What Is Atypia: When Cells Are Abnormal But Not Cancer

Atypia is a pathology term meaning that cells look abnormal under a microscope but do not meet the criteria for cancer. It sits in an uncomfortable gray zone: the cells have changed in size, shape, or organization compared to their healthy neighbors, yet they have not crossed the threshold into malignant growth. A diagnosis of atypia can show up in nearly any tissue in the body, from breast and thyroid to cervix, endometrium, prostate, bladder, and skin. What makes it confusing for patients is that the word itself reveals very little about what comes next, because the risk of progression to cancer varies enormously depending on where the atypia is found, what subtype it is, and what caused it.

What Pathologists Actually See on the Slide

When a pathologist examines a biopsy, they are comparing the cells in front of them against what healthy tissue from that organ normally looks like. Atypical cells deviate from that baseline in specific, recognizable ways. The nuclei may be larger than expected, irregularly shaped, or contain clumped or unevenly distributed chromatin (the material inside the nucleus that carries genetic information). In breast tissue, for example, a standardized scoring system grades nuclear changes on a scale: mildly atypical nuclei are only slightly enlarged with uniform chromatin, while severely atypical nuclei are markedly variable in size and shape, sometimes with bizarre forms and prominent internal structures called nucleoli.¹ Similar grading systems exist for other organs, though the specific features pathologists weigh differ by tissue type.

Beyond the individual cells, pathologists also look at how cells are arranged relative to one another. Normal tissue has an orderly architecture: glands form predictable patterns, layers of cells maintain their proper thickness, and cells stay within their expected boundaries. Atypia often involves architectural disorder alongside the cellular changes. In the endometrium, for instance, glands may crowd together in unusual density. In the skin, melanocytes may spread beyond their normal position at the base of the epidermis. The combination of cellular oddness and structural disorganization is what moves a diagnosis from “reactive change” (cells responding normally to injury or irritation) toward “atypia” or “dysplasia.”

Common Organs Where Atypia Is Diagnosed

Atypia is not one disease. It is a descriptive label applied across many different clinical contexts, and the implications shift dramatically depending on the organ involved. Here are the most common settings where patients encounter the term.

Breast

Atypical ductal hyperplasia (ADH) and atypical lobular hyperplasia (ALH) are among the most frequently discussed forms of breast atypia. These findings typically turn up on a core needle biopsy after a mammogram flags a suspicious area. ADH involves cells lining the milk ducts that have started to look and grow abnormally but have not filled the duct enough to qualify as ductal carcinoma in situ (DCIS). Women diagnosed with either form of atypical hyperplasia face roughly four to five times the breast cancer risk of women without atypia, and the cumulative incidence reaches about 30% over 25 years after diagnosis.² That sounds alarming, but it also means the majority of women with atypia will not develop cancer. Surgical excision of the biopsy site remains standard practice after a core needle biopsy shows ADH, primarily to make sure a more serious lesion was not missed by the needle.³

Cervix

The most common atypia finding on a Pap smear is “atypical squamous cells of undetermined significance,” or ASC-US. This is a catch-all label that means the cells look a bit off but the pathologist cannot say why. In many cases, the changes are caused by a transient human papillomavirus (HPV) infection that the immune system will clear on its own. High-risk HPV testing has become a key triage tool: if HPV is not detected, the chance of an underlying precancerous lesion is extremely low, and the patient can safely return to routine screening.⁴ If HPV is detected, a closer look via colposcopy is warranted. Most ASC-US results resolve without treatment.

Thyroid

When a thyroid nodule is biopsied with a fine needle aspiration and the cells do not fit neatly into “benign” or “suspicious for malignancy,” pathologists may assign the label “atypia of undetermined significance” (AUS) or “follicular lesion of undetermined significance” (FLUS). The Bethesda System, a widely used classification framework, originally estimated the cancer risk in this category at roughly 5 to 15 percent. Subsequent studies have found the actual malignancy rate to be substantially higher, with some reporting figures in the range of 25 to 55 percent.⁵⁶ That discrepancy has pushed clinicians to supplement cytology with molecular testing and ultrasound features rather than relying on the biopsy result alone.⁷

Endometrium

Endometrial atypical hyperplasia (EAH) is one of the highest-risk forms of atypia. The endometrial lining overgrows, and the glandular cells themselves display nuclear atypia. Long-term studies show the cumulative cancer risk climbing from roughly 8 percent within the first few years to nearly 28 percent over a decade or more.⁸ Making matters more complex, when women with EAH undergo hysterectomy, cancer is sometimes already present in the surgical specimen. Across multiple studies, roughly 20 to 50 percent of cases originally diagnosed as atypical hyperplasia on biopsy turn out to harbor concurrent cancer once the entire uterus is examined.⁹¹⁰ The architectural pattern matters a great deal here: one multicenter study found that when atypical glands formed a confluent, fused pattern, the carcinoma rate jumped to about 65 percent, compared with under 10 percent when the atypia was low grade.¹¹

Prostate

In prostate biopsies, the finding “atypical small acinar proliferation” (ASAP) means the pathologist sees small glandular clusters that look suspicious but fall short of a definitive cancer diagnosis. The suspicion could reflect either a genuinely pre-cancerous process that was only partially sampled by the needle or a benign mimic such as inflammation. The traditional response has been to recommend an immediate repeat biopsy, though some clinicians now question whether that urgency is always justified, since many repeat biopsies reveal either low-grade disease or no cancer at all.¹² The challenge is that high-risk disease cannot be excluded by the initial biopsy alone.

Skin

Dysplastic nevi, sometimes called atypical moles, are the skin’s version of atypia. These moles sit morphologically between ordinary moles and melanoma. The World Health Organization now classifies them into low-grade and high-grade dysplasia.¹³ Despite their unsettling appearance under the microscope, the chance that any single dysplastic nevus transforms into melanoma is very small. Their significance lies mainly in what they signal about the person’s overall melanoma risk rather than in the individual mole itself.¹⁴ For severely atypical moles that have been fully excised with clear margins, observation alone appears to be a safe alternative to additional surgery.¹³

What Makes Cells Atypical in the First Place

There is no single cause of atypia. Different organs develop atypical changes for different reasons, though a few themes recur.

Hormonal imbalance is one of the most well-established triggers. Endometrial atypia, for example, is strongly linked to prolonged exposure to estrogen without the balancing effect of progesterone. Conditions that create this imbalance, such as obesity, polycystic ovary syndrome, or estrogen-only hormone therapy, raise the risk of atypical endometrial hyperplasia.⁹

Chronic infection plays a parallel role in the cervix, where persistent high-risk HPV strains drive the cellular changes behind most cases of cervical atypia and dysplasia. Chronic inflammation more broadly can push cells toward atypical-looking changes in the bladder and other organs, sometimes mimicking true pre-cancerous atypia closely enough to confuse pathologists.¹⁵

Environmental toxicants are an area of growing interest. Pesticides, herbicides, industrial chemicals, and other exposures have been linked to epigenetic changes, alterations to how genes are read rather than to the DNA sequence itself, that may prime cells for abnormal growth.¹⁶ These changes accumulate over time in dividing cell populations and can nudge tissue toward atypia or, eventually, cancer.

The “Driver Mutation” Paradox

One of the more surprising findings in recent years is that many of the genetic mutations traditionally blamed for cancer also show up in cells that are clearly not cancerous, and sometimes at even higher rates. The BRAF mutation known as V600E, for instance, is considered a hallmark driver of melanoma. Yet it turns up in roughly 80 percent of completely benign moles, compared with about 60 percent of dysplastic nevi and only 40 to 45 percent of actual melanomas.¹⁷ Similar patterns hold for other mutations in other tissues: HER2 overexpression is more common in ductal carcinoma in situ than in invasive breast cancer, and FGFR3 mutations are more frequent in low-grade bladder tumors than high-grade ones.¹⁷

This does not mean the mutations are harmless. It means that a mutation alone is usually not enough to cause cancer. The tissue context, the surrounding cells, the immune environment, and additional genetic hits all play a role in determining whether an atypical cell stays put or progresses. In breast tissue, atypical ductal hyperplasia frequently carries the same PIK3CA mutations and chromosome 16q losses seen in low-grade breast cancers, which supports the idea that ADH can be a direct precursor to cancer, but many women with ADH never progress.¹⁸ In dysplastic nevi, the absence of the BRAF mutation in a high-grade lesion may actually be the more worrisome sign, as it could indicate the lesion has picked up other, more aggressive genetic changes closer to those seen in early melanoma.¹⁹

The takeaway for patients is that finding a “cancer gene” mutation in an atypical lesion does not mean cancer is imminent. It means the cell has taken one step on a long road, and most cells with a single driver mutation never take the rest.

Why Pathologists Sometimes Disagree

Atypia sits in the murkiest part of the diagnostic spectrum, and one of its most frustrating features is that expert pathologists frequently disagree about what they are seeing on the same slide. A Gynecologic Oncology Group study illustrates the problem starkly: when a panel of pathologists reviewed cases originally diagnosed as atypical endometrial hyperplasia, the panel agreed with the original diagnosis only 38 percent of the time. In 29 percent of cases, the panel’s majority opinion was that the tissue actually represented cancer, and in 18 percent, they thought it was non-atypical hyperplasia. The overall agreement among panel members was poor.²⁰

This is not a failure of individual skill. It reflects the inherent difficulty of drawing a line through a biological continuum. The boundary between “atypical but not cancer” and “early cancer” is not a bright line in nature; it is a judgment call that depends on how much architectural disorder is present, how abnormal the nuclei look, and which criteria the pathologist is weighting. In prostate biopsies, structured peer-review processes have been shown to substantially improve agreement among pathologists, bringing the level of consensus from “fair” to “moderate” on standard reproducibility scales.²¹ But even with those improvements, “moderate” agreement leaves a meaningful gap.

Physical artifacts add another layer of difficulty. Bladder biopsies taken with electrosurgical tools, for example, can produce thermal damage that distorts cells in ways that closely mimic true atypia or malignancy.²² Inflammation from prior treatments, radiation, or infection can produce reactive cellular changes that overlap visually with pre-cancerous atypia in the bladder and other organs.¹⁵ In these settings, a pathologist is essentially trying to decide whether cells that look abnormal are genuinely on a worrisome trajectory or simply responding to injury, and the appearance under the microscope can be nearly identical.

What Happens After an Atypia Diagnosis

There is no one-size-fits-all response to atypia, because the risk it represents depends on the organ, the subtype, and the patient’s individual circumstances. Management options generally fall along a spectrum:

Surveillance: Repeat imaging, biopsies, or Pap smears at scheduled intervals. This is the typical path for low-risk findings like ASC-US on a Pap smear with a negative HPV test, or a low-grade dysplastic nevus fully excised with clear margins.
Repeat biopsy: Particularly common for thyroid AUS/FLUS and prostate ASAP, where the initial sample may not have captured enough tissue for a definitive diagnosis. Molecular testing is increasingly used alongside repeat biopsy in the thyroid to better stratify risk.
Surgical excision: Standard for atypical ductal hyperplasia found on core needle biopsy of the breast, because the needle samples only a small area and more significant disease can lurk nearby. Hysterectomy is often recommended for atypical endometrial hyperplasia in women who have completed childbearing, given the high rate of concurrent or future cancer.
Hormonal treatment: For women with atypical endometrial hyperplasia who want to preserve fertility, progestin therapy can sometimes reverse the atypical changes, though close follow-up is essential.

The decision about how aggressively to respond almost always involves weighing the statistical risk against the patient’s age, preferences, and other health factors. A 35-year-old woman with endometrial atypia who wants children faces a very different calculation than a 60-year-old with the same biopsy result.

The Emotional Weight of an Ambiguous Result

Hearing that your cells are “atypical” or “abnormal” but not quite cancer creates a particular kind of distress that a clean diagnosis, even a bad one, does not. You are left in limbo: not sick enough for a treatment plan, not healthy enough to stop worrying. Research on women who receive abnormal Pap smear results gives some scale to this. In one Swedish study, nearly half of women notified of an abnormal result reported clinically meaningful anxiety levels, compared with about 20 percent in the general population.²³ Depression rates, interestingly, were not elevated, which suggests the dominant emotional response is worry about what the finding could become rather than despair about a current illness.

Patients also report frustration with the process of management itself. The experience of repeat testing has been studied in the context of Pap smear follow-up, where women consistently prefer strategies that resolve uncertainty quickly, even when the clinical outcome is statistically the same either way.²⁴ Paradoxically, immediate HPV testing, which resolves the question faster, was valued somewhat less than repeat Pap smears in one study, possibly because the finality of a definitive test carries its own anxiety. The broader point is that how patients experience the diagnostic process matters alongside the medical outcomes, and clinicians are increasingly recognizing this.

Artificial Intelligence and Diagnostic Consistency

Given how much atypia diagnosis depends on subjective visual judgment, there is obvious interest in whether machine learning can reduce the variability among pathologists. Deep learning algorithms trained on digitized pathology slides have shown promise in distinguishing between benign, atypical, and malignant tissue patterns. The appeal is straightforward: an algorithm gives the same answer every time it sees the same slide, which human pathologists demonstrably do not.²⁵

The reality, however, is more complicated. AI systems are only as consistent as their training data, and if the training slides were labeled by pathologists who disagreed with each other, the algorithm inherits that disagreement. Unsupervised approaches, where the algorithm identifies patterns without being told which category a slide belongs to, offer one potential way around this by letting the machine discover natural groupings that may not map neatly onto existing diagnostic labels. For now, AI tools are entering clinical practice mostly as decision-support aids rather than standalone diagnosticians, flagging cases that deserve closer human review and providing a second opinion that is at least free from fatigue and anchoring bias. In the specific domain of atypia, where the diagnostic gray zone is widest, these tools may end up having their greatest practical impact.