A detected variant is not an actionable one.

Overview

Seven statements that are not interchangeable

Each step requires different evidence and a different kind of judgement. A variant can stop at any step. Moving from one to the next is never automatic.

1. 01
   Detected

   The sequencing assay observed a difference from the reference sequence that passed quality filters.

   Laboratory assay
2. 02
   Annotated

   The variant is described: gene, transcript, predicted protein change, population frequency and database entries.

   Computational
3. 03
   Biological or functional significance

   There is evidence that the variant changes how a gene or protein works.

   Evidence-derived
4. 04
   Associated with disease or phenotype

   The variant, or variants like it, are associated with a cancer type, inherited condition or phenotype.

   Evidence-derived
5. 05
   Associated with treatment sensitivity or resistance

   Evidence links the variant to response or resistance to a therapy, in a specific tumour context.

   Evidence-derived
6. 06
   Potential clinical actionability

   The strength and relevance of that evidence may justify considering a clinical action for this patient.

   Expert-reviewed
7. 07
   Reviewed and incorporated into a clinical decision

   The treating oncologist weighs the finding with everything else they know and decides.

   Treating oncologist

01

A variant is detected

Detection means the assay observed a difference from the reference sequence that passed its quality filters. It says nothing yet about what the variant does.

02

A variant is annotated

Annotation describes the variant: which gene and transcript it affects, the predicted change to the protein, how common it is in the general population, and whether it appears in curated databases.

03

It has evidence of biological or functional significance

Functional significance means there is evidence the variant changes how a gene or protein works — from laboratory experiments, recurrence at known hotspots, or structural knowledge. Computational predictions of effect are supporting evidence, not proof. Many variants have no functional evidence at all.

04

It is associated with a disease or phenotype

An association links a variant — or variants like it — to a cancer type, an inherited condition or a phenotype. Associations are context-specific: a variant well characterised in one tumour type may be poorly understood in another.

05

It is associated with treatment sensitivity or resistance

Some variants are linked by evidence to response or resistance to particular therapies. That evidence varies hugely in strength — from regulatory approvals and guidelines to clinical trials, case reports and laboratory models — and it applies in a specific tumour context.

06

There is evidence for potential clinical actionability

Potential actionability means the strength and relevance of the evidence may justify considering a clinical action for this patient. It is still a potential: approval status differs between countries and changes over time, and the patient's whole situation matters.

07

The finding is reviewed and incorporated into a clinician's decision

The final step belongs to people, not software. The treating oncologist weighs the finding alongside the diagnosis, stage, prior treatment, comorbidities, the patient's goals and the Molecular Tumour Board's discussion — and decides.

Two contexts

Somatic and germline interpretation are different questions

A somatic variant was acquired by the tumour. A germline variant is present from birth in every cell and may be inherited. They are interpreted with different frameworks and have different implications, so one classification framework is never applied as though it covered both.

Uncertainty

Variants of uncertain significance (VUS)

A VUS is a variant whose effect is not yet known. Uncertain findings are reported as uncertain, not upgraded to fill a gap.

• A VUS does not establish a diagnosis.
• A VUS should not, on its own, direct therapy.
• A VUS may be reclassified as evidence grows — in either direction.

Context

Clinical, molecular and phenotype context

Interpretation depends on the cancer type, stage, prior treatment, sample type and the patient's situation. Twin shows which context an interpretation depends on and flags when that context is missing.

Biomarkers

Biomarker and therapeutic associations

Some biomarkers are not single variants but composite measurements — such as tumour mutational burden, microsatellite instability or homologous recombination deficiency, all listed among Twin's integrated analytical features. Their values and thresholds depend on the assay and the indication, and are reported with that context.

Trials

Clinical-trial eligibility information

A biomarker can make a trial worth reviewing. It does not make a patient eligible: eligibility depends on many criteria and is confirmed only by the trial team after screening.

Display

How evidence is shown

Every finding in a Twin report is intended to carry its interpretation, source, strength, supporting and conflicting evidence, required context, limitations, the date it was last checked and its review status. The panel below shows the structure using a fictional example.

Over time

Reassessment when evidence changes

Evidence changes: variants are reclassified, new therapies are approved, guidelines are updated and trials open and close. During the six-month journey, relevant evidence changes can be surfaced for the treating oncologist, who decides whether reassessment is warranted.