Cancer of unknown primary

CUP means metastatic cancer has been confirmed but its original site remains unidentified after an appropriate work-up. The place where a metastasis is found is not necessarily where the cancer began. A liver biopsy showing carcinoma does not automatically mean primary liver cancer. If a primary is subsequently established, care can move to that cancer’s pathway.

Symptoms depend on the sites involved, such as a new lump, persistent pain, breathlessness, weight loss or abdominal swelling. These symptoms are not specific to CUP. The useful clinical question is not only “where did it start?” but also “which type is it, how far has it spread, and what can be treated now?”

An adequate tissue sample enables morphology and immunohistochemistry to distinguish carcinoma from lymphoma, melanoma, sarcoma and other malignancies. Further stains can suggest an organ, but no isolated stain should outweigh the complete clinical picture. Where feasible, preserve tissue for molecular testing rather than exhausting a small biopsy with indiscriminate tests.

Ask whether a specialist pathology review would change classification. A second biopsy can be useful if the original material is insufficient or discordant with the scans, but repeating an invasive test needs a specific expected benefit.

The work-up commonly includes a detailed history and examination, blood tests, and CT of the chest, abdomen and pelvis. Endoscopy, breast imaging or specialist examinations are selected from symptoms, metastatic pattern and pathology. Broad tumor-marker panels can produce misleading clues; useful markers answer a particular clinical question.

PET–CT is particularly useful when finding another site would change a potentially radical local plan, such as selected neck-node or limited metastatic disease. It is not a guaranteed primary-finding test. Before another procedure, ask how a positive or negative result would change treatment and whether waiting would delay needed care.

Some presentations resemble a known cancer closely enough to justify its treatment approach. Examples include isolated axillary adenocarcinoma with a breast pattern, serous peritoneal disease with a gynecologic pattern, or cervical-node squamous cancer requiring a head-and-neck assessment. Selected limited metastases may be considered for definitive local treatment after staging.

A favorable group is a treatment category, not a guarantee of cure. Neuroendocrine carcinoma with an unidentified primary is now classified and managed as a neuroendocrine malignancy rather than simply placed in an old “favorable CUP” list. Precise classification is more useful than a broad label.

A confirmed NEC diagnosis has a separate guide.

Tissue-of-origin assays use a tumor’s expression pattern to predict its likely source. Comprehensive genomic profiling searches for alterations that may guide drugs, regardless of the organ. A prediction is not direct visualization of a hidden primary, and a mutation is not automatically actionable. Results should be reconciled with pathology and reviewed for the strength of treatment evidence.

Discuss tissue adequacy, test turnaround, the probability of finding an actionable result, and whether the proposed drug is realistically accessible. Blood-based profiling can be considered in selected circumstances, but a negative blood result may reflect limited tumor DNA in the sample rather than absence of all useful targets.

CUPISCO enrolled selected patients with unfavorable, nonsquamous CUP who achieved disease control after three cycles of platinum chemotherapy. Molecularly guided treatment then improved progression-free survival over continued chemotherapy. The September 2026 update reported median progression-free survival of 6.1 versus 4.4 months; the overall-survival difference was not statistically significant. These are results from randomization, not a prediction of an individual’s life expectancy.

Fudan CUP-001 tested a different strategy: first-line treatment chosen using a 90-gene tissue-of-origin assay. In this single-center randomized study, median progression-free survival was 9.6 versus 6.6 months compared with empirical chemotherapy. It supports the potential value of a validated origin-directed approach, but does not prove that every commercial test or every predicted organ leads to better treatment.

If no specific treatment group or actionable result is found, an empirical platinum-based combination may be offered when fitness and organ function permit. “Empirical” means a regimen selected for the clinical and pathological situation without a confirmed primary; it does not mean random treatment. Agree on the intended benefit and when scans will determine whether to continue.

Radiation, surgery or another local intervention may relieve a painful lesion, obstruction or another immediate problem even when systemic treatment is also needed. A rapidly worsening symptom should be addressed while the diagnostic review proceeds. For extensive disease and poor fitness, the burden of chemotherapy may outweigh its likely benefit.

A qualifying MSI-H/dMMR or other supported biomarker can open a tumor-agnostic treatment pathway, subject to the drug’s exact indication. For example, the US pembrolizumab MSI-H/dMMR solid-tumor label includes prior-treatment and lack-of-satisfactory-alternative conditions. A generic “immune-positive” report is not enough; obtain the actual test and result.

Japan has a nivolumab indication for CUP. The September 2026 Japanese label requires an appropriate systemic and pathological work-up confirming an epithelial malignancy of unknown origin and no recommended treatment based on location or histology; combination use for CUP is not established in that label. This regional approval is not a worldwide CUP approval or a guarantee of access through Chirona.

Pain, nausea, breathlessness, fluid accumulation and nutritional problems can be treated while anticancer options are assessed. Palliative care means active symptom and quality-of-life care and can accompany treatment. Ask for a named contact so new symptoms do not get lost between different organ specialists.

Follow-up imaging should have a purpose: measuring response, identifying complications or planning another useful treatment. The schedule depends on the treatment and disease course. A new symptom can justify earlier review; repeatedly searching for the primary without a likely management benefit can add burden without improving care.

Bring the pathology report and available tissue, all immunostains rather than only the conclusion, original scans, endoscopy and examination reports, molecular results, and a treatment timeline. Include any previous cancer diagnosis: a recurrence may need to be distinguished from a new CUP.

Ask which diagnoses have been excluded, whether a recognizable treatment group is present, and whether the proposed molecular test seeks an origin or a drug target. For each additional test or treatment, clarify the expected decision, turnaround, evidence, cost and access. A coordinated plan can proceed even when the primary remains unknown.