What Is Anatomic Pathology and How Does It Work?

Anatomic pathology is the branch of medicine in which doctors examine tissue, cells, and organs to diagnose disease. It involves both the naked-eye (gross) and microscopic study of biological specimens, correlated with a patient’s clinical and imaging history, to arrive at an accurate diagnosis.¹ If you have ever had a mole removed, a lump biopsied, or surgery for cancer, an anatomic pathologist almost certainly examined the tissue and wrote the report your treating doctor used to decide what happened next. The field is broader and more layered than most people realize, touching everything from rapid mid-surgery consultations to molecular profiling that determines which drugs will work against a tumor.

From Patient to Slide

The journey of a tissue sample through an anatomic pathology lab follows a surprisingly regimented pipeline, and errors at any stage can compromise the final diagnosis. A surgeon, dermatologist, or other clinician removes tissue and places it in a fixative solution, usually formalin, to preserve its structure. The specimen is then labeled, logged, and sent to the pathology lab. One quality study found that the vast majority of processing errors happen before the tissue ever reaches the microscope: out of 113 non-conformities identified, roughly 93% belonged to this pre-analytical phase, with mislabeled slides being the single most common mistake.²

Once in the lab, a trained technician or pathologists’ assistant processes the tissue: trimming it to representative sections, embedding it in paraffin wax, slicing it into sections just a few micrometers thick, and mounting those slices onto glass slides. The slides are then stained, most commonly with hematoxylin and eosin (H&E), a combination that highlights cell nuclei in blue-purple and cytoplasm in pink. This is the workhorse stain of pathology, used on nearly every specimen. A pathologist then views the stained slide under a microscope, interpreting the patterns of cells, tissue architecture, and any abnormalities.

Surgical Pathology and the Operating Room

Surgical pathology is the largest and most familiar branch of anatomic pathology. Whenever tissue is removed during a procedure, it is sent to a surgical pathologist. The pathologist examines the specimen grossly first, measuring it, describing its appearance, inking the surgical edges with colored dyes to track margins, and cutting it into sections for microscopic evaluation.³ The gross exam matters more than people think: in breast cancer surgery, for example, pathologists examine the specimen while the patient is still in the operating room, looking for tumor near the cut edges. One study found that gross examination alone does not reliably reflect margin status in at least a quarter of partial mastectomy cases, which is why microscopic confirmation later is essential.⁴

Sometimes surgeons need an answer while the patient is still under anesthesia. This is where frozen sections come in. A small piece of tissue is rapidly frozen, sliced, stained, and interpreted within minutes, and the pathologist calls the surgeon with the result. The technique plays an important role in surgical management but has to be used judiciously: it sacrifices some tissue quality for speed, so it is not appropriate for every question.⁵ In ovarian cancer surgery, frozen section biopsy has shown strong diagnostic accuracy, around 94%, with perfect specificity in one study, meaning that when it called something malignant, it was right every time.⁶ In oral cancer, though, frozen section and gross examination performed comparably for margin assessment, with no survival benefit seen when frozen sections were added.⁷ The value of a frozen section, in other words, depends heavily on the clinical scenario.

Cytopathology

Not every diagnosis requires cutting out a piece of tissue. Cytopathology examines individual cells or small clusters of cells, often collected by scraping a surface (as in a Pap smear) or by inserting a thin needle into a mass. Fine-needle aspiration is the most widely used screening and diagnostic method for thyroid nodules, for instance, and is typically guided by ultrasound to ensure the needle hits the right spot.⁸ Because the sample is so small, pathologists increasingly face pressure to extract maximum information from these tiny specimens, running not just the standard stain but also molecular and immunohistochemical tests on the same limited material.⁹

Autopsy Pathology

The autopsy is anatomic pathology’s oldest tool, and it remains surprisingly relevant. When someone dies and the cause is unclear, a clinical autopsy can reveal diagnoses that were missed during life. A South African comparative study found that about 39% of autopsied cases had a major discrepancy between the clinical diagnosis and what the autopsy revealed, with pulmonary conditions being the most commonly missed cause of death.¹⁰ A separate analysis of 284 clinical autopsies found major discrepancies in about 30% of cases; the most frequently missed diagnoses included cancers, pneumonias, invasive fungal infections, aortic dissections, and pulmonary embolism.¹¹

These numbers sometimes surprise people. Modern imaging is excellent, but it has limits. A prospective study comparing post-mortem CT and MRI to a traditional autopsy found that even the two imaging modalities combined could not detect conditions like sepsis, certain blood cancers, liver fibrosis, heart valve disease, and small findings in general.¹² The hands-on examination of tissue under a microscope still catches things that machines looking at pictures cannot.

What Happens When the Basic Stain Is Not Enough

H&E staining answers many questions, but pathologists have a toolbox of additional techniques for cases where the standard slide leaves room for doubt. These ancillary tests can identify the specific type of tumor, predict how a cancer will behave, or detect disease processes invisible under ordinary light microscopy.

Immunohistochemistry (IHC) uses antibodies tagged with a visible marker to detect specific proteins in a tissue section. It is one of the most widely used techniques in diagnostic pathology, especially for cancer, because certain tumors express unique protein signatures that help distinguish one type from another.¹³ If a pathologist sees a poorly differentiated tumor under the microscope and cannot tell whether it started in the lung, the breast, or somewhere else, a panel of IHC stains can often resolve the question.

Electron microscopy (EM) magnifies tissue to a level far beyond what a light microscope can achieve, revealing the ultrastructure of cells. It is especially valuable in kidney disease: one study found that EM was essential for diagnosis in about 18% of kidney biopsies, clearly contributed in another 54%, and had no influence in only about a quarter of cases.¹⁴ In heart biopsies, EM can detect myofilament loss and endothelial cell changes that light microscopy commonly misses, while immunofluorescence can reveal immune deposits in heart tissue that would otherwise go undetected.¹⁵

The Pathology Report

Everything the pathologist finds gets distilled into a pathology report, a document that your treating doctor relies on to plan your care. For a cancer specimen, the report typically includes the type of cancer, its grade (how abnormal the cells look), the stage or depth of invasion, whether the surgical margins are clear of tumor, and the results of any ancillary tests. Histologic grade is a strong predictor of outcome that refines the prognostic information provided by staging alone.¹⁶

Grading and staging can be subjective, though. In bladder cancer, a combined analysis of five European trials found large variations in how pathologists classified tumor stage and grade. Pathology review downstaged more than half of cases originally classified as T1 (tumor invading just below the surface) to Ta (confined to the surface), and agreement on certain stage-grade combinations was as low as 50%.¹⁷ That level of disagreement is not unique to bladder cancer; it reflects a general challenge in pathology where human interpretation of complex tissue patterns can differ between observers.

When Patients Read Their Own Reports

With the rise of online patient portals, many people now see their pathology report before their doctor has a chance to explain it. This has created a genuine problem. Standard pathology reports are written for clinicians, packed with specialized terms like “high-grade dysplasia,” “perineural invasion,” or “immunophenotypic analysis.” Studies consistently find that patients, regardless of their diagnosis, struggle to understand these documents, mainly because of complex medical terminology.¹⁸ Simply giving patients portal access to the report, without additional explanation, does not improve their understanding.

Researchers have advocated for patient-centered pathology reports that use plain language alongside the technical version. A systematic review found that medical jargon in reports often leaves patients confused and uncertain about their diagnosis, and that the way information is shared with patients directly affects whether they can participate meaningfully in decisions about their care.¹⁹ Some institutions have started piloting lay-language supplements to standard reports, though the practice is far from widespread.²⁰

How Pathology Catches Its Own Mistakes

Because so much rides on a pathologist’s interpretation, the field takes diagnostic error seriously. Second opinions, peer review, and external quality assurance programs all serve as safety nets.²¹ The evidence is striking in breast pathology, where diagnostic disagreement is well studied. A simulation study of breast biopsy interpretations found that the overall misclassification rate dropped from about 25% to 18% when all cases received a second opinion. Having both the first and second reader be high-volume breast pathologists brought the rate down to roughly 14%. Cases of atypia, a borderline category between normal and precancerous, had the highest single-reader misclassification rate at over 52%, and it remained above 34% even in the best second-opinion scenarios.²²

Those numbers are not a sign that pathology is unreliable. They reflect the inherent difficulty of interpreting tissue patterns that exist on a biological spectrum rather than in neat categories. Observer variability is a recognized part of diagnostic histopathology.²³ The practical takeaway for patients: if you receive a diagnosis in a borderline or high-stakes category, requesting a second pathology opinion from a specialist is reasonable and increasingly standard.

Digital Pathology and Artificial Intelligence

For over a century, pathologists looked through eyepieces. That is changing. Whole slide imaging (WSI) uses high-resolution scanners to create a digital copy of the glass slide, which the pathologist can then view, zoom, and annotate on a monitor. The FDA has approved WSI systems for use in primary surgical pathology diagnosis, opening the door for broader adoption.²⁴ A large academic center’s randomized study comparing digital to glass-slide diagnosis found that the diagnostic agreement between the two was about 99%, though pathologists working digitally were roughly 19% slower per case.²⁵ Accuracy is there; speed is still catching up.

Where digital pathology gets genuinely exciting is in the application of artificial intelligence. Machine learning and deep learning algorithms can now analyze digitized histopathology images to perform automated tumor detection, classification, identification of molecular biomarkers, and prediction of treatment response.²⁶ The technology is being driven by advances in computer vision and transformer-based deep learning, with foundation models trained on millions of slides showing promise across multiple cancer types.²⁷ The realistic near-term picture is AI as a co-pilot rather than a replacement: flagging regions of interest, pre-screening slides for abnormalities, quantifying features that are tedious for a human to count, and freeing the pathologist to focus on the interpretive decisions that require clinical judgment.

Molecular Pathology and Targeted Therapy

Perhaps the most consequential shift in anatomic pathology over the past two decades is the rise of molecular testing. Pathologists are no longer just describing what tissue looks like; they are reading its genetic code. When you hear about “targeted therapy” or “precision medicine” in cancer, the pathologist is typically the one who determined that the tumor carries the specific mutation the drug was designed to attack. Companion diagnostics, tests that determine whether a patient’s tumor has a treatable genetic alteration, have become a core part of surgical pathology for cancers of the breast, colon, lung, skin, and others.²⁸

The practical challenge is that much of this testing must be done on small biopsies or cytology samples, since many patients with advanced cancer never undergo major surgery. A study of patients with non-small-cell lung cancer found that small tissue specimens could successfully provide the prognostic and predictive information needed for current guideline-recommended biomarkers in most cases.²⁹ That means pathologists have to be increasingly strategic about how they use a finite number of cells, prioritizing which stains and molecular tests to run so that the sample does not run out before all clinically important questions are answered.³⁰

Anatomic Pathology Beyond Human Medicine

The same skills that pathologists use on human tissue apply in veterinary and research settings. Comparative pathology, the examination of disease in animal models, is essential for drug development and basic biomedical research. Animal studies produce molecular and biochemical data, but without histopathology to provide the morphologic context, those data can be misleading.³¹

The stakes here are real. Some failures in translating animal research to human treatments may stem not from genuine species differences but from poor pathological interpretation of the animal model itself. Misclassifying an adaptive response as a disease, modeling the wrong stage of a disease, or missing background strain effects can make a model look like it failed when the real problem was how the tissue was read.³² This is an area where improved pathology expertise could have outsized effects on the pace of medical discovery, though it receives far less public attention than, say, AI in pathology.