Digital pathology and AI in diagnosis

From glass slides to digital images

Pathology is the branch of medicine that studies tissue and cells to diagnose disease. For over a century, pathologists have examined thin slices of tissue mounted on glass slides under a microscope. Digital pathology replaces that microscope with a high-resolution scanner that turns each slide into a very detailed digital image, sometimes called a whole slide image. The pathologist then views and analyses the sample on a computer screen instead of down a lens. The diagnosis is still made by a trained doctor; what changes is how the images are stored, shared and worked with. This shift is the foundation that makes computer analysis, including AI, possible.

Why going digital helps

Digital images bring practical advantages. They can be shared instantly, so a pathologist can get a second opinion from an expert anywhere in the country without posting fragile glass slides. This supports faster diagnosis and easier collaboration, especially for rare or difficult cases. Digital slides can be measured and annotated precisely on screen, magnified without a microscope, and stored electronically alongside a patient's other records. It can also help balance workloads between hospitals when some are busier than others. For patients, the hoped-for benefits are quicker, more consistent results and better access to specialist expertise, wherever they happen to live.

Where artificial intelligence fits in

Artificial intelligence refers to computer programs trained to recognise patterns in images. In pathology, AI tools can be trained on many examples of tissue to help flag areas that may contain cancer, count and measure features, or highlight the parts of a slide that most deserve a closer look. Used well, this acts like a careful assistant: it can speed up routine tasks, reduce the chance of a small area being overlooked, and bring more consistency to measurements that are hard to judge by eye. The technology is a support tool. It works alongside the pathologist rather than replacing the expert judgement that ties findings together into a diagnosis.

Benefits and limitations

The potential benefits are faster reporting, fewer missed details and more consistent results, which could ease pressure on a stretched workforce. But there are real limitations. An AI tool is only as good as the data it was trained on, and it may perform less well on samples that differ from that data, or produce confident-looking answers that are wrong. It cannot understand a patient's wider circumstances. That is why these tools must be carefully tested, validated and regulated before use, and why a qualified pathologist reviews and takes responsibility for the final result. Transparency about how a tool reaches its conclusions, and ongoing checking, are essential to keep patients safe.

What it means for patients

For most patients, digital pathology and AI happen quietly in the background and do not change the experience of having a biopsy. What may change over time is the speed and consistency of results and the ease of getting specialist input. In the UK, adoption is growing but uneven, and these technologies are introduced with careful evaluation to make sure they are safe and genuinely helpful. Importantly, a human specialist remains in charge of your diagnosis. If you are ever unsure how your sample was analysed or what your result means, you can ask your clinical team, who can explain the process and what happens next.