An early-detection blood test can make cancer statistics look better without helping a single patient live longer. That is not a cynical claim — it is arithmetic, and it has a name: lead-time bias. Two related effects, length bias and overdiagnosis, complete the picture.

Why this matters for anyone considering an EDIM test. The EDIM Apo10/TKTL1 test is marketed on the intuition that finding cancer sooner must be better. That intuition is often correct, but not automatically, and the history of cancer screening contains several programmes that detected far more cancer while changing mortality very little [2].

What this article does. It explains the three biases in plain terms, shows what randomised trials found when they tested the intuition directly, and applies the same reasoning to EDIM and other blood-based cancer tests offered in Bangkok and elsewhere.

What it does not do. It does not argue against cancer screening. Several screening programmes reduce cancer mortality in defined populations, and this article is not a reason to skip a mammogram, a colonoscopy, or a low-dose CT for which you are eligible.

Direct answer

Lead-time bias means a test can add time to the period you know you have cancer without adding time to your life. Overdiagnosis means a test can find cancers that would never have surfaced or caused symptoms. Both effects inflate survival statistics, so survival rates alone cannot prove that an early-detection test — including EDIM — is beneficial. Only mortality endpoints from controlled studies can.

Spectrum of tumour biology relevant to overdiagnosis in EDIM and early cancer detection — dormant to aggressive cell clusters

What lead-time bias actually is

Lead-time bias is the artificial gain in measured survival that occurs simply because a diagnosis was made earlier in the same disease course. Imagine a cancer that becomes symptomatic at year 8 and causes death at year 10. Diagnosed on symptoms, the patient records two years of survival. Now imagine a blood test detects the same cancer at year 5, and the disease follows the identical course to death at year 10. The recorded survival is now five years — a 150% improvement in the statistic, with the date of death unchanged.

This is not a subtle statistical artefact. It is the default expectation whenever a test moves the diagnosis date earlier, and it operates on every early-detection technology equally: imaging, tumour markers, ctDNA, and monocyte-based tests such as EDIM. Any survival figure quoted for an early-detection test that was not derived from a comparison against an unscreened group carries an unknown amount of lead time inside it [1].

The practical consequence is that "patients detected early lived longer" is not evidence. It is the expected result even when the test does nothing. To separate real benefit from lead time, researchers need to compare death rates between a group offered the test and a group not offered it, over the same calendar period — which is exactly what randomised screening trials do.

Length bias: why screening preferentially finds slow cancers

Length bias is the tendency of any periodic test to over-sample slow-growing disease. A cancer that takes eight years to become symptomatic spends eight years in a detectable pre-symptomatic window. A cancer that becomes symptomatic within six months spends six months there. If a test is applied once a year, it has many more opportunities to catch the slow one.

The screened population therefore ends up enriched with indolent tumours, which have a better natural prognosis regardless of treatment. Fast, aggressive cancers tend to appear between tests as symptomatic "interval cancers" and are counted in the non-screened comparison. The result is a screen-detected group that looks healthier for reasons that have nothing to do with the intervention.

Length bias and lead-time bias compound each other. Together they can produce impressive-looking case series for a test with no effect on mortality at all, which is why single-arm cohorts of "patients whose cancer was found by test X" cannot establish clinical utility [13].

Overdiagnosis: finding cancer that would never have caused harm

Overdiagnosis is the detection of a histologically real cancer that would never have produced symptoms or caused death within the person's lifetime. It is the extreme end of length bias: the tumour is so slow, so dormant, or the person's competing risks so significant, that the cancer never becomes clinically relevant [2].

Overdiagnosis is not a misdiagnosis. The pathology is correct. The harm is that a person who would have lived a normal life without ever knowing about the tumour instead becomes a cancer patient — with surgery, radiation, systemic therapy, insurance consequences, and lasting anxiety, none of which could have benefited them because there was no harm to prevent.

The clearest documented example is thyroid cancer in South Korea. After ultrasound screening became widely available, thyroid cancer incidence rose roughly fifteen-fold over two decades while mortality from thyroid cancer stayed essentially flat — the signature pattern of overdiagnosis rather than a genuine epidemic [5].

Breast cancer provides a quantified example. Analysis of three decades of United States screening mammography estimated that a substantial proportion of detected early-stage breast cancers represented overdiagnosis, because the expected reciprocal fall in late-stage disease was much smaller than the rise in early-stage detection [3]. A later analysis of tumour size distribution reached a similar conclusion and attributed much of the improvement in breast-cancer mortality to better treatment rather than to earlier detection [4]. An independent United Kingdom review concluded that screening does reduce breast-cancer mortality but at the cost of a real and quantifiable amount of overdiagnosis, and framed the decision as a genuine trade-off rather than a pure benefit [11].

Evidence checkpoint

An early-detection test that raises cancer incidence without lowering late-stage incidence and cancer mortality is most likely finding disease that did not need finding. Rising incidence with flat mortality is the fingerprint of overdiagnosis, not of success.

Why five-year survival is the wrong yardstick

Five-year survival measures how long people live after diagnosis, so it moves whenever the diagnosis date moves — even if nobody lives longer. A landmark analysis compared changes in five-year survival with changes in mortality across many cancer types and found the two were only weakly related; improvements in survival were substantially driven by increased diagnosis of early-stage and indolent disease rather than by fewer deaths [1].

Mortality is the honest endpoint because it is anchored to the calendar, not to the diagnosis date. Deaths per 100,000 people per year in a population offered a test, compared with a population not offered it, cannot be inflated by lead time or by overdiagnosis. This is why guideline bodies weigh randomised mortality data far more heavily than survival statistics from case series.

Two secondary endpoints also help. A genuine early-detection benefit should reduce the incidence of late-stage disease — if a test finds cancers earlier and this matters, advanced presentations should fall. And it should reduce cancer-specific mortality without a compensating rise in deaths from investigation or treatment. Neither pattern can be demonstrated by counting how many cancers a test detected.

What the randomised trials actually showed

When the intuition that earlier is better was tested directly, the results varied by cancer, by test, and by population. That variation is the whole point: benefit is specific, not general.

The neuroblastoma programmes deserve particular attention in any discussion of a novel blood test. They were biologically plausible, technically successful at what they set out to do, and ultimately abandoned because the outcome that mattered did not move. Plausibility and detection are necessary but not sufficient.

Clinical decision pathway after an abnormal EDIM or blood-based cancer detection result in Bangkok

How this applies to EDIM and other blood-based tests

Every published EDIM study to date reports detection or discrimination, not mortality. The core papers describe Apo10 and TKTL1 measured inside activated monocytes and compare scores between people with known cancer and control groups [16][17][18]. These are early-phase biomarker studies. They can support a hypothesis; they cannot answer whether testing changes outcomes, because they were not designed to.

Three specific gaps follow directly from the biases described above:

More extensively studied blood tests illustrate the same constraint. The DETECT-A study prospectively applied a multi-analyte blood test with PET-CT follow-up in women without known cancer and demonstrated that such a pathway is feasible, while explicitly not establishing a mortality benefit [19]. A large methylation-based multi-cancer test reported high specificity and tissue-of-origin prediction in validation cohorts, again as a step towards — not a substitute for — outcome evidence [20]. If tests with those resources behind them are still awaiting mortality data, a smaller investigational assay is certainly not further along.

What evidence would settle the question

Biomarker evaluation follows a recognised sequence, and EDIM currently sits near its beginning. The standard framework describes phases running from preclinical exploratory work, through clinical assay development and retrospective longitudinal studies, to prospective screening studies and finally cancer-control studies that measure the effect of screening on the population's disease burden [13].

Concretely, confidence in EDIM would increase with:

None of these are unreasonable demands. They are the same evidence ladder that lung-cancer CT screening climbed before it entered guidelines, and the same one that stopped infant neuroblastoma screening when the top rung was reached and the answer was negative.

Questions worth asking before paying for an early-detection test

A short list of questions separates a well-explained investigational test from an over-sold one. A clinic offering the test in good faith should be able to answer all of them without discomfort.

Structured decision support helps here. A Cochrane systematic review of decision aids (doi:10.1002/14651858.CD001431.pub6) found that they improve people's knowledge of their options and produce choices that better match their own values in screening and treatment decisions — which is exactly the situation created by an investigational cancer test with uncertain benefit.

Frequently asked questions

Does lead-time bias mean early detection never helps?

No. It means earlier detection helps only when there is an effective intervention that works better at the earlier stage, and when the disease found is one that would have caused harm. Low-dose CT in high-risk smokers is an example where randomised trials showed a real mortality reduction.

If my EDIM score is normal, can I skip my mammogram or colonoscopy?

No. A normal EDIM result does not exclude cancer and has not been shown to substitute for any organ-specific screening programme. Continue whatever guideline-recommended screening applies to your age and risk profile.

Why do clinics quote survival rates instead of mortality rates?

Survival rates are easier to obtain — they only require following people after diagnosis — and they respond favourably to earlier detection whether or not the test helps. Mortality rates require a comparison population and long follow-up. The easier number is also the more flattering one.

Is overdiagnosis actually common, or a theoretical worry?

It is documented and quantified in several cancers, most dramatically in thyroid cancer following widespread ultrasound use, and in a measurable fraction of screen-detected breast cancers. Its magnitude depends heavily on the cancer type, the test, and the population.

Does this mean the EDIM test is useless?

It means the test is investigational and its clinical benefit is unproven, which is a different statement. The underlying biology is a legitimate research question. What is not supportable is presenting an EDIM score as a cancer diagnosis, as a screening replacement, or as evidence that a patient's life has been extended.

Limitations and medical disclaimer

This article summarises well-established methodological concepts and cites primary studies, but it is not a systematic review of cancer screening, and the magnitude of overdiagnosis in any particular programme remains actively debated in the literature [11][12]. Estimates depend on modelling assumptions, follow-up duration, and the comparison population chosen.

EDIM is an investigational adjunct test. Published studies are small, largely case-control in design, and do not demonstrate that EDIM-guided testing improves survival or reduces cancer mortality. VELAR Center does not present EDIM as a stand-alone cancer diagnosis or as a substitute for guideline-recommended screening.

This article is educational and does not constitute medical advice, diagnosis, or a treatment recommendation. Decisions about cancer screening, further imaging, or biopsy should be made with a qualified physician who knows your history and your risk profile.

Related reading

For the underlying assay and its evidence base, see our reviews of EDIM evidence, accuracy and limitations and EDIM compared with liquid biopsy and standard screening. For the biology behind the markers, see Apo10 and TKTL1 and the Warburg effect. For what to do with a result, see understanding EDIM test results.

References

  1. Welch HG, Schwartz LM, Woloshin S. Are increasing 5-year survival rates evidence of success against cancer? JAMA. 2000;283(22):2975–2978. doi:10.1001/jama.283.22.2975 ↩
  2. Welch HG, Black WC. Overdiagnosis in cancer. Journal of the National Cancer Institute. 2010;102(9):605–613. doi:10.1093/jnci/djq099 ↩
  3. Bleyer A, Welch HG. Effect of three decades of screening mammography on breast-cancer incidence. New England Journal of Medicine. 2012;367(21):1998–2005. doi:10.1056/NEJMoa1206809 ↩
  4. Welch HG, Prorok PC, O'Malley AJ, Kramer BS. Breast-cancer tumor size, overdiagnosis, and mammography screening effectiveness. New England Journal of Medicine. 2016;375(15):1438–1447. doi:10.1056/NEJMoa1600249 ↩
  5. Ahn HS, Kim HJ, Welch HG. Korea's thyroid-cancer "epidemic" — screening and overdiagnosis. New England Journal of Medicine. 2014;371(19):1765–1767. doi:10.1056/NEJMp1409841 ↩
  6. Schilling FH, Spix C, Berthold F, et al. Neuroblastoma screening at one year of age. New England Journal of Medicine. 2002;346(14):1047–1053. doi:10.1056/NEJMoa012277 ↩
  7. Woods WG, Gao RN, Shuster JJ, et al. Screening of infants and mortality due to neuroblastoma. New England Journal of Medicine. 2002;346(14):1041–1046. doi:10.1056/NEJMoa012387 ↩
  8. National Lung Screening Trial Research Team. Reduced lung-cancer mortality with low-dose computed tomographic screening. New England Journal of Medicine. 2011;365(5):395–409. doi:10.1056/NEJMoa1102873 ↩
  9. de Koning HJ, van der Aalst CM, de Jong PA, et al. Reduced lung-cancer mortality with volume CT screening in a randomized trial. New England Journal of Medicine. 2020;382(6):503–513. doi:10.1056/NEJMoa1911793 ↩
  10. Andriole GL, Crawford ED, Grubb RL, et al. Mortality results from a randomized prostate-cancer screening trial. New England Journal of Medicine. 2009;360(13):1310–1319. doi:10.1056/NEJMoa0810696 ↩
  11. Independent UK Panel on Breast Cancer Screening. The benefits and harms of breast cancer screening: an independent review. The Lancet. 2012;380(9855):1778–1786. doi:10.1016/S0140-6736(12)61611-0 ↩
  12. Esserman LJ, Thompson IM, Reid B. Overdiagnosis and overtreatment in cancer: an opportunity for improvement. JAMA. 2013;310(8):797–798. doi:10.1001/jama.2013.108415 ↩
  13. Pepe MS, Etzioni R, Feng Z, et al. Phases of biomarker development for early detection of cancer. Journal of the National Cancer Institute. 2001;93(14):1054–1061. doi:10.1093/jnci/93.14.1054 ↩
  14. Bossuyt PM, Reitsma JB, Bruns DE, et al. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ. 2015;351:h5527. doi:10.1136/bmj.h5527 ↩
  15. Whiting PF, Rutjes AWS, Westwood ME, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Annals of Internal Medicine. 2011;155(8):529–536. doi:10.7326/0003-4819-155-8-201110180-00009 ↩
  16. Coy JF. EDIM-TKTL1/Apo10 blood test: an innate immune system based liquid biopsy for the early detection, characterization and targeted treatment of cancer. International Journal of Molecular Sciences. 2017;18(4):878. doi:10.3390/ijms18040878 ↩
  17. Saman S, Stagno MJ, Warmann SW, et al. Biomarkers Apo10 and TKTL1: epitope-detection in monocytes (EDIM) as a new diagnostic approach for cholangiocellular, pancreatic and colorectal carcinoma. Cancer Biomarkers. 2020;27(1):129–137. doi:10.3233/CBM-190414 ↩
  18. Stagno MJ, Schmidt A, Bochem J, et al. Epitope detection in monocytes (EDIM) for liquid biopsy including identification of GD2 in childhood neuroblastoma — a pilot study. British Journal of Cancer. 2022;127(7):1324–1331. doi:10.1038/s41416-022-01855-x ↩
  19. Lennon AM, Buchanan AH, Kinde I, et al. Feasibility of blood testing combined with PET-CT to screen for cancer and guide intervention. Science. 2020;369(6499):eabb9601. doi:10.1126/science.abb9601 ↩
  20. Klein EA, Richards D, Cohn A, et al. Clinical validation of a targeted methylation-based multi-cancer early detection test using an independent validation set. Annals of Oncology. 2021;32(9):1167–1177. doi:10.1016/j.annonc.2021.05.806 ↩