Leon van der Graaff and Sjoerd Stallinga published paper on “Multi-line fluorescence scanning microscope for multi-focal imaging with unlimited field of view”
Leon van der Graaff and Sjoerd Stallinga have recently published a paper in Biomedical Optics Express on “Multi-line fluorescence scanning microscope for multi-focal imaging with unlimited field of view”.
The paper proposes a new optical design for laser illumination of large area fluorescent tissue slides, in which Point Spread Function engineering using diffractive optics is used to create a set of parallel scan lines that focus at different depths inside the sample. In this way a multi-focal scan can be made of the fluorescent slide, with no moving parts other than the stage scan, and with confocality to suppress out-of-focus background.
The work has been done in collaboration with pathology experts from Leiden and Rotterdam and with support from Philips Digital & Computational Pathology and NWO-TTW.
Abstract
Confocal scanning microscopy is the de facto standard modality for fluorescence imaging. Point scanning, however, leads to a limited throughput and makes the technique unsuitable for fast multi-focal scanning over large areas. We propose an architecture for multi-focal fluorescence imaging that is scalable to large area imaging.
The design is based on the concept of line scanning with continuous ‘push broom’ scanning. Instead of a line sensor, we use an area sensor that is tilted with respect to the optical axis to acquire image data from multiple depths inside the sample simultaneously. A multi-line illumination where the lines span a plane conjugate to the tilted sensor is created by means of a diffractive optics design, implemented on a spatial light modulator. In particular, we describe a design that uses higher order astigmatism to generate focal lines of substantially constant peak intensity along the lines.
The proposed method is suitable for fast 3D image acquisition with unlimited field of view, it requires no moving components except for the sample scanning stage, and provides intrinsic alignment of the simultaneously scanned focal slices. As proof of concept, we have scanned 9 focal slices simultaneously over an area of 36 mm2 at 0.29 µm pixel size in object space. The projected ultimate throughput that can be realized with the proposed architecture is in excess of 100 Mpixel/s.