Blood-Based Early Cancer Detection: What the Evidence Shows

The premise

Tumours shed fragments of DNA into the bloodstream. If those fragments can be detected and traced to a tissue of origin before symptoms appear, screening could in principle extend to cancers that have no established screening pathway at all — pancreatic, ovarian, oesophageal. That premise explains the extraordinary attention multi-cancer early detection assays have received.

What has been demonstrated

Large validation cohorts have shown that methylation-based assays can distinguish cancer from non-cancer plasma with high specificity, and can predict the tissue of origin with useful accuracy in most positive cases. Sensitivity rises sharply with stage: performance in advanced disease is strong, while detection of stage I disease — the stage where screening earns its value — remains considerably lower for most tumour types.

What has not been demonstrated

No multi-cancer test has yet shown a reduction in cancer-specific mortality in a randomised trial. This is the outcome that matters, and it is the reason major guideline bodies have not recommended these assays for routine population screening. Detecting more cancer earlier is not automatically the same as preventing death from it, a lesson the field learned expensively in other screening programmes.

Overdiagnosis and the arithmetic of specificity

Even a highly specific test generates false positives when applied to a population where most people do not have cancer. Each false positive triggers a diagnostic cascade — imaging, sometimes biopsy — with its own cost, radiation, and anxiety. There is also the harder problem of overdiagnosis: identifying indolent disease that would never have caused symptoms, and treating it anyway.

How these tests are actually being used

In current practice, multi-cancer assays are largely used as a supplement to, not a replacement for, established screening. Guidance consistently emphasises that a negative result does not exempt anyone from mammography, colonoscopy, or lung cancer screening. That message is easily lost when a test is marketed directly to patients.

What to watch

Several large randomised trials are underway with mortality endpoints and multi-year follow-up. Their results, rather than any additional accuracy data, will determine whether these assays enter guidelines. Until then, the honest clinical position is that this is a promising technology with an incomplete evidence base — a distinction worth making carefully with patients who arrive having read the headlines.