Background Overdiagnosis is defined as the detection of a cancer that would not have caused symptoms or death if left undiscovered during the patient’s lifetime. Unlike false positives, where further testing may rule out disease, overdiagnosed cancers meet pathological diagnostic criteria and cannot be reliably distinguished from clinically relevant malignancies, often leading to unnecessary disease labelling and overtreatment. Overdiagnosis unnecessarily turns healthy individuals into patients and poses major challenges for clinical practice and for the long-term sustainability of universal health coverage. Overdiagnosis cannot be quantified at the individual-level and can only be assessed using population-level approaches.
Objectives Our objective is to provide evidence on the global magnitude of cancer overdiagnosis across cancer sites and settings, discuss the methodological approaches used to estimate overdiagnosis at the population-level, and explore its implications at the clinical  and health systems levels.
Methods We utilize high-quality population and surveillance data to infer the presence of overdiagnosis through its specific epidemiological signatures. These features include a fast increase in cancer incidence (affecting predominantly low-risk tumours) alongside stable or mildly declining mortality, large geographical heterogeneity, and a shift in age-specific incidence curves toward younger populations. We use the examples of thyroid and prostate cancer. To quantify overdiagnosis of thyroid cancer, we employ an ad hoc developed method based on the multistage model of carcinogenicity. Overdiagnosis is estimated by comparing observed incidence with projected historical age-specific incidence rates from the pre-imaging era; excess cases are considered overdiagnosed.
Results The examples of thyroid and prostate cancer demonstrate the massive scale of this issue. Thyroid cancer, relatively rare until the 1970s, was estimated to have been overdiagnosed in approximately 1.7 million individuals across 63 countries only during 2013-2017. Thyroid cancer overdiagnosis disproportionately affects young women, ranking as one of the top three cancers in 100 countries among females aged 15–39 in 2022. In China, it is already the most frequently diagnosed cancer in women of all ages. Similarly, prostate cancer incidence has inflated significantly in many European countries due to opportunistic PSA testing. We show that overdiagnosis is higher, and benefits smaller, at the population-level compared with randomized controlled trials.
Conclusions The magnitude of cancer overdiagnosis is large and increasing over time. Cancer overdiagnosis is driven by the rapid development and widespread use of increasingly sensitive diagnostic technologies (e.g. ultrasound and CT scans) capable of detecting small, slow-growing, or indolent tumours, and is further fuelled by opportunistic and non-indicated diagnostic scrutiny and screening in asymptomatic individuals, reinforced by technology hype, commercial incentives, and a prevailing culture that equates more testing with better care.