There is an urgent need for imaging procedures that can be used in the intraoperative recognition of tumor borders. Coherent anti-Stokes Raman scattering (CARS) enables label-free, high-resolution imaging. This paper describes a rigid, compact CARS endoscope with a spatial resolution of 800 nm and a field of view of 250 µm at a working distance of 187 mm.

By P. Zirak // G. Matz // B. Messerschmidt // T.  Meyer // M.  Schmitt // J. Popp // O. Uckermann // R. Galli // M. Kirsch // M.  Winterhalder // A.  Zumbusch

The intraoperative differentiation between diseased and normal tissue in the human brain is very difficult due to the lack of imaging techniques that can specifically visualize cancerous tissue. Although there are established methods such as magnetic resonance imaging and computer tomography, which can only rarely be used intraoperatively, and fluorescence-based methods, which do not provide a clear signal for every tumor type, a reliable assessment of the tumor’s border in particular is often only possible after the tissue has been removed and the subsequent histopathological examination performed. Stain-free and molecularly sensitive coherent anti-Stokes Raman scattering (CARS) microscopy, especially in combination with other non-linear optical imaging techniques such as second harmonic generation (SHG) and two-photon-excited fluorescence (TPEF), is a novel optical technique that can make intraoperative tumor edge detection possible.

Within the scope of the BMBF-funded joint project endoCARS, both multimodal imaging and especially CARS imaging were studied for their effectiveness as tools in distinguishing the tumor from healthy tissue and representing different parts of normal brain tissue. In order to establish the technology as a fast and reliable endoscopic procedure for diagnosis and therapy in everyday clinical practice, new endomicrooptics and laser sources were researched and subsequently integrated into a compact overall system shown schematically in Figure 1. Figure 2 shows a photograph of the entire system and of the GRIN optics that came into contact with a section of tissue during test measurements.

The CARS endoscopy system implemented as part of this project consists of a nonlinear endomicroscope optic with a gradient index (GRIN) optic that is color corrected for the near-infrared range and has a high NA of 0.5 at a total length of almost 20 cm and an outer diameter of only 2.2 mm. The GRIN optic was researched and developed by Grintech and consists of a compact fiber laser source from TOPTICA that can be electronically tuned in the range of the CH stretching vibration of 2700-3200 cm-1and a compact laser scanning microscope that was researched and developed in close collaboration with Leibniz IPHT, the University of Constance, and Grintech. The innovative GRIN optic has many advantages over conventional lenses: The surfaces are flat, and the refractive index profile can be varied in a large range via diffusion during production. This allows a very good optical image to be achieved despite a small diameter of 500 µm, for example. The CARS endoscope provides images with a resolution of 800 nm at a field of view of 250 µm, which is why this high-resolution endomicroscope is combined with a neurosurgical applicator (STORZ) for wide-field imaging in order to precisely select and approach the target region for the use of the endoscope. To demonstrate the excellent optical resolution, Figure 3 shows CARS images of adipocytes and brain tissue. To determine the resolution, CARS images of polystyrene spheres with a diameter of 1 µm (left) and 3 µm (right) at the CH stretching vibration at 3050 cm-1were taken in Figure 4. Since the spheres can be visualized with a diameter of 1 µm, a resolution of better than 1 µm is achieved as originally intended.

The long working distance of 20 cm and the small outer diameter also allow use in minimally invasive neurosurgical procedures. The demonstrator is currently being used in neurosurgery at TU Dresden for the creation of reference data sets and the pre-clinical evaluation of the functional pattern.

Funded by: BMBF