Diagnostics: Milan System for Reporting Salivary Gland Cytopathology
The Milan System sorts the results of a salivary gland needle biopsy into a small set of categories, each with an estimated risk of cancer. It gives pathologists, surgeons and patients a common language for these results, much as the Bethesda System does for thyroid biopsies.
The salivary glands and why lumps are checked
People have three pairs of major salivary glands:
Parotid glands, the largest, sitting in front of and just below each ear. The facial nerve, which controls the muscles of facial expression, runs through the parotid gland.
Submandibular glands, located under the jawbone on each side.
Sublingual glands, located under the tongue.
Hundreds of tiny "minor" salivary glands also line the mouth and throat. More about the glands and their common problems is on the salivary glands page.
A lump in a salivary gland can come from many things, including swelling from infection or inflammation, a blocked duct, a fluid-filled cyst, an enlarged lymph node, or a growth (tumor). Most salivary gland lumps, especially in the parotid gland, are not cancer. The most common salivary gland tumors are benign (non-cancerous), with pleomorphic adenoma and Warthin tumor the two most frequent. Salivary gland tumors are uncommon overall, are most frequent in people over 40, and become more likely to be cancer with increasing age. As a general rule, a tumor in a smaller gland (submandibular, sublingual or minor glands) is somewhat more likely to be cancer than a tumor in the parotid gland.
Because lumps can look and feel alike, doctors usually want a small sample of cells before deciding whether treatment is needed, and if so, what kind.
What is a fine needle aspiration (FNA)?
A fine needle aspiration, or FNA, is a quick, minimally invasive and relatively inexpensive procedure that is widely used as the first test for a salivary gland lump. A very thin needle, similar to or thinner than the kind used for blood draws, is placed into the lump to collect a small sample of cells and fluid.
Before the procedure. Most people need no special preparation. Patients are usually asked about blood thinners and bleeding problems. The procedure is typically done in a clinic or radiology department, and patients go home the same day.
During the procedure. The skin is cleaned. A numbing medicine may or may not be used, since the needle is so fine. The doctor usually passes the needle into the lump a few times over just a few minutes, moving it gently to collect cells. Many people describe the feeling as pressure or a brief sting.
After the procedure. Mild soreness, a small bruise or slight swelling is common and usually goes away within a day or two. Serious problems are rare. Uncommon complications include a larger collection of blood under the skin (hematoma), infection, or, with Warthin tumors in particular, temporary inflammation of the gland. Injury to the facial nerve from a fine needle is extremely rare. Spread of tumor cells along the needle track is a theoretical concern that has very rarely been seen in practice.
What happens to the sample. The cells are spread on glass slides and stained, and leftover material may be spun down into a small block of cells (a "cell block") that can be used for extra testing. A pathologist who specializes in looking at cells, called a cytopathologist, examines the sample under the microscope. When needed, the laboratory can run extra tests on the sample. These include special stains that identify proteins in the cells, tests for specific gene changes linked to certain tumor types, and, if lymphoma is a concern, tests that sort white blood cells (flow cytometry). Results often take several days, longer if extra testing is done.
The role of ultrasound guidance
Many salivary gland FNAs are now done with ultrasound guidance. The doctor watches the needle on an ultrasound screen as it enters the lump. Ultrasound guidance helps by:
Placing the needle precisely in small or deep lumps that are hard to feel.
Aiming at the solid part of a lump that is partly cystic, since fluid alone often contains few useful cells.
Avoiding blood vessels and other structures.
Showing the size, shape and borders of the lump, which adds to the overall picture.
Studies show better diagnostic results and fewer "non-diagnostic" (insufficient) samples when FNA is done with ultrasound guidance than when it is guided by feel alone. Some centers also have a pathologist or technician check the slides on the spot (rapid on-site evaluation), so more passes can be taken right away if the first sample is too thin.
Why a standard reporting system is used
In the past, salivary gland FNA reports varied widely from hospital to hospital. Different pathologists used different words for the same findings, and phrases like "atypical cells" or "cannot rule out a neoplasm" could mean very different things depending on who wrote them. That made it harder for surgeons, other doctors and patients to understand the result and agree on next steps.
The Milan System was developed by an international group of pathologists, organized by the American Society of Cytopathology and the International Academy of Cytology, and first published in 2018. The second edition (2023) kept the same basic framework and refined the risk estimates using systematic reviews of the large body of newer research. It also added a chapter on imaging, expanded guidance on extra laboratory testing, and updated some terminology.
The Milan System does three main things:
Common language. Every result is placed into one of a small number of defined categories, so a report means the same thing everywhere.
Estimated risk of cancer. Each category comes with an approximate risk of malignancy, meaning the chance, based on large studies, that a lump given that category turns out to be cancer.
General guidance on next steps. Each category is linked to commonly considered management options.
The American Society of Clinical Oncology (ASCO) guideline on salivary gland cancer calls for salivary gland FNA results to be reported with a risk-based system of this kind and names the Milan System as the current standard.
How to think about "risk of malignancy" numbers
The percentages on this page are averages taken from many studies. They describe groups of patients, not any single person. Actual figures vary between hospitals, and the numbers shift depending on how they are calculated. For example, studies that look only at patients who had surgery tend to show higher risks than the true overall risk. The care team combines the FNA category with other information, such as how long the lump has been present, how fast it is growing, pain, facial weakness, enlarged neck lymph nodes and imaging results, to judge the risk for an individual patient.
The Milan categories at a glance
The risk figures below are the benchmark values from the second edition of the Milan System.
| Category | Name | Approximate risk of cancer | Commonly considered next steps |
|---|---|---|---|
| I | Non-diagnostic | About 15% | Repeat FNA (often with ultrasound) or core needle biopsy; compare with exam and imaging |
| II | Non-neoplastic | About 11% | Treat the underlying condition; follow-up exam and imaging; re-check if the lump does not go away |
| III | Atypia of undetermined significance (AUS) | About 30% | Repeat FNA or core needle biopsy; sometimes surgery |
| IVA | Neoplasm: benign | Less than 3% | Surgery to remove, or monitoring in selected cases |
| IVB | Neoplasm: salivary gland neoplasm of uncertain malignant potential (SUMP) | About 35% | Usually surgery; sometimes core biopsy or extra lab testing first |
| V | Suspicious for malignancy | About 83% | Surgery, often with frozen section; sometimes core biopsy first for planning |
| VI | Malignant | More than 98% | Treatment planning by a cancer team; surgery for most salivary cancers; a different approach for lymphoma or cancer that spread from elsewhere |
The categories in detail
Category I: Non-diagnostic
What it means. The sample did not contain enough of the right kind of cells to make an interpretation, or the cells could not be seen clearly. This is a statement about the sample, not about the lump. It is neither reassuring nor alarming by itself.
What usually leads to this result.
Too few cells collected.
The sample is mostly blood, or the cells are poorly preserved.
The needle drew only clear fluid from a cyst, without cells from the cyst wall.
Only normal salivary gland tissue was collected when imaging clearly shows a separate lump. This can mean the needle sampled the gland next to the lump rather than the lump itself.
Approximate risk of cancer: about 15%. This number surprises many people. It reflects the fact that some cancers, particularly cystic or slow-growing ones, can shed few cells into a needle sample. The figure is highest in studies that include only patients who went on to surgery; across all patients with this result, the true risk is likely lower.
Commonly considered next steps. The result is compared with the physical exam and imaging. Often the next step is a repeat FNA, ideally with ultrasound guidance and aimed at any solid area. If a repeat sample is also insufficient, or if the lump looks worrisome on exam or imaging, a core needle biopsy is often considered. In some situations, especially when imaging and examination strongly suggest a tumor that would be removed anyway, the next step is surgery.
Category II: Non-neoplastic
What it means. The cells show no sign of a tumor. The findings point to a non-tumor condition such as inflammation, infection or a reactive process.
Findings and conditions usually placed here.
Sialadenitis: inflammation of the salivary gland, which can be sudden (acute, often from infection) or long-standing (chronic, often linked to blocked ducts or stones).
Sialadenosis: a non-inflammatory, painless enlargement of the glands, sometimes linked to diabetes, nutritional problems, alcohol use or certain medicines.
Reactive lymph node: lymph nodes within or next to the parotid gland that have enlarged in response to an infection or irritation.
Granulomatous inflammation: a particular pattern of inflammation seen with conditions such as sarcoidosis or certain infections, including tuberculosis.
Certain benign cysts, for example some lymphoepithelial cysts.
Approximate risk of cancer: about 11%. Although most lumps in this category are truly benign, a small number turn out to be a tumor or cancer that the needle missed. Examples include a cancer with a large cystic component, a slow-growing (low-grade) cancer, or a lymphoma that looked like a reactive lymph node.
Commonly considered next steps. Treatment is aimed at the underlying condition, for example antibiotics, hydration, warm compresses, gland massage or management of a salivary stone (see treatments for salivary gland conditions). Non-tumor conditions are generally managed without surgery. Follow-up with repeat examination and imaging is common. If the lump does not shrink or go away as expected, grows, or does not fit the picture on imaging, further testing such as a repeat biopsy or core needle biopsy is usually considered.
Category III: Atypia of undetermined significance (AUS)
What it means. The sample contains some cells that look a little unusual ("atypical"), but there are too few of them, or the changes are too mild, to tell whether they come from a tumor or from a harmless reactive process. In plain terms, the pathologist sees something that is not clearly normal but not clearly a tumor either. Low numbers of cells and the presence of lymphocytes (a type of white blood cell) are common reasons for this result.
Findings usually placed here.
A scant sample with a few atypical cells.
Cells changed by inflammation or healing that could imitate a tumor.
Thick, mucus-like cyst fluid without enough lining cells to identify the source. Some low-grade cancers can produce this kind of fluid.
An unusual population of lymphocytes in which lymphoma cannot be fully ruled out.
Cells showing squamous change (a skin-like cell type) without clear features of either a benign or a malignant process.
Approximate risk of cancer: about 30%. This category is one of the most variable between hospitals. One comparison across institutions found risks ranging from 0% to 73%, so the true risk in any given practice may be higher or lower.
Commonly considered next steps. The second edition of the Milan System treats an AUS result as a reason to get more or better tissue rather than simply accepting the uncertainty. Options include a repeat FNA, often with ultrasound guidance and with material saved for extra laboratory tests, or a core needle biopsy, which gives a larger sample. When imaging or symptoms raise concern, or when the lump would likely be removed anyway, surgery may be chosen. If lymphoma is suspected, additional sampling for specialized blood-cell testing is usually arranged.
Category IV: Neoplasm
A neoplasm is a growth or tumor, meaning an abnormal collection of cells that keeps growing on its own. "Neoplasm" does not automatically mean cancer. This category has two subcategories.
Category IVA: Benign neoplasm
What it means. The sample shows a tumor with the features of a recognizable benign (non-cancerous) type.
Tumors usually placed here.
Pleomorphic adenoma: the most common salivary gland tumor. It is benign but tends to keep growing slowly. If left for many years, it carries a small but real risk of turning into cancer. It can also come back if not completely removed.
Warthin tumor: the second most common benign parotid tumor. It is more common in older adults and in people who smoke, and it can occur in both parotid glands. It has a very low risk of becoming cancer.
Other benign tumors when their features are typical, such as oncocytoma, some basal cell adenomas, lipoma (a fatty tumor) and schwannoma (a nerve sheath tumor).
Approximate risk of cancer: less than 3%. This is one of the most reliable and consistent categories across studies.
Commonly considered next steps. Options typically include surgical removal or monitoring, depending on the type of tumor and the patient (see treatment of salivary gland tumors).
Pleomorphic adenomas are usually removed, because they tend to keep enlarging, can become harder to remove as they grow, and carry a long-term risk of turning cancerous.
Warthin tumors are often suitable for observation with periodic examination and imaging, especially in older adults, people with other health conditions that make surgery riskier, or people who prefer to avoid surgery. In one large study, FNA identified Warthin tumors correctly about 98% of the time, observed tumors grew on average only about 0.3 millimeters per year, and no observed patient was later found to have cancer. FNA can occasionally mistake certain low-grade cancers for Warthin tumor, which is one reason follow-up matters. Surgery remains an option if the lump grows, becomes painful or inflamed, or is bothersome in appearance.
Monitoring may also be reasonable for some other benign tumors in selected patients.
Category IVB: Salivary gland neoplasm of uncertain malignant potential (SUMP)
What it means. The sample clearly shows a tumor, but the pathologist cannot tell from the cells alone whether it is benign or cancerous. This happens because several benign and slow-growing malignant salivary tumors look very similar under the microscope. A small needle sample also cannot show how the tumor relates to surrounding tissue, which is one of the key features that separates certain cancers from benign tumors. A SUMP result may also be given when the clinical or imaging findings do not match what the cells suggest.
Findings usually placed here.
Tumors made of crowded cells with a similar look. Pathologists group these by cell appearance into three main patterns: "basaloid" (small cells with little surrounding cytoplasm), "oncocytic" or "oncocytoid" (cells with plentiful pink, granular cytoplasm) and "clear cell" (cells with pale, clear-looking cytoplasm).
Benign tumors, such as a pleomorphic adenoma, that show some unusual features in the sample.
Some low-grade (slow-growing) cancers that cannot be confidently recognized on a needle sample, such as certain low-grade mucoepidermoid carcinomas or basal cell adenocarcinomas.
Lumps with clearly aggressive-looking features generally do not belong in this category. They are usually placed in Category V or VI.
Approximate risk of cancer: about 35%. Individual studies have reported risks from roughly 23% to 38%. The large majority of lumps in this category are true tumors of some kind, whether benign or malignant.
Commonly considered next steps. Surgical removal is the usual next step, because the lump is known to be a tumor and the only way to be certain of its type is to examine all of it. Before surgery, some teams use extra laboratory tests on the FNA sample or a core needle biopsy to help identify the tumor type, especially when the answer would change the plan for surgery. During surgery, a frozen section (described below) may help guide decisions.
Category V: Suspicious for malignancy
What it means. The sample shows features that strongly suggest cancer, but not quite enough to make a definite diagnosis. Often the cells look clearly abnormal but are too few in number, are partly obscured by blood or inflammation, or fit cancer but not perfectly.
Findings usually placed here.
Small numbers of cells with clearly abnormal features suggesting a high-grade (faster-growing) cancer.
Samples that suggest a particular type of cancer but are incomplete.
Possible lymphoma when confirmatory testing could not be done.
Approximate risk of cancer: about 83%. This figure has been fairly consistent across hospitals.
Commonly considered next steps. Surgery is the usual next step. A frozen section during the operation is often used to confirm the diagnosis and help the surgeon decide how much tissue to remove. Imaging such as MRI or CT is commonly done to look at the extent of the lump and nearby lymph nodes. Some teams use a core needle biopsy before surgery to pin down the exact type and grade, which can affect surgical planning. If lymphoma is suspected, a core or surgical biopsy for specialized testing is often preferred over removing the gland.
Category VI: Malignant
What it means. The sample shows definite cancer. When possible, the pathologist also names the type of cancer and indicates whether it appears low-grade (usually slower-growing and less likely to spread) or high-grade (usually faster-growing and more likely to spread to lymph nodes or elsewhere).
Tumors usually placed here.
Primary salivary gland cancers (cancers that started in the gland). There are many types, for example:
Mucoepidermoid carcinoma, which can be low-, intermediate- or high-grade.
Acinic cell carcinoma and secretory carcinoma, usually low-grade.
Adenoid cystic carcinoma, which often grows slowly but persistently and tends to spread along nerves.
Salivary duct carcinoma, high-grade.
Carcinoma ex pleomorphic adenoma, a cancer arising in a long-standing pleomorphic adenoma.
Lymphoma: a cancer of the immune system that can arise in the gland or its lymph nodes.
Metastatic cancer: cancer that spread to the gland or its lymph nodes from another site. In the parotid region this is most often a skin cancer of the face or scalp, such as squamous cell carcinoma or melanoma.
Approximate risk of cancer: more than 98%. This is the most stable category across studies.
Commonly considered next steps. Care is usually planned by a multidisciplinary team, which may include a head and neck surgeon, radiation oncologist, medical oncologist, radiologist and pathologist.
For most primary salivary gland cancers, surgery is the main treatment (see treatment of parotid tumors and treatment of submandibular gland tumors). How much tissue is removed, whether neck lymph nodes are also removed, and whether radiation or other treatment is added afterward depend on the cancer's type, grade, size and spread. High-grade cancers more often need lymph node surgery and added treatment. Imaging (often MRI and/or CT, sometimes chest imaging or PET/CT) is commonly used to look for spread.
Lymphoma is usually not treated by removing the salivary gland. Further biopsy may be needed to classify it, and treatment is generally directed by a blood cancer specialist.
For metastatic cancer, treatment depends on where the cancer started and how far it has spread.
Other tests and procedures that may come up
Repeat FNA
A repeat FNA is commonly used after a non-diagnostic or AUS result, or when results do not match the clinical picture. A repeat procedure is often more useful when done with ultrasound guidance, aimed at the solid part of the lump, with on-site checking of the sample when available, and with extra material saved for laboratory testing. Simply repeating the same procedure the same way may produce the same unclear result, which is why teams often change the approach the second time.
Core needle biopsy
A core needle biopsy uses a somewhat larger, hollow needle, almost always guided by ultrasound, to remove one or more thin cylinders ("cores") of tissue. It is done with local numbing medicine.
Advantages.
The tissue keeps its natural structure, so the pathologist can see how cells are arranged and how the tumor relates to surrounding tissue.
It provides more material for special stains and genetic tests.
Insufficient samples, and the need for repeat procedures, are less common than with FNA.
In pooled research, it detects cancer more often than FNA does. One analysis found that it detected about 92% of cancers compared with about 65% for FNA; a more recent analysis found about 99% compared with about 68%.
Disadvantages.
It is slightly more invasive.
Minor bleeding or bruising is somewhat more common. Serious complications, such as facial nerve injury or tumor spread along the needle track, remain very rare when it is done under ultrasound guidance.
For lumps deep in the parotid gland close to the facial nerve, some doctors prefer a fine needle.
Practices differ. Some centers and recent analyses favor core needle biopsy as the first test. Others use FNA first and reserve core biopsy for situations where FNA is insufficient or unclear, or where the exact tumor type or grade would change the treatment plan, including suspected lymphoma.
Frozen section
A frozen section is a rapid examination done during surgery. The surgeon sends a piece of the removed tissue to the pathologist, who quickly freezes it, cuts thin slices and looks at it under the microscope while the patient is still under anesthesia.
Frozen sections can help the surgeon decide during the same operation whether to remove more of the gland, whether to remove nearby lymph nodes, and whether the edges of the removed tissue are clear of tumor. In experienced hands, frozen section is very good at telling benign from malignant tumors. It is less reliable at naming the exact type of cancer, and some results remain uncertain until the full final analysis is complete, which usually takes several days. Major decisions that cause lasting harm, such as cutting the facial nerve, are generally not made on the basis of an uncertain needle biopsy or frozen section result alone. Surgeons rely heavily on what they see during surgery, and a facial nerve that works well before surgery is preserved whenever possible. Frozen section is not available at every hospital and works best with a pathologist experienced in salivary gland disease.
Imaging
Ultrasound is often the first imaging test for a salivary gland lump. MRI or CT scans may be used to look at the whole gland, including the deep part of the parotid gland that ultrasound cannot always see, to check nearby lymph nodes, and to help plan surgery. The second edition of the Milan System includes a chapter on imaging, reflecting the view that imaging findings are always considered together with the FNA result.
The limits of needle biopsy
FNA is a valuable, safe and inexpensive first test, and it can spare many people unnecessary surgery. However, it has real limits.
Only a small sample is taken. A needle collects cells from a few spots, and a tumor can look different in different areas. The sampled cells may not represent the whole lump.
Some cancers are hard to detect. Slow-growing (low-grade) cancers and cancers with large cystic, fluid-filled areas are the most likely to be missed or underestimated. In one recent surgical study, FNA and the final diagnosis disagreed about whether a low- or intermediate-grade cancer was malignant in nearly 90% of such cases, compared with about 20% for high-grade cancers. Overall, research shows that FNA misses a meaningful minority of salivary cancers, with false-negative rates as high as about 20% in some studies. When FNA does call a lump cancer, it is usually correct.
Structure is lost. FNA shows individual cells and small clusters but not how the tumor is built or whether it is invading surrounding tissue. These are key clues for some diagnoses.
Salivary tumors are diverse and look alike. There are dozens of types of benign and malignant salivary tumors, and many share similar-looking cells. Even when a lump is correctly identified as cancer, the exact type and grade may not be clear until the whole tumor is examined.
Results depend on experience. Accuracy depends on the skill of the person taking the sample, the quality of the slides and the experience of the pathologist. Pathologists do not always agree on the same sample, especially in the in-between categories.
A "benign" result is not a guarantee. Because of these limits, a lump that grows, becomes painful, becomes fixed in place, causes facial weakness or numbness, or is accompanied by new lumps in the neck is usually re-evaluated even after a reassuring biopsy.
The final diagnosis comes from the removed tissue. When a lump is surgically removed, the full microscopic examination of the entire specimen (the "final pathology") provides the definitive diagnosis, and treatment after surgery is based on that result.
For these reasons, needle biopsy results are interpreted together with the history, physical examination and imaging rather than on their own.
Questions patients often ask their care team
Which Milan category was the result, and what does it mean in this particular situation?
Was the sample adequate? Was it taken with ultrasound guidance?
Were any extra laboratory tests done on the sample, or could they be?
Does the imaging fit the biopsy result?
Would a repeat FNA or a core needle biopsy add useful information before deciding on treatment?
If surgery is planned, how much of the gland is likely to be removed? Will frozen section be used?
What are the risks to the facial nerve and other structures with surgery?
If monitoring is an option, how often will checkups and imaging be done, and what changes would prompt an earlier visit?
Would a second opinion on the slides from a pathologist experienced in salivary gland disease be helpful?
More detail on the tumors themselves is in the Learning Center page on salivary gland tumors.