaging.

Standalone endoscopic en-face and 3D imaging OCT imaging probes are difficult to fabricate due to the complexity of miniaturising a 2-axis scanning system. 2D systems without miniaturised scanners have been developed based on optical fibre bundles, but the image quality is poor due to cross-talk between adjacent cores, poor efficiency, and the multi-mode behaviour of the fibre. In the approach of the Kent and Imperial researchers, a forward-viewing 1-axis line-scanning endoscopic probe was inserted into the robotic scanner and one axis of the robot’s motion was used as the second OCT axis, allowing en-face and 3D OCT images to be generated. This enabled direct and real time en-face visualisation.

The fluorescence endomicroscope features a multicore (30,000) fibre optic bundle, camera, dichroic, objective, and filters. The fibre bundle uses no distal optics, allowing for the diameter of this probe to be less than 1mm. However, having no distal optics means that the field of view is normally limited to the size of the bundle (600 µm-diameter area). Using the robotic scanner and a mosaicking algorithm it was instead possible to generate an effective field of view of 0.53 x 3.25 mm2 .

The en-face OCT images are acquired simultaneously with the fluorescence endomicroscopy images, meaning one sweep of the robotic arm yields a mosaicked endomicroscope image and an OCT image. The probe has been validated with ex-vivo tissue imaging, using both ovine kidney and porcine lung and oesophagus, and also in-vivoon human skin tissue using both a mounted and a hand-held configuration. For more information, please check the article page.