Cancer Screening: Weighing Earlier Detection Against Overdiagnosis

The best cancer screening test is not the one that finds the most tumors. It is the one that prevents the most deaths while treating the fewest people who were never going to be harmed. Those two goals pull against each other, and the tension has a name: overdiagnosis. Overdiagnosis is not a false alarm or a lab mistake. The tumor is real and meets every pathological criterion for cancer. It simply would have sat still for the rest of your life, never causing a symptom, if screening had not gone looking for it.

Key points#

• Screening detects cancer earlier, but it also detects real tumors that would never have caused illness or death; that second harm is overdiagnosis, not a diagnostic error.
• Survival figures can improve even when no lives are saved, because lead-time and length bias distort them, so only lower mortality in a randomized trial proves a program works.
• The body carries a hidden store of slow or silent tumors whose size varies widely by organ, which is why breast, thyroid, prostate, and lung screening each strike a different bargain.
• Focusing screening on people at genuinely high risk, as low-dose CT does for heavy smokers, improves the ratio of lives saved to harmless tumors found.

Why "survival went up" can be a mirage#

Screening is most often marketed with survival rates, and the National Cancer Institute is unusually blunt about why the number can deceive. Two biases push survival upward without anyone actually living longer.

The first is lead-time bias. Screening moves the date of diagnosis earlier without moving the date of death. Picture two people with the same slow tumor and the same fixed lifespan. One is diagnosed at 60 by a screening test; the other notices a symptom and is diagnosed at 67. If both die at 70, the screened person appears to survive ten years against three, yet not a single day was added. All that changed was how long each carried the label of patient.

The second is length bias. Screening tends to catch tumors that grow slowly, because a lesion that lingers in a silent phase for years offers a wide window to be caught, while one that turns lethal in months is often diagnosed by symptoms between screening rounds. Screen-detected cancers are therefore a favorably selected group before a single treatment begins. Overdiagnosis is the far end of that same spectrum: the tumor grows so slowly, or not at all, that you die of something else first.

Because all of these effects inflate survival regardless of true benefit, the only proof of a screening program's value you can trust is a drop in cancer-specific and overall mortality inside a randomized trial. When survival climbs but mortality stays flat, that pattern is the fingerprint of overdiagnosis, not of lives saved.

The silent reservoir#

Overdiagnosis is possible only because most of us carry tumors we will never know about. A peer-reviewed review indexed in PubMed Central describes a large hidden reservoir of lesions that satisfy the pathologist's definition of cancer yet behave nothing like the aggressive disease the word calls to mind. Autopsy studies of men who died of unrelated causes found prostate cancer in roughly a third to a half of older decedents: present, genuine, and completely silent. Comparable stores of dormant disease exist in the thyroid and, to a smaller degree, the breast and lung.

Screening does not create these lesions. It reaches into the reservoir and pulls some of them into the clinic, where each one becomes a diagnosis, a decision, and often a course of treatment. The size of the reservoir differs sharply from one organ to the next, which is exactly why screening cannot be judged as a single practice. It is better understood as four separate bargains.

Four cancers, four different bargains#

Thyroid: overdiagnosis you can see from space#

Thyroid cancer offers the clearest natural experiment. As described in a New England Journal of Medicine perspective, South Korea watched thyroid-cancer diagnoses climb to roughly fifteen times their 1993 level after ultrasound screening spread through routine check-ups, while deaths from thyroid cancer barely moved. An incidence curve that soars while the mortality curve stays flat is overdiagnosis made visible at the scale of a whole country. The authors attribute the surge to detection, not to any genuine rise in dangerous disease.

Prostate: the widest range of harm#

Screening with the PSA blood test carries some of the highest overdiagnosis estimates, ranging across studies from roughly 20 percent to 50 percent or more of screen-detected cancers. A large share of the men treated would never have been troubled by their disease, and treatment is far from harmless: incontinence and erectile dysfunction are common. That arithmetic is why the U.S. Preventive Services Task Force shifted in 2018 to a shared-decision approach for men aged 55 to 69, urging that the choice follow a real conversation about benefits and harms rather than a reflex blood draw, and recommending against routine PSA screening for men 70 and older.

Breast: real benefit, modest in absolute terms#

For breast cancer, the National Cancer Institute puts overdiagnosis at roughly 19 percent of screen-detected cases, and the review discussed here places some mammography estimates near a quarter of detected cancers. The benefit is genuine but small when counted in absolute numbers: if you are a woman in your fifties, more than a thousand women like you must be screened to prevent one breast-cancer death, and that single prevented death sits alongside many false positives, biopsies, and treatments of lesions that would never have progressed.

Lung: what happens when you screen the right people#

Lung cancer shows how the same tension resolves differently once baseline risk is high. Low-dose CT in heavy smokers cut lung-cancer mortality by about 20 percent in the pivotal trial, and in 2021 the Task Force widened eligibility to adults aged 50 to 80 with a 20 pack-year smoking history. Overdiagnosis still occurs here, estimated in the low tens of percent, but concentrating the test on people at genuinely elevated risk raises the ratio of lives saved to indolent tumors found. Targeting, more than the technology itself, is what makes the bargain defensible.

Reading a screening decision honestly#

A sound appraisal never asks simply whether a test can find cancer earlier, because almost all of them can. It asks whether earlier detection turns into fewer deaths, on what absolute rather than relative scale, and how many people pay for that benefit by being treated for a harmless lesion. Cervical and colorectal screening clear this bar with clear mortality reductions and the added advantage of removing precursor lesions before they can turn malignant. Others sit nearer the margin, where two well-informed people weighing the same evidence will reasonably choose differently, and you are entitled to be either of them. The honest version of the conversation puts absolute numbers on both benefit and harm, names overdiagnosis out loud, and treats the choice as one to make with you rather than one to impose on a population.