Longevity & Aging · Metabolic & Cardiometabolic
multi-cancer early detection MCED blood tests such as Galleri detects early-stage cancers with high specificity but modest sensitivity
In plain terms: Can a blood test like Galleri catch early cancer accurately?
Part of: • Cancer Screening
Specificity is genuinely high (about 99%) but sensitivity is low for early (stage I/II) disease and drops steeply by stage, so a negative result is weakly reassuring and no mortality benefit is yet shown.
📅 Last reviewed: 2026-07-14 ⓘ
Evidence ladder
How far up the ladder this claim has climbed. A high consensus on a low rung means "consistent so far," not "proven in people."
Top evidence so far: Population patterns (Observational)
How the studies fall
What the evidence shows
Specificity is genuinely high, about 99 to 99.5 percent, confirmed in PATHFINDER across 6,662 adults over 50. Sensitivity is the problem: about 51 percent overall and about 17 percent at stage I in the CCGA validation, weakest for exactly the early disease the tests are sold to find. In PATHFINDER 1.4 percent of those screened had a signal, and 62 percent of those signals were false positives.
The evidence (11)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Jamshidi A, Liu MC, Klein EA, Venn O, Hubbell E, Beausang JF, Gross S, Melton C, Fields AP, Liu Q, Zhang N, Fung ET, Kurtzman KN, Amini H, Betts C, Civello D, Freese P, Calef R, Davydov K, Fayzullina S, Hou C, Jiang R, Jung B, Tang S, Demas V, Newman J, Sakarya O, Scott E, Shenoy A, Shojaee S, Steffen KK, Nicula V, Chien TC, Bagaria S, Hunkapiller N, Desai M, Dong Z, Richards DA, Yeatman TJ, Cohn AL, Thiel DD, Berry DA, Tummala MK, McIntyre K, Sekeres MA, Bryce A, Aravanis AM, Seiden MV, Swanton C 2022 · Cancer Cell | observational | supports | moderate | CCGA methods substudy confirmed whole-genome methylation as best MCED approach at fixed 98-99% specificity, with sensitivity tracking circulating tumor fraction (low in early tumors). |
| Jeong S, Go D, Jeon Y, Kim YJ, Lee H, Kim YW, Woo HG, Park JK, Kim YT, Park YS, Lee SH, Han W, Lee HB, Kim KH, Lee S, Kim S, Song SH, Kim HP, Cha Y, Bang D, Kim TY 2026 · Exp Mol Med | observational | mixed | low | A whole-genome methylation multimodal cfDNA MCED assay tested on 1415 samples (8 cancer types) achieved 93.2% sensitivity and 95% specificity overall, with 92.3% sensitivity for stage I and 92.2% for stage II. |
| Schrag D, Beer TM, McDonnell CH, Nadauld L, Dilaveri CA, Reid R, Marinac CR, Chung KC, Lopatin M, Fung ET, Klein EA 2023 · Lancet | observational | mixed | high | PATHFINDER prospective cohort of 6,662 adults >=50 found 99.5% specificity but only 1.4% cancer-signal-detected rate and PPV of 38%, confirming high specificity yet limited overall yield. |
| Marlow LAV, Schmeising-Barnes N, Warwick J, Waller J 2023 · BMJ Open | observational | tested-null | low | Qualitative study documented substantial psychological harm and diagnostic-odyssey burden from cancer-signal-detected results, tempering claims that MCED testing is straightforwardly accurate/beneficial. |
| Nicholson BD, Oke J, Virdee PS, Harris DA, O'Doherty C, Park JE, Hamady Z, Sehgal V, Millar A, Medley L, Tonner S, Vargova M, Engonidou L, Riahi K, Luan Y, Hiom S, Kumar H, Nandani H, Kurtzman KN, Yu LM, Freestone C, Pearson S, Hobbs FR, Perera R, Middleton MR 2023 · Lancet Oncol | observational | mixed | high | SYMPLIFY (symptomatic patients): specificity 98.4% but sensitivity only 66% overall and 24% at stage I — confirming poor early-stage sensitivity even in enriched populations. |
| Bryce AH, Thiel DD, Seiden MV, Richards D, Luan Y, Coignet M, Zhang Q, Zhang N, Hubbell E, Kurtzman KN, Klein EA 2023 · JCO Precis Oncol | observational | mixed | moderate | CCGA substudy-3 symptomatic subgroup showed high specificity (~99.5%) but sensitivity rose steeply with stage, remaining low in stage I disease. |
| Khare S, Burton JC, De Asis F, Mekonnen S, Stewart-McLellan MJ, Potts D, Howard TL, Nabavizadeh N 2025 · Front Oncol | observational | supports | low | Single-institution PATHFINDER case series illustrated real-world true and false positives, reinforcing high specificity alongside meaningful false-positive workups. |
| Klein EA, Richards D, Cohn A, Tummala M, Lapham R, Cosgrove D, Chung G, Clement J, Gao J, Hunkapiller N, Jamshidi A, Kurtzman KN, Seiden MV, Swanton C, Liu MC 2021 · Ann Oncol | observational | supports | high | CCGA validation: targeted-methylation MCED test achieved 99.5% specificity but overall sensitivity only about 51%, and far lower at stage I (about 17%) — high specificity, modest/stage-dependent sensitivity. |
| Neal RD, Johnson P, Clarke CA, Hamilton SA, Zhang N, Kumar H, Swanton C, Sasieni P 2022 · Cancers | observational | supports | moderate | NHS-Galleri design paper for a 140,000-participant RCT powered on late-stage incidence reduction, underscoring that specificity is established but clinical-utility/mortality benefit remains unproven. |
| Nadauld LD, McDonnell CH, Beer TM, Liu MC, Klein EA, Hudnut A, Whittington RA, Taylor B, Oxnard GR, Lipson J, Lopatin M, Shaknovich R, Chung KC, Fung ET, Schrag D, Marinac CR 2021 · Cancers (Basel) | observational | mixed | low | PATHFINDER real-world implementation projected PPV only about 49% (many signals need extensive workup), tempering the high-specificity-solves-false-positives framing at population prevalence. |
| Held D, Bolland S, Freese R, Puskas R 2025 · Biomedicines | observational | mixed | low | A protein-based 5-cancer MCED test (not Galleri) achieved 100% sensitivity across all five cancer types (including 100% of Stage I cancers) and 97% overall specificity in 141 cancer patients vs 119 healthy controls. |
Disagree, or know a study we missed?
We grade by evidence, not opinions. The way to weigh in is to point us to a study we haven't cited (check the evidence table above first), or to flag a problem with one we have. Every submission is reviewed; if it holds up, the grade updates and shows in Science Changes Its Mind.
Opens a short form. You'll sign in with Google so submissions are tied to a real account — we don't display your identity, and we only accept a link we can verify (PubMed, DOI, ClinicalTrials.gov).
Educational only, not medical advice. Grades and scores reflect published evidence weighted by study design and quality; see the methodology.