Whole+Slide+Scanners

WHOLE SLIDE SCANNERS
Whole slide scanners allow the user to load a complete box of slides, e.g., 100 slides 1" x 3" (25 x 75 mm), into a slide carrier. The recent models of slide scanners will automatically collect a digital image in RGB, transfer it to a computer, and allow you to view the images over the internet. The most widely used units at this time (July 2010) are the Aperio Scanner and the Nanozoomer (Hamamatsu/Olympus). They are typically used with a 20X Olympus objective, with a numerical aperture (n.a.) of 0.75. Unfortunately, for reasons that are not adequately explained the manufacturers of these units have not used suitably matched condensers of similar n.a. As a result, those of you with experience with microscopes will readily note excessive refractility in the resulting images.

The slides are placed on a motorized stage by the slide autoloader. A low power image of each slide is captured, including the label on the slide. The slide is then scanned using one of several different technologies: this may consist of an array of three linear detectors, or a high resolution video camera. There is typically a separate detector for each major color, Red, Green and Blue. The linear detector is ca. 1,000 pixel width. The resulting pattern of color and brightness is then digitized and stored on a hard disk. Images are collected in vertical lanes, along the Y-axis of the slide. The stage then moves back to the top of the slide and over approx 960 pixels. This process is repeated until the desired region of the slide is fully digitized. The resulting image is assembled with overlapping lanes to assist in alignment and to prepare a nearly seamless image of the whole slide. However, Nanozoomer 2.0 apparently has sufficiently high accuracy in stage positioning that they claim that they do not have any overlap of lanes. [!]

I recently (October 25, 2010) attended the Pathology Visions meeting in San Diego, where several manufacturers of WSI (Whole Slide Imaging) devices exhibited their machines. Typical resolution is 0.5 micrometers/pixel. Most manufacturers use a 20X Objective.


 * Aperio**: Pioneer in the field. Separate units for Bright Field. 20X Olympus Objective. Autoloader for 120 1" x 3" slides. Manual loader for 2" x 3" slides. Reliable, good images. Free viewing software. They also make a flourescent scanner, but only handles five 1" x 3" slides. Unable to reliably read sections thicker than 15 micrometers. (We routinely use 30 micrometer thick sections). Separate filter cubes for each wavelength in fluorescence, results is lovely images, but multispectral scans are extremely slow. If you only need bright field, this is a reasonable unit to consider.

Scan time in bright field 20 x 20 mm area in 100 seconds.Changing fluorescent filter cubes is clumsy. Relies on Semrock triple cube for multiple simultaneous fluorescence; some breakthrough problems. Scan time for fluorescence depends upon brightness of fluorescent emission. Typical time for a 20 x 44 mm area with bright specimen is 10- 20 minutes. A new generation Hamamatsu will be on exhibit at the SfN 2010 meeting. It is a significant redesign, using filter wheel for fluorescence, with better spectral separation, and a newly designed fiber optic light source. This is a workhorse of a unit.
 * Hamamatsu/Olympus Nanozoomer 2.0**: Bright Field and Fluorescence. Enclosed unit. Autoloader only accepts 1" x 3" slides. Seven cassettes. Each cassette holds 30 slides for total of 210 slides. 20X Olympus Objective. Excite light source with fiber optic coupling for flourescence. Halogen for bright field. No dark field at present.


 * Leica - SCN** - Attractive all in one unit. New auto slide loader holds approx. 380 slides, 1" x 3". Bright Field only, but claim that they will have a fluorescent unit in 2011. Reported speed of scanning is 15 x 15 mm area in 100 seconds. Carrier for 2" x 3" slides, but slides are restricted to glass slides purchased from Leica. Cannot scan archived slides on 2" x 3" standard glass.


 * Olympus VS-110**: Microscope based platform. Bright Field, Fluorescence. Claim that they will soon have Dark Field. Multiple objectives. No autoloader at present, but they promise an autoloader for 100 slides, 1" x 3", in near future. Presently uses a stage that holds up to 4 or 5 slides. No exact measure of time to scan. Dependent upon magnification. Estimated time for fluorescent scan is about 4 - 5 times longer than Leica and Hamamatsu. Fluorescent images very clean, but acquisition is very slow. The representative indicated that there will be a new model in 2011, the VS-220. Scan times will be much faster, with autoloader. Good software.


 * Zeiss: Mirax**. Nice images, but very slow. They do not appear to be serious contenders in this game.


 * BioZIP**: Microscope based. Small company in San Diego. No further info. Looks like it would be worth getting more info.


 * CRI- 3DHisTech** - 1" x 3" fluorescent. Slow scan speeds. Need additional information


 * Huron Technologies** - formerly Confocal.com: Did not have an actual unit on display. They stand out for their claim of being able to scan a complete section of a human brain in 4 hours (presumably at 0.5 um/pixel). They claim that their next generation unit will allow scanning of whole human brain on 6" x 8" slide in 30 minutes. If so, this is a unit worth considering for our project.

My own experience is largely limited to the Aperio for Bright Field, and the Hamamatsu 2.0 for both Bright Field and Fluorescence. Throughput times are critical. The only current unit suitable for high throughput fluorescence is the Hamamatsu.

[More Later - HJK]

H.J.Karten; October 28, 2010