WebDoctor Encyclopedia

Oncology

Cancer of unknown primary (CUP)

Cancer of unknown primary (CUP) is a cancer that has spread in the body, but whose starting point cannot be found despite full tests. Treatment depends on the cell type, gene changes and where it has spread.

Medically reviewed Last reviewed October 21, 2026

What is cancer of unknown primary?

Cancer of unknown primary (CUP) is a cancer that has already spread in the body (metastatic cancer), but doctors cannot find where it started, even after a full set of standard tests. The place where a cancer first grows is called the primary tumour, and the deposits elsewhere are called secondaries or metastases.

In most people with cancer, the primary is easy to find. In CUP, the doctors find secondaries in places such as the liver, lymph nodes, lungs or bones, but no primary. The European Society for Medical Oncology (ESMO) defines CUP as a carcinoma or undifferentiated cancer for which a standardised work-up fails to identify the primary.

CUP makes up about 2-5% of all cancers and is becoming less common as scans and laboratory tests improve. About 1 in 5 people have a “favourable” type that is treated like a known cancer and can do well. The other 4 in 5 have an “unfavourable” type, which is harder to treat. Treatment is guided by what the cancer cells look like, their gene changes, where the cancer has spread and how fit the person is.

1. How common is it?

Feature Details
Share of cancers Under 5% of cancers (ESMO); about 2% of cancers in the UK
Numbers Around 7,900 people a year in the UK (Cancer Research UK)
Trend Falling. In Zurich, the rate dropped from 17.6 per 100,000 in 1997 to 5.8 per 100,000 in 2014, probably because primaries are now found more often
Age Rises with age; almost 60% of UK cases are in people aged 75 and over
Sex Slightly more common in men
Cell type About 50% well or moderately differentiated adenocarcinoma, about 30% poorly differentiated adenocarcinoma or undifferentiated carcinoma, about 15% squamous cell carcinoma and about 5% undifferentiated cancers

Sarcomas, melanomas, lymphomas, germ cell tumours and neuroendocrine tumours with no obvious starting point are not counted as CUP. These cancers are treated according to their own type, wherever they started.

2. Causes and risk factors

2.1 Why is the primary not found?

Nobody knows for certain why the primary stays hidden. The main explanations are:

  • the primary is too small to be seen on scans, while its secondaries grow quickly;
  • the primary may have shrunk or disappeared, possibly because of the immune system;
  • some cancers seem to have a built-in tendency to spread very early.

In older post-mortem studies of people who died of CUP, a small primary was found in about 73%. The most common sites were the lung (27%), pancreas (24%), liver and bile ducts (8%) and kidneys (8%), followed by the bowel, genital organs and stomach.

2.2 Risk factors

Risk factor What is known
Smoking Risk rises with the amount smoked: about 1.8 times higher at 1-15 cigarettes a day and 4.1 times higher at more than 25 a day (EPIC study)
Type 2 diabetes About 1.8 times higher risk
Some autoimmune diseases Higher risk after polymyositis or dermatomyositis, primary biliary cirrhosis and Addison disease
Family history CUP, or some other cancers, in close relatives
Other possible factors High body mass index, large waist size and lower income
Earlier cancer About 25% of people with suspected CUP have had cancer before, so a return of the earlier cancer must always be ruled out

3. How does it look?

CUP does not have one typical appearance. What doctors see depends on where the cancer has spread and on the type of cancer cell.

3.1 Where the secondaries appear

Many people have secondaries in several organs at once. The liver is the most common site, followed by the lungs, lymph nodes, the lining of the abdomen (peritoneum), bone and brain.

Site What may be noticed or seen
Lymph nodes A firm, painless lump in the neck, armpit or groin
Liver Enlarged liver; several round spots on scans
Lungs and pleura Nodules on a chest X-ray or CT scan; fluid around the lung
Peritoneum Fluid in the abdomen (ascites) and thickened, nodular lining
Bone Areas of bone damage on scans; in men, dense (“blastic”) patches in the pelvis and lower spine can suggest prostate cancer
Brain One or more round lesions on an MRI scan
Skin Firm nodules under or in the skin

3.2 On scans

A CT scan of the neck, chest, abdomen and pelvis shows how far the cancer has spread. Radiologists also look for clues that one of the lesions is in fact the primary, such as its size, shape and the pattern of nearby lymph node spread. A PET-CT scan shows active cancer throughout the body, but it finds the primary in only about a third of cases (37% in a meta-analysis).

3.3 Under the microscope

A biopsy is the key test. A pathologist first decides the broad cell type: carcinoma, lymphoma, melanoma, sarcoma or another cancer. Most CUPs are carcinomas, and adenocarcinoma (a gland-forming carcinoma) is the most common.

The pathologist then uses immunohistochemistry (IHC). These are stains for proteins that are typical of certain organs.

Stain What a positive result may suggest
Keratins (AE1/AE3) Carcinoma (cancer of the lining tissues)
CD45 Lymphoma or another blood cancer
SOX10, S100 Melanoma
CK7 and CK20 The pattern helps narrow the site; CK7-negative, CK20-positive is typical of bowel cancer
TTF1, Napsin A Lung (but about 40% of lung cancers are TTF1-negative)
CDX2, SATB2 Bowel
GATA3 Breast or bladder
PSA, PSMA, NKX3.1 Prostate
PAX8 Kidney, ovary or thyroid
Synaptophysin, INSM1 Neuroendocrine tumour
Calretinin, WT1, CK5/6, D2-40 Mesothelioma

The report may read something like “metastatic adenocarcinoma, CK7-positive, CK20-negative, TTF1-negative; primary site not identified”. A diagnosis of CUP is made only when the full work-up has not found a primary.

4. Symptoms

There are no symptoms specific to CUP. The symptoms come from the secondaries and depend on where they are.

Site of spread Common symptoms
General Tiredness, loss of appetite, unexplained weight loss, feeling unwell, sometimes fever or night sweats
Lymph nodes A lump or swelling in the neck, armpit or groin
Liver Pain or discomfort under the right ribs, swollen abdomen, nausea, yellow skin and eyes (jaundice)
Lungs and pleura Cough, breathlessness, chest pain, coughing up blood
Peritoneum Bloating, swelling of the abdomen, feeling full quickly, constipation or sickness
Bone Constant, deep bone pain, often worse at night; bone breaking after minor strain
Brain Headaches, fits (seizures), weakness, confusion, changes in vision or speech
Skin Firm lumps that may be painless

4.1 Urgent symptoms

Seek urgent medical help for:

  • new back pain with leg weakness, numbness or problems passing urine or controlling the bowels (possible spinal cord compression);
  • extreme thirst, confusion, constipation and sickness (possible high calcium level);
  • sudden severe breathlessness, chest pain or a first seizure.

Many of these symptoms are far more often caused by harmless conditions. Symptoms that last longer than a few weeks, keep getting worse or have no clear explanation should be checked by a doctor.

5. Diagnosis

5.1 The stages of diagnosis

The UK National Institute for Health and Care Excellence (NICE) describes three steps:

Term Meaning
MUO (malignancy of undefined primary origin) Cancer has been found in a secondary site after a few tests, but before full investigation
Provisional CUP A biopsy shows a carcinoma or neuroendocrine cancer, and initial tests have found no primary
Confirmed CUP Specialist review and any further tests have still found no primary

NICE recommends that every cancer centre has a CUP team, including an oncologist, a palliative care doctor and a specialist nurse or key worker.

5.2 Tests

Test Purpose
History and examination Including the breasts, lymph nodes, skin, genitals, rectum and pelvis
Blood tests Blood count, kidney and liver function, calcium and lactate dehydrogenase (LDH)
CT or MRI Neck, chest, abdomen and pelvis in everyone
Mammogram Recommended for all women by ESMO; NICE advises it only when features suggest breast cancer
Targeted tumour markers PSA in men; AFP and hCG if a germ cell tumour is possible; CA125 if ovarian cancer is possible. Markers are not used as a general screen
Endoscopy Camera tests of the stomach, bowel or airways only if symptoms, scans or the biopsy point there
Breast MRI For women with cancer in the armpit lymph nodes; it finds a breast primary in about two-thirds of those with a normal mammogram
PET-CT Recommended for neck lymph node CUP and when only one or a few secondaries are present; optional in other cases
Biopsy with IHC Essential; core biopsies are preferred because they give enough tissue for all tests
Gene testing (NGS) Looks for gene changes that point to a likely primary or can be treated with targeted drugs. ESMO says it may be done routinely

Tests that look for the “tissue of origin” by reading many genes at once (gene expression profiling) are not recommended outside research. They have not been shown to improve survival.

Further tests are not always right for everyone. NICE advises that more investigations should only be done if the results could change treatment and the person is fit and willing to have that treatment.

6. Favourable and unfavourable CUP

CUP has no stage, because it has always spread by the time it is found. Instead, doctors decide whether it belongs to a favourable subtype, which looks like a known cancer and is treated in the same way.

6.1 Favourable subtypes (ESMO 2023)

Subtype Typical features Treatment
Single site or oligometastatic One deposit, or up to five, all treatable Surgery and/or radiotherapy to every lesion
Breast-like Women with cancer only in armpit lymph nodes As for node-positive breast cancer
Ovary-like Women with serous adenocarcinoma in the peritoneum As for advanced ovarian cancer: surgery, carboplatin and paclitaxel, PARP inhibitor maintenance if suitable
Head and neck-like Squamous cell carcinoma in neck lymph nodes (not above the collarbone) Neck surgery and/or radiotherapy with or without chemotherapy
Prostate-like Men with dense bone secondaries and/or raised PSA As for metastatic prostate cancer
Colon-like CK7-negative, CK20-positive, CDX2-positive adenocarcinoma, mainly in the abdomen As for metastatic bowel cancer, e.g. FOLFOX or FOLFIRI
Kidney-like Clear cell or papillary carcinoma with kidney stains Kidney cancer drugs may be used

The 2023 ESMO guideline no longer counts neuroendocrine cancers or “poorly differentiated carcinoma of the midline” as CUP subtypes. Neuroendocrine cancers are now treated according to neuroendocrine guidelines. Young men with midline masses should be checked carefully for a germ cell tumour or the rare NUT carcinoma.

6.2 Unfavourable CUP

About 80% of people have unfavourable CUP. For these patients, doctors estimate the outlook with a simple score based on performance status (how active a person is) and the blood level of LDH.

Risk group Definition Results in the original study
Good risk Performance status 0-1 and normal LDH Median survival 11.7 months; 45% alive at 1 year
Poor risk Performance status 2 or more, or raised LDH Median survival 3.9 months; 11% alive at 1 year

7. Treatment

Treatment is planned by a multidisciplinary team and depends on the subtype, gene test results, symptoms and the person’s wishes.

7.1 Main treatment options

Treatment When it is used How it works and how it is given Main side effects
Site-specific treatment Favourable subtypes The same treatment as for the cancer the CUP resembles As for that treatment
Surgery and/or radiotherapy Single-site or oligometastatic CUP; isolated brain secondaries; neck or groin nodes Removes or destroys every visible deposit; PET-CT and brain MRI first Depend on the site; fatigue and skin soreness with radiotherapy
Platinum doublet chemotherapy Standard first treatment for unfavourable CUP in fit people Carboplatin with paclitaxel, or cisplatin with gemcitabine, through a drip every 3 weeks Tiredness, low blood counts and infection risk, sickness, hair loss, numb fingers (paclitaxel)
Targeted drugs Gene changes found by NGS Tablets or drips that block the faulty protein, e.g. larotrectinib or entrectinib for NTRK fusions; EGFR, ALK or ROS1 inhibitors; BRAF or RET inhibitors Vary by drug; often milder than chemotherapy
Immunotherapy MSI-high or mismatch repair-deficient cancers, high tumour mutational burden or high PD-L1, usually after chemotherapy Checkpoint drugs such as pembrolizumab or nivolumab, by drip every 2-6 weeks Tiredness, rash, diarrhoea, thyroid and other immune side effects
Palliative radiotherapy Painful bone secondaries, spinal cord compression, bleeding Short courses to relieve symptoms Usually mild and short-lived

7.2 Chemotherapy for unfavourable CUP

ESMO recommends a two-drug combination with a platinum drug. No regimen has been proven better than the others, and three-drug combinations are not recommended because of their side effects. Chemotherapy has not been compared with supportive care alone in a randomised trial, so its aims are to control the cancer, ease symptoms and extend life.

Choosing treatment by predicting the tissue of origin did not help. In a Japanese trial, 44.0% of patients given site-specific treatment were alive at 1 year, compared with 54.9% with carboplatin and paclitaxel. A French phase III trial (GEFCAPI 04) also found no advantage.

7.3 Gene testing to guide treatment

The CUPISCO trial tested a different approach in 436 people with unfavourable CUP whose cancer was controlled after 3 cycles of chemotherapy.

CUPISCO result Gene-guided treatment Further chemotherapy
Median time without progression (2024) 6.1 months 4.4 months (HR 0.72)
Updated median survival (2026) 15.2 months 12.8 months (HR 0.79; not statistically significant)

Based on these results, the trial team recommends comprehensive gene testing at diagnosis for people with unfavourable CUP.

7.4 Immunotherapy

In a Japanese phase II trial of nivolumab in people who had already had chemotherapy, the cancer shrank in 22.2%. The median time before the cancer grew again was 4.0 months, and median survival was 15.9 months. Responses were better in cancers with high PD-L1, a high mutation count or MSI-high status.

7.5 Supportive and palliative care

Good symptom control matters at every stage. Palliative care teams help with pain, breathlessness, sickness, fatigue and emotional support, alongside active treatment. A CUP specialist nurse or key worker can coordinate care. Clinical trials are encouraged for everyone with CUP.

8. Outlook

The outlook varies widely, from cure in some favourable subtypes to a short life expectancy in others.

Group Outcome
All CUP (registry data) About 20% alive at 1 year; median survival about 3 months; little improvement over time
Adenocarcinoma outside the lymph nodes 17% alive at 1 year; median 3 months (Swedish registry)
Adenocarcinoma limited to lymph nodes 41% alive at 1 year; median 8 months
Squamous cell carcinoma Better than adenocarcinoma (36% vs under 20% alive at 1 year)
Single-site or oligometastatic, treated locally Median survival 52.5 months; about a third stayed free of cancer during follow-up in one series
Unfavourable CUP in recent trials Median survival about 13-15 months in fit people whose cancer was controlled by initial chemotherapy (CUPISCO)

Registry figures include many elderly or frail people and are lower than results in clinical trials. Factors linked with a poorer outlook include poor performance status, raised LDH, low albumin, liver secondaries, many organs involved, male sex and adenocarcinoma. Cancer only in the lymph nodes and a favourable subtype predict a better outlook.

9. Follow-up and living with CUP

9.1 Follow-up

Situation Suggested follow-up (ESMO)
During and after treatment for unfavourable CUP CT or MRI about every 3 months, if the person is fit for further treatment
After surgery or radiotherapy for single-site or oligometastatic CUP CT or MRI every 3-6 months for 2 years, then every 6-12 months in years 3-5
Long-term survivors Normal cancer screening (bowel, breast, prostate, skin), as the risk of second cancers is raised
Possible inherited cancer risk Genetic counselling and testing

9.2 Coping with uncertainty

Not knowing where a cancer started can be very hard. Many people worry that something has been missed or that treatment cannot work without a primary. It helps to know that treatment is aimed at the cancer cells that are present, based on how they look and behave.

Support from the CUP team, a specialist nurse, counselling services and charities such as Cancer Research UK and Macmillan Cancer Support can make a real difference.

10. Frequently asked questions

Question Answer
Will the primary ever be found? Sometimes it shows up later, but in most people it is never found. Treatment does not depend on finding it
Is CUP a separate kind of cancer? It is a group of cancers that have spread from a hidden starting point. Some experts think CUP also has its own biology, with very early spread
Can CUP be cured? Sometimes, mainly when there is only one or a few deposits that can be removed or treated with radiotherapy, or in some favourable subtypes
Why do I need so many tests? To find a primary or a favourable subtype, and to look for gene changes that targeted drugs can treat. Tests are only done if they could change treatment
Is CUP inherited? Usually not, although having a close relative with CUP or some related cancers raises the risk a little
Can I take part in a trial? Yes. Trials are especially encouraged in CUP. Ask your team or search ClinicalTrials.gov for “cancer of unknown primary”

Resources

  1. Binder, C., Matthes, K.L., Korol, D., Rohrmann, S. and Moch, H. (2018) ‘Cancer of unknown primary – epidemiological trends and relevance of comprehensive genomic profiling’, Cancer Medicine, 7(9), pp. 4814-4824. doi:10.1002/cam4.1689.
  2. Cancer Research UK (2024) Cancer of unknown primary (CUP). Available at: https://www.cancerresearchuk.org/about-cancer/cancer-unknown-primary-cup (Accessed: October 2026).
  3. Culine, S., Kramar, A., Saghatchian, M., Bugat, R., Lesimple, T., Lortholary, A. et al. (2002) ‘Development and validation of a prognostic model to predict the length of survival in patients with carcinomas of an unknown primary site’, Journal of Clinical Oncology, 20(24), pp. 4679-4683. doi:10.1200/JCO.2002.04.019.
  4. Hayashi, H., Kurata, T., Takiguchi, Y., Arai, M., Takeda, K., Akiyoshi, K. et al. (2019) ‘Randomized phase II trial comparing site-specific treatment based on gene expression profiling with carboplatin and paclitaxel for patients with cancer of unknown primary site’, Journal of Clinical Oncology, 37(7), pp. 570-579. doi:10.1200/JCO.18.00771.
  5. Hemminki, K., Bevier, M., Hemminki, A. and Sundquist, J. (2012) ‘Survival in cancer of unknown primary site: population-based analysis by site and histology’, Annals of Oncology, 23(7), pp. 1854-1863. doi:10.1093/annonc/mdr536.
  6. Kaaks, R., Sookthai, D., Hemminki, K., Krämer, A., Boeing, H., Wirfält, E. et al. (2014) ‘Risk factors for cancers of unknown primary site: results from the prospective EPIC cohort’, International Journal of Cancer, 135(10), pp. 2475-2481. doi:10.1002/ijc.28874.
  7. Krämer, A., Bochtler, T., Pauli, C., Baciarello, G., Delorme, S., Hemminki, K. et al. (2023) ‘Cancer of unknown primary: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up’, Annals of Oncology, 34(3), pp. 228-246. doi:10.1016/j.annonc.2022.11.013.
  8. Krämer, A., Bochtler, T., Pauli, C., Shiu, K.K., Cook, N., de Menezes, J.J. et al. (2024) ‘Molecularly guided therapy versus chemotherapy after disease control in unfavourable cancer of unknown primary (CUPISCO): an open-label, randomised, phase 2 study’, The Lancet, 404(10452), pp. 527-539. doi:10.1016/S0140-6736(24)00814-6.
  9. Krämer, A., Bochtler, T., Pauli, C., Shiu, K.K., Cook, N., Janoski de Menezes, J. et al. (2026) ‘Molecularly guided therapy versus continued chemotherapy in unfavorable cancer of unknown primary: updated efficacy and safety from the randomized, phase II CUPISCO study’, Journal of Clinical Oncology, 44(30), pp. 2895-2902. doi:10.1200/JCO-25-02974.
  10. Kwee, T.C. and Kwee, R.M. (2009) ‘Combined FDG-PET/CT for the detection of unknown primary tumors: systematic review and meta-analysis’, European Radiology, 19(3), pp. 731-744. doi:10.1007/s00330-008-1194-4.
  11. National Cancer Institute (2021) Carcinoma of Unknown Primary Treatment (PDQ) – Patient Version. Available at: https://www.cancer.gov/types/unknown-primary/patient/unknown-primary-treatment-pdq (Accessed: October 2026).
  12. National Institute for Health and Care Excellence (2010, updated 2023) Metastatic malignant disease of unknown primary origin in adults: diagnosis and management (Clinical guideline CG104). Available at: https://www.nice.org.uk/guidance/cg104 (Accessed: October 2026).
  13. Pavlidis, N. and Pentheroudakis, G. (2012) ‘Cancer of unknown primary site’, The Lancet, 379(9824), pp. 1428-1435. doi:10.1016/S0140-6736(11)61178-1.
  14. Pentheroudakis, G., Golfinopoulos, V. and Pavlidis, N. (2007) ‘Switching benchmarks in cancer of unknown primary: from autopsy to microarray’, European Journal of Cancer, 43(14), pp. 2026-2036. doi:10.1016/j.ejca.2007.06.023.
  15. Pouyiourou, M., Wohlfromm, T., Kraft, B., Hielscher, T., Stichel, D., von Deimling, A. et al. (2021) ‘Local ablative treatment with surgery and/or radiotherapy in single-site and oligometastatic carcinoma of unknown primary’, European Journal of Cancer, 157, pp. 179-189. doi:10.1016/j.ejca.2021.08.019.
  16. Rassy, E. and Pavlidis, N. (2019) ‘The currently declining incidence of cancer of unknown primary’, Cancer Epidemiology, 61, pp. 139-141. doi:10.1016/j.canep.2019.06.006.
  17. Tanizaki, J., Yonemori, K., Akiyoshi, K., Minami, H., Ueda, H., Takiguchi, Y. et al. (2022) ‘Open-label phase II study of the efficacy of nivolumab for cancer of unknown primary’, Annals of Oncology, 33(2), pp. 216-226. doi:10.1016/j.annonc.2021.11.009.

Prepared by the editorial team, October 2026.