Cancer

The second-largest cause of death on Earth: 9.7 million deaths a year.

Updated 2026-09-11 · Download this data

The numbers

Updated 2026-09-11 · CSV · JSON

Cancer

new cancer cases per year

19.3M19.7M20M20M202020212022
19.3M19.5M19.8M20M20M202020212022
Data table
YearValue
202019,300,000
202220,000,000
IARC GLOBOCAN 2020 and 2022. WHO projects a 77 percent rise in cases by 2050. GLOBOCAN editions are not strictly comparable, so the series is an anchor, not a trend line · As of 2026-09-11

Neglectedness

2

score out of 10. Higher means more neglected.

NCI budget; IQVIA Global Oncology Trends 2025; Merck full-year 2024 results

As of 2026-09-11

Tractability

6

score out of 10. Higher means effort buys more progress.

WHO cancer fact sheet

As of 2026-09-11

What you can do

Open data

No checked action is listed for this yet.

Build

Startups, labs, open work

None yet.

Fund

Grants, prizes, funders

None yet.

Buy

Products to use or back

None yet.

Learn

Explainers and papers

None yet.

The burden behind it

1.9Mdeaths a year in the causes this problem addresses

Cancers, all sites: 9.7M (2022), shown for scale, not summed

Trachea, bronchus and lung cancersWHO #6 in 2021

Carcinogens, overwhelmingly from tobacco, accumulate mutations in airway cells until growth control fails. Symptoms arrive late, so most cases are found after spread.

The largest single cancer killer. Counted under the cancer problem; the all-cancers aggregate below is shown for scale and not summed.

1.9M2021

Deaths in the causes this problem addresses, from WHO 2021 and GBD 2023. Not deaths this problem would prevent. A cause shared by two problems is counted under both, so sums across problems overlap.

Sources: WHO Global Health Estimates 2021, top 10 causes of death · IARC GLOBOCAN 2022, global cancer statistics

Why it is not solved yet

Biggest blocker

Known prevention and treatment do not reach most people who die of cancer.

  • WorkPick one LMIC country and map which cancer treatments exist, where, and what they cost.
  • FundFund one radiotherapy or pathology training slot through an existing IAEA or St. Jude program.
  • RunOffer logistics or procurement help to a cancer center listed in a Rays of Hope country.
Why, and how we know

About half of cancer deaths trace to modifiable risks, and about 70% of deaths occur in low- and middle-income countries. The fixes are policy, machines, staff and drugs, which nobody profits from selling at scale. Binding choice is inference.

Wrong if: If modifiable risks explained well under a quarter of deaths, or survival gaps closed without new delivery, this is not binding.

Found: Looked for refutation. Found the opposite: ACS/IARC Cancer Atlas says about half of cancer deaths are attributable to modifiable risks; a 2026 WHO/IARC study puts 37.8% of new cases on modifiable risks. WHO AFRO says about 70% of cancer deaths are in LMICs. Weak point: the "half" figure is from an ACS-led atlas, and attribution is not the same as deaths avoidable in practice.

On it: WHO Global Initiative for Childhood Cancer, IAEA Rays of Hope, St. Jude Global

ASCO Post: nearly 50% of cancer deaths attributable to modifiable risks (Cancer Atlas) · ASCO Post: WHO/IARC analysis, 37.8% of cases attributable to modifiable risks (Feb 2026) · WHO AFRO: cancer

Tobacco still causes about one in five cancer deaths, and 1.2 billion people use it.

  • WorkWrite a one-page brief on tobacco tax levels for one country where use is rising.
  • FundGive to a tobacco control program in a country where absolute user counts are growing.
Why, and how we know

Tobacco use fell 27% since 2010, short of the 30% WHO target. Users in Africa are still rising with population growth. The levers are taxes and laws, which take politics, not science.

Wrong if: If tobacco-attributable cancer deaths fell sharply without further policy change, this would not be a live constraint.

Found: WHO October 2025 report: 1.38B users in 2000 to 1.2B in 2024; prevalence 33.1% to 19.5%; projected over 1.15B by 2030. Decline is real but slow. The "one in five cancer deaths" figure is from the Cancer Atlas, not a WHO death count; WHO/IARC gives 15.1% of new cases from smoking.

On it: WHO

WHO tobacco trends report, 6 Oct 2025 · ACS Cancer Atlas: risk factors

Children with cancer survive about 80% in rich countries and about 20 to 30% elsewhere.

  • FundFund one year of essential childhood cancer drugs for a single named hospital partner.
  • RunHelp one pediatric oncology unit set up a basic cancer registry to count outcomes.
Why, and how we know

More than 20 African countries have no radiotherapy at all. Pathology, drugs and trained staff are missing together. WHO targets 60% childhood survival by 2030, which means the gap is still open.

Wrong if: If LMIC childhood survival already approached 60% broadly, or radiotherapy gaps were closed, this child would shrink.

Found: UICC: about 80% five-year survival in high-income countries; WHO-linked sources say 20 to 30% in LMICs, St. Jude says about a third. A 2019-data study of 197 countries found no low- or lower-middle-income country at 60% at baseline. IAEA says more than 20 African countries lack radiotherapy; the exact current count was not found.

On it: WHO Global Initiative for Childhood Cancer, IAEA Rays of Hope, St. Jude Global

UICC: Global Initiative for Childhood Cancer · St. Jude: Global Initiative · IAEA Rays of Hope story · IAEA/WHO joint statement on Rays of Hope

Most cancer deaths come from cancers found late, with no proven screening test.

  • WorkTrack the peer-reviewed NHS-Galleri paper and summarize its mortality data when published.
  • FundFund an independent analysis of stage-shift evidence for one cancer with no screening.
Why, and how we know

About 70% of US cancer deaths come from cancers with no recommended screening. The leading multi-cancer blood test missed its main trial goal. Stage shift does not reliably predict fewer deaths, so proof is slow.

Wrong if: A randomized trial showing a multi-cancer test cuts cancer deaths would turn this into a delivery problem.

Found: NHS-Galleri (ASCO, May 2026) missed its primary endpoint: stage III/IV incidence ratio 1.03 (95% CI 0.92 to 1.14). Stage IV fell 14% overall, a secondary endpoint, company-reported and not yet peer reviewed. A 2025 meta-analysis found late-stage vs mortality correlation of 0.69, but varying by cancer. In LMICs, 40% to 80% of new cases are already advanced at diagnosis.

On it: GRAIL, NHS England

GRAIL: full NHS-Galleri results at ASCO 2026 (company release) · AJMC: cancers without screening and multicancer detection · Late-stage outcomes as surrogates for mortality: meta-analysis · Late-stage presentation in LMICs (Clin Exp Metastasis 2023)

Proven screening is underused: under one in five eligible US adults get lung screening.

  • WorkBuild a free eligibility checker for lung screening that tells people what to ask their doctor.
  • RunGet one primary care clinic to record pack-year smoking history for every adult patient.
Why, and how we know

Eligibility depends on smoking history that clinics rarely record. Uptake varies threefold across states, lowest where lung cancer deaths are highest. Only about 14% of diagnosed cancers are found by screening.

Wrong if: If eligible adults were mostly up to date and deaths still did not fall, uptake would not be the limit.

Found: ACS, JAMA, 2024 NHIS data: 12.8M eligible, 18.7% up to date, state range 9.7% to 31%. An older ALA figure was 5.8%, so the trend is up. ACS models 62,110 lung cancer deaths prevented over five years at full uptake. Global screening uptake was not researched.

On it: American Cancer Society

AuntMinnie: ACS, only 1 in 5 eligible US adults screened for lung cancer in 2024 · AARP: lung screening guidelines (ALA 5.8% figure) · Duke Cancer Institute: 14% of cancers found by screening (NORC)

Cancer trials are slow because few patients enroll and many trials never finish.

  • WorkList active cancer trials within 100 miles of one rural county and publish the gap.
  • FundCover travel costs for patients enrolling in one cancer center trial program.
Why, and how we know

About 7% of US cancer patients join treatment trials, and most US counties have no active cancer trial. About one in five adult cancer trials fails to finish enrollment. Sites, eligibility rules and costs are not coordinated.

Wrong if: If most patients offered trials refused, the limit would be willingness, not access or design.

Found: Refutation check: a 2020 meta-analysis found 55% of patients agree when offered a trial, so willingness is not the main limit. Enrollment estimate is 7.1% (Unger, JCO 2024), up from the older under-5% figure, and modeled from accreditation data. Cost per patient: only a 2013 industry estimate ($74,800, phase 3) and a 2019 single-center Italian estimate (EUR 11,379) were found; current cost is unknown.

HealthDay: enrollment in cancer treatment trials about 7% (Unger, JCO) · Healio: many patients lack trial access where they live · Healio: one in five cancer trials fails to complete enrollment · PLoS ONE 2019: per-patient trial cost, Italian center · Business Wire 2013: phase 3 oncology per-patient cost

Poor accrual is the leading reason cancer trials stop early.

  • WorkAnalyze terminated oncology trials on ClinicalTrials.gov from the last year and tag stop reasons.
Why, and how we know

Trials open at too few sites, with narrow eligibility, and compete for the same patients. Poor accrual caused 38.7% of early terminations in one analysis and 33% in another.

Wrong if: If terminated trials mostly stopped for safety or futility, accrual would not be the bottleneck.

Found: Estimates differ by dataset: 20% risk of failing to complete enrollment (2005 to 2011 adult trials); 8.5% terminated or withdrawn across 12,875 oncology registrations; 37.9% of NCI phase III trials missed minimum accrual. Poor accrual is the top cause in each.

PMC: premature termination in oncology trials · Healio: one in five cancer trials fails to complete enrollment

Who is on it

New science

All 10

Who is moving on Cancer

A human checks every entry. Nothing here can be bought.

Open starts

0 willing to start

0 so far. Be first.

Open joins

0 want in

Roche

public

0 people want in

I want in

Merck & Co.

public

0 people want in

I want in

AstraZeneca

public

0 people want in

I want in

Abbott

public

0 people want in

I want in

GRAIL

public

0 people want in

I want in

Open capital

0 allocators watching

0 so far. You can stay anonymous.

Act on Cancer

What exactly

0 / 600

A human reviews every submission. Nothing publishes itself.

Cite this

9.7 million cancer deaths and 20 million new diagnoses in 2022 are both annual flows. Dividing them is not an annual individual mortality risk and is not scored. A prevalence denominator was not opened, and none is invented here.

  1. optimism.fun. (2026). Cancer. https://optimism.fun/p/cancer
  2. IARC GLOBOCAN 2022 (UICC summary). https://www.uicc.org/news-and-updates/news/globocan-2022-latest-global-cancer-data-shows-rising-incidence-and-stark
  3. IARC GLOBOCAN 2020 (ASCO Post summary). https://ascopost.com/news/december-2020/globocan-2020-database-provides-latest-global-data-on-cancer-burden-cancer-deaths
  4. WHO, global cancer burden growing, February 2024. https://www.who.int/news/item/01-02-2024-global-cancer-burden-growing--amidst-mounting-need-for-services
  5. IQVIA Institute, Global Oncology Trends 2025. https://www.iqvia.com/insights/the-iqvia-institute/reports-and-publications/reports/global-oncology-trends-2025
  6. Merck, fourth-quarter and full-year 2024 results. https://www.merck.com/news/merck-announces-fourth-quarter-and-full-year-2024-financial-results/
  7. NCI budget and appropriations. https://www.cancer.gov/about-nci/budget
  8. WHO cancer fact sheet. https://www.who.int/news-room/fact-sheets/detail/cancer