What Is Digital Pathology?

From the glass slide to the calibrated screen — how pathology went digital, and why color fidelity is now part of the diagnosis.

For more than a century, the diagnosis of disease at the tissue level looked the same: a pathologist seated at a microscope, moving a glass slide under the objective lens. Digital pathology replaces that physical act of looking with a digital one. The glass slide is scanned into a high-resolution image, and the pathologist reads it on a screen instead of through an eyepiece. The tissue does not change — the medium through which it is examined does.

That shift sounds simple, but it reorganizes the entire workflow of a pathology lab: how slides are captured, stored, shared, analyzed by software, and ultimately reported. This paper explains what digital pathology is, how it works end to end, who the major players are, and why one detail — color — deserves far more attention than it usually receives.

Before: The Analog Era

Traditional pathology is an analog discipline. A tissue sample is fixed, sectioned into slices a few micrometers thick, mounted on a glass slide, and stained — most commonly with hematoxylin and eosin (H&E) — to make cellular structures visible. The pathologist then examines the slide directly under a light microscope.

This method is reliable and has underpinned diagnostic medicine for generations, but it carries real constraints. Glass slides are physical objects: they must be transported, filed, and retrieved by hand. They can break or fade. A second opinion means physically shipping the slide to another expert. And because only one person can look down the eyepiece at a time, collaboration is slow. The analog slide is, in effect, a single copy that lives in one place at one time.

Now: Pathology Goes Digital

Digital pathology removes that bottleneck by turning each glass slide into a whole-slide image (WSI) — a complete, high-resolution digital scan of the entire specimen. Once a slide exists as data, it can be copied, viewed simultaneously by colleagues anywhere in the world, archived without physical storage, and — critically — analyzed by software.

The clinical credibility of this approach is well established. In 2017 the U.S. FDA cleared the first whole-slide imaging system for primary diagnosis, supported by a pivotal multi-center study of 1,992 cases that found digital reading non-inferior to conventional microscopy across a wide range of specimens, sample types, and stains. Digital pathology is no longer an experiment; it is an accepted way to practice.

How a Glass Slide Becomes a Digital Image

1. Acquisition: the slide scanner

The process begins with a slide scanner. The scanner moves a high-quality microscope objective across the glass slide and captures the tissue as a sequence of many small, overlapping image fields — because no single camera frame can hold an entire slide at diagnostic magnification (typically 20× or 40×). Modern scanners can process racks of slides automatically, capturing each one in seconds to minutes.

2. Stitching: assembling the whole-slide image

The scanner’s software then stitches those individual fields together into one seamless, navigable image. The result is enormous — a single whole-slide image is routinely a gigapixel image and can run to several gigabytes. To make it usable, the image is stored as a multi-resolution pyramid: zooming and panning load only the tiles needed at the current magnification, so the pathologist experiences smooth navigation despite the file’s size.

3. Storage and the imaging application

The finished image is then made available in a dedicated viewing application designed for pathology. Unlike a generic image viewer, these applications support the pyramid format, very high magnification, annotation, measurement, and side-by-side comparison of cases. From there, images flow into longer-term clinical storage.

Where the Images Live: Viewers, PACS, and File Formats

Borrowing the model that radiology adopted decades ago, digital pathology stores and distributes images through a PACS (Picture Archiving and Communication System) — or, increasingly, a vendor-neutral archive (VNA) that can hold images from many different scanners. The PACS connects to the laboratory and hospital information systems so that each image is tied to the right patient, case, and report.

File formats remain a practical challenge. Most scanner vendors write their own proprietary format — Leica/Aperio’s .svs, Hamamatsu’s .ndpi, and Philips’ iSyntax, among others. To keep images portable across systems, the industry is converging on the DICOM standard for whole-slide imaging — the same family of standards that already governs radiology — alongside open formats such as OME-TIFF. Standardization is what lets a hospital mix scanners, viewers, and AI tools without being locked to a single supplier.

The Market: Who Leads Digital Pathology

The digital pathology market spans scanner hardware, image-management software, and increasingly AI. A handful of established companies anchor the field:

• Leica Biosystems (Aperio) — a long-standing leader in slide scanners and image management; its .svs format is one of the most widely used in research and clinical labs.
• Philips — producer of the IntelliSite Pathology Solution, the system behind the first FDA clearance for primary diagnosis.
• Hamamatsu — widely used NanoZoomer scanners and the .ndpi format, strong in both research and clinical settings.
• Roche / Ventana — combining staining and digital pathology workflows, with a focus on integrated diagnostics.
• 3DHISTECH, Sectra, and others — scanner makers and enterprise imaging vendors that supply the viewing, workflow, and archive layers.

Around these incumbents sits a fast-growing ecosystem of specialized AI companies and open-source tools (such as QuPath and OpenSlide) that build on the images the scanners produce.

Why AI Is So Compelling in Pathology

Pathology is unusually well suited to artificial intelligence, and the reason is the data itself. A whole-slide image contains billions of pixels of densely packed biological detail — far more information than any other routine medical image. That scale is exhausting for a human to review exhaustively, but it is exactly what modern deep-learning models thrive on.

AI in pathology is already being used to flag regions of interest, quantify what was once counted by hand (mitoses, stained cells, tumor extent), grade tumors, and even estimate molecular biomarkers directly from the image. As of mid-2025, roughly fifty digital pathology AI tools had been cleared for diagnostic use in the EU alone. The promise is not to replace the pathologist but to handle the tedious, high-volume measurement work and surface the cases and regions that most need expert attention.

The Detail That Cannot Be Lost: Color

There is a quiet assumption running through everything above: that the image on the pathologist’s screen faithfully represents the tissue on the glass. In pathology, that assumption is diagnostic. Stains encode meaning in color. The intensity of a hematoxylin nucleus, the precise hue of an immunohistochemistry reaction, the subtle pink gradients of eosin — these are not decoration. They are the signal a pathologist interprets to reach a diagnosis.

In the analog world, the pathologist looked at the actual stained tissue through glass. In the digital world, several devices sit between the tissue and the eye: the scanner’s camera and color processing, the file format and color space, and — most variable of all — the display. Each step can shift, compress, or distort color. An uncalibrated monitor can render the same image too warm, too cool, too dark, or with crushed tonal detail, subtly changing what the pathologist sees — and potentially what they conclude.

This is why color management is not a finishing touch in digital pathology but a clinical requirement. Radiology solved its version of this problem long ago by calibrating displays to the DICOM grayscale standard. Pathology raises the bar, because pathology is color. Displays must be calibrated and quality-assured for accurate, consistent, and reproducible color — so that an image looks the same on the scanner operator’s screen, the reporting pathologist’s screen, and a remote consultant’s screen, today and a year from now.

This is precisely the gap QUBYX addresses. Reliable digital pathology depends on every screen in the chain being measured, calibrated, and continuously verified — so that no diagnostically relevant color is lost between the slide and the diagnosis. As pathology becomes fully digital and AI models are trained on these same images, consistent color is no longer just a viewing comfort — it is part of data quality and patient safety.

Conclusion

Digital pathology takes a discipline that was physical, local, and analog and makes it data-driven, shareable, and ready for computation. Slides are scanned, stitched into gigapixel whole-slide images, stored in PACS and vendor-neutral archives, viewed in specialized applications, and increasingly interpreted with the help of AI. The leaders — Leica, Philips, Hamamatsu, Roche, and a widening field of AI specialists — are building the infrastructure for a fully digital practice.

But the value of all that infrastructure rests on a single, often-overlooked condition: the image must be true to the tissue, and color must survive the journey from slide to screen. Get the color right, and digital pathology delivers on its promise of speed, collaboration, and intelligence. Get it wrong, and even the best scanner and the smartest algorithm are reading the wrong picture.

QUBYX — calibration and quality assurance for medical displays.

Learn more at qubyx.com