Found Late: Stage at Diagnosis, Emergency Diagnosis and Why Timing Matters

How far a cancer has grown by the time it is found is one of the strongest predictors of what happens next. In the United States, half of pancreatic and lung cancers are diagnosed after they have already spread to distant organs^[1], and across six high-income countries between a quarter and two-fifths of eight common cancers were diagnosed only after an emergency hospital admission^[2].

Behind each of those figures is a person whose illness was found later than anyone would have wanted: often someone older, someone whose symptoms were vague or easily mistaken for something common, someone living in a poorer area, or someone in a country where diagnostic services are thin. A late diagnosis is rarely a single person's delay. It usually reflects how hard a cancer is to detect, whether there is a screening test for it, and how quickly a health system can turn a worrying symptom into a diagnosis.

Five-year survival can fall from over 90% to under 20% depending on stage

The US Surveillance, Epidemiology, and End Results (SEER) programme publishes five-year relative survival by stage for people diagnosed in 2016–2022. For bowel (colorectal) cancer, five-year relative survival was 91.3% when the cancer was localised (confined to where it started), 75.2% when it had reached nearby lymph nodes, and 16.9% when it had spread to distant organs^[4]. For lung cancer the figures were 65.5%, 38.2% and 10.5%^[5]. For women with breast cancer, localised disease had five-year relative survival of 100.0%, meaning no measurable excess deaths compared with women of the same age in the general population, against 33.8% once it had spread to distant organs^[6]. Even for pancreatic cancer, where survival is lowest of the four cancers shown here, it was 43.6% when localised and 3.4% when distant^[1].

England's registry shows the same pattern. For men diagnosed with lung cancer in 2018–2022, age-standardised five-year net survival was 59.6% at stage 1 and 4.2% at stage 4^[7]. Across 21 cancer types, age-standardised one-year net survival in England (people diagnosed 2016–2020, followed up to 2021) was highest when the cancer was diagnosed at stage 1 and fell with each later stage^[3].

These comparisons need one caution. Stage is not the only thing that differs between people diagnosed early and late: the two groups can differ in age and general health, and cancers found early include some that grow slowly. Screening can also find cancers that would never have caused harm (overdiagnosis), and finding a cancer earlier lengthens measured survival even if death is not delayed (lead-time bias). England's figure for prostate cancer shows why this matters: five-year net survival was 102.3% at stage 1, against 54.8% at stage 4^[7]. Net survival above 100% means people in that group survived better than the general population of the same age, which is usually read as telling us something about who tends to be diagnosed at stage 1, not as a sign that a diagnosis improves life expectancy. The gap between stages is still real and large; it is simply not all caused by timing.

Half of US lung and pancreatic cancers are found after they have spread

The share of cancers found late differs hugely by cancer type. In SEER's data for 2016–2022, 51% of lung cancers and 51% of pancreatic cancers were diagnosed at a distant stage, compared with 23% of bowel cancers and 6% of breast cancers in women (shares of all recorded cases, including the 2–6% whose stage was unknown)^[5][1][4][6].

England's 2023 registrations, the latest published, show a similar spread. Among staged cancers, 91% of testicular cancers, 91% of melanomas in women and 87% in men, and 85% of breast cancers in women were diagnosed at stage 1 or 2^[8]. There were also gaps between the sexes for the same cancer: 51% of laryngeal cancers in men were diagnosed early against 40% in women, and 77% of bladder cancers in men against 68% in women; for breast cancer the gap ran the other way, 85% in women against 75% in men^[8].

Why some cancers are so often found late is partly biology and anatomy. England's 2026 National Cancer Plan technical annex notes that tumours in deep locations such as the ovaries or pancreas are difficult to examine without advanced imaging or invasive procedures, that they typically cause few or non-specific symptoms early on, and that there is no effective population screening test for them, unlike breast, bowel or cervical cancer^[3]. How symptoms turn into a diagnosis, and where the delays build up, is covered in our article on delays from first symptom to diagnosis.

Up to two in five of eight common cancers are diagnosed after an emergency admission, and outcomes are worse

England pioneered "Routes to Diagnosis", linking registry, hospital and screening records to place every cancer diagnosed in 2006–2008 into one of eight routes, from screening and urgent GP referral to emergency presentation; people diagnosed through emergency routes had substantially lower one-year relative survival^[9]. The series has since been extended to later years^[10].

The International Cancer Benchmarking Partnership (ICBP) applied a common definition, a cancer diagnosed within 30 days of an emergency hospital admission, to 857,068 people with eight cancers (oesophagus, stomach, colon, rectum, liver, pancreas, lung and ovary) in 14 jurisdictions across Australia, Canada, Denmark, New Zealand, Norway and the UK, diagnosed 2012–2017^[2]. Taking the eight cancers together, the share diagnosed as an emergency ranged from 24.0% in the lowest jurisdiction to 42.5% in the highest. Pancreatic cancer had the highest share on average, 46.1%, ranging from 34.1% to 60.4% by jurisdiction; rectal cancer had the lowest, 12.1% on average, ranging from 9.1% to 19.8%^[2].

The same study found that emergency diagnosis was most common among the oldest people: the share was highest in those aged 85 or over in 110 of 112 jurisdiction-and-cancer combinations, and highest for cancers that had spread to distant sites in 98 of 101 combinations with stage data. People diagnosed as an emergency had substantially higher odds of dying within 12 months than people diagnosed by other routes, with an odds ratio above 1.9 in every one of the 112 combinations. At the level of whole jurisdictions, a 10% increase in the percentage of emergency diagnoses was associated with a decrease in one-year net survival of between 2.5% and 7.0%, for colon, stomach, lung, liver, pancreatic and ovarian cancer^[2]. The authors concluded that emergency presentation probably contributes to international differences in survival, a theme our article on how survival varies between countries explores further.

In England, emergency presentation is still the route to diagnosis for a fifth of cancers, and it is more common for people living in the most deprived areas^[3]. How people reach an emergency diagnosis has also changed. Among 554,621 people diagnosed as an emergency in England in 2006–2015, 62% arrived through accident and emergency and 24% through an emergency referral from their GP. Over that decade, the GP-referred share fell from 31% to 17% while the A&E share rose from 55% to 68%. The authors linked the fall in emergency diagnoses over the period to the steadily growing use of urgent two-week-wait GP referrals and fewer emergency GP referrals^[10].

England's early-diagnosis rate was flat for a decade, and the deprivation gap is only starting to close

England measures early diagnosis as the share of staged cancers diagnosed at stage 1 or 2. According to the technical annex to the 2026 National Cancer Plan, that share was flat for nearly a decade, from 2013 until the pandemic in 2020, at around the mid-50% range, with a persistent gap of around 8 percentage points between the most affluent and the poorest areas^[3]. It fell sharply in 2020, when services were disrupted, then recovered to around 60% by mid-2025. In the 12 months to September 2025 the early-diagnosis rate was 3.5 percentage points above its pre-pandemic level, which the annex equates to around 10,000 more people being diagnosed at stage 1 or 2^[3]. These recent figures come from rapid cancer registration data, produced ahead of the final annual registrations and so liable to differ from them; the annex plots both series side by side, and the latest final registrations cover 2023.

The annex attributes much of the recent progress in the poorest areas to the rollout of lung cancer screening. In the year before the pandemic (March 2019 to February 2020), 51.6% of staged cancers in the most deprived fifth of areas were diagnosed early, against 59.8% in the least deprived, a gap of 8.2 percentage points. By October 2024 to September 2025 the figures were 56.0% and 62.2%, a gap of 6.2 points^[3]. The gap is narrower but has not gone. The wider pattern of cancer outcomes by income and area is the subject of our article on deprivation and cancer.

In sub-Saharan Africa, most breast cancers are still found at stage III or IV

In many low- and middle-income countries, population-based staging data barely exist, so the evidence comes mostly from hospital studies pooled in systematic reviews. Breast cancer has the most data. A 2016 meta-analysis of 83 studies covering 26,788 women in 17 sub-Saharan African countries, published before 2014, found a median of 74.7% diagnosed at a late stage (stage III or IV), with a range of 30.3% to 100% between studies. The late-stage share among Black African women had fallen over time (by 10.5 percentage points for studies from 2000 or later compared with 1980 or earlier) but around 2010 was still higher than it had been for Black and white women in the US 40 years earlier^[11].

An updated review published in 2026, pooling 125 studies published between January 2014 and December 2025 and covering 42,689 people with staged breast cancer, found that 63.1% were diagnosed at a late stage (stage III or IV), ranging from 55.2% in Southern Africa to 74.5% in Middle Africa^[12]. Its authors caution that differences in study populations and reporting mean this should not be read as proof that stage at diagnosis has improved. Population-based registry data point the same way: in a random sample of breast cancers diagnosed in 2008–2015 from 14 registries in 12 countries, 64.9% of those with known stage were diagnosed late and 18.4% had already spread to distant organs. Three-year relative survival was 78% for early-stage disease against 40.3% for stage III or IV^[13].

The WHO Global Breast Cancer Initiative, launched in 2021, aims to cut breast cancer mortality by 2.5% a year through early detection, timely diagnosis and comprehensive treatment^[14]. Its first benchmark, set out in the initiative's key performance indicators, is that at least 60% of breast cancers are diagnosed at stage I or II^[15]. In a hospital-based cohort of women recruited in Nigeria, Uganda, Zambia and Namibia in 2014–2017, achievement of that benchmark and of timely diagnosis ranged between 22% and 49% across population groups, and the group that came closest also had the highest survival^[15]. The wider gap in cancer outcomes between richer and poorer countries is covered in our article on the global cancer gap.

What the evidence asks of us

The registry evidence points in a consistent direction. For services, commissioners, funders and policymakers, it suggests:

• Measure emergency diagnosis everywhere, and act on it. The ICBP authors call monitoring emergency presentations, and identifying and acting on the behavioural and health-care factors behind them, a global priority for cancer control^[2].
• Report stage completely and by group. England publishes stage only for cancers with a staging system and enough complete records, and most brain tumours have no staging system at all^[8]. Breaking stage and route to diagnosis down by age, sex, deprivation and ethnicity shows where the gaps sit.
• Target the people most often diagnosed late. Older people and people in the most deprived areas are consistently over-represented among emergency diagnoses^[2][3]. England's experience with lung screening suggests well-targeted programmes can narrow a deprivation gap that had barely moved for years^[3].
• Build fast routes for vague symptoms and hard-to-detect cancers. For pancreatic, ovarian and oesophageal cancers, with no screening test and non-specific early symptoms^[3], diagnostic capacity and research into earlier detection matter most.
• Fund diagnosis and treatment together in lower-income settings. The WHO Global Breast Cancer Initiative pairs early detection with timely diagnosis and complete treatment^[14]; a hospital-based cohort study measuring its benchmarks in four sub-Saharan African countries found all three still well short^[15].

Earlier diagnosis is not the whole story; treatment access shapes survival too. But a cancer found before it has spread gives treatment the best chance, and whether that happens should not depend on someone's age, postcode or country. More evidence on this and related topics is collected in our cancer research library.

Language: this piece says "people with cancer" and "people living with and after cancer", and describes the illness and its treatment plainly, without combat metaphors. Clinical terms such as "stage", "incidence" and "net survival" are used in the sense the cited source defines them.

Why we're publishing this

People affected by cancer, and the charities, support groups and hospices that stand with them, often hear that early diagnosis matters without seeing the numbers behind it. Those numbers make a strong case: for better diagnostic services, for fair access to screening, and for attention to the groups most often diagnosed late. Every figure here links to its source, so organisations can use it in their own advocacy. A late diagnosis is a gap in the system, not a failing of the person.

How to read this data

Stage systems differ. SEER uses summary stage (localised, regional, distant); England uses numbered stages 1 to 4 based on TNM. The two are not interchangeable, so this article does not set US and English percentages side by side as a comparison. The denominators differ too. England's percentages are of staged cancers only: cancers that cannot be staged or have no recorded stage are left out. SEER's percentages are of all recorded cases, with unknown stage shown as its own category (2% of female breast, 4% of lung, 5% of pancreatic and 6% of colorectal cases in 2016–2022). The sub-Saharan African figures are shares of people with a known stage.

Survival measures differ. SEER figures are five-year relative survival; England's are age-standardised net survival. Both compare people with cancer with the general population, so values can exceed 100%. Figures describe people diagnosed years ago (2016–2022 for SEER, 2018–2022 for England) and do not predict any individual's outcome. Differences by stage partly reflect who is diagnosed at each stage, overdiagnosis and lead-time bias, not only timing.

Emergency diagnosis here means a cancer diagnosed within 30 days of an emergency admission (ICBP) or via England's Routes to Diagnosis emergency route. The ICBP study covers six high-income countries and eight cancers; its jurisdiction-level link between emergency diagnosis and survival is an association, not proof of cause.

Where evidence is thin. Most population-based stage data come from high-income countries. Evidence from sub-Saharan Africa comes largely from hospital studies, which may not represent everyone diagnosed, and mostly concerns breast cancer. England's most recent early-diagnosis figures (to 2025) come from rapid registrations, produced ahead of the final series and liable to differ from it; the latest full annual registrations cover 2023.

Use this data

Free to cite with attribution to Better Health For Everyone. Charities, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Important: This article summarises published research for general information. It is not medical advice and cannot be used to diagnose, treat or predict anything about an individual. If you have concerns about your own or someone else's health, please speak to a qualified health professional.

Sources

1. National Cancer Institute. Cancer Stat Facts: Pancreatic Cancer (SEER 21, 2016–2022). Surveillance, Epidemiology, and End Results Program, accessed 2026. seer.cancer.gov/statfacts/html/pancreas.html
2. McPhail S, Swann R, Johnson SA, Barclay ME, et al. Risk factors and prognostic implications of diagnosis of cancer within 30 days after an emergency hospital admission (emergency presentation): an International Cancer Benchmarking Partnership (ICBP) population-based study. Lancet Oncology 23(5):587–600, 2022. europepmc.org/article/MED/35397210
3. Department of Health and Social Care. Technical annex to the 2026 National Cancer Plan. GOV.UK, February 2026. assets.publishing.service.gov.uk/…/national-cancer-plan-technical-annex.pdf
4. National Cancer Institute. Cancer Stat Facts: Colorectal Cancer (SEER 21, 2016–2022). Surveillance, Epidemiology, and End Results Program, accessed 2026. seer.cancer.gov/statfacts/html/colorect.html
5. National Cancer Institute. Cancer Stat Facts: Lung and Bronchus Cancer (SEER 21, 2016–2022). Surveillance, Epidemiology, and End Results Program, accessed 2026. seer.cancer.gov/statfacts/html/lungb.html
6. National Cancer Institute. Cancer Stat Facts: Female Breast Cancer (SEER 21, 2016–2022). Surveillance, Epidemiology, and End Results Program, accessed 2026. seer.cancer.gov/statfacts/html/breast.html
7. NHS England. Cancer Survival in England, cancers diagnosed 2018 to 2022, followed up to 2023: Cancer survival by stage (adults). NHS England Digital, April 2026. digital.nhs.uk/…/cancer-survival-by-stage
8. NHS England. Cancer Registration Statistics, England, 2023: Cancer incidence by stage. NHS England Digital, October 2025. digital.nhs.uk/…/england-2023/cancer-incidence-by-stage
9. Elliss-Brookes L, McPhail S, Ives A, Greenslade M, Shelton J, Hiom S, Richards M. Routes to diagnosis for cancer: determining the patient journey using multiple routine data sets. British Journal of Cancer, 2012. europepmc.org/article/MED/22996611
10. Herbert A, Abel GA, Winters S, McPhail S, Elliss-Brookes L, Lyratzopoulos G. Cancer diagnoses after emergency GP referral or A&E attendance in England: determinants and time trends in Routes to Diagnosis data, 2006–2015. British Journal of General Practice, 2019. europepmc.org/article/MED/31455644
11. Jedy-Agba E, McCormack V, Adebamowo C, dos-Santos-Silva I. Stage at diagnosis of breast cancer in sub-Saharan Africa: a systematic review and meta-analysis. Lancet Global Health, 2016. europepmc.org/article/MED/27855871
12. Fall AD, Veses V, Sheth CC. Stage at diagnosis and molecular subtype distribution of breast cancer in sub-Saharan Africa: a systematic review. Cancer Reports, 2026. europepmc.org/article/MED/42266142
13. Joko-Fru WY, Miranda-Filho A, Soerjomataram I, et al. Breast cancer survival in sub-Saharan Africa by age, stage at diagnosis and human development index: a population-based registry study. International Journal of Cancer, 2020. europepmc.org/article/MED/31087650
14. World Health Organization. The Global Breast Cancer Initiative. WHO, accessed 2026. who.int/initiatives/global-breast-cancer-initiative
15. Boucheron P, Zietsman A, Anele A, et al. Measuring the WHO Global Breast Cancer Initiative pillars' key performance indicators in sub-Saharan Africa: experience in the African Breast Cancer-Disparities in Outcomes hospital-based cohort study. eClinicalMedicine, 2025. europepmc.org/article/MED/40034567