Virtual Slides

A whole-slide scanner acquires the tissue at high resolution, 20x or 40x, and that scan is the base level of the image pyramid. Each lower-resolution level comes from a 2× downsample of the one before it. A 40x scan of 100,000 × 50,000 pixels yields levels of 50,000 × 25,000, then 25,000 × 12,500, and so on down to a thumbnail.

The file divides each pyramid level into tiles that load independently, in tile-based formats such as SVS, NDPI, CZI and MRXS. Slide viewer software streams the tiles for the current zoom and loads adjacent tiles as you pan, so the multi-gigabyte file never has to sit in memory whole.

The scanning resolution sets how much information the virtual slide holds. At 40x with 0.25 µm pixels, the Nyquist limit allows detection of features down to 0.5 µm, enough for nuclear morphology and subcellular localization. At 20x with 0.5 µm pixels, features below 1 µm can't be resolved. Match the choice to the analysis: nuclear detection and biomarker quantification work at 20x, while subcellular localization, mitotic figure identification and nuclear grading may require 40x.

Most navigation needs very little resolution ("where is the tumor?"), while detailed analysis needs a lot ("what does this nucleus look like?"). The pyramid takes advantage of that: it computes the low-resolution views ahead of time instead of on the fly, trading storage space for viewing speed. Neighboring levels differ by exactly 2× in linear resolution, which gives the standard 2× zoom steps.

Virtual slide formats use lossy (JPEG) or lossless (JPEG2000, LZW) compression. Lossy compression shrinks files 5-10× but introduces artifacts, most visibly blocks around sharp edges at high compression. Lossless compression preserves the exact pixel values that intensity measurement and spectral unmixing depend on, while moderate JPEG compression is acceptable for morphological assessment and detection. Pawley's advice for data intended for quantitative analysis is short: "avoid lossy compression."

Suppose you are building a digital pathology archive for a study of 300 patients with 2 slides each, 600 virtual slides in all. Scanned at 20x with 9 multispectral channels, each slide takes about 5 GB, or 3 TB in total, stored on a networked server with DICOM-compliant metadata. Expert readers review the slides remotely for quality control and ROI annotation. StrataQuest analyzes all 600 slides with the same parameter set, for consistency.

The archive serves study audits, re-analysis with updated algorithms, and publication. Every analysis can be reproduced, every result verified against the original image, and every slide re-analyzed with future methods. Glass slides can fade or break, while the digital archive persists.