3D-Pathology of human tumors by light-sheet microscopy

We intend to develop a microscopical method which allows one to investigate the threedimensional histology of pathological specimens from human tumor tissue. Up to now tumors removed during cancer operations are investigated by generating thin histological sections to make e.g. sure that the whole tumor has been removed. Due to the high workload by the preparation and investigation of the histological sections in larger tumors the pathologist can investigate in general only a rather limited sample. On the contrary the light sheet microscopy of cleared samples we developed allows one to obtain quickly thousands of optical sections: First the tissue from pathology is cleared by a quick chemical procedure which makes it transparent while during the preceeding fixation process the autofluorescence of the tissue has been strongly increased. Then the transparent tumor is illuminated from the side with a very thin light sheet under the microscope and the single optical sections are recorded from above. When the preparation is now raised from below in very small steps, thousands of optical sections can be recorded with a high resolution camera. In these sections the single cells can be clearly delineated by their autofluorescence. We could already show by comparison with standard histological stainings that we can clearly differentiate between normal tissue and tumor tissue in colon cancer. In addition it is very easy to generate 3D reconstructions oft the tumors with cellular resolution. In future these 3D reconstructions should help the pathologist in his evaluation of the tumor. By this and the thousands of optical sections obtained the speed and quality of evaluation of human tumors removed during cancer operations should be markedly improved.

The aim of the project was to develop a method for imaging human tumor specimens in 3D with cellular resolution, thereby enabling safer and more accurate cancer diagnostics. We achieved this by developing a rapid clearing protocol. The imaging was performed using so-called light sheet microscopy, in which the optically cleared tumor samples are illuminated from the side with a very thin sheet of light, allowing them to be imaged from above with a microscope. This produces optical sections without the need to physically cut the specimen. These optical sections are then used to reconstruct a 3D model of the tumor on the computer, which can subsequently be used by pathologists for cancer diagnostics. This approach allows pathologists to utilize thousands of optical sections for their diagnosis, whereas conventional histology typically only evaluates a single 4 m-thick section every 5 mm. In addition, our method enables assessment of the three-dimensional structure of the tumor tissue, which was not possible with previous routine diagnostics. We therefore believe that our method of 3D pathology will be increasingly used in the future. A key factor in the success of the project was our ability to develop a rapid chemical dehydration protocol for the specimens, which allowed dehydration within a few hours, whereas previous clearing methods often required weeks due to the high cell density in tumors. Furthermore, our developments in the optics of light sheet microscopy were important, as they enabled us to generate very thin light sheets, resulting in high axial resolution within the specimen. Additionally, we identified dyes whose fluorescence is preserved even after organic tissue clearing. This allowed us to specifically stain cell nuclei and cytoplasm and, using a computer program we developed, transform these into the familiar H&E colors for pathologists. Thus, the resulting 3D images can be used directly for pathological diagnostics. Since we also identified a clearing-resistant dye for blood vessels, our method can also provide information about the tumor's blood supply. This is particularly important, as it enables assessment of the extent of neovascularization, which is of great prognostic significance but could not previously be visualized.