Multi-cancer early detection tests read tumour-derived fragments in circulating cell-free DNA — chiefly methylation patterns — to flag a shared cancer signal and predict its tissue of origin across dozens of cancer types, most with no screening test of any kind. The appeal is structural: current programmes screen for four or five cancers one organ at a time, while the majority of cancer deaths come from unscreened types, often diagnosed late through emergency presentation. A single annual blood test covering the unscreened majority is the biggest idea in cancer control in a generation — which is precisely why it demands the biggest evidence, and why one company (GRAIL, with its Galleri test) ran the largest randomised screening trial ever attempted to get it.

The NHS-Galleri trial randomised over 142,000 English adults aged 50-77 to annual Galleri testing plus standard screening versus standard screening alone, across three yearly rounds. The topline, reported in February 2026 and detailed at ASCO in May: the primary endpoint — a significant reduction in stage III and IV cancers combined — was missed; stage IV diagnoses fell substantially; screen-detection of cancer roughly quadrupled, with reductions in cancers diagnosed through emergency presentation; around half of positive tests led to a cancer diagnosis — an extraordinary positive predictive value for a screening test; no serious test-related adverse events occurred; and an unexpectedly high stage III count in the intervention arm complicated the primary endpoint. Effects strengthened after the first (prevalent) round, and follow-up is being extended. Both enthusiasts and sceptics left the readout with ammunition — the mark of a genuinely informative trial.

The entire debate compresses into one question: does shifting stage at diagnosis translate into fewer deaths? The optimist reading: stage IV reduction plus fewer emergency presentations is exactly the leading indicator a screening benefit shows first, and mortality data need only time. The sceptic reading, armed with the overdiagnosis hub's toolkit: stage shift can be manufactured by lead time and by preferentially detecting slower tumours (length bias), survival and detection counts are structurally untrustworthy, and screening history is littered with stage-shifting tests that never moved mortality — the missed primary endpoint keeps that possibility fully alive. Both readings are legitimate today; only extended follow-up and mortality analyses can arbitrate, and no honest party claims that verdict is in.

MCED's designers aimed the technology away from classic overdiagnosis: methylation signals skew toward aggressive, high-cell-turnover tumours, and the tests deliberately underdetect the indolent reservoir (low-grade prostate, papillary thyroid) that plagues organ-specific screening. The NHS-Galleri PPV — roughly half of positives being cancer, against the single-digit PPVs of urgent-referral pathways — supports the engineering. What no design can dodge: some detected cancers in elderly cohorts would still never have caused death; false positives, though proportionally few, trigger whole-body diagnostic odysseys when the signal's origin prediction is wrong; and a negative test cannot license skipping proven screening, since MCED sensitivity for early-stage disease is modest. The test is additive by design — and additive means adding costs too.

MCED now forces the hardest question in screening policy: act on stage-shift evidence or wait the decade mortality data requires? England's national programme decision rests on NHS-Galleri's interpretation; in the US, Galleri sells at roughly a thousand dollars a year as a lab-developed test while FDA approval, Medicare coverage legislation and the NCI's own Vanguard trial proceed in parallel; and health economists note that even a modestly priced annual test for everyone over 50 would rank among the largest line items in preventive medicine — money with proven alternative uses in vaccination, blood pressure and the screening programmes that already have mortality receipts. Between revolution and rerun of screening's oversold past, the honest position is unusually simple: the trial everyone demanded was run; its answer arrives in instalments; the mortality instalment is not yet here.

For individuals: MCED tests are commercially available, plausibly beneficial, unproven on mortality, and never a substitute for established screening — anyone buying one should budget emotionally and financially for the roughly one-in-a-hundred chance of a positive requiring a full diagnostic workup, and treat a negative as no licence to skip mammography or colonoscopy. Highest defensible use today: elevated-risk individuals inside trials or structured programmes. For health systems: the NHS-Galleri readout justifies neither rollout nor abandonment — it justifies the extended follow-up now underway. And for readers of the coming headlines: the claims to trust will cite mortality, not detection counts, survival percentages or stage statistics alone — this site's standing rule, applied to its biggest test case yet.