How digital technology and AI are revolutionizing disease diagnosis at the cellular level
In a laboratory, a scientist examines a sliver of tissue, thinner than a human hair. Through the microscope, they don't just see cells—they read a story of disease, a narrative of cellular rebellion that holds the key to someone's survival. This is the world of cellular pathology, the medical science dedicated to studying diseases at their most fundamental level: our cells. Often operating behind the scenes, cellular pathologists serve as medicine's silent detectives, examining tissue and cell samples to unravel the mysteries of illness.
For decades, cellular pathology relied on microscopes and stained glass slides to visualize cellular structures and identify abnormalities.
Today, the field is embracing digital scanners and artificial intelligence, creating powerful synergy between human expertise and technological precision.
Cellular pathology, also known as histopathology, is the field of microscopic examination of tissue samples to detect abnormalities and diagnose disease while simultaneously guiding treatment decisions 1 . This discipline forms the cornerstone of diagnosis for conditions ranging from cancer to inflammatory disorders, providing visual evidence of disease processes at the cellular level.
The field rests on two main diagnostic approaches:
The examination of whole tissue samples (biopsies or surgical specimens) to study the architecture of disease. This method preserves the spatial relationships between cells, allowing pathologists to see how diseases affect tissue structure.
The study of individual cells or small cell clusters, often obtained through less invasive methods like Pap smears, fluid aspirations, or fine-needle biopsies.
| Feature | Histopathology | Cytopathology |
|---|---|---|
| Sample Type | Solid tissue sections | Individual cells or small clusters |
| Invasiveness | Generally more invasive | Often less invasive |
| Tissue Architecture | Preserved | Not preserved |
| Primary Applications | Tumor typing, staging, grading | Cancer screening, lesion evaluation |
| Preparation Time | Longer (hours to days) | Shorter (minutes to hours) |
Cellular pathology is particularly indispensable in oncology. When a surgeon removes a tumor, it's the cellular pathologist who determines whether the margins are clear of cancer cells—a critical factor in predicting recurrence.
The traditional world of glass slides and light microscopes is undergoing a seismic shift toward digitization. Whole slide imaging technology now allows pathologists to convert physical slides into high-resolution digital files that can be viewed, shared, and analyzed on computer screens 1 . This transition has paved the way for artificial intelligence to enter the pathology laboratory, creating what is now known as computational pathology 8 .
Research presented at ASCO 2025 showed that AI assistance boosted diagnostic agreement among pathologists for challenging HER2 cases 4 .
AI models analyzing spatial relationships between cells have significantly outperformed traditional methods in predicting immunotherapy responses 4 .
AI algorithms can now predict molecular alterations directly from standard H&E-stained slides, potentially bypassing costly molecular tests 4 .
Slides can be scanned and stored in cloud-based platforms, enabling remote consultation and collaboration regardless of geographical constraints.
Digital transformation brings automation to pathology workflows, freeing pathologists to focus on complex diagnostic challenges.
This technology is especially valuable given the global shortage of trained pathologists, which causes diagnostic delays in many regions 1 .
To understand how computational pathology works in practice, let's examine a landmark study presented at the 2025 ASCO meeting that demonstrates the powerful synergy between pathologists and AI algorithms.
Background: Accurate assessment of HER2 protein expression is critical in breast cancer, as it determines eligibility for targeted therapies. However, distinguishing between HER2-low, HER2-ultralow, and HER2-negative cases has proven challenging, with significant inter-observer variability among pathologists.
Breast cancer tissue samples processed using standard immunohistochemistry techniques
Stained glass slides converted into high-resolution whole slide images
Pathologists scored HER2 status without AI assistance initially
Same pathologists re-evaluated cases with AI algorithm assistance
The AI assistance yielded remarkable improvements in diagnostic consistency and accuracy:
| Assessment Type | HER2-Low Agreement | HER2-Ultralow Agreement | HER2-Null Misclassification |
|---|---|---|---|
| Without AI | 73.5% | 65.6% | Baseline |
| With AI Assistance | 86.4% | 80.6% | Reduced by 65% |
Modern cellular pathology relies on a sophisticated array of reagents and technologies that enable precise visualization and analysis of cellular components. Here's a look at the essential tools driving today's diagnostic and research advances.
| Tool/Technology | Primary Function | Application Examples |
|---|---|---|
| Immunohistochemistry (IHC) | Detects specific proteins in tissue sections using antibody-antigen interactions | Tumor typing, predictive biomarker testing (e.g., HER2, PD-L1) |
| Whole Slide Scanners | Converts glass slides into high-resolution digital images for computer analysis 1 | Creating digital pathology archives, enabling AI-based analysis |
| AI-Based Image Analysis Software | Automates quantification of cellular features and patterns in digital pathology images 4 | HER2 scoring, tumor cell counting, spatial analysis |
| Molecular Pathology Assays | Analyzes DNA/RNA alterations in tissue samples using techniques like PCR and next-generation sequencing | Identifying targetable mutations, detecting microsatellite instability |
| Special Stains | Highlights specific tissue structures or microorganisms using histochemical methods 2 | Identifying connective tissues, detecting fungi |
| Circulating Tumor DNA (ctDNA) Tests | Detects tumor-derived DNA fragments in blood samples 2 | Monitoring treatment response, detecting minimal residual disease |
The cellular pathology toolkit has evolved dramatically from the basic hematoxylin and eosin (H&E) stain that reveals tissue architecture in shades of pink and blue. Immunohistochemistry, which entered diagnostic practice in the 1980s, revolutionized the field by allowing pathologists to identify specific proteins that characterize different cell types and disease states .
As we look toward the horizon, several emerging trends and innovations are poised to further transform cellular pathology in the coming decade. The market itself is projected to grow at a compound annual growth rate of 8.1%, rising from USD 41,536 million in 2025 to approximately USD 90,508 million by 2035 1 .
The future lies in combining multiple data types. Researchers are developing models that integrate whole slide images with clinical variables and genomic data to create more powerful prognostic tools.
Similar to large language models, foundation models in pathology—trained on hundreds of thousands of whole slide images—are enabling faster development of specialized AI tools 4 .
New technologies are moving beyond simple cell identification to analyzing complex spatial relationships within the tumor microenvironment 4 .
Cloud-based platforms are facilitating cross-border consultation and second opinions, making specialized diagnostic expertise accessible regardless of geography 1 .
There is a global shortage of trained pathologists, causing diagnostic delays in many areas. While AI tools can help increase efficiency, they cannot fully compensate for the lack of human expertise 1 .
Digital pathology generates enormous files—a single whole slide image can require several gigabytes of storage. Managing, storing, and securing this data presents significant technical and financial challenges 1 .
The initial setup costs for digital pathology systems, including slide scanners, storage infrastructure, and computing resources, can be prohibitive, especially in low- and middle-income countries 1 .
Cellular pathology stands at a remarkable crossroads, where traditional diagnostic expertise converges with cutting-edge computational power. What began as a discipline focused on the microscopic examination of stained tissues has evolved into a high-tech field that leverages digital imaging, artificial intelligence, and molecular biology to unravel the complexities of disease at an unprecedented resolution.
As Andrei Ivanov, PhD, Program Chair for Pathobiology 2025, notes, the field is focused on "understanding of cellular and molecular mechanisms of disease" that will lead to "disease diagnosis, prevention, and treatment" 7 .
The future of cellular pathology will likely see even deeper integration of human expertise with algorithmic assistance, more sophisticated multi-analyte diagnostics, and increasingly accessible technologies that democratize precision medicine. While challenges remain, the continued collaboration between pathologists, scientists, engineers, and clinicians ensures that this once-hidden specialty will continue to illuminate the dark corners of disease, guiding medical decisions and shaping therapeutic strategies for generations to come.
In the intricate patterns of cells and tissues, we find not just evidence of disease, but the blueprint for healing—and with each technological advancement, we learn to read that blueprint with greater clarity and insight than ever before.