Cancer of Unknown Primary
In cancer of unknown primary, metastatic disease is confirmed but the site of origin is not established after standard evaluation. When anatomical origin cannot answer the question, biology carries proportionally more of the weight.
Disease categories
Major disease categories
- Adenocarcinoma of unknown primary
- Poorly differentiated carcinoma
- Squamous cell carcinoma of unknown primary
- Neuroendocrine carcinoma of unknown primary
- Favourable-subset presentations
Molecular
Molecular considerations
Immunohistochemistry panels are the first analytical step towards tissue of origin. Gene expression and DNA methylation-based classifiers are studied as further tools. Comprehensive genomic profiling serves two purposes here: it may suggest a likely origin, and it may identify alterations relevant irrespective of origin.
Markers commonly discussed in cancer of unknown primary
- IHC panel (CK7, CK20, TTF-1, CDX2, GATA3, PAX8 and others)
- MMR / MSI
- TMB
- NTRK fusions
- HER2
- Methylation / expression classifiers
Not exhaustive, and not every marker is assessed in every patient. Which of these is relevant depends on the individual case and should be discussed with the treating team.
Testing
Genomic testing considerations
Comprehensive profiling is a central part of the discussion in cancer of unknown primary, since tissue-agnostic findings may be relevant even where origin remains unresolved.
Germline
Germline considerations
Considered where the molecular or clinical picture suggests it.
Molecular monitoring
ctDNA and liquid biopsy considerations
ctDNA may contribute where tissue is limited, though it does not resolve tissue of origin on its own.
Single-cell & computation
Single-cell and algorithmic approaches
Cancer of unknown primary is a computational problem as much as a pathological one. Tissue-of-origin classification from gene expression and DNA methylation data is the principal algorithmic contribution — methods trained across large reference cohorts that assign a probable origin where immunohistochemistry has been equivocal. Single-cell analysis contributes by characterising cellular composition where the histology is ambiguous and a mixed population is suspected. Alongside this sits evidence-synthesis work: where origin remains unresolved, the reasoning has to run on tissue-agnostic alterations, and assembling that evidence reliably is its own retrieval and ranking problem.
In practice
What a review of this disease typically surfaces
Cases usually arrive after a broad but unsystematic workup. The immunohistochemistry panel has often been performed piecemeal rather than as a structured algorithm, and the results are difficult to interpret as a set. Comprehensive genomic profiling is frequently absent despite being one of the few things that can both suggest an origin and identify tissue-agnostic options. A recognised favourable subset — disease that should be managed as a specific cancer rather than as CUP — is sometimes missed.
Patterns observed across precision oncology case reviews. Every case differs; these are the questions worth checking, not conclusions about any individual patient.
Decisions
Treatment decision questions
Questions worth raising with a treating team when a decision is being made:
- Has a systematic immunohistochemistry panel been performed and documented?
- Has expert pathology review been obtained?
- Has comprehensive genomic profiling been performed?
- Does the presentation fit a recognised favourable subset that would be managed as a specific disease?
Resistance
Resistance and recurrence questions
- Has the diagnosis been revisited as the disease pattern has evolved?
- Has a primary become apparent on later imaging?
Trials
Clinical trial considerations
Tissue-agnostic and biomarker-defined trials are of particular relevance in this setting.
When a precision oncology review may help
A review is most useful where the standard pathway has been exhausted or where the picture is more complicated than a protocol accounts for — rare presentations, treatment resistance, complex or apparently contradictory molecular findings, or a decision between several defensible options.
The output is a structured analysis and a set of questions for the treating team. It is not treatment, and it does not replace the treating oncologist.